Ethical Issues of Health Management Predictive Modeling

Predictive Modeling

Predictive Modeling (PM) is defined by Jonathan Weiner, PhD of Johns Hopkins Bloomberg School of Public Health as ‘a process that applies available data to identify persons who have high medical need and are at risk for above-average future medical service utilization’ (Carlson, 2003). Predictive Modeling is currently a diverse field and generally utilizes algorithms derived from regression analysis, decision trees, rule-based systems or neural networks to identify high risk individuals. The overall goal is to improve health status while reducing cost (Axelrod & Vogel, 2003). Uses include (Tremblay, 2005):

• 1.

  Identifying healthy persons whose health is likely to decline;

• 2.

  Reducing inconsistency in care;

• 3.

  Predicting who is likely to be non-compliant with prescribed treatments and medications;

• 4.

  Identifying patients with high risk for actionable chronic diseases such as diabetes, congestive heart failure (CHF), asthma, chronic obstructive pulmonary disease (COPD), and depression.

Factors such as age, gender, total co-payments at drug initiation, total medication burden, and initial compliance with prescribed care are predictive of long term compliance (Carlson, 2003). Limitations of PM are that it must use readily available data and generally does not include affective constructs such as attitudes and beliefs. As adoption of PM applications becomes more widespread it is incumbent upon health management professionals to use ethical guidelines as a framework in deciding when to use it and how to use it. A review of the basic tenets of ethics in relation to PM provides a guideline for recommending and designing programs.