+ Targeting Zero

The problem

The DHHS collects a lot of rich and comprehensive data about harm and near-misses in hospitals, however we've never reported back to the sector. There are various technical and design barriers that have caused this

Proposed Solution

An Open-source database that enables:

• Reporting back to the sector meaningfully
• Abiility to extract meaning from unstructured data
• Provides real-time insights into emerging trends

+ Health Management Tool

The problem

Why?

For people with multiple conditions, it is difficult to manage each aspect of their health and to capture the medical advice they are given by various health professionals.

This health management tool will comprise a checklist of questions, a record of symptoms and blood test results at each interaction with a health professional, and a summary of the discussion with details of prescriptions and dosage.

Ideal solution for this problem?

I envisage an online diary with sections for a checklist, a record of symptoms, blood test results, a summary of the discussion, details of prescriptions and dosage. It should be password-protected and accessible via desktop and all mobile devices.

What can be done in 48h

Web-based app.

What are the current solutions for handling this problem?

Currently patients have to juggle management of every aspect of their own health, they have to know what questions to ask their doctor, what symptoms to monitor.

Blood test results are presented in different formats by different labs, and patients are not always given copies, which they do not know how to interpret anyway.

Health professionals may use jargon in speaking with patients and due to the short consultation times, speak so quickly that patients cannot remember everything that is said. It would be helpful for patients to remember what prescriptions and dosages they have both in the past and present.

+ Identifying and Counting Objects on a Photo

The problem

Overview

Detecting, counting, and measuring highly similar objects on an image is a surprisingly hard problem to automate, especially when their numbers vary and the objects touch or even partially overlap. Think about ‘counting people on a group image’. It is a very simple task for the human eye and brain, however, it is a very hard problem to ask a computer to do; so I have been told. I don’t know! I am a biologist. not a computer scientist. I would like to have a computer program that takes images, identifies objects, counts them and measures each one. Ideal output is summary statistics of size, color and texture.

Details

Applications are many and not restricted to medical images. As test data, I have prepared highly similar photos of highly similar objects. All photos have the same layout. The objects are highly similar in shape, but differ in size (length/width), color, and texture, such as rice seeds that will use as a working example.

The solution

At this Hackathon I would like to find skilled people who explore this problem with me. An approximate solution would be a sufficient outcome. The approach can be partially supervised, in a way that a human expert teaches the algorithm one or several typical objects. Success is guaranteed: Either we show that it can be done, or we find and that it is much harder than it seems. If it turns out to be possible, we will prepare a scientific publication, and I am also willing to push development towards a mobile app.

+ Build a test for depression based on time perception

The problem

I’m doing a psychology PhD that looks to build a test for depression based on time perception. I’ve built a version already, but it doesn’t work on all devices and browsers. It’s very simple in principle, but it’s getting it to work flawlessly for clinical purposes, that’s the problem. All it needs to do is measure the time between a user pressing start and stop buttons in their web browser.

Sounds simple. Isn’t. For example, Apple devices do weird things at around 40 seconds if you haven’t touched the screen or moved the mouse. To be effective, the screen will need to be idle for about 90 seconds.

Why do you want to solve this problem?

I’m very keen to have people troubleshoot and build a solution because I’ve already spent my allocated funding on development. With a working web app, I will be able to conduct a population-wide test and hopefully prove it up for clinical use. If it works, it will help identify symptoms of depression and anxiety like a thermometer measures fever. Simple, quick, cheap, accessible and, most importantly, difficult to fake.

What do you envision as the ideal solution for this problem?

I already have a version that works fine on PC and on Apple up to around 45 seconds. Ideally, the web app will function equally well on all devices and in all browsers. The current developers have ideas about how that might be done, but I'm keen to see how others approach it.

The ideal solution would also include an element of feedback after each section of the test so that it could be used as a therapeutic treatment as well as a screening test.

What sort of Open Source solution do you think can be created in 48 hours, by a small team of developers, designers and data analysts?

It's such a simple program, I imagine that a fully functioning version could be built and tested in 48 hours. If a quick technical solution is found that works independently of the browser/system, then work could be done on upgrades to the basic version that gamifies of provides clever feedback to the user. At the moment, I am asking test subjects to make sure that they move the mouse while doing the 40s-90s intervals, which is not ideal. There is currently no workaround in terms of general release for people to use at home on any/all devices and browsers. My research is at a standstill until the problem is fixed.

+ Improving MRI with Signal Processing

The problem

Overview

MRI uses a very interesting technology (that won a Nobel Prize) for determining the composition of tissue inside the human body. It is especially attractive as it causes no harm the the individual (unlike X-Ray). There are on-going challenges with this technology around resolution and overcoming noise and distortion within the image reconstruction.

Solution

We are interested in working on this problem to see what we can achieve in terms of improving MR Imaging. Programming skills and signal processing skills could be a good addition to the team.

+ Health Data Management Solution Searching for a Problem

The problem

The Solution

I would like to address a problem in the health industry - data management. The solution I want to suggest for this problem relies on technology like machine-to-machine data transfer, which can be easily done in a hospital environment. This solution would include general patient care, infrastructure maintenance with a possibility of remote calibration and control of relevant devices.

Overview

I am seeking a domain expert from the health sector with a real problem to which this type of solution could be applied.

+ Managing Citations

THe problem

The reason I decided to attend HealthHack 2016 in Canberra is to try to solve a problem of my every day work: time spent on managing and comparing citation libraries.

As a medical doctor and recently started PhD student at Research School of Population Health at ANU, one part of my work includes systematic reviews. A systematic review involves searching and importing a large number of specific references from different online scientific libraries to our computer and using citation manager to find the most relevant references which give us useful publications. As these libraries are independent of each other, sooner or later we get some duplicates. Unfortunately, none of the existing citation managers is able to find and remove duplicates efficiently.

Why? Because references of the same article from different libraries are not consistent (spelling errors, different author order, punctuation) so they are not recognized as the citation for the same article. Consequently, we end up with a final library that is much bigger than it should be and we have to find the rest of duplicates manually. To be done correctly, manual sorting of duplicates usually takes 2-3 weeks if you are dealing with for example 9000 articles like me. Also, there is no automatic way of comparing two different libraries (on our computer) if there are more people working on the same project.

Solution

Therefore, my idea is to come up with a program or a tool which would recognize such duplicates with great accuracy and therefore save the precious time looking for these duplicates manually instead of actually doing research. Also, I wanted to have a tool which would allow me to compare different citation libraries stored on my computer.

Who I need?

I need a group of people with some knowledge of programming to help me implement my ideas.

+ Improving Techniques in Medical Radiation

The problem

I am the Chief Medical Physicist for ACT Health, managing a team of 20 (including 15 medical physicists and 4 radiation engineers/technicians) many of whom develop code to improve safety, quality and efficiency of health care delivery using medical radiation. Several opportunities to develop front-line code exist to control or collect and analyse data on these technologies.

Proposition

We would like to be involved in HealthHack Canberra both as problem owners and and problem solvers. We have the skills to help other participants with their challenges, but we are also looking for collaboration with other hackers to help us with our own problems.

+ Registry of Research Study Participants

The problem

Overview

Many research studies, especially in medicine, require participation of people in the research experiments. For example, many studies are trying to understand the relationship between lifestyle habits, genetic predispositions and individual well-being, ranging from physical to mental fitness. In many cases, those participants are volunteers and the final study results often go unreported to the study participants.

The solution

We would like to create and maintain a registry for study participants that will manage their involvement and feedback with the research project dynamically through mobile apps or an online platform.

+ The Sensory Box

The problem

The idea

A service to cater for early developmental educational age (2-7) who are pre-mid-post a diagnosis that includes Sensory Processing Disorder.

Solution?

A subscription box for parents to access.

More info to follow...

+ Better X-ray Image Quality for Less Radiation

The problem

Overview

In medical diagnostic X-ray industry, X-ray image clarity is improved using a hardware device that increases radiation dose. Better quality images help clinicians make better diagnosis and trust more to their decisions, which makes this device important tool in the process. However, the higher dosage of radiation can lead to increased probabilty of radiation-induced cancer for people who undergo X-ray examinations. This danger is also reflected in careful considerations of patients when ordering the test.

The solution

In our lab, we have developed a software program relying on complex algorithms for image reconstruction and image processing which can replace the radiation device described above, while retaining the image quality. We are happy to recruit more skilful candidates to expand our development concurrently to computer tomography (CT), fluoroscopy, radiography, portable radiography, and mammography.

+ Making Decisions with the Future in Mind

The problem

Overview

People are greedy. We want stuff now, we don’t want to wait until later.

This kind of “cake now” (rather than “skinny later”) decision-making leads us into all kinds of trouble – obesity, smoking, booze…

In the lab we’ve got a pretty simple activity that can turn this around, helping people make decisions that benefit the long-term.

But can we get this to work in the real world? Just because we can get you to resist a cookie while some jerk in a white coat glares at you doesn’t mean it’ll work when it matters.

The Solution?

I want to develop a delivery method that can take this activity out of the lab and to people in real life.

Apps, Games, Simulations, the floor is open.

+ Annotations for Scientific/Medical Papers for Non-experts

The problem

To help non-expert audiences understand scientific and medical papers using electronic annotations of specific sections of the papers.

Why do you want to solve this problem?

Genetics and genomics are currently highly specialised fields. To allow for genomic medicine to be main-streamed and for non-experts to be able to understand the outcomes, methods and limitations of current knowledge, we want to annotate current papers in a way that medical specialists, other healthcare professionals and patients/family members can understand.

Different texts could also allow genetic/genomic literacy to be developed as part of school science programs, where teachers use papers, scientific articles and other texts (that could be saved and served as PDFs) to develop students' scientific literacy.

What do you envision as the ideal solution for this problem?

Ideally, 2 outcomes:

1. An online solution or app that allows us to annotate and enable access for specific audiences to see a PDF paper with highlighted or annotated sections (via annotation marks or an annotation frame/viewer) to see the information we provide (eg. a glossary scientific and medical papers so that the annotations can be shared with different audiences (via a closed community unless copyright issues could be surmounted).
2. A tool that could be shared with and used by healthcare educators and school teachers to upload their own resources, modify existing annotations and add their own, according to their group's interest and ability.

What sort of Open Source solution do you think can be created in 48 hours?

The solution would focus mostly the capability to view and use annotations through building a tool that allows upload of a document, visual signposting of interpreted sections and a way of seeing/accessing the annotation. If there were any thoughts about how annotations could be shared (eg. by export) or modified too, that would be wonderful.

What are the current solutions for handling this problem?

Fermat's library, Crocodoc/Goodreader and PDF professional all enable annotations of PDFs in some way, but few are intuitive or shareable and none are open source.

+Curated Information for Health Professionals and the Public

The problem

To effectively display curated Internet links on genetics and genomics for health professionals and/or patients and members of the public.

Why do you want to solve this problem?

Educational resources about genetics and genomics are prolific throughout the world, but there is no effectively-displayed curated list of resources for either health professionals or patients/members of the public. Many of these resources are very useful for people to understand the promise and pitfalls of genomic testing , but hard to find.

What do you envision as the ideal solution for this problem?

A online portal that presents curated resources -- down to specific media objects or individual web pages -- that can be filtered by audience (health professional, teacher, patient, interested non-expert) visually with a small amount of information (name, author, web address, community ratings, audience information) that would extract images directly from the site to avoid issues with copyright and contract negotiations.

What sort of Open Source solution do you think can be created in 48 hours?

User interface that provides an intuitive visual display, with a back end that allows additional information to be tagged and displayed both as an immediately-viewable record and as part of a longer information record. If this could be filtered according to a specific audience, that would also be amazing.

What are the current solutions for handling this problem?

At this stage, we just put these links on a webpage. It is neither intuitive or filterable.

http://www.garvan.org.au/research/kinghorn-centre-for-clinical-genomics/clinical-genomics/learn-about-genomics/useful-links

+ Eliminate Errors from Genome Sequencing Data

The Problem

Overview

To improve genome modelling workflows to help identify genetic disease

We are interested in the discovery of very rare mutations that cause autoimmune diseases. These mutations are acquired during the lifetime of the patient and are only present in a small fraction of their cells. At the moment the frequency with which these mutations are expected to occur at frequencies below the error rate of the technology used to generate the data. To improve our ability to detect rare mutations we augment the data with molecular identifiers that allow us to exploit redundancies in the data to correct errors and improve the signal to noise ratio sufficiently to enable the detection of rare mutations.

Why

Currently there are no tools that are capable of fully exploiting this information. Once created this solution will form a crucial part of a pipeline for the identification of rare mutations in patients with autoimmune diseases. Ultimately this will lead not only to a better understanding of these diseases but may also enable better, more targeted treatments for patients.

Solution

An easy to use command-line tool that identifies redundant sequencing reads (that originated from the same DNA molecule in the sample) and replaces them with their consensus sequence, thereby substantially reducing the error rate.

The accuracy of the sequence clustering, and especially of the resulting consensus sequence, is critical.

Tools for clustering of sequence reads based on sequence similarity exist but these do not take molecular tags into account. The existing solutions typically rely on a greedy clustering strategy that generally seems to work well.

A similar approach, customised to incorporate molecular tags, could be adopted here.

Outcome

Devising a successful strategy to tackle the sequence clustering problem and the implementation of an initial prototype seem possible.

+ Custom GMS Framework Hack

The Problem

At a high level, I want to be able to be able to hack into the GMS** framework and implement customized changes onto the variant calling workflow which has a direct impact on both pre-clinical and clinical research.

There are two immediate low level problems at hand that I intend to solve (in the order of priority)

1. utilising the BWA MEM option for read alignment.
2. importing new tool sets for SV detection (LUMPY, DELLY)

The latest release of GMS implements BWA ALN(hardcoded) in its workflow making it difficult to invoke the BWA MEM option which is better suited for aligning of reads longer than 75 bases.

Another challenge is the repertoire of available for Structural Variant (SV) detection tools. SV detection tools like DELLY and LUMPY have demonstrated superior performance compared to the existing tools and would be immediate candidates for implementation on GMS workflow.

http://dx.doi.org/10.1371/journal.pcbi.1004274

Why do you want to solve this problem?

GMS provides a user-friendly interface for end-to-end somatic variant calling that utilises an integrative approach by combining different omics datatypes (WGS, WES and transcriptome) and contextualizing the outcome for therapeutic intervention by leveraging databases like dbGAP and DGIdb. The framework for data-management and options for customisation of analytical experiments make GMS an ideal choice as a ""hackable framework"" and greatly improve the scope for Clinical Translation. The field of genomics is growing fast and it is imperative to keep up with the latest algorithms and variant databases given their impact on the clinical translation. However, it requires expertise in software engineering to enhance the repertoire of tool tree offered by GMS.

This effort will contribute towards addressing the problem of improving the scope of GMS tool set.

What do you envision as the ideal solution for this problem?

If you had unlimited resources

Implement and test of multiple algorithms released by the PANCAN groups.

Implement new code base to perform downstream analysis using the command classes*.

Implement data import module for SNP array data and external BAMs.

https://github.com/genome/gms/wiki/Writing-a-New-Command-Class-in-the-Genome-Modeling-System

What sort of Open Source solution do you think can be created in 48 hours, by a small team of developers, designers and data analysts?

Please be aware that the code will need to be made available to the community.

Implementation and testing of

1. BWA MEM, LUMPY (day1) and
2. DELLY (day2)

What are the current solutions for handling this problem?

Or your current workflow for this problem

BCBIO* – lacks pipeline for clinical translation

GALAXY* – misses out on data management and integrative analysis

https://docs.google.com/spreadsheets/d/1o8iYwYUy0V7IECmu21Und3XALwQihioj23WGv-w0itk/pubhtml

+ Epilepsy Diary

People with epilepsy have many more seizures that they are typically able to report using seizure diaries. This can lead to difficulty in understanding their condition and prescribing the best medication and care for them. In addition, it is difficult for pharmaceutical companies to evaluate new medications as there can be a lack of reliability in the reporting of seizures in the clinical trials.

The Problem

People with epilepsy have many more seizures that they are typically able to report using seizure diaries. This can lead to difficulty in understanding their condition and prescribing the best medication and care for them. In addition, it is difficult for pharmaceutical companies to evaluate new medications as there can be a lack of reliability in the reporting of seizures in the clinical trials.

Are there any existing solutions?

Patients usually record their seizures on paper or in electronic form, but this requires them or their carers to do this.

What would be the ideal solution, given a big budget and lots of time?

A mobile phone or other device that can tell if a person is having a seizure and record it for reporting to the patient's clinician either immediately or on a regular basis. Ideally, this device should record the situation in a way that allows for determination of false positives, which can thus enable higher sensitivity.

What can be done in 48 hours at a hackathon?

A prototype that the detects shaking, turns on the microphone for a while to record local activity and logs the event would be a great proof of concept that would help further our work and be useful in short term use.

+ Pathology Report Text Mining for Infectious Diseases

The problem

We extract a large volume of pathology reports from GP patient management systems, that each use different formatting standards (or none!). Translating between any given pathology report from one of ~20 (!) different laboratories, needless to say, introduces various inconsistencies.

We need to pull out relevant information from the pathology reports in a standardized manner - everything from disease name, assay type to the result (frequently multiple in the one record) - in order to perform productive and meaningful analyses on it.

Solution

The solution should parse the raw text into four to five variables indicating disease, assay, test result, specimen type and specimen site. We are quite comfortable with a command line style utility.

We anticipate that the solution may need to take advantage of a reasonable amount of computing power.

Some exploration of disease ontologies may be useful in order to improve the accuracy of the system.

Our current system is written in the statistical package STATA - however, we are open to other systems, particularly Python or R.

+ Addressing Vaccination Hesitancy with Serious Games

A minigame to educate the public about vaccination, how it works and why they should get vaccinated

The problem

Background

With the continual improvement of telecommunications and information technologies, eHealth has a great potential to address many health challenges across the world.

One of the greatest health challenges today is vaccination with vaccine hesitancy as one of its main concerns.

In our recently published review on serious games addressing vaccination hesitancy, we found that none were specifically designed to address vaccine hesitancy determinants and none were formally evaluated regarding their impact on vaccination uptake.

The challange

As the pharmaceutical industry shifts towards creating services to match their products, serious games may be a cost effective and efficient service solution for vaccines.

We believe that serious games addressing vaccination hesitancy may be an innovative public health approach and we would like to create adaptable and customizable serious games that will demonstrate an impact on vaccination behaviors accustomed to both the public population and healthcare professionals.

We are looking for people that can help bring our mock-up idea of serious video games to life. We are in need of developers, graphic/creative designers and marketing experts to help turn this idea into reality.

Outcome

The solution can be used by both public health organisations and private companies in order to promote and improve vaccination at local, national and international levels.

+ 3D Pain Scale App

A pain journal app to assist patients to communicate their condition to doctors.

The Problem

When in a very short appointment (sometimes 10 minutes) with your doctor, it's hard to explain the breadth and depth of a chronic pain condition, this hinders diagnosis and treatment. Currently the 10 point scale of ""how are you feeling right now"" is used as a representative pain scale. This is not in-depth or detailed enough to explain a chronic pain condition.

The solution

Research exists about different types of pain we can feel and describe for example MPQ (and it's short form), as well as adult measures of pain, among others. These existing systems (or any other in the literature) can be turned into an app based pain journaling solution for tracking your daily pain levels and reporting your pain to your doctor at the next appointment.

+ Patient Journeys in the Victoria Health System

A patient's journey through the healthcare system can be varied and there is currently no visually intuitive way to view the most common journeys or measure the time it takes to complete these journeys.

The Problem

We would like to visualise the most common patient journeys, accounting for different cohorts of patients. What services do they use, how often do they have contact with providers and how long does it take to complete? How do these journeys differ for patients who live in various areas? For example, is the total journey time for a public patient who needs a hip replacement the same if they live in Sale as it is for a similar patient who lives in East Melbourne? Does the length of time vary depending on the starting locations e.g. different providers, ED presentation vs specialist clinic or by other characteristics such as age, location, gender, disease profile?

Solution

We would like a prototype web enabled visualisation tool, utilising open source D3.js visualisation libraries and the unit record data supplied by DHHS (most likely .csv formatted files).

+ Morgana - A data Visualisation tool to Inspect Complex Datasets.

The Problem

Researchers working in clinical studies collect a lot of patient samples, laboratory and clinical data, and these are often collected independently of each other.

A major challenge is to easily see what data we have on whom. For example, we might have genetic data on 15% of our cohort, epigenetic measures on 35%, we will have collected flow cytometry data on 60% and serology on 100% of the cohort.

How can we aim to curb life-threatening diseases when the laboratory data we collect is not even remotely connected to clinical outcomes, and specimen data?

How can data stored in different incompatible systems be of any use to solving a research question?

There’s a severe need for an intuitive and interactive way to visualize relationships between different types of data collected in clinical studies.

We need to be able to examine this data in a myriad of scales and contexts, from a bird's eye view through to specific subgroups containing a half dozen specimens can inform this critical stage of research.

This project from the Murdoch Children's Research Institute will utilize actual data from one of the worlds leading cohorts of paediatric food allergy, to develop a flexible, intuitive, visual solution to research data management.

The solution

Ideally, the solution would consist of an easily queried database (preferably SQL-based), providing data to a web frontend. Several excellent libraries exist for this purpose (e.g. d3.js, plot.ly), though as is to be expected, the devil is in the details - gluing the various components together into a cohesive, user-friendly whole is the real challenge for this project.

If you think you can help us develop this tool, you’ll be working on a project that will make research teams more effective in combating life-threatening disease.

Data Available

All data for the project is available at Github

+ Mouse Movement Detection in Videos

Identify periods of activity in videos of mice with epileptic disorders

The Problem

Overview

Mouse models assist in obtaining insights into epileptic disorders caused by genetic mutations. Mice with specific mutations are bred and monitored 24/7 through video.

Going through these videos is a tedious process, when there are long periods where mice are inactive. Therefore, we need to extract periods of activity, in order to increase the efficiency of our analysis pipeline.

Technical details

Part 1:

• Identify periods of the video where mice are active (ex: walking, eating, drinking)

Part2:

• Develop an app (web or offline) which allows a user to quickly skip the video to identified periods
• Allow users to mark time periods where mce are having seizures using the GUI

Data Available

• 3 video clips of varying time lengths which monitors 2 mice simultaneously

+ Web-based Metadata Annotation Tool

Create a framework to automate the update of descriptions quickly while still providing good feedback to the user about invalid data?

The Problem

Overview

In research datasets there is a need to describe your data (called metadata) using multiple fields.

Stemformatics has datasets where there are over 1,500 samples and 9 mandatory fields and they are only going to get bigger!

Create a framework to automate the update of descriptions quickly while still providing good feedback to the user about invalid data?

Technical details

It should be:

1. Web based
2. Stand alone
3. API driven
4. Fast internationally

Want to be able to handle:

1. Simple validation (eg. nunber only had digits)
2. Validation against defined words
3. Validation of multiple fields
4. Customised validation

Data Available

• Dataset with 1549 samples and mandatory 9 fields that has been annotated with errors.
• Dataset with 6 samples annotated correctly.
• Synthetic dataset with 50,000 samples annotated with errors.

+ Helping Transgender Youth with their Voices

Having a voice that does not match one’s gender identity can cause significant distress for transgender individuals, especially during puberty when the voice can change.

The problem

Having a voice that does not match one’s gender identity can cause significant distress for transgender individuals, especially during puberty when the voice can change.

Transgender health professionals can assist with this in several ways, including the use of voice training and hormonal interventions.

We would like to develop an app that assists transgender youth attain and maintain a voice which is congruent with their gender identity.

Two ideas for projects that can be done in 48 hours:

• An app that will track vocal range (pitch and volume) over time to monitor pubertal changes and to monitor the effects of interventions.
• A game that will provide a fun experience that helps the user practice modulating the quality of their voice to achieve the goal of a gender congruent voice.

+ Improving the data collection of ticks that bite people in Australia

To collect and analyse data on ticks that bite people in Australia to provide current and relevant information for the public and medical professionals. The purpose of this project is to create an app that allows people to log the geographical location of their tick bite, obtain information about the tick species they were bitten by and document symptoms associated with their bite. The data will provide greater insight into ticks and tick-borne disease in Australia.

The Problem

While tick-borne disease is a global threat to the health and well being of animals and humans, very little is known about the ‘bugs’ (bacteria, viruses and/or protozoa) transmitted by Australian native ticks. Our lab has used innovative and advanced genetic techniques to unlock the bacterial diversity found within Australian ticks, which has led to the identification of new and putative disease-causing bacterial species.

Accurate information pertaining to where and when a person was bitten by a tick (e.g. geographical location, season etc) is often missing from the data sheet upon arrival of the tick specimen in our lab.

With around 70 tick species, including 65 native and 5 introduced ticks in Australia, knowing precisely where a person was bitten, when they were bitten, and by which tick species, will provide the necessary information to help us understand the geographical location of ticks that bite humans, along with tick-borne microorganisms in Australia. This will also provide valuable information about the geographical areas of disease risk (‘hotspots’) to be identified, which is essential for informing medical professionals and the general public about ticks and tick-borne disease.

+ Improving the Anaesthetic Experience for Paediatric Patients

We'd like to build an imaginative virtual space using Google Cardboard, to help children undergo anaesthesia without fear or distress.

The Problem

Why?

Approximately 2.5 million Australian patients undergo surgery each year. A significant proportion of these will require general anaesthesia. Currently, anaesthetic medications can be given in only two ways: by injection, or by inhalation. To “put a patient under” (“induce”) anaesthesia, injections are more reliable, but most younger children find needles distressing. The alternative is to administer anaesthetic vapours through a mask, but this takes longer and children often have trouble tolerating the face masks. Skilled paediatric anaesthetists devise ingenious stories and cunning games to reassure and develop rapport with the child, but if trust and communication break down, the result is invariably traumatic for everyone involved.

Many children in Australia these days play games on smartphones or tablet computers. They may even have their own device, complete with headphones, and maybe even a virtual reality headset. Many studies describe the use of cartoons and hand-held games to relieve anxiety around the time of surgery. Hordes of children, playing zombie-themed games on their electronic hardware until they are almost mindless zombies themselves, are already a common sight in hospital waiting rooms. Time to enter the VR arena of the Operating Theatre!

The solution

Sophie is 6 years old, and wants to be an astronaut when she grows up. She has already started the rigorous physical training process required, going to the park and climbing monkey bars every weekend until her hopes came crashing down to Earth this Sunday, when she slipped and broke her right arm. The emergency doctor organised for some X-rays (which she thought were pretty cool), and has some bad news: both of her forearm bones are broken, and she will need to have surgery to get her arm back into shape. Those astronaut plans will have to wait — but not for long!

When the time comes to go into surgery, little Sophie - her head full of fun stories the nurses have been telling her about outer space - feels a stab of anxiety about the coming operation... but instead of a scary injection, a VR device and mask is waiting for her. ""Are you ready to go into space?"" the nurses ask her. She nods, distracted by the new toy, and they help her put it on. Once she does, the anaesthetic bay is instantly transformed into the cockpit of a rocket! A few small bumps in the ground could be the clamps on the spaceship being released as Sophie is wheeled into the operating theatre, filled with excitement.

“Are you ready to put your breathing mask on and start flying the rocket?” Sophie can’t wait.

“Delivering oxygen now. Take deep breaths in an out.”

Is that beeping coming from the heart monitor, or is it the countdown for launch? 3... 2... 1...

“Commencing rocket ignition... this might smell a bit funny. Keep breathing normally...”

As the rocket starts to take off, there is a scent like jet fuel, and Sophie begins to feel really light... or really heavy... she feels herself really take off into the stars.

Before she knows it... “Sophie? Sophie! Wake up Sophie! How are you feeling?”

She's back on Earth, and - happily - the whole procedure has gone by unnoticed, as she dreamed of her adventures in outer space.

To make Sophie's scenario the new norm for kids undergoing surgery, we'd like your help to:

• Create a virtual space in Google Cardboard, designed to ease children's fears during pre-surgery.
• Create some interactions or animation to hold the children's interest.
• Consider and/or prototype other possible scenarios or design ideas, e.g. pupil tracking, other forms of physiological or psychological monitoring.

We can also provide:

• Anaesthetic masks.
• Amateur video of a simulated anaesthetic session.
• Close-up view / details of anaesthetic machines.

+ Abnormal Growth Charts

Build a mobile app that allows parents with children who are born with bone dysplasias to record their height and plot it on a growth chart appropriate for their condition.

The problem

Why?

Plotting a child’s height on a growth chart is an easy way to monitor a child’s developmental health. If a child’s height falls outside of the normal range, this could be an early warning sign that there is a serious medical problem.

Children with bone dysplasias have a genetic defect that prevents their bones from growing normally. The most common types of bone dysplasias are Achondroplasia, Pseudoachondroplasia, Diastrophic Dysplasia and Spondyloepiphyseal Dysplasia Congenita (SEDC). As these children grow differently from other children, they need to be plotted on growth charts specific to their condition. Since bone dysplasias are rare, it is difficult to collect sufficient measurements from lots of children with these conditions to produce a high quality growth chart. Low quality growth charts based on a small amount of measurements do exist, but they are rarely used as many doctors are not aware of their existence. The goal of this project is therefore twofold:

1) promote the use of the existing growth charts and

2) collect height measurements from patients using the app to produce better growth charts with the power of citizen science!

Patients with diastrophic dysplasia (left) and achondroplasia (right).

Solution

• Native iPhone and Android apps
• Use growth charts from the publications provided
• On first launch register new patient:
  • enter patient's name, sex, date of birth and diagnosis (dropdown for achondroplasia, pseudoachondroplasia, diastrophic dysplasia, SEDC)
  • Message: ‘the height measurements you are entering in this app will be used to produce better growth charts for your condition in the future. If you do not want your height measurements to be used this way, please tick the box below'
  • Message: 'to produce better growth charts for your condition in the future, it is important that we confirm your diagnosis and get some more details on your medical history. We would therefore like to contact you by email to ask you a few questions about your condition. Do you agree to be contacted by email?’ Tickbox ‘Yes, I agree to be contacted’ if ticked yes, enter email address
• Option to register multiple patients and choose who is currently shown on the growth charts
• Show growth chart for the active patient. Previously entered measurements are plotted on the growth chart as dots and connected with a line. Choose correct growth chart based on diagnosis and patient’s sex (if sex-specific growth charts are available)
• ‘+’ button to enter new measurement. Default to today’s date, but date can be changed by user (e.g. to record older measurements). Choose unit of measure (cm of feet/inches) in preferences.
• option to add a comment to each measurement (for example to record that patient was sick at the time of measurements, or other factors that might influence the accuracy of the measurement).
• Option to export the growth chart with recorded measurements as a PDF (e.g. to put into medical records)
• Store entered measurements on a server: use unique ID assigned by app to identify individual patient without using their name. Also need to store patient’s sex, date of birth, diagnosis, email address (if given).
• A way to push updated growth charts back to user (e.g. through App update?)

What can be done in 48 hours?

if you cannot make native iPhone and Android apps, a web-based app that works on mobiles would work as well.

What is the current solution to the problem?

Patients or doctors make photocopies of the disease specific growth charts and plot their patients on these (if they are aware they exist and can get their hands on these very old papers).

Growth charts for achondroplasia 1978 Horton et al. Growth charts for Diastrophic Dysplasia, Spondyloepiphyseal Dysplasia Congenita and pseudoachondroplasia.

+ Health Checkup Calendar

We want to make it easy for members of the public to remember all the various regular health appointments they should be booking in.

The Problem

Why?

The best practice for regular checkups is not well known by the public, leading people to only find out about them once it’s too late.

The solution

A web interface which asks the users relevant demographic data, and then generates an .ics , .ical, .ifb, .icalendar or other various calendar files that can be imported to their calendar.

Calendars already exist for other science interests like this: http://cantonbecker.com/astronomy-calendar/

Some example reminders might include:

• Yearly doctor appointment
• 6 monthly dentist appointment
• 2 yearly optometrist appointment
• Ear checkups
• Cardiac health (over 50)
• Bone health (over 50 or after several recent breaks)
• Weight check? Gym/health review. “It’s time to check on my health”
• Colonoscopy
• STI test (possibly more relevant for younger demographic but research is needed to confirm)
• Breast exam
• Regular vaccinations
• Reminder to evaluate quality of sleep and get a new mattress
• Mental health checkups

Any other regular tests supported by the literature.

Other considerations might include:

• When their last appointment was (i.e. doctor was last week, might not need to go back for a while)
• How old they are
• Their gender
• Their overall health (self-assessed)

If it’s an app: option to add a photo of their ailments over time to track recovery.

+ Art from Personal Health Data

We want to create something artistic using data from genetics or wearable health trackers to engage the public with science, quantified health and encourage citizen science.

The Problem

Why?

Science communication is arguably the most important part of science. Not only does it help inform people about the latest discoveries, possible applications, as well as help them make sense of the world around them, it also inspires them. By bringing creativity and art to the world of genomics, bioinformatics and quantified health, I want to push the boundaries of what can be done with the vast amount of data produced in our research. A secondary goal, is to give a brief introduction to bioinformatics using a different, and more creative approach.

The solution

An application that can automatically create interesting art (music, visuals, etc) from genetic or other health data and can be incorporated into multiple platforms.

Stretch goals would be to make it easily sharable through social media, and encourage people to use it in an interactive way.

Consider collecting data from:

• Genomes i.e. the human genome project
• Wearable trackers i.e. fitbit
• Research journals

Some interesting data from the Oxford Nanopore Sequencer, MinION, will be available.

+ ADHD Info and Assistance App

The purpose of this project is to create a solution that helps people with ADHD and aprents of children with ADHD find reliable information and service son ADHD as well as assist them with organisation for their appointments and medication

The Problem

The Problem Statement

ADHD is the commonest mental health condition in children which affects 6% of the population although only 1.2% are treated with medication. Currently there are over 18,000 children and adolescents who are prescribed stimulant medication in Western Australia.

Problems patients and their caretakers face

Finding information about ADHD health professionals

Parents with children affected by ADHD have no clear source of up front information regarding professional links and allied services which makes it hard and frustrating to get the right help for their children.

Currently they have to rely on google searches for information or the government website which only provides medication information and generic lists of doctors that parents have to go through and vet for themselves. Paediatricians and other healthcare professionals have access to various documents with information on ADHD and lists of specialists in WA but this information is currently passed around manually.

Preparation for appointments with health providers

It is often frustrating for parents who need to advocate and negotiate the common barriers when trying to assist their child. The wait time in some areas to see a paediatrician has increased to over 6 months for assistance with a developmental, behaviour or learning difficulty. Hence it is best to be well prepared in order that you and your child get the best information at the time of your appointment

Monitoring treatment

ADHD patients need to keep track of various biometric data for continuous treatment. Often patients don't do this or when they do it is manual and inconsistent.

Solution

A central and informed source of ADHD information for parents and patients and a solution that also provides assistance for organisation for of various tasks like preparation for doctors appointments, health monitoring, and a reminder system for taking/renewing prescription medication.

Existing solutions

• Pdf documents
• Google
• A few apps that set reminders for medication but not tailored for ADHD patients needs

+ Helping Young People to Find Suitable Counseling Services

The aim of this project is to assist young people (including at-risk/vulnerable communities such as LGBTIQ, Indigenous Australian, people with disabilities, etc) to make informed decision about which counselling services they should seek.

The problem

The Telethon Kids Institute recently surveyed 12 - 25 year olds in their opinions about health services. One of the key issues identified in the survey is the lack of freely available and reliable sources for counselling services in Perth. Other barriers to accessing these services include:

• Knowledge of services (with the exception of GPs and school health services)
• Discomfort in disclosing health concerns
• Accessibility of services including transport, characteristics of services such as opening hours
• Flexibility of appointments

For example, if a young person is in need of a counselling service specifically for LBGT issues, they would enter this search term in Google, resulting in literally thousands of hits.



How would they know which website is reliable?"Reliable" meaning; most up to date information including a correct website address, opening hours and contact information.

• Is there any reviews on the centre?
• How would they know if it's in close proximity to where they are located?

There is a lot of additional research they would need to conduct to eventually choose a service.

The Challenge

To come up with a solution that helps young people find and make informed decisions about which counselling services they should seek so that they are not discouraged from getting help.

The solution should:

• Be easy for users to identify services in their area/specific area
• Be easy for users to rate and write reviews of services.
• Leave users with certainty about which counselling service they should book.
• Allow for the improvement of the data set, for example, if a particular service is not listed, information relating to the service should be able to be added easily.

Ideas

The solution could be an app aimed at risk communities in particular young adults that allows the user to rate services.

+ Helping the Deaf Enjoy a Cinema Experience

To create an app for a tablet or phone that plays back a subtitle file for a given movie in an easily digestible manner.

The Problem

There are existing subtitles for just about every film, and when watching a movie at home on a computer, having them displayed is a trivial matter. However most cinemas out there do not show the subtitles unless it is a foreign film. In order for the deaf to fully enjoy the cinema experience they currently need to go to special venues that provide a device called the CaptiView upon request. This solution has several downside, as described in the existing solutions section.

This application would replace the need to use specially designed equipment, and would allow the deaf to enjoy the movie experience wherever they are, enabling a greater level of inclusion in mainstream society.

The Challenge

The challenge here is to come up with a solution that aligns the subtitle text with the audio.

The solution could:

• be capable of nudging the subtitle text forwards/backwards
• have a sync capability that uses some form of voice recognition
• be extended to:
  • involve an AR system that overlays the subtitle text on top of a video feed
  • allow other languages to be loaded up so the film can be enjoyed by non-deaf people of other
  nationalities

Existing Solutions

Currently, some cinemas offer a device called the "CaptiView"

There are several issues associated with having this specialized piece of equipment

• Availability
CaptiView sessions only at specific times & location and even then, the dongles are not always available
• Reliability
The dongles don't always work, and occasionally have flat batteries
• Discretion
These devices are fairly obvious and draw attention which is often undesirable

There are also a few apps out there that run subtitles, however they do not have an automatic syncing function and rely on the user manually pressing play when they hear the first sound. This is not incredibly useful to a profoundly deaf user.

+ Data Collection for Respiratory Health Survey and Results Analysis

To collect and analyse information relating to respiratory health from patients and healthy individuals so that the causes to common problems such as asthma and allergies can be better understood.

The Problem

Asthma, idiopathic pulmonary fibrosis (IPF) and COPD are complex respiratory diseases that involve many different genetic and environmental factors. Through our clinical trials unit we collect samples from patients and simultaneously gather a large amount of medical information and family history from patients, but feel we are currently not utilising this as much as we could. As geneticists primarily working in the laboratory, identifying patterns and trends in this clinical data would help us to enhance our experimental design and data interpretation in order to better understand these complex diseases.

The Challenge

To build an application targeted at patients that collects information relating to respiratory problems and create a means for researchers to navigate the results.

The information is currently gathered through two questionnaires that are manually filled out by the patient and then entered into a database by a researcher. This is very time-consuming and has the potential for introducing errors, and there is a need to have this process improved.

Due to comprehensive nature of the information collected, there is also a need to come up with ways of grouping the data and displaying trends so that the results can be more easily understood by researchers.

The solution should:

• Be easy to use for the users/patients
• Collect the data and populate relevant database tables
• Ensure patient anonymity by separating contact information and health data
• Allow additional questions to be easily added or modified

Analysis should:

• Make it easy for researchers to navigate and understand the results

Existing Solutions

There are already many existing applications out there that can build surveys, so these could be used as a starting point. The solution should focus on moving the collection process to an offline environment, and ensure that the format matches the required back end.

+ Helping GPs Make Choices about Treatment Options for Diabetes

To support GPs to make informed decisions about the many available treatment options for people with diabetes.

The Problem

Making treatment choices for people with diabetes is complicated. There are many factors to consider such as medication protocols, patient specific factors and Pharmaceutical Benefit Scheme (PBS) restrictions. Doctors are faced with the challenge of combining these factors in a short time frame. They may also be unfamiliar with the many new drugs available.

Optimal blood glucose control often depends on a doctor being ready and able to upscale treatment. The current situation is leading to treatment inertia for many patients. Good diabetes management reduces the risk of long term complications that inhibit a person's quality of life, productivity and life expectancy.

The Challenge

To come up with a solution that assists general practitioners to make timely and informed decisions on a patient's treatment plan.

The solution should:

• Take into consideration
  • Medication protocols
  • Patient specific factors
  • Pharmaceutical Benefits Scheme
  • Be very easy, and quick to use
• Leave users with certainty about which treatment options are available
• Leave users with all the information they need to take the next step

Ideas

The solution could be an app aimed at GPs

Existing Solutions

Doctors currently use combine a flow chart (from RACGP), information from specific drug companies, and online drug specific medication guides which includes the PBS information.

Having chosen a medication, the doctor then needs to check whether an authority code is required. There are two ways to obtain the information. The medication brochure which many doctors do not retain, or phoning the PBS while the patient is present which is time consuming and inefficient.

+ Collecting Experience Reports from Meningitis survivors

The purpose of this project is to create an application that enables patients recovering from meningitis to document their experiences with the disease which can then be compiled to assist researchers in understanding what support mechanisms are required for patients.

The Problem

The sequelae from infection with Neisseria meningitidis

Neisseria meningitidis is a bacterium that is an exclusively human pathogen and is a major cause of bacterial meningitis in children and young adults. It has a fatality rate of about 10% and in survivors causes serious impairment. Meningitis and sepsis can result in long term disabilities such as amputation, chronic kidney disease or epilepsy. However, other issues such as post traumatic stress syndrome and learning difficulties are less well known. At the moment researchers need more data from survivors of meningitis so that they can understand what support mechanisms are required for these patients.

Meningococcal disease is currently at very low levels in the population. Although meningococcal disease can be vaccinated against, the serogroup B vaccine is not free on the National immunization program. This is because the economic argument has not been clearly made regarding the burden on survivors and their families recovering from this disease. By gathering as much information as possible on the burden of this disease on Australian families, and collating it, the data could be used in future applications to introduce meningococcal vaccines to eradicate this disease in Australia.

The Challenge

A perfect solution would be an app that would allow patients to document their experiences in real time. The data would be collated in real time to a server and artificial intelligence would be used to score the responses into different categories from severe to less severe. This data would then be used to call in patients for remedial intervention particularly if it involves mental health issues or helps in making other appointments for specialist treatments regarding their disabilities.

For this event, an ideal solution would be to create an application that collects data on symptoms that Meningitis sufferers are experiencing and aggregates them in a machine and human readable format so that the data can be used in future applications. If that can be achieved then a larger version could be built later on and beta tested with pilot groups.

The solution should

• Be easy for the user to navigate and use. Keep in mind that users may be experiencing symptoms of the condition while using the solution.
• The data collected should be readable by researchers but also be usable by other applications.
• The solution should provide some context on the common symptoms so that users can better articulate their experiences.

Ideas

An app that lists common Meningitis symptoms and lets the user report on their experience. The app then takes this data and combines multiple entries into a csv file for further use.

Existing Solutions

Meningitis Now: An information app designed to spread awareness of the symptoms. Doesn’t allow the user to submit a report.

https://www.meningitisnow.org/fight-for-now/wtf-meningitis/identifying-disease/meningitis-move/

Additional information

Dr Kahler is Associate Professor, Deputy Director of the Marshall Center for Infectious Disease as the University of Western Australia. She is involved with two charities; the Meningitis Australia and Amanda Young Foundation who have a large contact base with survivors. She also is a fellow of the Telethon Kids Institute which is also conducting a retrospective data linkage project on sequelae in meningitis patients. They are interested in this app as a real time tool for future research trials.

Neisseria

Neisseria meningitidis

Dr Charlene Kahler

http://health.thewest.com.au/news/2867/robbies-struggle-puts-vaccine-in-the-spotlight

http://www.news.uwa.edu.au/marshall

Amanda Young Foundation

+ Help Researchers and Clinicians Identify Suitable Patient Samples for Research

There is a lack of ability to access critical clinical information for research purposes, because of sensitive patient information that needs to be protected. An app or a web-based form to help researchers and/or clinicians identify suitable clinical samples for research can solve this problem.

The Problem

Medical research often makes use of clinical samples taken from patients. We have had difficulty selecting clinical samples suitable for such research studies. In the current system, there are multiple people, such as researchers, clinicians and clinical nurses involved in the sample selection process. This increases the chances of missing out key clinical information, such as inflammatory diseases or infections at the sample identification or selection stage. The clinicians or clinical nurses may not be aware of all the relevant criteria for a research study, and as a researcher, I also have no direct access to the clinical information until the samples were de-identified, which happens subsequent to sample selection. This means that some key information may be lost between the sample identification and selection stage, which I may not be aware of. Therefore, I would like to propose an app or a form that can be used to filter out samples by particular fields that are suitable for a particular research purpose.

Solution

A standardized manner of entering clinical samples in the hospital database would be ideal. All the relevant information is stored in the same format and filtering out samples for research needs can be achieved easily using an app or a web-based form.

An interface that can gather the relevant clinical information from a database and a systematic exclusion protocol for the researcher allowing them to include/exclude samples suitable for the study will be useful.

Current Solution

Currently, I provide details of the clinical samples required for my study to a clinical nurse who then provides me a list of samples that fit the criteria.

+ Automating Systematic Reviews

I want to be able to auto-generate multi-disciplinary review summaries on any health topic that I specify. The reviews should include information about the study characteristics used (study type, sample size, etc.).

The Problem

Systematic reviews are considered the cornerstone of evidence-based medicine. Significant advancements in database technology and the refinement of reviewing protocols have greatly sped up the process of reviewing material. Several protocols exist (e.g., PRISMA, Cochrane Collaboration, Joanna Briggs Institute etc.), however, even though these protocols make the task largely algorithmic, the processes of data extraction and synthesis of journal article content continues to be done manually (and by multiple raters). Systematic reviews usually begin by identifying a body of literature through database searches. This literature is reduced down to a set of studies that meet strict 'inclusion criteria', including basic levels of quality. The number of included and excluded studies at each step, needs to be recorded. From this set, study characteristics and data pertinent to the question are extracted into a sheet for each article. These data may then be synthesized, or in some cases (meta-analysis), quantitative data may be re-analysed.

While systematic reviews are considered an evidentiary gold standard, they can be limited if one wants to assess the scope of a field or synthesise knowledge on a topic that crosses multiple fields. Syntheses of this type tend to be in the form of 'narrative reviews' which are not considered as rigorous as systematic reviews.

Solution

Ideally it would be nice to conduct informational censuses with a press of a button on a given health topic that synthesizes all information available online and produces interactive visualisations that can be explored as well as readable summaries. Information would be classified according to disciplinary perspective and study type that would also provide indications of study quality. It may be that a protocol needs to be established so that all journal articles produce a machine readable summary of findings that can be imported into a software to make this possible?

Create software that can reliably extract information on study characteristics for a sample of research articles of different types and return it in a table. Study characteristics would include: field of study; study type; population characteristics; sample size; type of intervention and intervention effectiveness (if applicable).

Current Solution

There is no tool as described that I am aware of that is in widespread use. There are bits and pieces of innovation that are relevant to this question, but none have been brought together into a cohesive and user-friendly product. What has been done before, that I am aware of, is as follows:

• Google and Google Scholar
• Endnote and Excel are currently the tools used by most researchers doing reviews - however, there are significant downsides to these softwares (e.g., Endnote doesn't allow you to save iterations of a database - a required part of systematic review reporting process; Excel can easily get unwieldy to manage)
• Systematic review software (e.g., ) exists but is only an interface for managing the review process and reporting requirements, it does not automate key tasks
• Articles on systematic review automation (e.g., http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4100748/ and http://www.cochranelibrary.com/editorial/10.1002/14651858.ED000091)
• Articles on priority ranking algorithms, for identifying article relevance (http://systematicreviewsjournal.biomedcentral.com/articles/10.1186/s13643-016-0263-z)
• Articles on the use of text mining for identifying relevant studies (https://systematicreviewsjournal.biomedcentral.com/articles/10.1186/2046-4053-4-5)
• Text analytics (e.g., Leximancer) exists for language processing but does not read from or to endnote, and has not been specifically adapted for the project described
• The journal Systematic Reviews has the most up to date discussions in this area

+ 16 Panel Multiview Graph Visualisation

The Problem

Allow users to view up to 16 graphs on the one page and on selection display a chosen graph larger than the rest. The aim is also to increase the speed of rendering multiple graphs while increasing the quality of the figures rendered.

Problem

Multi-omics datasets are the "Big data" of expression data. Because there is no current way of integrating these datasets, showing related data side by side is the only option. This is what the multiview was designed for.

The aim is to improve the user experience as it is time consuming to open multiple graphs and switch between open pages to draw correlations between data. This is particularly important when analysing multi-omics datasets as the data can't be displayed within the same graph. Users can analyse a specific reaction by displaying a range of graphs with multiple datasets comprising of different genes, metabolites etc and gauge the effect/changes on the reaction.

Solution

A client side JavaScript module which renders graphs as needed. These would ideally use a visulisation library such as d3 to give the user a more dynamic experience. It would be good to give the user the ability to save their graph and gene choices and download all, or individual graphs from the page.

An open source JavaScript solution could be developed by a team within 48 hours. Layout and general approach to the solution will be provided. The graphing tools are already written in JavaScript so it would essentially be building a dynamic wrapper to house multiple scripts.

Current Solutions

Currently there is a multi view option. The graphs are rendered server side. Each graph is the same size and a maximum of four graphs can be displayed at a given time.

+ Over-treatment in the Market for Dental Services

The Problem

In privatised health care markets, such as the market for dental care in Australia, doctor-patient relationships are characterised by an information asymmetry and a conflict of interest. The dentist diagnoses the patient's problem and often also provides the treatment. While both the dentist and the patient has an incentive for the problem to be correctly diagnosed and sufficiently treated, the dentist can make higher profits by providing an unnecessarily complicated treatment to the problem (over-treatment), doing excessive diagnostic tests (over-testing) or charging for an expensive treatment while actually providing a cheap but sufficient treatment (over-charging). Lacking the necessary expertise, patients are unable to recognise when they are over-treated or over-charged. And if the patient is fully insured, she may not even care much as her problem is resolved while the cost is shifted to the insurer.

The problem negatively affects the patients and the insurance providers due to increases in health costs. However, many dentists are also adversely affected as many patients avoid having regular check-ups due to their mistrust of the recommended treatments. Solving this problem will reduce health care costs, increase rewards for honest high quality care providers, and increase efficiency in the healthcare system.

Solution

Separating diagnosis from treatment would help solve the problem but is usually prohibitively costly. Making diagnostic tests/information easily accessible by alternative dentists would help but there are technical challenges as well as potential privacy issues.

Unfortunately, standard reputation/online review systems and competition (second opinions) alone don't work in these markets. An anonymous feedback system where patients can report their suspicions about being over-treated/over-charged/over-tested and solicit alternative diagnosis can assist the solution. A rare example of a solution to a similar credence goods market is one by Uber, which uses a combination of GPS (which helps identify over-treatment), standardised pricing and a feedback/reputation mechanism. A similar solution would need to gather information from patients and dentists, solicit automated or manual third party diagnosis and hence give more confidence and power to the patients in identifying whether they are over-treated/over-charged.

Current solutions

Emphasis on medical ethics and self-regulation by industry associations and complaints handling by The Office of the Health Ombudsman or National Health Practitioner Ombudsman and Privacy Commissioner are some of the current solutions. However, while these methods are somehow effective for under-treatment they don't adequately address over-treatment or overcharging.

+ Securely Sharing Medical Photos to Assist Patient Referrals

The Problem

Every day doctors around Queensland refer injured patients to Brisbane hospitals. Being able to securely share photos of injuries during this referral process greatly assists patient assessment and care. As a rule remote doctors have private mobile phones with cameras. However, Queensland Health doesn't have a web service to handle or receive these photos securely from personal (i.e. non-Queensland Health) mobile devices. Also lacking is the ability to add this information to patient's official records. The ability to securely receive, log and store medical photos sent from doctors' personal mobile devices would improve patient quality of care during emergency situations and subsequently.

In many cases one photo or video instead of long explanation can save time and probably the life of the patient.

Solution

There is a necessity to make the process of communication simpler, faster and more efficient e.g. start QH mobile application or visit QH website using private mobile phone, take and submit photos and videos, call to QH On-call staff and continue consultation on phone (as it is now) when QH On-call staff got pictures/videos reviewed.

It can be web or mobile application, which securely (e.g. via SSL or any other form of encryption) transfers pictures to QH server and allowing QH On-Call staff to review and annotate this photos and videos e.g. verify identity of the person who submitted it, identity of the patient as well as a summary of provided consultation.

Messengers e.g. WhatsApp (not limited to) can be used for client site and Open source API e.g. WhatsApp API for QH server site handling of encrypted pictures e.g. WhatsAPINet. Received data can be stored in Microsoft SQL Server 2016 Express

IT Security assurance and risk assessment will require to use this application in production.

Official process of image transfer to QH PACS requires access to QH network as well as access rights to PACS system, but not all involved staff has required access rights, knowledge of how to do it, and a time to proceed.

+Helping those with Neurodiversity to Navigate Built Environments.

The development of an app that may assist those with neurodiversity to navigate the built environment

The Problem

Neurodiversity not only includes individuals on the autism spectrum, but also individuals with acquired brain injury/traumatic brain injury, mental health concerns and other neurological conditions. As we age, our senses used to make sense of our world alter. In Australia, as we have an ageing population the ability to navigate built environments in the absence of taking into consideration altered sensory experiences will have a significant impact on a large part of the national population.

Solution

The development of an app that would enable individuals to navigate their route through built environments successfully.

As this is a complex problem I would envisage that what could be achieved in 48 hours may include discussions with relevant individuals to assist us to continue exploring what other elements would need to be completed by our research team to refine the design of such an app.

Current solutions

As far as we are aware, there is not currently a solution for this identified problem.

Partnered with: Queensland Centre for Intellectual and Developmental Disability

+ Real-time Tracking of Antibiotic Resistant Bacteria

The Problem

The health system is currently witnessing a dramatic increase in the number of antibiotic resistant bacterial infections. Drug resistant bacteria are projected to account for 10 million deaths globally per year by 2050. Identifying the presence of an antibiotic resistant bacteria, as well as to which antibiotics it is still susceptible currently takes more than 48 hours. However, this procedure is expensive and often inconclusive. Moreover, tracking the outbreak of drug resistant bacteria is challenging, and usually only occurs retrospectively on suspected cases collected during the outbreak. With the use of a new DNA sequencing technology and novel analysis pipeline, we aim to solve this problem and provide a real-time antibiotic resistant bacteria tracking solution. This system will rapidly alert clinicians to the presence of a drug-resistant infection, and how it relates to other infections observed in Qld and Australia, all while ensuring each clinician has the information needed to prescribe the safest, and most effective antibiotic for each patient.

By solving this problem, we will be able to help limit the spread of antibiotic resistant bacteria. Developing this tool will help clinicians prescribe the most effective antibiotic to best combat the bacteria infecting each patient. Storing information about each patient's infection and looking at the data from multiple patients will allow us to track an outbreak in real-time. This can identify hotspots of infection and provide healthcare services the information they need to minimise transmission and fight an outbreak in real-time.

Solution

The genome (DNA) of a bacteria encodes all the information necessary for it to evade antibiotics. Moreover, by comparing genomes from different bacterial infections it is possible to reconstruct relationships between bacteria and the chain of transmission. The Centre for Superbug Solutions (CSS) have been working with a miniature, portable nanopore sequencing technology (called MinION) which generates sequence data in real-time, and which can be streamed to the cloud.

Thus, we envision the ideal solution to be a web-database application for streaming analysis of real-time bacterial DNA sequence streamed from MinION sequencers in the clinic. This platform would allow the clinician to record clinically relevant information, including allergies and history of antibiotic usage, and would connect to the nanopore sequencing data-stream. It would update the clinician in real-time with a prediction of the strain and species of bacteria, responsible for a patient's illness, and would predict the best antibiotic treatment. This information can also show the relationship to other infections as well as a potential chain of transmission. By recording the patient response to treatment, the platform would update its model of drug resistance to continually improve future predictions. A global view of emerging outbreaks could be provided to hospitals as well as regional, state and national authorities.

+ Identifying Mental Health "filter bubbles"

Do mental health ’filter bubbles’ exist for information seekers from different backgrounds and, if so, what do these look like?

The Problem

Mental disorders, including substance use disorders, are the third highest contributor to the burden of disease in Australia. One in two Australians will display symptoms of mental illness at some point in their lifetime, and one in five will display symptoms over a 12 month period. In spite of this burden, mental illness often goes untreated or is inadequately treated.

Given the stigma of mental illness as well as cost constraints on health care, digital mental health care, ranging from educational websites to internet-delivered treatments, offers cost-effective solutions to ensuring population mental health. However, there are risks involved in encouraging direct-to-consumer health care online.

One of these risks is the risk of consumers receiving inadequate or harmful information and 'treatment' or being exploited by online purveyors of health products and services. Consumers without the skills and knowledge to appraise the information they see are at particular risk.

There are also systems-level SEO and service provision questions, not to mention philosophical questions, that surround mental health info-diversity. These questions could be informed by an understanding of filter bubbles as a kind of 'info-scape' that people judge and navigate over time.

Solution

A graphical user interface that displays the mental health filter bubbles for people of different backgrounds to identify whether, and for whom, differences exist. Filter bubbles would be defined as the frame of information that a user perceives based on a combination of platforms' algorithmic selections and personal choices. Thus, filter bubbles would be based on differences in key words used in searches, the information that is prominent in google searches and on social media channels, and the way that different consumers navigate to and through sites. Ideally, the display would classify mental health information according to type based on text analysis of content, provide a quality assessment of information, and identify salient differences in vernacular. Identifying the 'terrain' of these bubbles, in terms of how they are navigated and utilised over time would also be valuable.

This technology could be expanded to other health conditions, such as cancer and cardiovascular diseases, and be used to investigate (and compare) filter bubbles in other countries and languages.

A software could be developed to collect data on how people of different age, cultural identity, educational and economic status search out, navigate and appraise mental health information online via different platforms and search engines. An approach to analysing content of mental health websites, using natural language processing, that could be correlated with self-reported assessments of that site, along with user-engagement metrics (e.g., viewing time) could also be developed.

Current solutions

Appraising and visualising online content has been a prominent thread of recent research in e-mental health. However, most of this research is focussed on

(i) screening or diagnosis of mental illnesses based on naturally occurring text data or

(ii) appraising and improving the quality of mental health information that is available online.

Facebook has collaborated with headspace and beyondblue to target and assist people who may be in need of mental health care. More generally, Google has announced that it is implementing protocols to promote quality and relevant health information via its platform. CSIRO and BlackDog Institute have been engaged in analysing Twitter feeds to assess population emotions.There have been failed attempts reported in the UK, of charitable organisations attempting to monitor and assist people who are at risk of suicide via Twitter. Individual organisations collect and use web analytics with limited data made publicly available. Quality of websites and treatments available via google and app stores have been appraised using (manualised) systematic review methodologies.

There have been debates among internet activists and researchers regarding the existence and significance of 'filter-bubbles' and 'echo-chambers' in relation to information flows. Some studies have been conducted to test the nature and significance of these phenomenon, however, results appear equivocal and may be subject and platform dependent.

To my knowledge there is no research that has attempted to identify and visualise the existence of different mental health filter bubbles, or their terrain, for consumers with different backgrounds, taking into account the differences attributable to search terms and navigation decisions to or away from websites. Limited trial searches indicate that peak bodies may have been successful in their SEO, in terms of getting their sites to appear prominently in Google and Facebook, but it is unclear whether and how this prominence is perceived by consumers from different backgrounds. Variation in search terms may be an important determinant of results.

+ A more efficient database for cardiology patient management

To improve the efficiency of managing pateint admission and follow-up for certain cardiac surgery procedures.

The Problem

The demand for cardiac surgery to be performed via the groin is soaring. These procedures give older patients unfit for open heart surgery a second chance. Despite the rapid increase in number of procedures performed hospitals currently use an excel spreadsheet to track patient data and determine procedural eligibility. A specialized database system to support decision making is desperately lacking.

The rate of growth in the number of procedures performed is outstripping the ability to manage patients using an excel spreadsheet. It has become a workflow and efficiency problem. The number of people across the cardiology, radiology and surgical departments with access to the excel spreadsheet is growing and data integrity is an issue. It is also impacting on arranging timely follow up for these patients.

The ideal solution is a .NET application with a SQL server backend. The database would be accessible by multiple users from multiple computers within a healthcare network.

Using Microsoft's .NET framework, the most common platform used in the healthcare system, building a database within 48 hours is a realistic target. If time allows a simple text messaging service that asks the patient to rate their breathlessness in the weeks following their procedure with the responses incorporated into the database would allow for improved patient follow up.

As described above the current solution uses an excel spreadsheet to manage data.

+ 3D annotation of Terrascale Biomedical Imaging Data in a 2D Web

Web based visualisation and annotation of 3D imaging data is difficult, the TissueStack project (www.tissuestack.org) is an open source attempt at this via HTML5 canvas elements akin to 2D mapping websites. We now want to delve into annotation of datasets and need to develop new methods of annotation beyond dropped pins and 2D polygons.

The Problem

Neuroimaging datasets are large, too large for desktop computers, the answer is the web. We can currently view multi GB and TB imaging datasets via a HTML5 + JavaScript interface based around tiling in a Canvas Element www.tissuestack.org. A detailed overview of the interface and how it works can be seen here. Full description of the challenge can be found on our GitHub repository.

Challenge

Our (international) users have now become very familiar with this style of interface but want more. The #1 request we have is for annotation. Even the ability to add simple elements on top of the image would be a great outcome. Landmarks, simple polygonal lines and text.

How this data is stored is a vexed question, currently all data on a TissueStack server is open, so the tracings and annotations would also be open. For now everything being open is a reasonable approach.

Thought needs to be given as to how tracings and annotations work beyond a 2D plane as is more typical in traditional web interfaces. A common approach it to treat a point as a sphere and show smaller circles in adjacent slices. Failing that a user will only be able to find annotations by chance when scrolling through the volume. User cues should feature heavily here but the mechanism for doing this requires a group of coders with some novel visual design ideas for web interfaces.

Stretch Goal

Collaborative annotation across multiple instances of TissueStack. We have diverse groups of researchers working from multiple sites in Australia (and internationally). Ideally each volume should be traced by the combined knowledge and skill of multiple experts. Being able to collaboratively annotate images would be a fantastic boon to science ala google docs and other collaborative tools.

+ Public Health Spending Optimization

Optima Nutrition is used by countries to inform their public health spending on nutritional interventions. The optimisation routine samples many different spending scenarios. However, currently only the ‘best’ scenario is presented as a recommendation for spending. It would be useful for the user to be able to explore the parameter space more interactively, as there may be other spending allocations which fit better with their practical or political constraints.

The Problem

Background

Undernutrition causes 45% of child deaths per year and leads to 159 million stunted children. Early investment in nutritional interventions can build human capital and boost shared prosperity. The United Nation’s Sustainable Development Goals (SDGs) specify a 40% reduction in the number of stunted children by 2030. The World Bank estimates that this target will not be met without nutrition-specific interventions.

Methods

Optima Nutrition is a tool to inform policy decisions in child nutrition. It consists of an epidemiological model which is integrated with economic and financial analysis frameworks and a formal mathematical optimisation routine. The model is a partial cohort model written in Python which tracks the health of children from birth to age five. Cost and coverage information about nutritional interventions are used to predict health outcomes such as death and stunting. A unique feature of Optima Nutrition is the optimisation function that is used to calculate the optimal allocation of resources to different program areas to minimise adverse outcomes. Additionally, it has capability to optimise funds across geographical regions, as well as being used for scenario forecasting.

+ Medical Reminders

Help us help the nearly half a million Australians aged 85 years and over maintain good health by building a better diary, calendar, social interaction and medication reminder solution.

The Problem

Ageing well is an urgent and pressing issue across Australian society. Over the past two decades, the number of persons aged 85 years and over increased by 148%, compared with a total population growth of 32.1% over the same period.

Cognitive decline, and in particular short term memory loss, occurs for most people in their late 80s and 90s, and even earlier for those who do not enjoy good physical heath. Moreover, the majority of older people retain the ability to care for themselves, even with the onset of dementia.

However, there is a major problem challenging the ability of older Australians to care for themselves. Many older people are not able to maintain good health because:

• remembering,
• attending medical appointments,
• taking medications and
• keeping socially interactive

are more challenging.

As people begin to lose cognitive ability, many will use a calendar or diary to help remember and organise their outings and appointments. The diaries of people with dementia and memory problems often become cluttered with notes as they attempt to keep all the details that they may forget. Over time, the ability to plan diminishes and diary and calendar notes disappear, or are written by others.

Help us help the nearly half a million Australians aged 85 years and over maintain good health by building a better diary, calendar, social interaction and medication reminder solution.

Possible Solutions

A tablet that allows multiple agencies and family members remote access

A Display at the clients home that has clear information on what will be happening today

Ability for the care team to access the tablet and avoid booking appointments that will clash with other events

An easy alert system with a reminders for medications

Ability to dial the client and have a ‘facetime’ conversation

A clear contact list display with a phone call made at the touch of the screen

+ Antibiotic Usage Data: Assisting clinicians to improve prescribing

Improving the use of antimicrobials is a critical element in the national and international strategies to address the problem of antimicrobial resistance. Australia has one of the best programs for collation and analysis of antimicrobial usage data from 73 hospitals across the nation. Currently antimicrobial usage data is submitted bi-monthly to the National Antimicrobial Utilisation Surveillance Program (NAUSP). The NAUSP process analyses the data for each hospital and then issues a report that enable the institutions to benchmark their usage against similar comparator hospitals. This information provides great insights into patterns of use across Australia but is not optimal for targeting which units with in the hospital are likely to benefit most from intervention. A project is currently underway at a Perth metropolitan hospital to bring the analysis of the data in-house. This has the advantage of not just being able to examine the information in more detail but it has been able to increase the number of antibiotics that are included, examine whether the agents are given intravenously or orally and include emergency department prescribing that is not included in the existing national program.

The Problem

The goal of this challenge is to develop the data analytical tools to eliminate the current manually intensive process and produce smart reports of monthly antibiotic usage for feedback to the hospital clinical teams as well as create an interactive interface to enable dynamic queries. We intend this tool to help clinicians take greater ownership of the information about their prescribing practice with the aim of improving antibiotic use for better patient outcomes and reducing collateral effects from antibiotic misuse. If the project proves that this strategy is effective, it could be converted to a generic platform to be used more widely nationally and internationally.

Hospital monthly antimicrobial dispensing and hospital occupancy data is to be converted to the internationally recognised drug usage rate of defined daily dose (DDD) per 1000 occupied bed days for analysis over time, by agent, route of administration and department. The output will be smart visually attractive reports and an interactive platform for analysis on demand.

Methods

• supply by tablet and dosage will be converted to supply in grams
• the correct World Health Organisation defined daily dose (DDD) will be assigned to each agent
• the standardised dose will by calculated by dividing the supplied grams by the DDD
• the standardised dose will be divided by 1000 bed days to allow benchmarking of doses of agents used by department

+ An Integrated Quality control system (IQ system)

We believe there is an untapped potential for building a quality control systemthat is integrated with our existing genomics analysis capabilities (pipeline) and variant curation capability (database) – an Integrated Quality control system (IQ system). This will allow the pipeline and database to continually learn to discriminate between good and bad quality. Sometimes this will help us discover unusual patterns in the data that turn out to be important to a patient.

The problem

The Melbourne Genomics Health Alliance brings together ten leading healthcare, research and academic organisations to tackle the challenges of putting genomic medicine into practice – for the benefit of patients. The Alliance’s vision is for Victoria to be a world leader in the use of clinical genomics in healthcare. But to do this, we need to be smart about quality control (QC). That’s what has inspired the concept of the IQ system, a database and analysis capability to allow efficient quality control for genomic testing. We want to build an intelligent system that can use all the patient genomic data accumulating in Victoria to greatly improve genomic testing and boost performance and efficiency. We believe there is an untapped potential for building a quality control system that is integrated with our existing genomics analysis capabilities (pipeline) and variant curation capability (database) – an Integrated Quality control system (IQ system). This will allow the pipeline and database to continually learn to discriminate between good and bad quality. Sometimes this will help us discover unusual patterns in the data that turn out to be important to a patient. Secondary data that is output from our pipeline are currently stored in flat files rather than in a database structure to allow analysis of trends and summaries of historical performance. The long term vision of the IQ system is that it will

(i) improve clinical interpretation of patient genomic variant data,

(ii) meet quality control needs for monitoring the pipeline,

(iii) evaluate pipeline updates and

(iv) help establish quality assurance programs for genomic testing.

Melbourne Genomics will provide anonymised pipeline data generated from genomic testing of 300 patients for HealthHack 2015, and we hope this will ensure that the foundations of an IQ system can be designed and built (beginning with items i and ii).

The complexity and sheer volume of the data means that the IQ system needs to be fast and sophisticated. The IQ system we envisage, will be an exciting innovation that will greatly improve genomic testing and bring better results to patients.

+ A Simple Online Clinical Notes Program

Chiasma Health has recently been approached by a community psychology practice with a brief to create a medical record system for small scale deployment. We are keen on creating a robust system that meets data security and retention laws but is also easily implemented and extensible.

The problem

Electronic health records are utilised almost universally by General Practitioners in Australia, however, the penetration of electronic record keeping systems into small psychologist practices has been poor. There have been a number of factors identified as barriers to adoption of electonic records, these include:

• expense
• difficulty of installation and use
• concerns around legal requirements
• practitioner acceptance

Despite the benefits to practitioners and a government policy supportive of electronic records, the regulatory framework has been slow to adapt to electronic health record keeping, meaning that these records are covered by the Privacy Act 1988. In responding to the needs of small psychology and counselling services, we have identified a significant need for a secure, easily deployed and accessible health record system.

Aims

To create a web-based medical record with the following features (in order of priority):

• Secure authentication and sharing of records between local practitioners
• Backup and audit trail for data
• Export templates for letter writing and referrals
• A patient record:

– Demographic Information

– Patient Summary (including medical conditions and medications), for example see Case Histroy Summary from Monash Uni. Ferris and Giorlando Page 2 of 3

– Journalled entries for each visit

– Non-blocking and non-locking of documents to allow flexible use by multiple users and user types.

+ Mapping social relationships, alcohol and drug use and other risky behaviours.

Our current paper-based tool uses sticky notes, stickers and markers to enable people to map out their social group memberships and patterns of alcohol and drug use. This visual method can simplify an often complex reality enabling people to better understand how their substance use impacts their social lives and vice versa. We are looking to develop a prototype of this tool, as an app for mobile devices or interactive website, focusing on intuitive interaction and useful data visualisation.

The problem

Social relationships can have a powerful effect on our health and wellbeing although their effects are typically complex: they can either heal or harm, (or they can do both or neither). For instance research has shown that adding one friend to your social network confers similar benefits to stopping smoking exceeding other well-established risk factors including obesity and physical inactivity. Using this ‘social cure’ paradigm we have a developed novel social relationship mapping intervention we are currently piloting at a range of alcohol and other drug (AOD) rehabs across Australia which we’d like to develop into a tool for mobile devices. Our current paper-based tool uses sticky notes, stickers and markers to enable people to map out their social group memberships and patterns of alcohol and drug use. This visual method can simplify an often complex reality enabling people to better understand how their substance use impacts their social lives and vice versa. We are looking to develop a prototype of this tool, as an app for mobile devices or interactive website, focusing on intuitive interaction and useful data visualisation. While this tool is currently being used in an AOD setting, this intervention could be adapted to apply to any range of risky behaviours for use by clinicians and consumers alike. We are looking to collaborate with data visualisation experts and developers who like us are excited by this opportunity to create a novel open-source intervention that would allow people to visualise their social relationships and provide a platform to empower them to better understand their lives and improve their health and wellbeing.

The problem

An interactive digital visual platform that allows people to audit, visualise and map their social relationships and substance use. While primarily this will be a tool to assist people to account for the way social and environmental contexts affect their lives (and in particular their AOD use), this tool could also provide a platform for data collection and help answer research questions about the link between social connectedness, problematic substance use and health and wellbeing. We are a group of researchers in psychology, criminology, alcohol and drug studies and neuroscience working on a national research project about how social connectedness impacts recovery prospects among people with alcohol and drug problems. We’d like to prototype a data visualisation methodology / platform to assist people to audit / visualise their social relationships and alcohol and other drug (AOD) use which could dovetail with our current work. While this is an ambitious goal, at its heart this is a simple but potentially powerful idea and we are hopeful that this work will result in the development of an app for mobile devices and/or a website for use by clinicians and consumers alike. The tool we have in mind would log the user’s:

• Social relationships and group membership
• Personal AOD use
• AOD use among social network

These data would be used to create a visual representation (a map) of the user’s social world with an overlay showing the location, intensity and type of AOD use (or other risky behaviours).

In short, we’re seeking to create a novel visualisation / mapping tool to allow app users to better account for and understand the complex relationship between their social lives and their AOD use, which could be used as an intervention to empower them to make more informed choices about their lives.

We eagerly look forward to the opportunity to pursue this new e-health and research platform.In particular, we are really excited by the prospect developing an open-source tool that would provide a platform for people to affect positive change in their lives and increase health and wellbeing while simultaneously leading to improvements in understanding about the link between social connectedness, problematic substance use and health and wellbeing.

+ Wheelchair controller

Clients with very significant physical and/or cognitive disabilities can have major difficulties using a conventional electric wheelchair interface. In particular, joystick interfaces can require significant gross arm movement and very fine control when navigating confined spaces. Curtin Mechatronics in collaboration with disability services company Rocky Bay have developed an sensor assisted “smart wheelchair” concept, where a client can be assisted by a semi-autonomous wheelchair control system to ensure that desired trajectories can be achieved while also ensuring that collisions are avoided. The HealthHack challenge is to develop a smartphone or small tablet based user interface, which enables control instructions to be delivered to the wheelchair, while also displaying mapping and navigation information provided by the wheelchair control system.

The problem

The wheelchair is equipped with three Hokuyo scanning laser sensors, which provide a rich representation of the environment around the wheelchair. A mapping subsystem integrates the data from the lasers into a robot-centred local map, which shows apparent surrounding objects in relation to the wheelchair’s position. A navigation subsystem will receive position instructions from the user interface, plot a safe path within the visible map, and execute local wheelchair speed and turnrate instructions through an open wheelchair control subsystem which has been developed previously.

The system is being developed using the Robot Operating System (ROS), an open source robotics middleware package which provides a broad range of device drivers, generic datatypes and communication interfaces, and libraries of basic functions for mapping, navigation and visualisation. The user interface should be designed to subscribe to a map and a path, and should visualise these live as the wheelchair moves through an environment. The user interface should enable a user to specify desired goals within the map, either by touching the map or by interacting with a virtual joystick control located at the bottom of the app window. The selected goal will be communicated to the ROS server through a published odometry topic.

The user interface will communicate with a laptop or other PC running Ubuntu and the ROS server. A wired interface is preferred, but we can also experiment with Blutooth or Wifi connectivity. Android tablet or phone app is preferred to other options but not essential.

References: http://wiki.ros.org/

+ Visualising risky teen drinking

We are investigating risky alcohol use amongst Australian teenagers. These young people are overrepresented in harm statistics such as hospitalisations, while being underrepresented in national health surveys that examine the Australian population as a whole. We are building a system to investigate their drinking context and alcohol-related issues such as harms experienced, in order to minimise alcohol-related harms. This survey/interview system is intended to be launched on an annual basis to provide trend data on patterns of use.

The problem

We have collected pilot data with more than 900 teenagers from four Australian cities describing their last drinking session when they drank more than seven standard drinks in a single sitting. Our dataset includes a large number of variables about the context of this drinking session and its outcomes. Variables include: the quantity of alcohol consumed, whether safety strategies were engaged in, types of harms that may have occurred following consumption, other drug use, motivations for choosing selected beverage types, location of drinking, demographics etc.

We seek to combine some of these diverse of variables within a single ‘last drinking session’ visualization to share with other researchers, members of the general public, and policy makers. There is even the potential to present personalized visualisations such as ‘what your night out was like compared to others your age’ to the young research participant themselves.

HealthHackers are also welcome to make suggestions for other potential improvements that we can use in our upcoming data collection phase (e.g. improving methods of generating anonymous self-generated identification codes used to link participant data over time).

Dataset: SPSS file that can be converted into Excel spreadsheet or similar. Over 900 cases over 800 variables. More information on the dataset here: http://onlinelibrary.wiley.com/doi/10.1111/1753-6405.12326/abstract

+ Unsteadiness Detector

As a Senior Citizen, I want a falls prevention solution that is proactive, personalized and tailored to my individual needs. I want a device or technology that would alert my carer(s) to my unsteadiness even before I fall. This would help me get the right treatment, at the right time and prevent me from falling

+ Molecular Dynamics Database (MDDB)

A database to store, share and visualise molecular dynamic simulation data.

The problem

Project background - Aims and outcomes

Proteins are large molecules responsible for proper functioning of the cell – they act as transporters, catalysts, signaling relays, supporting scaffold etc, and their malfunctioning may lead to a disease development or cell death. From that point of view, it is very important to understand the underlying relationship between the protein structure and its function and how this can be affected by changing the external conditions or binding o small molecules. The latter is especially important in drug design.

Molecular dynamics simulations can be considered as a “computational microscope” to study how atoms and molecules move in real time using physical models and supercomputers. In that process, the large data files (trajectories) are generated, which contain the information about position of each atom at time t and sometimes the velocities are included, too. MD simulations can be used to study various problems, from biological systems (proteins, nucleic acids, membranes) to materials (nanotubes, solid state matter).

However, the biggest limitation of MD at the moment is accessing the relevant timescales at which proteins operate (microseconds to seconds), while at the moment the routinely available scales are hundreds of nanoseconds. To generate a trajectory for a very large system (~500 000 atoms) in duration of 200 ns would take weeks on a supercomputer (or GPU) or longer, depending on the hardware and software. That means that most of the MD simulations are computationally expensive and time consuming, which makes the resulting trajectories valuable assets in protein studies. However, in most of the cases the obtained trajectories are not made available to public for further use after the results have been published. That implies that each system of interest is repeatedly simulated every time when there is a need for it because the existing data is not available, which is not the most efficient way of using human and computational resources.

Having a platform that would support sharing of the data generated:

• increased transparency of publications relying on MD simulations;
• long-term storage potential (this would be a very useful feature!);
• new perspectives on old data: alternative analytical approaches and data mining possibilities;
• availability of ensembles of conformations instead of a single-structure representations given in PDB;
• improved structural models for ligand docking etc.

The existence of this vast amount of data organised in one place would definitely lead towards better understanding of protein structure and function, which is crucial for understanding the origin of diseases on molecular level, cellular regulatory functions and specific drug design.

Technical challenges include:

Choice of database: storing, indexing and serving large binary files (typically > 4GB). Solutions envisioned could for instance include SQL database for indexable fields (simulation parameters, etc.), filesystem for large binary files while ensuring data integrity. Network issues relative to uploading large files over HTTP(S). How to handle limited bandwidth and long upload times, etc.

User interface (input):

User registration and metadata information required to allow submission.

Interactive session/Download:

• Visualisation options: It would be useful to have a quick look how the system trajectory looks like before you download the entire file. Possible to sample only a few snapshots from the entire trajectory.
• Atom selection and analysis: Another good thing to have would be a set of tools to do some basic and common data analysis and plot the results. MDAnalysis toolkit (python library) already offers the tool able to deal with this kind of tasks.
• Download counter: Only registered users can download materials and downloading should be logged.
• 3D structure alignment: There are some tools already provided at: http://www.ncbi.nlm.nih.gov/Structure/MMDB/mmdb.shtml, http://nar.oxfordjournals.org/content/42/D1/D297.long

Curation and user rating system.

Users can report suspicious files, but they have to justify their flagging. If multiple users flag the same entry, it should be removed from the database. Each upload could have a discussion board where users can leave a comment, ask a question either directly the uploader of other users, ask for an advice and similar.

+ YuGene Graph

Realtime graph simplification/scaling.

The problem

We have a graph that shows all the samples in the database for a particular gene. In one example, the gene Gapdh has around 57,000 elements and a download of 5.5MB data. We would like to find a way to make this graph appear much faster than it currently is in terms of finding a smart way to reduce the size of the data downloaded or to make it more like google maps where a rougher but still overall accurate graph can be achieved. We would look at using flot but we also are interested in d3.js. This would also be using jQuery although other technologies would be OK.

+ Clever multi-view graphs

Stemformatics is a portal to a series of public experiments describing mouse and human stem cells and how they differentiate to become mature cells, tissues and organs.

The problem

We would like to develop a visualisation technique/tool to display four graphs that we have for four different datasets in a way that you can both see the overview and so that you can see the graphs big enough to see the nuances.

A screenshot of the current multiview graph implementation at a high level (multiview.png).

Technology wise, we would look at using d3.js graphs and jQuery.

+ Real-time pain management tool

A holistic pain management app involving a diagnostic component, a management component and an information component.

The problem

Pain is one of the great public health issues facing societies around the world, afflicting an estimated 20% of the global population with millions of new diagnoses every year (Goldberg and McGee, 2011). Pain also comes with a great social, mental, and financial burden with patients often having strained relationships with friends and family, suffering from depression (1 in 5; Gureje et al 1998), and struggling to work efficiently. Despite these grave effects, Pain remains an under-assessed, and under-treated condition. Thus, pain treatment and management is a significant health problem requiring urgent tackling.

Our app, is a holistic pain management application that aims at bringing patients and their families, researchers and clinicians together to address the problem collectively. Our team has sought inputs and continues to seek inputs from clinicians, scientists and most importantly patients to address their unmet needs. It will feature three core components, a diagnostic component that evaluates pain, a management component that tracks pain and the usage of therapies or medication and an information component that will provide patients with access to information about pain. At HealthHack, we are aiming to make an early version of this app that we hope one day will form an indispensible part of any pain management routine.

+ Cloud Computing Carpentry (CCC)

Develop lessons to teach researchers how to use cloud and/or high performance computing environments.

The problem

Context Automation and rapid technology advances are driving research in the biomedical sciences. This is rapidly becoming digital as more research analysis needs to be performed using high computational power. It is then that most researchers are required to be users of high performance computing (HPC) resources or/and cloud environments. As part of the volunteer non-for-profit organisation Software Carpentry (https://software-carpentry.org) we have been exposed to the transformative change that access to digital skills provides to research scientists in different fields.

Problem

Cloud services currently are within reach, but scientists may be scared to try the benefits of the cloud infrastructure for lack of basic knowledge on how to use it. On one side commercial clouds such as Amazon Web Services, (https://aws.amazon.com/) are available to everyone. Such resources have already extensive self-training courses: https://aws.amazon.com/training/intro_series/. On the other side, Australia’s research community has access to the National eResearch Collaboration Tools and Resources project (NeCTaR) (http://nectar.org.au/) which provides an online infrastructure to collaborate, access and store data. This is a valuable resource but it needs to be exposed to scientists in a reachable fashion. Australia’s researchers will be often pointed to the ‘Nectar’ research cloud or to a commercial cloud offering. However there is a gap on providing basic skills for them to be user of these resources. Health science researchers will have basic questions on how to use, install software, maintain, keep secure and scale cloud solutions.

Challenge

To develop a set of interactive lessons to teach researchers basic system/infrastructure administration skills. The focus is to be able to use and maintaining resources of their projects using research clouds such as Nectar or commercial clouds such as Amazon Web Services. We suggest to have a structure similar to the Software Carpentry lessons:

http://software-carpentry.org/lessons.html

The new cloud computing module should cover the following points, but it is not limited to:

Introduction, and motivation for using cloud services.

Defining concepts:

• image, instance, allocation, users, permissions, etc
• Setting up a local resource for accessing the cloud
• Manage permissions of files and folders, to address security
• Moving data to and from the research cloud
• Perform regular file archival and purge
• Managing data storage and file systems (when storage is allocated)
• Installing applications (where and how)
• Configuring services
• Repair and recover from hardware or software failures

+ Allergy and consumption related monitoring

Monitoring personal consumption of food, drinks and other consumer goods for allergies and consumption-related diseases.

The problem

Most of the food, drink or cosmetic product we consume, named as fast-moving consumer goods, are rigorously controlled and each marketed product can be traced via unique barcodes. Using portable scanning technology such as a mobile app, we digitalize our daily consumption similarly to MyFitnessPal.

Monitoring our personalised consumption can help diagnose and prevent health issues such as allergies that can not be identified by traditional methods. Coupled to a web or platform-based diary relating symptoms (e.g. ache, rash, inflammation), the long-term monitoring would enable to identify (via pattern recognition) the ingredient(s) associated to your symptoms. Such technology could also prevent associated risks to cardiovascular diseases, obesity or diabetes.

+ Identify and visualise the cost of treating cancer

A tool to identify and visualise differences in the cost of treating cancer patients.

The problem

Cancer patients are all different. Some cases are more complex or more challenging to treat than others. As a result, the time and effort it takes to consult with patients and plan and deliver their care varies from one case to another. For instance, it often takes fewer overall resources to successfully treat a young, early stage breast cancer patient than an older, frail patient with more advanced lung cancer.

State-based funding for hospital services generally does not reflect the complexity of care; in a sense, treating every patient is considered to require the same amount of resourcing. Although the ebbs and flows of this approach may even out across a system, the impact in particular hospitals will vary. Some hospitals provide a greater volume of complex care than others; in these situations, the average funding value may be insufficient.

Our current IT systems tend to focus on the information needed for safe clinical care and funding reporting obligations. This means gathering data to influence changes to funding approaches can be arduous (e.g. using paper-based collection processes). We figure there are much smarter ways of tackling this kind of challenge.

Create an app that will help us collect, analyse and visualise real-time information about each individual patient interaction with a cancer care team (doctors, nurses, allied health and admin professionals) so we can better understand and respond to patient variability. Optimally, the solution would work in iOS, Android and Windows environments. The app needs to span multiple settings and users and deal with situations where patients are present, as well as activity that occurs when patients are absent but is essential to proper care (e.g. planning for radiotherapy treatment). Ideally, the app could identify patients using barcode recognition (attached to current medical records) and allow complete extraction so that information collected can be linked with data in other systems for holistic analysis.

The app is proposed to be deployed in four large Queensland hospitals. The data captured will underpin advocacy for better funding allocations and assist local service planning. A key current objective is to make sure hospitals are funded fairly to treat the different groups of cancer patients they care for. This will increase overall efficiency in the cancer system and encourage innovation and quality. Ideally, the app could be ‘retrofitted’ to support similar processes in other clinical areas (e.g. diabetes care) as needed.

+ Clinical exam rotation rosters

Ever tried organising 3,000+ doctors around Australia on the same day? We do, the long way.

The problem

Please help us reduce the need for 40+ Excel spread sheets and 300+ tabs by designing an application that optimises our exam rostering system. We are a not-for-profit organisation that assesses unsupervised Australian General Practice at the workforce level, among other things. Perhaps you’ve heard of the Royal Australian College of General Practitioners (RACGP)? That’s us.

Our clinical exam rostering system is crucial to co-ordinate the candidates and examiners in the correct location and to populate the marking sheets with demographics. Each group of about 16-20 candidates (called a rotation) needs about 30-35 examiners. We anticipate examining over 1,000 candidates in future. The State Faculty administration team use rotation roster templates to populate candidate and examiner information manually. Sounds simple but it has many complexities! One is that rotation rosters need to be finalized in a short period of time (over a few days).

Examiners are stationary while candidates move from station to station (generally room to room). The clinical station criteria needs to match examiner demographics. As well as examiner gender, age, indigenous status, location and availability constraints, examiners may have conflict of interest with candidates or other examiners. Candidates may have conflict of interest with each other. Other complexities include:

• 6 rotation roster templates available – not enough flexibility for faculties.
• One rotation roster template = 10 to 12 hours to create and test.
• One rotation roster = 16 to 20+ candidates = 30 to 35 examiners = 1+ hour of admin work.
• Stations vary in time: long case (2 per rotation), short case (12 per rotation) or rest (varies).
• Rotation rosters have various start times and locations.

I imagine a world where we could input the constraints of the exam and output a first draft of every master timetable for every location with no clashes. Several staff could further tweak, check and finalize. When the rosters are final, we want to output (print) each individual roster for every person and every station. It would be great if the program could also tell us “You need X number of examiners with Y and Z attributes in locations ABC” so we can target/ recruit specific examiners. Our nice to have is implementation in R.

+ Exploring hospitalisation data

Produce a tool to visualise and interrogate inpatient hospitalisations for child health admissions.

The problem

Overview

This dataset (859,000 records) is a record of inpatient hospitalisations, a data type that is commonly analysed in health research. The International Statistical Classification of Diseases and Related Health Problems (ICD) coding system is used in this dataset to code both the diagnosis and any procedures performed. In this sample, each record has up to 10 different diagnoses. The ICD version used in this dataset is ICD-10, for the purposes of Health Hack a sufficient reference (look up) for codes is available on Wikipedia:

https://en.wikipedia.org/wiki/ICD-10

(Further information is available at http://www.icd10data.com/ICD10CM/Codes).

Example Themes

A typical theme would be identifying all the records for a particular disease/disorder. This may be looking up a single code as the primary diagnosis (diagcode_1) or it could be looking for all codes within a band (I40.00 - I44.99) across all diagnoses. You may then be interested in how much time passed between their first hospitalisation and their next hospitalisation, whether how long they were in hospital for varied based on the their co-morbid conditions etc. You may simply want to extract the first occurrence of a disease for each individual.

+ Visualising the complex home-care services ecosystem

When someone needs home-based services, the choices are overwhelming. What services do I need? What am I entitled to? How much will it cost? The choice has been aimed at health professionals, but this solution needs to be used by the people who wil consume the services. The aim of this challenge is to communicate the complexity of making choices and accessing the available services that meet consumers' multi-dimensional health and wellbeing needs, and ultimately empower them to better manage their own health and independent living.

The problem

Consumers of home care services including the aged, terminally ill and people with a disability currently receive services through multiple referral points and may be eligible for government allowances for home-based health services. They can use this allowance or pay out of pocket for home care services they choose to meet their needs. The challenge is to provide the simplest means to finding the right services that meet their needs and providing access to them.

An example to illustrate:

A person has a major operation planned for next week and will be at home in bed for several weeks afterward. What services are available for their condition? What are they entitled to? Where should they look? How can they acquire the services to ensure their peace of mind before surgery?

Most people in this situation feel overwhelmed, and the current solutions that exist aren’t really better than using google - you have to know what you’re looking for. Health care professionals, and the government assessment gateways provide advice on the services that exist and what options available. Unfortunately this advice is often incomplete and consumers will end up with services that may not be ideal or may not be aware of the more appropriate solutions that are available.

The Challenge

To communicate the complexity of this situation visually based on the underlying data and processes, to build understanding and inform policy and technology in this area.

Datasets: https://www.dss.gov.au/our-responsibilities/ageing-and-aged-care/tools-and-resources/aged-care-service-list

+ Improving diagnosis to better manage antimicrobial resistance

Antimicrobial resistance is a national and international priority area demanding action. To date most efforts in the human health arena have been targeting raising awareness of the problem and measuring the rates of antimicrobial resistance and antimicrobial use. This project is part of a strategy to focus on improving the diagnosis of infection as one of the key underlying reasons for factors for antimicrobial misuse. The goal is to use data within the pathology laboratory information system to identify where microbiology testing is being used optimally to diagnose infection and assess if there is a correlation between improving diagnosis and better antibiotic use.

The problem

In order to achieve this we are looking to develop a simple yet meaningful method of visualising all relevant data in a format that is easy for clinicians and decision makers within the organisations to interpret and guide future policy development. We wish to develop a dashboard which has the capability to show:

• Microbiology test ordering patterns for blood, urine, wound swab and sputum culture samples
• Compliance with recommendations for test ordering
• Proportion of samples with suboptimal collection
• Proportion of positive to negative results
• Benchmark best practice across hospital departments
• Efficiency of reporting the laboratory results by turnaround time and time of day results is reported

Dataset: Excel or CSV file. Pathology data for all urine, wound swab, sputum and blood microbiology culture and susceptibility testing request orders at a deidentified Perth Metropolitan Hospital for 2014. Includes description tables for data comments and coding. Confidential and de-identified patient, hospital and/or requesting clinician details. Sourced from PathWest Laboratory Informations Management System- ULTRA.

+ Navigating the home-care service ecosystem

When someone needs home-based services, the choices are overwhelming. What services do I need? What am I entitled to? How much will it cost? The choice has been aimed at health professionals, but this solution needs to be used by the people who wil consume the services. The aim of this challenge is to support consumers of home care services in making choices about the available services that meet their multi-dimensional health and wellbeing needs, make it easy to access these services and ultimately empower them to better manage their own health and independent living.

The problem

Consumers of home care services including the aged, terminally ill and people with a disability currently receive services through multiple referral points and may be eligible for government allowances for home-based health services. They can use this allowance or pay out of pocket for home care services they choose to meet their needs. The challenge is to provide the simplest means to finding the right services that meet their needs and providing access to them.

An example to illustrate:

A person has a major operation planned for next week and will be at home in bed for several weeks afterward. What services are available for their condition? What are they entitled to? Where should they look? How can they acquire the services to ensure their peace of mind before surgery?

Most people in this situation feel overwhelmed, and the current solutions that exist aren’t really better than using google - you have to know what you’re looking for. Health care professionals, and the government assessment gateways provide advice on the services that exist and what options available. Unfortunately this advice is often incomplete and consumers will end up with services that may not be ideal or may not be aware of the more appropriate solutions that are available.

The Challenge

To come up with a solution that helps consumer of home care services to find, choose and access the home care services that meet their multi-dimensional health and wellbeing needs.

The solution should consider that users:

• May be older & unfamiliar with technology
• Have eyesight issues
• May be feeling emotionally distressed
• May have physical disability
• Users have variable and multi-dimensional needs
• Be very easy to use - simplify finding, selection and accessing of services
• Provide users with certainty about which health and wellbeing services they should acquire
• Leave users with all the information they need to take the next step
• Make them feel empowered about managing their own health by improving consumer experience, flexibility, choice and control
• Let them feel independent in their own homes and regarding their health

Ideas

The solution could be aimed at consumers of home care or it could be to visualise the problem and the different factors that contribute to making these decisions (see the related visualisation challenge). It may also consider the development of a self-learning algorithmic matching function or app that can match common home care consumer search terms with services that can solved their home care needs.

Datasets:

https://www.dss.gov.au/our-responsibilities/ageing-and-aged-care/tools-and-resources/aged-care-service-list

+ Scan your skin to manage your sun exposure

We live in a very sunny environment, so being able to take control of personal sun exposure is important. Too much and you could put yourself at risk of skin cancer…too little and you risk a vitamin D deficiency. However, the amount of sun exposure that’s optimal for you depends on your skin type. This app will help you get the get the most out of our beautiful climate and the benefits of sunshine, without putting your health at risk.

The problem

We're thinking of an app, including the following essential elements:

• A mobile app that is able to be used on most smartphone devices (iOS, Android).
• People will take a photo of their skin (following specific instructions) with the app.
• The app will determine the person’s skin type (using image analysis, not yet developed) against the Fitzpatrick Classification
• The app can store skin types for each member of the family (e.g. parents and kids).
• The app can determine the current or predicted UV index for the user's location.
• The app will provide simple sun exposure management (management guidelines not yet developed) for a given person (skin), location and time.

There are numerous extensions possible.

+ Visualization of sequential study design properties – saving lives and costs in first-in-human clinical trials

Sequential study designs are an important approach in pioneer (first-in-human) drug safety and effectiveness research. A sequential trial typically includes a number of stages. Each stage is defined by a pre-specified number of individuals who is exposed to a new medication or a next higher dose-level of the medication under investigation. Depending on the number of individuals who experience a certain critical (or beneficial) event at the end of a stage, the trial is either stopped for futility or proceeds with the next stage (inclusion of further individuals and re-evaluation of pre-defined stopping criteria). If a substance (or dose) successfully passes all predefined stages of the sequential trial, the medication (or dose) is claimed to be acceptably safe and/or efficient. The project targets the development of a tool (online GUI / app) which allows the visualisation of the stochastic properties of an arbitrary sequential study which will help investigators to design and plan first-in-human trials under simultaneous consideration of patient safety and overall cost-effectiveness of the investigation.

The problem

Key variables (to be altered by the investigator when designing a sequential study): The key parameters (variable components) of a such a sequential study are: i) the number of stages j=1,…,k (k ≥ 2) ii) the number of individuals included in each stage nj (note that this number can vary with j) iii) the critical number of events cj not to be exceeded at stage j in order to proceed to the next stage j+1 (note that this number can also vary with j). iv) the number of ‘circles’, i.e. how many times is the sequential study repeated / the maximal number of dose-levels to be investigated v) a priori knowledge about a plausible range of possible event rates for a substance / dose.

Key outcomes to be produced / visualized by the developed tool: The following particular question arise at the design and planning phase of a sequential trial:

1. Given a certain minimum true event rate (i.e. toxicity between 0% and 100%) of a drug / dose, how likely is it that the trial is stopped early (at a certain stage j)?
2. Given a certain maximum true event rate (i.e. toxicity between 0% and 100%) of a drug / dose, how likely is it that the trial is successful (successfully passing all stages)?
3. Given the solution to question 1, what is the expected [as well as minimum and maximum] number of patients that will be exposed to a potential harmful drug / dose until the trial is stopped?
4. Given the solution to question 2, what is the expected [as well as minimum and maximum] number of patients that will be exposed to a potential harmful drug / dose for a successful trial?
5. Given a substance or dose passes all k stages of the trial, how likely is it that the true underlying event rate does not exceed a certain threshold (posterior probability distribution for the true event rate)?

Computational requirements:

The answers to the key questions can be derived via stochastic data simulation. The only software feature required is the ability to sample random variables from a Binomial distribution and the ability to store and process matrices that contain all observed (within a simulation) outcomes. This allows, for example, the calculation of the proportion of early stops / no early trial stops among all simulation runs, given a certain parameter distribution. The key idea is that we can interpret relative frequencies (achieved via a large number of repeated simulations) as probabilities.

+ Serious games teaching emotion recognition for Autism Spectrum Disorders, with biofeedback and Unity game engine

The aim of this project is develop an innovative biofeedback interface able to extract and deliver electroencephalogram (EEG) and eye tracking (ET) biomarkers, and linking this interface with a computer game to improve the emotion recognition skills of children with Autism Spectrum Disorders (ASD).

The problem

While it is clear that children with Autism Spectrum Disorders (ASD) experience significant impairment in emotion recognition skills the most effective intervention modality to improve the emotion recognition skills of these children remains unclear. Children with ASD have demonstrated improvements in emotion recognition skills following computer based training. Computer based intervention provides a structured and predictable environment for learning, making them particularly suitable for children with ASD who often experience distress when confronted with unpredictability. While adolescents with ASD may be internally motivated to engage in social communication and interaction, younger children with ASD are less socially motivated, and need more support in initiating and sustaining attention during emotion focused training. We aim to develop a prototype for what’s known as a “serious game platform” – games that educate and immerse users in realistic scenarios - to teach emotion recognition skills to children with high functioning Autism. In incorporating novel biofeedback measures to stimulate the behavioural and cognitive level of the player thereby guiding neuroplasticity to improve performance. The gaming platform will incorporate accessible and low cost electroencephalogram (EEG) and eye tracker (ET) devices which will allow the therapy to be delivered at the home environment. The EEG and ET devices and APIs will be provided at the HealthHack event. The chosen software development platform for game design is Unity, which provides key toolsets to realise the objectives of the serious game. A perfect team for this project should comprise of participants from the following domain areas; game design, programming, animators, human-computer interface, and psychology.

+ connectED: Creating opportunities to teach and learn in the health care setting

An app that a practitioner (teacher) can use to log potential teaching opportunities in real time, so learners can find out about specific learning opportunities in that healthcare setting.

The problem

Many of the daily duties of every heath professional working in a hospital or healthcare setting is an opportunity for a student practitioner to learn. For example, a radiographer interpreting an x-ray from an interesting patient brought into emergency, a nurse using a novel approach to dressing a wound or a speech therapist assessing speech in someone who has had a stroke all represent moments to teach and to learn about a technique, a procedure or a condition.

The issue is that there is no real time way of communicating these opportunities between practitioners (teachers) and learners within the hospital and as such many important opportunities to teach and to learn are lost.

The challenge

My idea is to develop an app that a practitioner (teacher) can use to log potential teaching opportunities in real time and for learners to be able to search for or have alerts set up for specific learning opportunities in that hospital or healthcare setting.

In this way student practitioners can tap into many of the clinical events that occur every day and learn from them equally practitioners can build their teaching portfolios whilst undertaking daily duties.

How it might work

At minimum this app could be used within one professional body, let’s say medicine, alerting medical students to some of the learning opportunities going on within the hospital. At best however, this tool could be used to connect across disciplines creating teaching and learning opportunities in a way that reflects the multidisciplinary approach to care. For example medical students on an orthopaedic rotation observing a physiotherapist perform a knee assessment or nursing students attending a dietician’s clinic.

Things to consider

• Needs to be very quick and easy to use “on the run”
• Confidentiality would have to be the biggest issue, the app would have to be closed to anyone outside the health service and so a specific means of logging into the app perhaps using only network email addresses such as is used with Yammer app would need to be considered
• A means of the teacher labeling learner opportunities so that they could be searched for and found easily
• A search function that allows learners to search for specific areas of interest
• An alert system would be useful so that opportunities in areas the learners was particularly interested could “pop up”
• A means for the learner to book themselves into a learning opportunity
• A means for alerting learners if an opportunity is cancelled or changed
• A means to restrict the learning opportunity to only a specified number of learners after which it’s closed.
• A means to record of the opportunities logged by a practitioner (teacher) and taken up by a learner, this record could form and important log or professional development activities for both. This is already a requirement of professional registration

+ A dashboard for building clinical laboratory networks

Build an online dashboard to help Human Variome Project Country Node Coordinators build local data-sharing networks with clinical laboratories.

The problem

The Human Variome Project is an international organisation working to build capacity in responsible genomics around the word and improve human health through facilitating the free and open sharing of genomic knowledge as part of routine clinical and research practice. We work to establish local centres of excellence in data sharing in individual countries (HVP Country Nodes) and work with them to build local data sharing networks with clinical laboratories and map genomic testing activity in-country.

We want to provide better tools for our HVP Country Node members to help them find clinical laboratories and work with them to encourage free and open data sharing of putative disease-causing genomic variants.

One such tool, delivered through a web-based application, would be a space where individuals working on behalf of an HVP Country Node could sign in and enter information about their country, including, but not limited to:

• what clinical laboratories are currently operating in their country
• what genetic diagnostic tests these labs offer
• the genes and diseases covered by these tests,
• and the contact details of the people working at these laboratories that control the ability to share the results of these tests.

Currently there are two online services that provide similar functionality: GeneTests and OrphaNet. Both of these services collect data from laboratories that volunteer information in order to advertise their services, so are incomplete; they also do not cover every country. We would like to be able to import the data from these services, to pre-populate our tool with existing data for the relevant countries. The tool would need to support bi-directional data flow from each service, to enable us to contribute new data to these services and for them to send updated data to us. We are currently negotiating these arrangements.

The smallest thing that would be useful to us is a web app that lets people that we assign credentials to log in an enter the information outlined above. Beyond this, we'd eventually like to be able to do the following:

• Use common user credentials with the members section on our website (Joomla & Community Builder)
• Users are able to delegate access to other users to spread the load of data entry
• Non-logged in users can suggest updates to the information; suggestions must be approved by a logged in user
• View contact details of HVP Consortium members in each country - data drawn from our member database (Joomla & Community Builder)
• Send EDM to HVP Consortium members in their country
• Provide a simple space on our website for each HVP Country Node to provide their own content or redirect to their own website

+ Searchable funnel plots for performance comparisons

When comparing the performance of health care providers, a typical approach is to plot “performance indicators” for a group of providers on a funnel plot. I’d like a tool which constructs a funnel plot for a set of performance indicators, and provides some interactive capability, such as high-lighting the performance indicators of particular subgroups of providers.

The problem

In order to maintain high standards of care, it is necessary to compare the performance of health care providers, such as intensive care units, surgeons, or dialysis centres, to determine if there are any with unusually poor (or good) performance. Performance indicators for each provider are estimated, along with some measure of the variability of these performance indicators (such as the standard error of the estimated performance indicators). Funnel plots are often used to identify those providers with unusual performance. The funnel plot is essentially a scatter plot, with the performance indicator of each provider plotted on the y-axis, against the measure of variability on the x-axis. If a provider has usual performance, its performance indicator is expected to lie close to some “null” value. Standard formulas are used to compute statistical confidence intervals around this null value, within which performance indicators corresponding to usual performance are expected to lie. These confidence intervals form “funnels” around the null value, hence the name. I’d like a tool that constructs a funnel plot for a given set of performance indicators and can then be explored. For example, the ability to highlight the performance indicators of specific sub-groups of providers, such as those within particular states or territories of Australia, would allow patterns within the performance of providers to be explored.

+ Trees to display changing treatments

There are many situations in which patients with particular illnesses switch between different treatments over the course of their illness. A specific example is in patients with kidney failure, who may elect to undergo different forms of kidney dialysis. I’d like a tool which uses a dataset to construct a tree diagram displaying the number of patients on each treatment type at a sequence of time points.

The problem

In many situations, patients may switch between a number of treatment types, die or stop treatment at any time over the course of their illness. Knowing how and when patients change treatments over the course of their illness can provide the insight that is required to compare different treatments. A tool allowing appropriate visualisation of this information is thus necessary.

I’d like a tool that provides a visual summary of how treatments change over time for a group of patients, where each patient may choose between 2 to around 5 treatment options. A tool that constructs a tree diagram for treatments from a given dataset, displaying the number of patients on each treatment option at a sequence of time points (e.g. at treatment start, day 50, day 100, etc), would be useful. A simple example of this kind of diagram is given in the figure. Since there are many situations in which patients may change treatment type at any point, the time points should be able to be specified when using the tool (e.g. looking at treatments in use at treatment start, day 50, day 100, etc, or at treatment start, day 10, day 20, etc.). It would also be useful to display the number of patients on each treatment type who have died or stopped treatment between each time point. The ability to select certain sub-groups of patients (e.g. those with a given characteristic such as male/female, or in a particular age range) and construct the treatment tree for these patients would also be useful.

+ Creating a framework to build interactive web experiences around scientific illustration & animation

Tagging DNA - mislabelling the cancer genome (epigenetic and methylation in cancer), video files rendered from after effects,layers, etc. potential to use Maya files, at least during the hack - HTML elements on top of video.

The problem

The Dream Some kind of tool that the researcher can build on to create educational materials out of web video - a way for the user to have a video be automatically stopped and annotated, allowing people to 'explore' within sections of video and learn about annotated areas. Also hidden functionality that allows tracking of how users interact with the tools.

How this help us? Education for a broad audience: schools, scientific peers, patients etc.

I would like to have at least four main features in this system:

1. Click/touch object and popup box appears. Popup box should have capacity to contain links, videos, text and images.
2. Navigation bar / menu on side or bottom of main video so you can select areas of interest.
3. Scrubbable video the user can easily control (user friendly menu or navigation system).
4. Click on aspect of an image and have the rest of the image greyed out (I can create the images to overlay for this) and have annotations appear to give more information on the object.

Ideally I would love a system that can be modular - ie I can build resources like these for future movies I make and animations. It should be able to be run from a web page and ideally within the Garvan website which is plone (not essential).

+ Feedback mechanism for young diabetes patients to promote healthy balance of activity and diet

To design a mobile app with a fun and friendly user interface suited for young patients with Diabetes Type 1 to help them manage their condition through the use of notifications (or emergency messages) based on data from glucose/activty/carb intake readings.

The problem

We would like to build a tool to help young adults with diabetes Type I to manage their condition so as to help them live an independent life. The solution is tracking their activity through wearables such as a FitBit which can communicate with a smartphone application or integrated with other carbohydrate app's intake (such as MyFitnessPal) and glucose readings.

The Data

Daily/Weekly carbohydrate intake and activity tracking data towards a balanced intelligent integration.

literature review

device or dummy data

+ Best way to display fracture risk for the general public and for doctors

To design an intuitive user interface that will allow people to understand and care about fracture risks.

The problem

The existing bone fracture risk calculator does not have a very easy interface for users as the input terms were not well understood by doctors (why there are 2 T-scores) and how the data was calculated into fracture risks is all behind the scene, and what kind of treatment should follow a 37% risk, for example.

The calculator should be designed for 2 kinds of audience: Doctors - give them recommendation General public - allow them to understand the risk, how it’s derived, the consequences and treatment

There is a webservice that returns fracture risks: https://k9.gimr.garvan.org.au/ws/FractureRiskCalculator/risk.xml - returns results as xml

https://k9.gimr.garvan.org.au/ws/FractureRiskCalculator/risk.json - returns results as json

Without parameters (as above) it returns documentation including example calls.

+ Mining tool for discovering commonalities in gene expression data of independent experiments.

Creating a tool that mines public databases for gene expression data by keywords of experimental conditions, stores results of differential expressed genes and allows discovering common patterns between independent experiments.

The problem

Hack public databases that store results for gene expression data with a defined search term such as “exercise” or “breast cancer” and produce a toptable of differential expressed genes for each mined dataset. The desired outcome is a list of genes including their respective pathways or their direct regulators (data that is stored in further public databases) and information on commonalities and similar patterns between the mined, independent datasets. Outcome format could be a simple list/table e.g. the top five genes affected in all experiments or the top five pathways affected and/or visualization of these pathways including the discovered genes and their interactors and whether they are up- or downregulated.

The Data

Microarray raw data in public databases such as ncbi, gene/protein interaction database and pathway database such as kegg or reactome

How this helps us

Producing genomic or proteomic data via microarray, RNA-seq or proteomic experiments is expensive and when researchers make the results of their data analysis available it is biased toward their interests. However, they often store the raw data of their experiments in public databases. This wealth of data could be the key to substantial discoveries in other fields, and any tooling that allows it to be mined without further expense would be invaluable to patients and researchers.

+ Automated tuning with control samples

Coming up with an algorithm to improve the BWA tool (http://bio-bwa.sourceforge.net/bwa.shtml) for sequence alignment

The problem

An accurate quantification of variability in bioinformatics data analysis is absolutely critical. To control for these sources of variability, we have developed a set of synthetic standards, known as sequins. The sequins emulate a diverse range of genomic features commonly encountered in bioinformatics. By comparing experimental data with the sequins, we are able derive a range of performance metrics, such as the sensitivity and specificity.

We will need an algorithm to use the sequins for improving existing bioinformatics tools. The algorithm will need to optimize the input parameters, with the aim to improve the performance of the tools. In this problem, we will focus on sequence alignment with the BWA tool.

The Challenge

The input space is highly dimensional, some of the parameters that we will need to optimize:

• mismatch penalty
• gap penalty
• clipping penalty
• number of threads
• penalty for an unpaired read pair
• substitution matrix
• maximum number of gaps
• maximum edit distance
• ......

It is not sufficient to tune each parameter individually as the joint-effects must be accounted. A typical BWA alignment could take hours to complete, it is easy to see a brute force solution would not be feasible.

A straightforward machine learning approach would also be insufficient, because we don't want to train a model once. We would like to do it in real-time.

The sequential probability ratio test is a possibility. But what stopping rule to use? How to apply it to sequence alignment? Is there a way to test for significance? What about bayesian? What about deep learning?

Can you come up with an efficient approach?

+ SenseS - A sensor for stress

"It’s like a pedometer, but for stress."

The problem

SenseS is a self-monitoring tool for people who perform in very complex, isolated and challenging environments (e.g. frontline health workers, construction workers etc) or a chronic disease where stress is a big risk to their health. Unlike (and in addition to) existing wearable sensors, our product improves your ability to manage stress in real-time. The problem that we’re working on is how to identify stress in different people and present the information in a meaningful way.

The dream / desired future state A mobile app / responsive site that frontline health workers can use to see and manage their stress accurately and in real-time.

The status

• We have collected annotated raw data collecting data using a medical grade sensor - Shimmer3 - which measures Galvanic Skin Response (skin conductivity) and Optical Pulse (heart rate).
• We have developed low-fi prototypes for the front-end app (see below)
• We have a draft architecture for SenseS

The hack

• Option 1: Develop a front-end visualisation for presenting the data to help users understand and manage stress (see above for low-fi prototype)
• Option 2: Set up the architecture for SenseS - pushing data live to the cloud with some form of security / privacy protocol
• Option 3: Develop algorithm for calibrating and identifying stress from Galvanic Skin Response and Optical Pulse data collected using the Shimmer

+ Arthritis Monitoring App

Design an Arthritis monitoring application for patients and/or doctors.

The problem

Arthritis affects 3.85 million Australians (15 million in the UK, 50 million in the US!). It costs the Australian economy $24 billion each year. It is painful, debilitating and can lead to permanent disability.

Yet arthritis receives little media coverage and even less dollars. Despite annual growth in sufferers, government funding is minuscule and shrinking. Treatment is mainly limited to surgery and prescription medicine, although research shows more holistic approaches considerably reduce symptoms.

Arthritis is not a disease limited to ‘old people’. Over half the people with arthritis – almost two million – are of working age. The cost to the community is huge, and it’s one of the most common causes of work absenteeism. Arthritis sufferers are also 30% more likely to suffer from depression.

Despite arthritis being a National Health Priority Area, and the huge, rising cost to the community, arthritis funding has declined. The National Health and Medical Research Council (NHMRC) reports funding was at its highest between 2008 and 2010 with over $26 million and has dropped to $21 million and $23 million respectively in 2013/2014. While funding to most other National Health Priority Areas has increased, arthritis has seen the highest decline.

There is little awareness about how bad the situation is and very little support for arthritis sufferers on how to manage their condition. An ageing and increasingly overweight population only exacerbates the problem.

The challenge

Although there currently is no cure for arthritis, patients can improve significantly by taking their medications on a regular basis, following a healthy diet, maintaining a healthy weight, regularly taking part in light to moderate exercise and taking anti-inflammatory supplements such as fish oil.

Managing a complex treatment plan and the resulting lifestyle changes can be challenging if not impossible for patients. On the other hand, it's challenging for doctors to provide the capacity, time or skills to deliver a holistic approach and monitor all aspects of a patient's wellbeing.

Here is where a technology solution could come in.

Can you create an online tool (app?) that allows patients to monitor their condition on a daily basis, tracks symptoms such as pain levels, mobility, fatigue as well as diet, mood, weight, exercise, medications etc.?

The app would need to have motivational components (alerts, goals, community, videos) and should be easy to use (e.g. consider users with arthritis in their hands or users who are not very tech-savvy). Medication reminders and script reminders should be a part as well (direct connection with doctor/pharmacy?).

Users should be able to create data records they can share with their doctors & pharmacists - even better, doctors have a corresponding website, where they can see data records and monitor their patients! Patient data could be collected anonymously to create the world’s largest arthritis study? Doctors could even submit survey questions.

The app can also feed patients with healthy recipes, tips and tricks to manage their condition and individual challenges on a daily basis. Any other ideas building on the above will be appreciated. (Online consultations with dietitians, doctors etc.? Wearable integration?)

While not directly related to research, this is an opportunity to help millions of sufferers across the globe manage a complex and painful disease and provide relief. The challenge is to develop a tool that is easy to use and creates results which serve both patient and doctor at the same time.

+ Explore your Genome and Epigenome in an Immersive Virtual Space

To build an “Engine” of sorts for an educational biological journey, with some kind of input in Unity or WebGL.

The problem

To create a prototype 3D environment depicting the inside of the nucleus of a cell that can be adapted for the Oculus Rift. Ideally I would like to create a 3D virtual world that has structurally accurate DNA renderable in real time. It should also have the capacity to add protein structures and directed animation in the future. Initial problem is just to get realistic DNA structures into an ‘engine’ that can be driven by the user. Current limitations include the balance between renderable animated structures and computer / graphics power to render this in real time.

The Data

I have several versions of DNA in Maya made from nurbs, particles and also driven from the new plugin dsDNA kit from Clarafi - molecular Maya. These are not all directly exportable via FBX to Unity due to having multiple meshes, complex animations or being particle systems. I need to get a lot of DNA into the engine so it needs to be driven in a way that’s super efficient. Ideally I would like to be able to get PDB files into Unity or WebGL that can then be rendered with a surface shader rather than the traditional atom appearance.

There are some existing tools for this type of thing, none are perfect: https://www.glmol.com/ https://biasmv.github.io/pv/ http://threejs.org/examples/webgl_loader_pdb.html http://portal.nersc.gov/gateway_demos/webgl/ http://www.inka3d.com/documentation.html

Here are some papers/projects that deal with the problem: http://genome3d.org/ http://pubs.rsc.org/en/content/articlepdf/2014/FD/C3FD00152K http://spie.org/Publications/Proceedings/Paper/10.1117/12.2007594 (Employing WebGL to develop interactive stereoscopic 3D content for use in biomedical research Johnson et al 2013)

These WebGL chrome experiments with particles makes me think one solution might be to convert the DNA PDB to a particle cloud and animate that with three.js https://www.chromeexperiments.com/particles?page=6&sort=newest

+ Genome Browser loader

A data upload tool that guides the user through the process of installing either a new reference genome, or genome annotation file.

The problem

Climate change is posing challenges to the persistance and adaptive potential of species around the world. In Australia, tiny vinegar flies (Drosophila) have provided us with heaps of ecological and genetic information about how species respond to a changing climate. As part of a bigger project to try and understand climate adaptation at the genomic level, we've taken advantage of recent advances in DNA sequencing technology to sequence the genomes of 23 Australian Drosophila species.

We and our interstate collaborators need to browse and search these genomes easily so we set up a local instance of the University of California (Santa Cruz)'s Genome Browser software. (Example genome)

The GB software is monolithic, somewhat finicky, and is used by researchers worldwide. It does something sufficiently complex that is hard to rewrite. In short, it's a problematic but important open-source software package. Uploading new data to display in the browser requires command-line skills that some researchers don't have, and so that this resource can continue to be used by everyone in the lab, we'd love some help creating a nice front end.

We are envisaging a data upload tool that guides the user through the process of installing either:

a) a new reference genome from a different Drosophila species

The reference genome itself is a ~150MB text file in fasta format.

This step also requires some metadata input, either in two small text files or ideally to be entered through the GUI. or

b) (optional) a genome annotation file to display on an existing reference genome.

This is a ~2MB text file in BED format, which contains information about the locations of genes, microRNAs, transcription factor binding sites or various other features of interest on the reference genome.

This tool will integrate with existing scripts that automate database creation and other steps in adding data to the genome browser, give feedback on success/failure, and display progress for the user. Some nice extras could include emailing an admin when new data is uploaded; validation, collection of feedback and stats; eye candy!

Key skills: (some combination of...)

• linux sysadmin on command line (ubuntu 14.04.3)
• bash
• web technologies (html, css, js, php, python, UI, UX - we are agnostic re tools used to complete the job)
• git

+ An online pedigree drawing tool that can be linked to patient data

My proposition for the HealthHack is to build a prototype of a tool which will allow to draw pedigree charts and export the drawn information in a standard file format. The tool should be easy to use and also easy to integrate into existing frameworks. There are a number of existing pedigree drawing open source projects but none of them offer the functionality described above. Maybe we can hack some of those into doing what we need.

The problem

Tracking clinical features through families and linking them to genetic variants is a frequent task in clinical genetics and genetics research. This is frequently done through pedigree drawings (family trees). Unfortunately, due to the lack of easy to use pedigree drawing programs, this is usually done manually on a piece of paper. This makes pedigree drawings difficult to handle and also precludes any computational analysis. For example, you cannot click on an individual in a paper-drawn pedigree to open the associated database record.

My proposition for the HealthHack is to build a prototype of a tool which will allow to draw pedigree charts and export the drawn information in a standard file format. The tool should be easy to use and also easy to integrate into existing frameworks. There are a number of existing pedigree drawing open source projects but none of them offer the functionality described above. Maybe we can hack some of those into doing what we need.

Update: The Phenotips Project has a nice pedigree drawing tool, but its not freestanding. The Phenotips Team has been planning to make it freestanding, there is already an open issue on their Github page. I have contacted the Phenotips Team and they would be pleased if we could help them with this task. They will participate remotely and answer questions.

+ Automated way of comparing sequencing results

Our lab needs an automated way of comparing sequencing results to a template, identifying the point mutations and amino acid substitutions, and presenting the analyses in an appropriate table format.

The problem

After performing flow cytometry based experiments that sort single spleen cells from mice responding to an infection, we PCR amplify the DNA from a single cell’s B Cell Antigen Receptor using a nested PCR approach, and obtain sequencing results covering the region of interest, which is 113 amino acids long.

We would love to have a tool or application that allows us to fully automate our analyses, which currently costs us several days of researcher time, which we imagine could be reduced to several minutes.

The manual analysis currently involves the following steps:

1) Compare sequencing results (.seq file) to wild-type template

2) Identify where the mutations lie

3) Check the validity and quality of the read by studying the chromatogram trace (.ab1 file)

4) If the quality is sufficient, trim the sequence result such that only the region of interest remains, and translate the nucleic acids into amino acids.

5) In an Excel spreadsheet (with the help of some basic macros), we copy the amino acid sequence from ApE or DNA Strider into Excel, convert text to columns (to get 1 letter in 1 cell) and then copy /paste the 113 amino acid sequence into a separate Excel spreadsheet.

6) Using some basic formulae, these sequences are placed into a table that only shows the amino acid substitions and not those amino acids that match the original template (germline sequence). Please see attached example table.

7) This table is then converted to pdf format to allow manipulation of colours, line weights etc in Adobe Illustrator, ready for manuscript submission or conference presentations.

In an Ideal World…

We would have the ability to:

1) Upload/ attach our sequencing results

2) Have the software tool do the rest, including the ‘human’ part of checking the quality of the read by studying the chromatogram trace (.ab1 file).

3) Perhaps even apply this to the nucleic acids and not just the amino acids, although the amino acid sequences are the main priority.

What this achieves, how it would help us, and challenges

A tool such as this would save us a huge amount of researcher time! We would generate approximately 100 sequences per treatment group, and on average we’d have 6 groups per experiment.

We do individual sequences one by one, and this becomes incredibly tiring (especially for the eyes) which makes human-error highly likely. We have been fortunate to have had some help on this problem before, but sadly we’ve only managed to get some of the way. The main challenge has been identifying a high quality sequencing read from a poor-quality read.

+ Building a Platform to Aggregate and Visualize Genome Sequencing Metrics

We’d like to be able to integrate, visualise and interpret the information that the sequencing pipelines generate in real-time, to substantially improve data quality & prevent poor data being used to guide clinical decision making.

The problem

We find ourselves in the era of personalised medicine, in part due to amazing advances in DNA sequencing technology. Using this technology, we aim to improve patient healthcare, through improved diagnosis, and tailoring treatments to match the unique genetic makeup of a patient. The problem is that the analytical pipelines to go from raw sequencing data to medically actionable information are complex, and consist of dozens of software components and databases. We’d like to be able to integrate, visualise and interpret the information that the sequencing pipelines generate in real-time, to substantially improve data quality & prevent poor data being used to guide clinical decision making.

The dream

We would love to develop a genome pipeline visualisation platform, which tracked and monitored NGS analysis pipelines in real-time, presenting the key metrics at each stage of the process to the analysts, via informative visualisations. This would aggregate and compare data over time, so that deviations can be easily identified. It would cross reference the metrics and meta-data associated with each sample, so that anomalies could be tracked back to a bad batch of reagents, a particular sequencing machine, or a particular operator for instance.

The hack

We would love to build a solution to import the metrics from key steps of the sequencing analysis pipeline, ingest these metrics into a database, and visualise these in an interactive fashion. We already have code for parsing most of the metrics, and a data-model for tracking the relationships between objects (like sequencing runs and samples). We are missing the bits that harvest this information in real-time, ingest into a database, and visualise this information; to this end, we have been considering building our own D3-based visualisation platform, or perhaps configuring a fluentd, splunk or tableau platform.

The future

We see clinical genomics becoming a mainstream part of healthcare, where for babies with inherited genetic diseases, or patients with cancer, it will make more sense to sequence a patient’s genome & use this information to guide improved healthcare. As we scale up to reach tens of thousands of patients per year, this platform will become an indispensible piece of the clinical genomics puzzle.

+ Life expectancy calculations for Aboriginal and Torres Strait Islander

Life expectancy calculations for Aboriginal and Torres Strait Islander men and women are subject to a large number of assumptions. I would like a model that shows the differences that differing underlying assumptions make. The ABS and AIHW both have different assumptions of under-identification in different age groups, they don’t put errors around that but we could! We could look at a variety of models and see what we come up with, and if it is different from the official statistics. Something graphical that lets people get away from the life tables.

+ Visualising what PubMedSaid

We want to build a visualisation tool to understand what these large gene sets have in common and how they link to diseases such as cancer and diabetes.

The problem

The rapidly growing body of scientific literature contains a wealth of information that is specifically written for a human audience. Reading through this information to extract links between data entities is however a time-consuming effort that requires expert knowledge.While previously biologists have focused on understanding a small handful of genes, new experimental techniques can provide sets of hundreds of genes that can be overwhelming to comprehend. We need to understand what these large gene sets have in common and how they link to diseases such as cancer and diabetes.

The dream

The dream is to understand what makes a given large set of genes special. We have performed text mining against all public full text articles and we want an intuitive system to understand and query relationships contained within this data. As scientific discoveries are made by drawing new links between previously unconnected components such a framework would achieve both: facilitate as well as accelerate scientific discoveries across a broad range of scientific disciplines.

The status

We have processed more than 2 million full-text scientific articles on biological and medical sciences contained within Pubmed Central, extracted meaningful keywords and associated these with categories such as disease, genes, biological pathways, cellular function and many more. The resulting 6 billion data points have been deposited in a document database (MongoDB), that is available through a Web API.Using Map-Reduce techniques we have created a table where each human protein-coding gene is mapped to every PubMed article it appears in along with every term found in these papers. We have algorithms for aggregating multiple genes and for performing basic statistics for over-representation.

The hack

Curiosity - having an question and going after the answer - is what drives science. To facilitate this process we require an interactive web front-end that allows intuitive querying across the various categories contained in our ‘little’ database as well as a visual presentation layer to convey the returned information and illustrate the connectivity between the discoveries.Ideally, we let the data speak for itself such as getting ranked recommendations via collaborative filtering or consensus via aggregations but also link back to the original publications to enable further in-depth studies.

How this helps us

Not being able to go after the answer at the opportune moment an idea strikes can stop new discoveries in its tracks. Thus, a framework that can reveal the connections between distinct scientific fields or discoveries on the spot has the potential to facilitate future Eureka moments.

+ Cut and paste - the precise way

We'd like to build a tool to allow life scientist to select the target site for their gene of interest and tune the compromise between on-target efficiency and off-targets risk as required for the application at hand.

The problem

The human reference genome has been sequenced more than a decade ago but to this date the insight into the function of any particular segment of the genome sequence is very limited. Fortunately, recent technological advances provide biological research and life sciences with an incredible set of new tools that enables us to modify single genomic positions with high accuracy and observe the effects of these perturbations on cellular processes thus providing insight on how life functions on the molecular and cellular level.

There is two sides to this coin though. On one hand, a genome modification should be introduced with high efficiency at the targeted side of interest. On the other hand, genomic positions that have a very similar DNA sequence will also attract the modification. Such off-targets need to be avoided as much as possible in order to guarantee a clean experimental outcome.

Most of the time the best target site will not comply with both objective functions simultaneously, i.e. maximal on-target efficiency and minimal off-target abundance, thus requiring a compromise between the two. This compromise will be depend on the application at hand.

The dream

The dream is to build a web based tool would facilitate genome science with optimal accuracy, precision and efficiency.

The status

A prototype for genome analysis is currently under development. It’s output comprises of a rich data set that contains a set of candidate target sites within the genomic region of interest, as well as a set of putative off-targets for each candidate all of which come with their own predicted efficiencies. In addition all regions will be annotated with known functional elements such as genes. These tasks are computationally demanding and will run on a High Performance Compute (HPC) cluster. Resulting output files will be either in plain text or Json format.

Examples on how such data might look like can be found at bioinformatics

The hack

The challenge is to guide scientist through the steps of designing their experiment and choosing their optimal genomic target. Off-target search will be done on a HPC cluster. The multidimensional output then needs to be collated and presented in a clean, slick and functional web-frontend indicating the pro and cons of each candidate.

How this helps us

Precise genome editing is changing the way of how biomedical science is performed. It is increasingly the focus of scientific discussion, e.g. at a recent Symposium in Melbourne. Precise genome editing not just enables insights into cellular behavior, it also provides new therapeutic opportunities for the treatment of heritable diseases as well as Cancer, Alzheimer and many more. Most importantly, a web-app that facilitates the intuitive design and assessment of genome editing tools is a prerequisite to achieve the precision, accuracy and efficiency gene therapy demands.

+ One Disease - Proof of concept for Health Hack

We proposed HealthHack to build a app to help Indigenous Australians to monitor and control their disease better. The app will provide visual cues and media—using images and pictographs rather than words or numbers to complete the diagnostic algorithm.

The problem

Background

Indigenous Australians live 1–2 decades less than their non-Indigenous counterparts. Scabies affects 7 out of 10 Indigenous children at least once before their first birthday. What presents as an innocent itch can lead to heart and kidney failure and ultimately premature death. One Disease aims to eliminate scabies, because we believe no Australian should die of a preventable disease. Currently, prevalence rates of scabies in Indigenous communities are established through screening events, or based on clinic presentation data. In screening events, a questionnaire developed from a sales app (DoForms) is used. In this, demographic information is preloaded onto the form, data is either typed or entered numerically during the screening, and the information is downloaded to an excel spreadsheet at the completion of the screening. This process has served the team well thus far, however there are limitations with accessibility due to the high level of English literacy and numeracy required to complete the form.

A proposed solution

If we were able to develop an app that relied more strongly on visual cues and media—using images and pictographs rather than words or numbers to complete the diagnostic algorithm—it would become more accessible to the Indigenous workforce and open the app up for greater use. Increased uptake would lead to increased screening, a better informed picture of the size of the scabies problem, and ultimately a more coordinated and comprehensive response to it. A visually based app would allow community members to take part in disease control efforts—effectively empowering them to crowd-source management of their own health. An open access app will also open the door for crowd-sourced surveillance of scabies if it is adopted for use by Government and Aboriginal Community Controlled Health Clinics.

The proposed app would require the following technologies:

• Data collection will take the form of a tablet-based app with GPS capability. The app will need the ability to function offline (due to connectivity restrictions in the remote areas we operate) and community demographic data (including names, dates of birth) will need to be preloaded.
• The interface must cater for English as a second language and low technology consumers and use strong visual inputs, including the ability to draw/finger paint onto images, image capture and storage capability, options for selecting images from a photo gallery (eg, skin sore gallery to drag and drop onto a diagram of a body)
• Collected records require storage on a highly secure cloud device, employing encryption and security technologies. Data is initially identifiable, therefore security is critical. Access will be from multiple sites, by multiple users, and content will need to be downloaded and stored for analysis.

From a data analysis perspective:

• Patient information needs to be able to be uploaded to the form from an excel or csv file (name, DOB, address, age at screening, sex, etc)
• Group results need to be downloaded to an excel or csv file following screening for analysis.
• Diagnostic algorithm that currently sits behind screening table needs to be preserved in visual translation.
• Conditionality needs to be built into the form- i.e. certain questions arise as a result of answers to previous questions.

+ Let’s create an easy-to-use “viewing platform” for the NHMRC funding landscape

If we could quickly generate some user-friendly, easy-to-read graphics that clearly explain where the funding goes each year, it would allow individuals, institutes and other interested parties to readily view the funding landscape.

The problem

Each year investigators patiently await the announcements of NHMRC funding. After submitting an application in March and rebutting the reviewers’ comments mid-year, it will not be until October that they find out if they have received funds or not. Keep in mind that academic research is largely funded through grants such as these. There is a lot of publicly available data on investigators and teams awarded NHMRC funding, along with a lot of statistics available. It would be terrific to be able to quickly visualise these results without poring through a large number of pdf files and spreadsheets. If we could quickly generate some user-friendly, easy-to-read graphics that clearly explain where the funding goes each year, it would allow individuals, institutes and other interested parties to readily view the funding landscape.

Why?

As an individual researcher or an organisation, such a tool would allow you to work out the demographics of successful applicants, at what level and in what research areas, without having to pore through the data. It will also inform policy development at the institute and national level since any group or association will be able to readily access this information. Just enter the terms you wish to define and ‘voila’, a graph with your desired information will appear.

+ A flexible smartphone app platform for research questions that link mood to behavioural decision-making

We are a group of Monash researchers in psychology, psychiatry, and statistical time-series analysis that would like to develop a smartphone app that logs location, time, emotional state (through questionnaires), and some measures of alertness (e.g., through a simple reaction time task embedded in the app).

The problem

We are a group of Monash researchers in psychology, psychiatry, and statistical time-series analysis that would like to develop a smartphone app that logs location, time, emotional state (through questionnaires), and some measures of alertness (e.g., through a simple reaction time task embedded in the app). We are hoping to use this app to answer a diverse set of research questions of interest to our group, including how mood and alertness effect decision-making, how addiction behaviors manifest in the real world, and how sleep quality affects daytime alertness and behavior. We are particularly interested in developing new time-series analysis techniques to develop individual prediction models for these unique datasets, information that could be use to inform interventions and to cluster individuals based on their behavioural responses. The potential size of datasets that we can collect using such an app, and the richness of the temporal sampling is unprecedented in our field, and we are really excited to have the opportunity to pursue this new paradigm in data collection and analytics!

+ Prioritising gene variants as they churn off the genomics pipeline

Toward this goal, we are keen to develop a novel, user-friendly module that automates a series of sequence- and structure-based protein analyses, and systematic review of the literature.

The problem

The ‘omics’ era is upon us and genome sequencing is increasingly being integrated into clinical practice and used to develop personalised medicines. In searching for new genes responsible for a given disease, large cohorts of individuals will have their DNA sequenced and evaluated by qualified geneticists and bioinformaticians, to determine which genes are most likely to contribute to the disease state.

This can be particularly challenging, however, when exploring multigenic complex diseases such as autism or schizophrenia. After a series of analyses and filtering processes, sometimes >200 gene variants will be identified as possible candidates. Researchers must then decide the best candidate genes to move into their basic and translational research programs. Laboratory studies and disease models are costly in both time and funding, so it is essential to maximise our ability to prioritise candidate genes from the genomics pipeline.

Toward this goal, we are keen to develop a novel, user-friendly module that automates a series of sequence- and structure-based protein analyses, and systematic review of the literature. This will complement existing filters to better prioritise potential pathogenic gene variants. It will also enhance the ability of clinicians to predict disease outcome and trajectory.

Once this module is optimised and demonstrated to work well, it could be incorporated into almost any hospital/institute’s ‘omics’ pipeline.

+ Health and Society Expeditions

We want to develop a data-journalism tool that can be used to facilitate content coding of online videos and visualise patterns in that content. Such a tool would provide virtual communities a broad perspective of a curated media landscape.

The problem

The internet is a quagmire of health information, health misinformation and health disinformation. Public health information services, both in government and the broader civil society, are interested in communications activities which inform people of how to approach this information critically. Improving public critical thinking in the use of these information sources should increase the efficiency and outcomes of human/health-system interactions. Critical thinking is a verb. Top-down dissemination of information will not support this sort of learning. Developments in interactive journalism may be able to provide a way to impact upon public critical thinking.For example, a person who has been recently diagnosed with a mental illness might be interested in understanding the media around their diagnosis. A not-for-profit working to relieve mental illness might be interested in facilitating that conversation about media perceptions. In response to a disease outbreak, (such as ebola), an individual may be interested in understanding how the media representations shape our perceptions. An international relief agency may be interested in monitoring the response to inform their media campaigns for donations.

Example Video Coding Tasks:

Purpose of creator Imputed Target Demographic Medical Treatments referenced Geocoding (place references) Demographics of video ‘Actors Sentiment Open Description Validation of the above

Examples of ‘Citizen Science’ interactive journalism projects in Australia

http://volunteer.ala.org.au/

http://exploretheseafloor.net.au/

http://www.ozdocs.climatehistory.com.au/

+ Understand how species adapt to climate stress

We'd like to improve the usability and flexibility of our existing tool, which processes and analyses raw DNA sequence data. With the help of the tool, identify the types of species most in need of protection under climate change.

The problem

Thanks to next-generation DNA sequencing technologies, evolutionary biologists can now regularly tackle important questions like “What are the genes involved in adaptation to climate?” and “What genomic factors prevent climate-sensitive species from adapting?”

The bottleneck in our research is the computational work involved in processing huge amounts of raw DNA sequence data. VLSCI researchers have developed a pipeline, Rubra, built on the Ruffus Python library, that allows us to process and analyse genomic data on the powerful VLSCI computing cluster. However this pipeline needs improvement to make it more generally usable and flexible. Within our research group, this will make Rubra accessible to biologists with limited programming skills, and allow us to easily customise our workflow for each new experiment. This will take us closer to understanding the genomic basis for climate adaptation, which in turn will let us identify the types of species most in need of protection under climate change.

+ A visualisation tool to simplify a person’s epilepsy story

Epilepsy is a disorder of brain functions that takes the form of recurring seizures. 3-4% of the Australian population will be diagnosed with epilepsy. There are a range of seizure types that people can experience, a variety of seizure triggers, a wide range of antiepileptic drugs and therapeutic doses, and over the course of a person’s epilepsy story the impact of epilepsy on a their life will vary. Patient history information is usually stored in narrative text making longitudinal analysis difficult. The tool that we would like to developed will be used by clinicians to visualise patient data, and find correlations between seizure frequency, medication, severity of impact on life, and any potential surgery.

+ Superbug Analytics

We'd like to build a tool to collect the twitter tweets the related to microbial, antimicrobial and superbug topics. With the collected data it allows the establishment of a baseline, the exploration of trends, and the detection of unusual events.

The problem

The rise of antibiotic-resistant bacteria, or Superbugs, has become a serious threat to public health and the economy worldwide. This is exemplified by Barack Obama recently announcing a new Executive Order together with a National Strategy for combating antibiotic-resistant bacteria (18/09/2014). In this HealthHack challenge, we will explore if social media data mining can be used for microbial prevalence monitoring.

The key steps will involve:

• Filtering Twitter tweets based on a list of keywords related to microbial, antimicrobial and superbug topics
• Analysing the tweets using advance data analytics techniques
• Interactive reporting/li>