Wednesday, November 18, 2015

CMS Quality Measures and PSA Testing

On an obscure webpage, CMS issued a request for public comment for a physician performance metric that would financially penalize physicians who order PSA testing.  Physicians would be penalized for providing testing allowed to the patient by the Medicare statute, and even for a range of purposes that are recommended by the CMS NCD for diagnostic PSA testing.

From that lead-in, you can guess I view this as a flawed idea for multiple reasons.   These are discussed, and links are also provided to four brand new JAMA articles on PSA testing.

The JAMA articles on PSA testing policy and whither it should be going, were picked up by the New York Times and other major media (here).  The day after this blog appeared, the WSJ (here) ran an article on the new anti-PSA metric that this article discusses.

UPDATE: February 26, 2016:  CMS will not continue work on the proposed measure, according to the AUA (here).


The fight for quality in healthcare has resulting in colossally confusing and proliferating metrics (see quotes here).

In October/November 2015, a very confusing and simply poorly written PSA metric has been proposed for public comment by CMS.  Even the route to making public comment on it is confusing (involving registration and log in and several search menus) -- and the comment period closes November 20.

The main CMS Quality Measures website is here.  The webpage for public comment on future measures is here.  11 measures have recently been under comment, with variable closure dates.  The PSA proposed metric is found here:

  • Electronic Clinical Quality Measures for (1) Functional Status Assessment and Target Setting for Patients with Congestive Heart Failure and (2) Non-Recommended Prostate-Specific Antigen (PSA)-Based Screening [Comment October 26-November 20]

Scrolling two-thirds of the way down the very long webpage, you reach a text section about this metric, although it has very little substantive detail.  Then, the reader is directed to one of a dozen ZIP files buried at the very bottom of the page.  There, you can download the ZIP and read what the measures are about in three PDF documents.  Then, you can come back up to the text section of the page where there is a link to another government website where you can make a comment on the metrics (the link being  ) 

The ZIP file is here:  
Two measures, CHF and PSA, are embedded together here.  Regarding CHF, the measure would recommend that patients with CHF be measured twice a year with one of several endorsed rating scales for CHF.  Not too controversial.

For PSA, the PDF document is in the cloud (my cloud) here.   The metric will produce a "bad" rating (meaning zero is ideal) for "men who were screened unnecessarily for prostate cancer using a PSA based screening test."  CMS states the metric (meaning the text and proposed rules) were created by Mathematica Policy Research, and the PDF is so branded.  Specifically the measure could be extracted from EHRs alone.

The PDF notes that "The U.S. Preventive Services Task Force (USPSTF) has reviewed evidence on PSA-based screening and concluded that the harms of screening outweigh the benefits in a general (not at-risk) population of men."  As is well known.  However, other major policy bodies have taken different positions.   It goes one, "The intent of this measure is to discourage the use of PSA-based screening in the general population of men. A lower rate on this measure indicates better performance."

There are three problems.

First, men in Medicare have a statutory benefit to PSA screening.  It is unfair to pin doctors between the patient's legal and specific benefits - and this one is crystal clear - and financial penalties imposed by CMS.

Second, there are significant differences among authorities on this issue, as witnessed by no fewer than four articles on the topic in the November 18, 2015 issue of JAMA (here, here, here, here).   This is not a dead issue, as reflected also in NYT coverage (here).  In addition, interventions including follow-on new test protocols to markedly reduce the unnecessary biopsies and surgeries are constantly evolving.  An important new study appeared, for example, in the December 2015 issue of Lancet Oncology (here; also here).

Third, and most worrisome, the measure is in conflict with Medicare's own NCD for PSA testing (here).   This NCD does not cover screening (that is a separate benefit).  However, the NCD covers PSA testing annually in nearly all men with any kind of urinary tract symptom - prostate enlargement, hesitancy, nocturia, etc.   Let's say conservatively that's 95% of men over 65.

In contrast, the metric has very narrow exceptions when a PSA test could be used: only in men with an active diagnosis of prostate cancer, men with an active diagnosis of "dysplasia" - histologic? - and men with a 5 alpha reductase inhibitor prescription or PSA >4 during the year.   Thus, many men with allowable PSA under the practice recommendations of the PSA would fail to have an allowable PSA under the tighter limits of the performance metric.   This makes no sense.

Should anyone care to comment before November 20, you have to go the separate website (here), register with name, email, and password, and then go through a fairly elaborate process to find the comment area.  Even when I found it, it provided a long form with about 8 boxes, some of which I found confusing and it was difficult to finalize my comment due to system controls.  (Instructions here).

A test nerd would point out, the PSA metric is against "unnecessary screening" and provides exceptions for certain conditions and symptoms, but in the presence of those conditions and symptoms, the test is no longer "screening" at all - leaving you a bit cross-eyed.

For a January 2016 American Urological Association press release condemning the new measure idea, see here and for an earlier warning from AUA see here.

My comment was posted, here.