Medical testing: is it about quality?

In her New York Times “Well” blog of April 13, 2015, physician Abigail Zuger comments upon “Quantifying Tests, Instead of Good Care”. She discusses a variety of factors that lead to physicians ordering more and more tests. In part, this is the simply the result of the advance of technology and thus the profusion of tests that can be done --  laboratory tests, imaging (like x-rays and their ever more complex cousins), and even parts of the physical examination (“Yes, little blood pressure cuff over there in the corner, that means you”) --  which sometimes can be seen as meaning they should be done. Other factors are about money. The tests cost money, a negative for patients and insurers, but they make money for the laboratories and imaging facilities and the hospitals which often own them, and frequently employ the physicians doing the ordering. Dr. Zuger notes that “Ordering tests, chasing down and interpreting results, and dealing with the endless cycle of repeat testing to confirm and clarify problems absorb pretty much all our time.”

Another impetus is our desire – patients and physicians alike – to find out what is wrong when there is something wrong. Or, when there is nothing really wrong, to perhaps prevent later disease by finding something hidden lurking inside. There are all these tests we can get! The patient may have had many of them before, ordered by another provider, or know people who have had them. They hope that something will give them an answer. The doctor may feel that there may be an answer out there that s/he might miss if a test is not ordered, or if s/he is a primary care physician that a specialist might order these tests so maybe we should do it first; after all, it sometimes seems as if one thing specialists do when we refer patients to them is order a bunch of tests that we could have. Or, and this is a big focus of Dr. Zuger, that their “quality ratings” may suffer if s/he doesn’t order these tests.

Quality ratings are measures of things that doctors should do (or, sometimes but more rarely, should not do) for patients with certain conditions or complaints; increasingly, insurers, employers, and third-party groups are assessing the quality of care that a physician delivers in this way. Are these recommendations not based on evidence? In many cases, they are. For prevention, the US Preventive Services Task Force assesses the degree to which tests may be indicated or not for different people, often based upon their age and sex. It gives them ratings: ‘A’ is strongly recommended based on good evidence, ‘B’ means there is some, but weaker, evidence to support doing it, ’C’ means there is conflicting evidence and patient preference should be taken into account, ‘D’ means not recommended (might even be harmful), and ‘I’ means insufficient evidence to assess. A smart phone app, ePSS, can be downloaded by anyone; put in age and sex and it tells you which preventive tests are recommended.

Of course, not all doctors do all the recommended screening tests, and frequently they do tests that are not recommended (e.g., PSA to screen for prostate cancer) or more frequently than recommended (Pap smears to screen for cervical cancer), or in people for whom the test is not recommended (mammograms in younger women). When it comes to diagnostic testing rather than screening, looking for the cause of something when the patient is actually ill, there is even less consensus, and often a prevailing attitude of “more (testing) is better”. One effort to combat this is the Choosing Wisely campaign begun by the American Board of Internal Medicine Foundation. This campaign asked specialty societies to identify 5 or more commonly-done tests or interventions in their specialty that should not be done. However, the same recommendations are not always endorsed by every relevant organization. For example, the American Academy of Family Physicians recommends “Don’t do imaging for low back pain within the first six weeks, unless red flags are present.” Great, but neither the American College of Radiology nor the American Academy of Orthopaedic Surgeons, arguably the groups that could make money from this, has such a recommendation.

And then there is the risk. Some of the tests receiving USPSTF ‘D’ recommendations are more than not helpful; they can lead to harm, and this is even more true of the many potential tests that can be done looking for disease. Discussing the work of a collaboration at the Dartmouth Medical School that looks at the utility of tests, Dr. Zuger notes that such iconic treatments as controlling blood pressure can lead to bad outcomes in many older people where the benefits (avoiding stroke and other complications) begin to be outweighed by the risks:  “One study found that nursing home residents taking two or more effective blood pressure drugs did remarkably badly, with death rates more than twice that of their peers. In another, dementia patients taking blood pressure medication with optimal results nonetheless deteriorated mentally considerably faster. Yet no quality control system that I know of gives a doctor an approving pat on the head for taking a fragile older patient off meds.”

This is just the tip of the iceberg. Tests cost money, and often show results that lead to more tests, and then maybe to procedures (biopsies, for example) that maybe show abnormalities that lead to bigger interventions. And, sometimes these interventions are life-saving, but not always. Or maybe, even, not often.  And they do often have side effects, some of which are not insignificant. The person in whom a series of tests identifies cancer and is then treated and then survives is almost always grateful for the testing and the treatment, although in many cases they would have survived just the same with no treatment, and without the costs both in dollars and in side effects. The classic example is prostate cancer, where the evidence of any treatment making a difference in survival is weak at best; it is either a devastating disease from which one dies, painfully, or more commonly a benign condition that one dies with but not from but which treatment seems to have little influence on. Except for the cost, and the non-trivial side effects of impotence, incontinence, and the effects of radiation on the rectum.

Increasingly, data is showing that much breast cancer is similar. As discussed in my December 2012 blog post “More on mammography: just because you don't like the results doesn't make research junk science”, a large study by Bleyer and Welch showed that, with new and more sensitive mammographic screening, many more women are being diagnosed with early-stage breast cancer. However, despite treatment, we are not seeing anything like that reduction in late-stage breast cancer, which means most of those early stage cancers would not have progressed at all. And yet every woman diagnosed and treated sees herself as a survivor who has been cured, even in those cases where progression would not have occurred. In his wonderful recent New Yorker article “Overkill” (May 11, 2015), Atul Gawande notes that we treat all “cancer” as if it were rabbits in a pen who have to be chased down before they escape, when in fact many cancers are “birds” that have flown away and are gone before we can do anything and many others are “turtles” who are not going anywhere whatever we do.

The goal of medical care is now to achieve the “Triple Aim” of high quality, patient satisfaction, and lower cost. As with many things that come in threes, it is not that hard to achieve two but very challenging to achieve all three. Patients satisfaction is important, but not at any cost. Cost control is good, but not if it seriously limits quality. One place to start, then, is with quality, and specifically not doing what is useless or even harmful as assertively as we do what is likely to be of value.

And, of course, to make sure that it is the disease or risk of disease that determines what tests are done, not people’s insurance or socioeconomic status, ethnicity, gender or personality.