The recent publication draft of the United States Preventative Health Service Task Force (USPHSTF) recommendations for breast cancer screening represents a step backward rather than forwards in empowering women to make an educated choice regarding the advisability of mammograms. This is not a reactive position for radiologists that provide such services but rather an evidence-based rejection of distorted scientific analysis and repeated contradictions.
Following the last set of similar recommendations in 2009, that was intentionally disregarded in providing recommendations to women under the Affordable Care Act, the Institute of Medicine drafted guidelines for future reports such as the recent USPHSTF statement. The Institute, recognizing the shortcomings of the methods used in the 2009 report, stated that no such committee should be constituted without experts on the panel that have knowledge and experience in the field. The current USPHSTF abandoned that advice and the work of the committee excluded, by design, such expertise.
The technical errors of the draft are complicated and beyond the scope of this summary. A few salient points can be made to expose the lack of validity of the USPHSTF proposals.
The panel did recognize a benefit from screening, but considered the benefit too small to recommend a rating beyond “C.” While the panel indicates it desires to provide women a choice, the “C” recommendation in fact preempts that choice by advocating that screening mammography not be funded for women under 50 or over 64 and that even this group is entitled to only a screening mammogram every two years. Studies have demonstrated that screening every two years instead of every year invites the detection of larger cancers when discovered, worse prognosis, and fewer lives saved.
Using short trends to justify their conclusions, the panel found insufficient evidence to avoid significant decrease in advanced cancers and lower mortality rates. Trend analysis is both deceptive and substandard when long term data is available. Rather than the few percentage points recognized by the panel for decrease mortality rates, the Unites States has in fact observed a 31% decrease since the 1990s, when increased incident rates of breast cancer are taken in to account, an account ignored by the panel.
Two types of large scale data bases are available for review. Randomized controlled trials look at populations to see if those undergoing screening do better than those who are not screened. Only one study has compromised this perspective, the Canadian trial which inexplicably placed women with advanced cancer into the “screening group” and employed mammography technique deemed unsatisfactory by outside international observers and even Canadian physicists. When even this study is included, these trials, based on old techniques, show a clear 20% benefit for women screened. Using current mammography techniques, this rate is likely much higher.
The other source of large data involves “service screening” where smaller controlled populations are observed. Decreased mortality rates of 43% from breast cancer were observed after 15 years in Norway and between 31 and 45% in Sweden. Swedish investigators have one of the longest follow up studies ever done at 29 years. Of particular note is the in British Columbia where service screening is performed in a more deliberate manner than the national screening trial noted above, the mortality reduction exceeded 40% for women under and over the age of 50. When the USPHSTF panel looks at
trends of shorter duration, it is not surprising that they miss the forest through the trees.
As a social policy it is important for any panel to reckon how many women must participate in order to show a benefit. The Panel looked at the number of women invited to screening to determine the ratio of screened women to cancer found. This is a notoriously inaccurate measure, because what investigators really want to know is how many were screened and how many lives saved. Thus the Panel concludes that between 377 and 1300 are need to be invited for one life saved where the more accurate Euroscreen report of 2012 calculated, based on actual attendance, which 111 women need to be screened to save one life.
The panel was concerned with over-diagnosis, where detected cancers would have no impact on lives saved. First, this concern is obviated by the results noted above. Second, the diagnosis of ductal carcinoma in situ, non-invasive cancer has raised concerns that there may be overtreatment, but this is not the same as over diagnosis. One does not know the nature of a malignancy until it has been detected, biopsied, and studied under the microscope.
Moreover, long-term studies indicate that nearly half of the cases of untreated ductal carcinoma in situ progress to invasive carcinoma.
There are innumerable other faults in the analysis of the recent USPSHTF but the purpose of this short essay is to expose the obvious lack of science behind the recommendations. Like those who deny the majority of scientific evidence showing climate change, the panel members sought misleading information to arrive at conclusions than cannot be justified by a wider view of the science. Indeed, new technologies are already showing evidence of better cancer detection and few imaging evaluations and biopsies required to find those cancer.
Read the Task Force’s Recommendation here http://www.uspreventiveservicestaskforce.org/Page/Document/RecommendationStatementDraft/breast-cancer-screening1