Hunting Down Hidden Healthcare Data
Summer of Research project by Kieran McCullough, University of Auckland, supervised by Associate Professor Robyn Whittaker.
Since the implementation of the New Zealand National Health Index in 1993, many different electronic health systems have popped up around the country. In his summer of research project Kieran McCullough looked at some of these systems and noticed that often the data that had been collected over the years was left undocumented and underused. There was a wealth of information in these databases which was often hidden in the back end of the software, and Kieran was confident he could extract and make sense of it.
Kieran worked with the Waitemata District Health Board, the biggest DHB by population in the country, and was able to access data from MedChart, the largest electronic system currently in use. MedChart provides healthcare facilities with an electronic medication prescription and administration system which helps reduce clinical risk and improve safety. Simply put, it tracks the medications given to patients in approximately 1000 beds across the Waitemata DHB.
Much of the information within the MedChart database was buried deep within the software, and was not available to everyday users. By mapping and documenting the database, Kieran turned the raw data into a format that can be easily understood by those wishing to analyse it. In doing so, it also becomes possible for this new data to be integrated with existing information in other health records. This allows for predictive analytics to help inform healthcare organisations, and in turn, create a more efficient healthcare system. It is important to note that the data is all anonymised and cannot be traced back to individuals.
Medication-related questions were already being asked at Waitemata DHB as part of various strategic projects, so the work that Kieran completed over the summer has helped answer questions like:
- What are the current usage patterns of sleep medications, and what are their effects during a patient’s stay in hospital?
- Is there any relationship between different types of pain medications administered, and length of stay in hospital?
- What are the relationships between prophylactic antibiotic administration and surgical site infections with different prescribing patterns?
Answering these questions will have benefits such as improving sleep in hospital, improving prescribing behaviour, reducing infections, and reducing the length of stays. All these benefits improve the hospital environment for both patients and staff, as well as allowing the healthcare system in general to operate more efficiently.
As part of his project, Kieran paid special attention to the data he collected around medication administration events. This includes times when medicines were given (administered) on time, withheld, delayed, refused, or missed.
Some of his results are displayed in the graphs below:
With this type of information, Kieran has made it possible to pinpoint exactly what time of day the majority of medications are distributed. Hospitals can use this to make sure there are enough staff to distribute said medications. Of course, healthcare staff are aware of these needs, but tracking this data over time in detail will allow analysts to discover new trends and give advice accordingly.
Tracking missed doses is also important, and as depicted in the above graph showing missed dose trends throughout a day, missing a medication dose happens more often in the middle of the night. This is most likely because patients are asleep, and therefore either unable or unwilling to take medication. Again, hospital staff are generally aware of this trend, but do they know why there is another spike in missed doses at 3pm? Tracking this data over time will help analysts discover the reasons behind missed doses, and then advise healthcare institutions on how to reduce these events.
The data that has been shared here is fairly simple, but it serves as a good example of what’s possible. However, we can delve deeper. Kieran’s project categorises everything available in the MedChart system. It’s possible to look at trends around everything from patient ethnicity, to which specific medication is being prescribed, and in what dosage.
Overall, Kieran’s research is likely to raise more questions as analysts pick up on certain trends, but that’s exactly the point. This project paves the way for more in-depth research into the data collected by electronic healthcare systems. The Precision Driven Health initiative aims to collect and integrate the combined data from all these systems and use it to provide patient-specific treatment to every individual. The future of healthcare is precise.
Kieran McCullough is among a group of students who took part in the summer of research programme funded by Precision Driven Health. This month we are featuring a blog series examining these projects. While at an elementary stage and considered to be a ‘proof of concept’, these projects offer fresh insights into what the world of healthcare will look like when precision medicine is fully implemented.
PDH is New Zealand’s unique health data science research partnership.