We live in the information age where data is a highly valued resource. With all the data available today, organisations are very busy “mining the data” for meaningful insights and clues to inform solutions to problems we have previously not been able to solve. This is nowhere more relevant than in the field of Health Data Science where years of patient and population health data can be analysed to reveal trends and patterns that can inform better health treatments and services.

The Ingham Institute has made a strong commitment to developing its Heath Data Science capabilities through appointing Dr Tim Churches as the inaugural Research Fellow in Health Data Science at the Institute. Dr Churches works across all groups at the Institute offering his services to any researcher who might require assistance in collecting and analysing health data, as well as conducting his own research program into key data science problems.

Dr Churches recently led a team of data scientists in the Sydney HealthHack 2017 event. The event ran over 24 hours, during which time the team was presented with a real world health data problem. The challenge the team was presented with was to enhance the existing Australian and New Zealand database of clinical trials so that it can be efficiently searched for trials involving drugs that target specific gene mutations or gene pathways. The ability to find such clinical trials is vital for patients in whom usual treatments have proven ineffective, and who have subsequently had their genome sequenced to determine whether they have specific (and possibly very rare) gene mutations which could be targeted by particular (and possibly experimental) cancer drug therapies. Currently ,searching for clinical trials which might target specific genetic mutations is a manual and time-consuming process.

“This kind of search engine would be used by cancer specialists primarily for patients who have not responded to initial treatments and are looking for new treatment avenues to pursue. For some patients, this information can literally be life saving,” said Dr Churches.

The team members were Dr Churches, Dr James Farrow, an independent computer scientist who has collaborated on several projects with the Ingham Institute, Dr Eugene Lubarsky Chief Technology Officer of local health software start-up, PAT Software, (www.patsoftware.com.au),  and Gabrielle Provost, a data scientist with Uber Australia (www.uber.com/en-AU/).

The team’s approach to this problem was to create program code that firstly automatically identified gene names, gene symbols, pathway identifiers and drug names in the free-text fields of cancer trial descriptions  contained in the Australian New Zealand Clinical Trials Registry (ANZCTR). They did this using advanced NLP (natural language processing) computing methods. They then set about creating a process to identify drug-to-gene-target connections using DrugBank.ca, Ensemble and other internet genomic data resources. They then wove these elements together to create a search engine in which clinicians can use gene variant/mutation information for their patients to search for suitable clinical trials.

By the end of the 24 hour event, the team had produced a working prototype, and received second prize overall and the Data Science award. Two of the team members, Dr Churches and Dr James Farrow, will continue to work on refining and validating the prototype in conjunction with the Garvan Institute problem owners, Drs Mark Cowley and John Grady, prior to publishing the solution as an open-source package for the R programming environment for use by Australasian doctors.

“Once this program is available, Australasian doctors will be able to quickly find clinical trials across Australia and New Zealand for their patients who might have run out of standard therapy options.” Dr Churches explains.

For more information on clinical trials available in South West Sydney visit: www.clinicaltrials.org.au