Wednesday, March 25, 2009

Follow-Up on Vermedx

The Disease Management Care Blog is happy to report that it received clarification about the Vermedx study mentioned in a prior post:

  • Many of the patients reported in the American Journal of Managed Care article were not involved in the original randomized clinical trial (RCT) in New York and Vermont. It turns out that the physician hospital organization - Vermont Managed Care - indirectly learned of the RCT and installed Vermedx independently. Not all VMC patients had access to Vermedx, which was based on geography, not primary care assignment.

  • Yes, it's possible that Vermedx did not, or only partially, accounted for the observed difference in claims expense - an effect described a 'treatment assignment bias.' The authors argue that is less likely because the claims trend prior to the Vermedx install was similar at baseline and then diverged once the intervention was launched.

  • The low number of patients was based on the selection of persons with two years of continuous enrollment. Many of the PHO's patients were only newly enrolled and didn't have two years of claims data to draw on.

  • Generalizability is being tackled by seeking funding for a repeat study in an environment different from Vermont: inner city with a greater proportion of non-white individuals.

  • In the meantime, the larger RCT has been submitted for publication and is underoing review.

The DMCB thanks the folks in Vermont for the feedback and salutes them for their transparency as well as willingness to subject their system to the travails of peer review. The DMCB also thinks the Editors over at AJMC could have done a better job of getting more of the facts into the original manuscript prior to publication.

Based on this additional input, the DMCB would have to say its opinion of Vermedx kicked up a notch. It's possible that the difference in claims expense was due to factors associated with geography and this involved a newly enrolled PHO population. There is still a lot we don't know about the participants in the trial or about the PHO's practice setting. It's also a stretch to believe anything involving a relatively small patient population when it comes to trending. However, this is a good start and the DMCB is looking forward to reading more about Vermedx in the years to come.

Last but not least, the authors are to be commended to simultaneously a) initiating a care system and b) studying the impact with an appreciable amount of scientific rigor. This is what is necessary if we're going to understand how initiatives like this work in the real world and is an important lesson for the disease management/population care community.

No comments: