Following our recently announced expanded partnership with the Institute for Health Metrics (IHM), we spoke with IHM’s Chief Medical Officer, Dr. Anita Karcz, on the goals of the partnership, the importance of social determinants of health data, and how synthetic data fits into the broader health data landscape.
Dr. Karcz is a physician and leader in various segments of healthcare, particularly in healthcare information technology. She spent several years as a practicing emergency physician, and was active in hospital and professional society leadership roles. She was vice president of clinical product development at InterQual Inc, has been on the founding team of three medical device companies, and served as a management team member or business advisor at several other early biotechnology, medical device and healthcare IT companies.
Tell us more about the newly launched data offering between IHM and Syntegra and the goal of this partnership.
The IHM team is excited about our partnership with Syntegra because data sharing has traditionally been impeded, and rightly so, by concerns about patient privacy. With synthetic data, that concern is removed by providing data derived from a real data set, statistically representative of the real data set, but containing no Protected Health Information (PHI). Now EHR data can be shared more freely for research, accelerating insights into clinical practice, and used to develop algorithms to improve early diagnosis and treatment.
The inclusion of social determinants of health variables in this data is particularly notable. Why was this so important to include?
Healthcare workers on the front lines have known for years that social determinants of health (SDOH) play a more important role in a patient’s health than the care provided by the healthcare system alone. If someone has to decide, “Do I spend my money on food or on my medicine?” — the answer will be food. The Covid-19 pandemic highlighted the socioeconomic differences in infection, hospitalization and death rates. Linking SDOH to EHR data is critically important for understanding the effects of education, housing, employment, food access, income and health insurance on clinical care and outcomes.
A key capability of Syntegra’s synthetic data engine is the ability to augment data, such as expanding population sizes for rare cohorts or shifting the demographic makeup to address areas of bias. What types of analyses can be done on this data, and in partnership with Syntegra, that wouldn’t normally be possible or would be difficult to achieve?
Expanding or adjusting populations for rare disease research or to account for bias are prime examples of the power of synthetic data. Synthetic data enables areas of analytic inquiry that either weren’t possible previously, or greatly improve analytical power and significance. In addition, the portability of synthetic data accelerates analytic innovation by reducing the overhead and restrictions required for working with PHI. While it’s not easy for the industry to make the mental leap from the idea of simply copying data to the paradigm of creating new synthetic data, we are entering an exciting era in data science.
How do you view synthetic data fitting into the broader healthcare data landscape?
While some still view the potential of synthetic data as another way to deidentify data, that notion is changing rapidly. Just a year ago, many people in healthcare hadn’t even heard of synthetic data. The terms “healthcare” and “early adoption” don’t often appear together, and for good reason, as the well-being of patients ultimately depends on the decisions informed by analytics. However, I believe adoption of synthetic data will grow exponentially over the next few years as researchers add to the statistical proof that synthetic data accurately represents real-world data. Synthetic data will be much more widely accepted as the evidence base grows, enabling rapid access to data for researchers and for algorithm development.