Pfizer Health Solutions, Florida: A Healthy State & more on Open Source
Posted Dec 02 2008 3:08am
In the last two Disease Management Care Blog (DMCB) entries, we impertinently suggested the industry jettison the short-term competitive advantage of their clinical/claims data bases and adopt an Open Source “business model.” The chief advantages are 1) increased credibility, thanks to inviting independent confirmation of any conclusions, 2) the likelihood of new discoveries, thanks to more scientists (and eventually amateur-scientist lay persons) “mining” the data and 3) achieving the fullest manifestation of what Kaiser Permanente is teaching us: that the advent of these colossal patient/enrollee/consumer data bases in the disease management industry are an incredibly valuable resource that exceeds the biggest baddest electronic health record. The potential benefit to the general knowledge of mankind is great. So, by the way, is the good will it would generate.
For a good example, first congratulate the folks at Pfizer Health Solutions (PHS) for getting some important clinical outcomes into the peer review literature. Most readers of the DMCB will recall PHS is providing disease management services to portions of the Florida Medicaid population. PHS funded a study on the impact of their “Florida: A Health State” Program on a number of outcomes including blood pressure, lipids, blood glucose and asthma control among persons with chronic illness who were enrolled in their program. Morisky and colleagues at the UCLA School of Public Health found there was a clinically and statistically significant improvement across multiple care domains.
The improvement was detected by comparing the study population (with available data) to an age, gender, ethnicity, poverty level and time-point matched National Health and Nutrition Examination Survey ( NHANES ) data set using odds ratios.
While that sounds pretty easy (not), it seems to the DMCB that it was the control population that was the critical factor in the reaching the study’s conclusions. In addition, odds ratios, compared to other approaches, can be difficult to fathom. Stakeholders not familiar with research techniques (for example, the human resource person in charge of a TPA, or a staffer in a politician’s office) may have difficulty understanding the more subtle dimensions of interpreting these findings. Last but not least, the study was honorably, but let’s face it, commercially funded.
You may say I’m a dreamer, but imagine for a moment if the data from the “Florida: A Healthy State” were available gratis to other researchers who were welcome to confirm the results using another control group or another comparative approach. The DMCB thinks that the conclusions would hold up, that the transparency would reassure stakeholders in academia and policy circles that the findings are for real and, who knows, other important insights about the needs or care of persons with chronic illness would be closer to discovery.
Florida and its citizens would be better off for it.