Humility and medical guidelines

We have LOTS of guidelines and recommendations. We we need is more humility.

The Advisory Committee for Immunization Practices (ACIP) issued a recommendation in September of 2014 that pneumococcal vaccine naive individuals 65 and over get a Prevnar13 followed by a Pneumovax 6-12 months later, and that those who have had a Pneumovax get a Prevnar13 a minimum of 12 months after their Pneumovax.

As a result of this recommendation, institutions are setting up protocols and adjusting their immunization forms and metrics to conform to this new 'standard' and clinicians have started telling patients they are 'due' for the new immunization. (This is true where I work, and I experienced it recently with a relative being seen at the Geriatric Clinic of the Massachusetts General Hospital.)

Austin Frakt and Aaron Carroll  wrote recently about how inaccurate the perceptions of patients are when it comes to risks and benefits. (I think clinicians are not much more accurate, sad to say.) One reason is the automatic cascade from limited evidence, to guidelines, and then to clinical practice work flows that get presented to patients without nuance or ambiguity. Instead of "We have some evidence that suggests..." the approach too often is "You are due for..." or "The treatment is..."  It is worth stepping back a moment to ask ourselves: "How sure are we? What does the data actually say?"

In the instance of Prevnar13, the evidence underlying the recommendation is actually quite limited:

  • There are 2 published studies of immunogenicity of the vaccine. These show that Prevnar13 causes a more robust antibody response in the short term (4 months) and medium term (4 years), compared to the PPSV23 (Pneumovax). Interesting and suggestive, but not proof of benefit.
  • There is one unpublished randomized clinical trial (done by Pfizer, the manufacturer of the vaccine) involving 85,000 community dwelling adults over 65 in the Netherlands (CAPITA) which shows that 1 invasive serious illness like meningitis or sepsis (IPD) is prevented for every 27,800 adults vaccinated with Prevnar13 and 1 hospitalization for community acquire pneumonia (CAP) is prevented for every 1620 adults vaccinated with Prevnar13. 

Based on this, a review in JAMA (2/17/2015) notes that the impact of the vaccine 'can only be inferred' at this point.

Perhaps we should not be telling patients they are due for a vaccine or incentivizing clinicians to pressure them to have it. Instead we should be saying that, based on fairly limited evidence, it appears that we have an improved vaccine that will protect 1 in 27,000 from a potentially lethal infection and 1 in 1600 from a hospitalization for pneumonia. 

Shared decision making means helping patients make autonomous choices based on current best evidence and personal values and preferences. We should be incentivizing clinicians to offer patients appropriate and balanced information about benefits and risks, not incentivizing clinicians based on whether or not the patient wants accepts the values and goals of the institution, an insurance company, a panel of experts, or CMS.


Links to more on this topic::