Those of us working in the life sciences suffer in a world of probabilities, uncertainties, and incomplete evidence. Students of the “harder” sciences, physics and math, may be right in their assessment that the only immutable truths are in pure mathematics.
On October 7th 2011, the U.S. Preventive Services Task Force, which advises the government on health prevention measures, downgraded its recommendation on prostate cancer screening to a “D,” which means it recommends against the service because “there is moderate or high certainty that the service has no net benefit or that the harms outweigh the benefits.” The task force based this on evaluation of the current evidence and determined that the small reduction in cancer deaths was overshadowed by serious adverse events related to the testing and diagnostic procedures triggered by an abnormal PSA.
The same panel caused caused a huge dust up in 2009. The USPSTF, using mathematical modeling, recommended against mammographic screen for women age 40-49, and for reducing the frequency of screening for older women. They estimated that, after a decade of screening, in a cohort of 1900 women aged 40-49, the following outcomes would be provoked:
1. 1330 call backs for reassessment
2. 665 breast biopsies
3. 8 cancers diagnosed
4. life saved
For an individual woman age 40, annual mammography would provide a 0.42% chance of having breast cancer diagnosed, and only a 0.05% chance of having her life saved by screening (1 in 20). Current estimates are that while breast cancer mortality has dropped 25-30% in the past 20 years, only 1/3 of that drop can be attributed to mass screening. The remainder of that reduction comes from improvements in cancer treatments, something I mentioned in an earlier blog.
The improvements in breast cancer treatment accrue not only from better chemo and radiation, but from the demise of the disfiguring and morbid radical Halstead mastectomy. Between 1895 and the 1970s, almost 90% of women treated for breast cancer in the USA underwent a Halsted radical mastectomy. The belief in radical surgery was so ingrained in our cancer community, a field dominated by surgeons, that a trial comparing breast conservation methods to radical mastectomy could not be done here. The first such study was done in Italy by Umberto Veronisi, another of my heroes, who continues to challenge the conventional wisdom about this disease. His most recent work has demonstrated that sentinel lymph node biopsy is adequate for staging breast cancer and, if negative, obviates the need for more extensive completion axillary lymph node dissection.
And so it is with PSA screening. A typical response to the recommendations of the USPSTF is to assert that PSA screening saves lives, and then trying to link PSA screening to the 30 percent decrease in death rates since 1991, a claim surprisingly similar to the claim that mammography save lives. But again, other factors may be in play and studies have found that after 10 years and three or four rounds of screening, PSA does not provide any significant improvements in mortality. Like mammography, PSA carries a substantial false positive rate of 12 to 13 percent.
In response to the recommendations of the USPSTF, other agencies and advocacy groups were fast to issue contrary and contradictory recommendations. In response to the breast cancer screening recommendations, the American Cancer Society issued a policy recommending continued annual screening for women 40-49. The American College of Ob-Gyn issued a similar directive.
Just as string theory, dark matter and neutrino oscillation have disrupted the overarching fundamentals within the conventional Standard Model of Particle Physics, so changes in the data available for epidemiological modeling may disrupt our conventional medical wisdom.
But old theories die hard. In the mid 19th century, Ignatz Semmelweis proposed asepsis to counter the extraordinarily high maternal mortality rates in the First Clinic at the Vienna General hospital, where medical students attended deliveries coming straight from the dissecting room. He introduced treating all items that came in contact with patients with chlorinated lime (basically household bleach) and, within weeks, mortality dropped 90%, the rate in April 1847 being 18.3%. After hand washing was instituted in mid-May, the rates in June were 2.2%, July 1.2 %, and August 1.9%.
In the following year, unprecedented zero mortality was evident in two of the months in the study. Yet in hospitals today, health workers continue to ignore and resist infection control measures including hand washing. Neck ties, wrist watches, lab coats and rings all have been implicated in hospital acquired infections. But many doctors refuse to give up their white coats, a traditional symbol of their authority and status. Some go so far as to claim that the “white coat” enhances patient satisfaction and security (and risk of multiple drug resistant staph). Can you believe any physician would say something like this 165 years after Semmelweis’s experiment?!
And vested interests often want to maintain the status quo. Surgeons want to do surgeries. Cancer groups want donations. Gynecologists want to avoid law suits. Radiologists want to read scans.
“One reason why mathematics enjoys special esteem, above all other sciences, is that its propositions are absolutely certain and indisputable…”
Certainty in the life sciences has a finite shelf life. New methods and new discoveries have radically changed the landscape in the past and will continue to do so in the future.