Scott MacStravic, Ph.D.
[The measurement challenge, cont'd]
Fortunately, new computer technologies and analytical tools have made predictive modeling, based on analysis of much more data in much less time than was true just a few years ago, far more reliable and credible. MedAI, for example, a predictive modeling firm in Orlando, FL, has won many awards for the proven accuracy of its forecasts of health costs. But since there is only the predicted future not experienced to compare to the actual future experienced, employers may simply not be willing to accept “on faith” anybody’s prediction if it results in making them pay more money for HPM services.
And when the differences made by HPM efforts represent those between estimates of productivity impacts, since productivity is so difficult to accurately measure, employer clients may be that much more uncertain about the results. Comparing estimates of actual reality to predictions of different realities will surely sound like a lot of guesswork, rather than a scientific basis for evaluating HPM effects and paying the providers that deliver such effects.
There are ways of making such comparisons more “scientific”. The random controlled trial, the gold standard in medical research, could, in theory, be used in HPM cases. Employers or the providers they hire could identify HPM prospects based on existing conditions or risks, then randomly assign some of them to being “treated” with a particular HPM intervention, while the rest are left “untreated” for comparison purposes. But consciously depriving some employees of the benefits of HPM participation, often including financial incentives, could easily cause labor unrest.
Moreover, history suggests that employers hate to limit something that works to only a portion of the population at risk. A random control trial involving different school districts in Wisconsin, for example, was ended after the first of three planned years. The results for the “treatment” group were so positive that the sponsors of the test enrolled all the controls in the treatment group, so there was no scientific comparison of the second and third year results.
And with HPM, results tend to increase over time. It takes a while for most people to change their risk behaviors, for example, or to adopt greater compliance with medications. Gordian Health Solutions, for example, has reported that the average ROI ratios for its employer DM clients have been $1.69:1 in the first year; $2.00:1 in the second, and $2.46:1 in the third. [“When It Makes Cents to Back into the 80/20 Rule” Gordian Health Solutions Oct 3, 2005 (www.healthleadersmedia.com) Lacking a random comparison control group, employers could not be sure that such results were scientifically valid.
Part 4 will appear tomorrow in HFMA Views. To read Part 1 of this post, click here. To read Part 2 of this post, click here.
