Why BRCTs Matter: The Case of HPV Immunization
When is 64% the same as 4/1000? Very frequently, as least when it comes to the presentation of medical data. Relative risk/benefit can be misleading on many levels, aplifying a tiny benefit from medical intervention into much more profound appearing changes. The recent report of HPV vaccine's efficacy is emblematic of a problem embedded into our medical culture: using relative benefit numbers to make a story seem more sensational. Unfortunately, hyperbole does not enable effective shared decision making, while translating such numbers into BRCTs is more meaningful.
When a link was discovered between Human Papillomavirus (HPV) and cervical cancer, a campaign was started to immunize teenagers and young women against HPV in an effort to curb cervical cancer. Largely this thurst has been successful, although there is some resistance to and doubt about universal vaccination for HPV. In February 2016 an NPR story stated that in fact HPV immunization has cut the risk of HPV infection by 64% according to recent evidence. The problem with the vaccine's success, stated the story, is that too many doctors are negligent about giving it.
Like all relative risk/benefit numbers, 64% reduction tells us little about the efficacy of the vaccine. What is the actual risk reduction in HPV infection with the vaccine? And what is the actual reduction of cervical cancer with the vaccine? Those numbers were not discussed either in the NPR story or in many of the articles written about HPV immunization.
The importance of using actual numbers and understandable visual aids in discussing the risks and benefits of medical interventions is illustrated by HPV vaccination. Patients and doctors need real information presented in a format that will facilitate shared decision making. Using relative benefit numbers like 64% and neglecting clinically important outcomes when assessing a medical intervention like HPV vaccine is not helpful to doctors or patients.
What is the actual benefit of HPV vaccination? Out of 1000 people vaccinated, 43 avoid being infected with HPV strains 16,18 (the forms of HPV associated with cancer) compared to 1000 people not vaccinated. Since only chronic infection with HPV 16,19 is a risk factor for cancer, and the CDC estimates that 10% of HPV infection becomes chronic, the following is the actual efficacy of HPV vaccination: Out of 1000 people who receive HPV immunization, 4.3 will avoid chronic infection with HPV compared to 1000 people not vaccinated (see BRCT at the start of this blog)
A larger and unresolved question is even more significant: How many people vaccinated against HPV avoid cervical cancer? This has not yet been studied, only inferred. A review of the literature is minimally revealing. The best evidence is that of the cases of people with chronic HPV, at most 12/1000 may develop cervical cancer. Therefore, using these numbers, it would take 20,000 vaccines to prevent one case of cervical cancer, which translates in the BRCT model to one seat being filled out of 20 theaters.
What about the harm of HPV vaccines? There are many anecdotal reports of adverse reactions. As of 2015, VAERS received 37,000 reports of adverse events and 209 deaths from several million vaccines, none of which could be positively ascribed to the vaccinations. In fact, a compellation of all studies looking at risk found equal harm from HPV vaccine and placebo vaccine, including that of death. So in the BRCT model, the number of seats filled for serious adverse events would be zero.
This way of presenting information allows doctors and patients to have a meaningful conversation about risks and benefits of vaccination. Each patients will interpret these results based in his/her own assessment of personal risk and benefit. Rather than stating that there is a 64% reduction in the risk of acquiring HPV infection with vaccination, as was done in the press, the BRCT model facilitates shared decision making by providing actual information in an easily understandable format.
