Advances in health information technology significantly improve our ability to identify population based health status and clinical need through more accurate, timely and clinically relevant measures. The tools and instruments that are now used most often by private and public insurance sponsors and others interested in adjusting financial or clinical quality and performance data rely on proxy measures of health status such as diagnoses, pharmacy data or procedures, usually captured from claims or other administrative data sources. These instruments and the risk assessments derived from them are used to adjust health plan or provider payments and physician profiles for quality assessments among other uses. The foundation of each model is that diagnoses or pharmacy dispenses generated from claims or administrative data serve as a signal of underlying health status and can be used to explain current or predict future health care use. Models that rely on these data were developed because until recently it has been impractical or too costly to capture markers of actual clinical status from clinical records on a population basis. However adoption of electronic medical records, which include clinical data previously not available for large populations has the potential to change the ways we assess population based risk and apply these assessment to adjusting clinical performance and health plan and provider payments. We propose to develop and test a risk assessment model that uses real time clinical data from an electronic medical record. We will test the hypothesis that the ability to assess population risk using clinical expressions of medical need will be more accurate than measures derived purely from diagnostic or pharmacy data and be more relevant in guiding clinical practice.
We propose to develop, test, and validate a population based risk assessment model using primarily real time clinical and diagnostic data obtained from electronic medical records, supplemented with self reported information on health behaviors. Advances in health information technology create opportunities to estimate population based risk using more complete data on health status and health outcomes. Risk models based on real time clinical data will be more accurate with respect to their ability to explain and predict medical care need as well as more closely aligned with clinical practice.
