The underlying construct that birthed the guideline movement was evidence based medicine.  Performance measures followed from guidelines.  Like many concepts in medicine, I believe we have too quickly accepted the construct that guidelines will help patient care.  We place great confidence in performance measures, yet one must ask whether any evidence supports that confidence.

Last week I blogged briefly about the primary care paradox.  As I understand that paradox, primary care physicians consistently perform less well on disease specific performance measures (when compared with the appropriate subspecialists.)  Yet, studies make clear that more access to primary care improves health outcomes at a lower cost.

“The good physician treats the disease; the great physician treats the patient who has the disease” – Sir William Osler

I suppose that implicitly those who promote guidelines and performance measurement assume that we can deconstruct patients into their various ailments.  We then should treat each ailment well, and the patient will necessarily benefit.

We should never deconstruct patients, even when considering the management of a single disease.  Our data sources too often examine patients with only one disease, so as not to contaminate or confuse the research.  As a clinician, I care for very simple, one disease, patients.  I care for patients with multiple diseases, social disadvantages, and habits of self destruction.  How do I balance these complexities.

Now I am not totally against guidelines – I just want restraint and common sense.  I am not against selected performance measures as a means for self improvement.

I am against the use of performance measures to rank my performance, or to “incentivize” my clinical care.  Performance measures should always remain a measure that I can personally use to analyze my own patient care.

Guidelines should only become performance measures after we do prospective effectiveness studies.  Currently we base our measures on efficacy, and efficacy studies should not influence medical care without subsequent effectiveness studies, unless (as in the initial penicillin testing) the effectiveness is so obvious.  We do not need effectiveness studies of parachutes when jumping out of airplanes.  We do need effectiveness studies of nasal swabs for MRSA, the 4 hour rule for pneumonia, using a third drug to move the HgbA1c below 7.  To understand the difference between efficacy and effectiveness:

The Differences between Efficacy Studies and Effectiveness Studies

In testing the effectiveness of psychotherapy, the efficacy study is the more popular method and the one traditionally more trusted by researchers (Seligman, 1995). Efficacy studies are highly controlled and methodological: the psychotherapy being tested is used with one group of patients, while other control groups undergo either no treatment or treatment with credible placebos. The patients selected to participate in an efficacy study must be suffering from a particular disorder, and any patient with multiple disorders is usually not included. Participating therapists are given detailed instructions and continuing guidance on how to perform the therapy and patients are seen for a relatively small (twelve, for example) number of sessions. Although the therapists and patients involved are of course aware of what type of treatment they have been through, the diagnosticians who test, interview, and evaluate are blind to which group a patient belongs to. The diagnosticians continue to evaluate the patients for some time after therapy has ended. Not surprisingly, efficacy studies are time-consuming and expensive.

In contrast to the efficacy study, an effectiveness study looks at how much benefit “actual” patients gain from “real-life” therapy. Patients who have already begun (and possibly completed) therapy are surveyed by researchers and asked detailed questions about their treatment and its effectiveness. Unlike the efficacy study, in an effectiveness study the researchers have no say in how therapy is performed, nor can they select which patients undergo which type of therapy or which therapist a patient sees. Therefore, the researchers cannot create a control group to use as a baseline and there are no placebo therapies. Patients with multiple disorders cannot be excluded from an effectiveness study. Since the patients who are surveyed for the study have already begun and possibly also completed therapy, much (if not all) of the therapy takes place with neither the therapist nor the patient even knowing that they will be part of a study.

So beware of experts telling one how to practice medicine.  Experts too often care for a disease.  We need consideration from those who care for patients, not diseases.

Related posts:

1. Performance indicators need testing
2. My position on performance measures
3. Natural history, guidelines and performance measures
4. Improved performance – P4P or consistent primary care?
5. Guidelines not rules

Related posts brought to you by Yet Another Related Posts Plugin.