Update – managing PT in workers comp

I received several calls about my post a couple weeks ago regarding using peer review to manage PT.
Generally, there appears to be some confusion over the article I cited in the post – specifically about its conclusion that peer review “may have had an impact” on the number of visits.
Here are the key paragraphs:
“It seems that peer review may have had an impact – possibly by reducing the number of claims with more than 24 visits (this wasn’t apparent from the article). Complicating the analysis was the underlying data; it wasn’t possible to determine objectively if there were jurisdictional differences or claim severity differences (e.g. there is a very wide range of ‘severity’ associated with lumbago). The article noted that more severe claims were probably more likely to have peer review, but that was based on the assumption (a reasonable one in my view) that lost time claims were more likely to have requests for peer review than medical only claims…if the peer review program did result in fewer cases with more than 24 visits, how many of those were still excessive (the average number of visits for PT in comp is much lower than 24). And what was done during those visits, were the claimants ‘shaked and baked’ or was there actual work hardening and therapy designed to increase the patient’s functionality?”
In actuality, despite a close read of the study itself there was no conclusive evidence that peer review had any impact on the number of visits.
The underlying data did not have enough detail to provide any meaningful comparisons; for example it wasn’t possible to determine if a specific claim was a med only or lost time, or how severe the injury was.
And the analysis of a smaller number of claims from one large employer (reportedly UPS) was not directly comparable as the claims may have been in different jurisdiction (or more or less severe).
The net is this. It was not possible from the article to determine if peer review had any impact on duration of treatment. Yet some readers (at least a few who contacted me) drew that conclusion.
As I noted in the conclusion, “more questions were raised than answered.” Here are a few:
– why was the article published if the conclusions were so…inconclusive? would it not have been more useful and interesting if some of these issues were addressed more thoroughly?
– the data was not corrected for jurisdictional variations. If that wasn’t possible, I’d suggest that another data set should have been sought, as there are huge differences among and between states regarding utilization and cost of PT, as well as changes within a state due to regulatory changes (see Florida, New York, Texas and California). It could be that no other data set was available.
– the severity issue is quite important; the data were evidently not sufficiently robust to make some assessment of differences in severity. Without some measure of severity, it is impossible to make comparisons and draw conclusions about those comparisons, especially as there can be wide ranges of severity for conditions such as lumbago (one of the diagnoses featured in the report).
I applaud the author (Janet Jamieson, PhD) for her research and effort to add to our understanding of this critical issue. While it would be best to wait until all the data are in, and tested, and all follow up questions asked and answered, that’s not realistic nor possible. Instead, it is usually more helpful to publish what you have, recognizing that it will likely spur additional research.
It is incumbent on those who read reports and analyses to think critically, and not read more into the results than they should.
That is a disservice to the researcher, and may well lead to inappropriate conclusions.