Expanding the electronic health record to enable predictive medicine

This is the second blog in our latest series on the 4 Ps of Medicine, focusing on the power of predictive medicine.

Imagine if you could predict a health issue before it became a problem. The way healthcare currently works means when a person develops a health issue, they see a doctor and are treated for their symptoms, costing the health system and the patient time, money and resources. What if we had already predicted that person is at high risk of developing that condition, put in place preventative measures and put them on a path to a positive outcome before they need costly treatment?

Predictive medicine is the branch of precision medicine that evaluates the probability of an individual to develop a disease in the future. The goal is to detect those patients at risk of developing a condition, and adopt preventative measures or significantly reduce the impact of the disease.

Let’s walk through the example of Phenylketonuria (PKU)[1], a rare inherited condition that increases the level of phenylalanine in the blood. A mutation in the gene responsible for breaking down this amino acid, causes a dangerous build-up of it, leading to serious health problems, including intellectual disabilities[2]. Newborns can be tested for PKU (not necessarily via a genetic test) and when assessed at high risk, they can follow a diet low in phenylalanine (usually found in protein, aspartame and artificial sweetener) to prevent the emergence of major health issues.         

There are different approaches to predictive medicine, reasoning on a variety of data. Even when evaluating the susceptibility to diseases, these methods are not able to predict with certainty that a specific health issue will occur. They are expected to be more effective though, when they make use of genetic data, for instance when applied to polygenic multifactorial diseases, such as diabetes mellitus, hypertension and myocardial infarction, or even better single-gene diseases, like cystic fibrosis.       

With technological advancements and the increasing variety of data that can be collected, health solutions can enable predictive medicine by integrating prediction tools that can identify individual risk for a number of health issues. By doing so, the health platform can contribute towards the paradigm shift of precision medicine.

How can such an integration be successful?

First of all, a health solution needs to support the contribution of data from a variety of settings, with particular attention to those data sources that can generate genomic data. Careful integration from data contributors such as EMRs, laboratory systems or other hospitals is crucial, as well as the appropriate modelling of brand-new data that can expand the current electronic medical record. Genomic data, for example, would require the definition of appropriate information modelling in case it would not fit the traditional observation profile.

HIT vendors should lead meaningful and exciting collaboration with biotechnology companies, laboratory systems and existing governance bodies to determine the standard formats and protocol to share this new type of information.

The involvement of patients and providers should not be neglected, and the appropriate presentation of these predictions is fundamental. Communicating risk scores and prognostic model results can have a different impact on people, some might overreact, while others might overlook them. The choice of relevant prediction tools, that make sense locally for the population (see the example of nzRISK), together with corresponding education and training material will help ensure effectiveness. The health solution needs to be flexible enough to offer integration with the variety of risk scores and the related documentation to interpret the results.

Finally, the ability to automatically generate predictions based on the clinical records that are available has to be enabled. Calculations can be executed asynchronously as more details are fed into the system. In fact, health networks might already share all data points that predictive methods require, the next step is to select and combine them to predict different risks. The result of such calculations should be kept as new pieces of information to look back on, to trend and to action against.           

The next blog in this series will focus on the next in the 4 Ps of medicine: preventative medicine.

Orion Health Amadeus is an integration platform that can support you in your journey towards precision medicine. Learn more about Amadeus below.