Screening programmes seem so simple in their objectives. How, theoretically, could early detection not be beneficial? How is it possible that taking asymptomatic people and classifying them as ‘likely’ or ‘unlikely’ to have a disease be anything other than good? How could treatment before symptoms develop not be better than treatment after symptoms develop?
Surprisingly, though, we struggle with screening programmes. As with many issues in healthcare and health policy, we struggle because we have conflicting actors who all bring different viewpoints to this debate. And we struggle with screening programmes for a variety of other reasons too – most notably of which is the fact that screening programmes have led to increases in the diagnosis of diseases but haven’t necessarily led to decreases in mortality. And that is the real reason we struggle with screening programmes: because they’re not always successful.
Recently, the United States Preventive Services Task Force (USPSTF) sought public opinion on prostate cancer screening and also reviewed the evidence available.The USPSTF’s draft conclusion was that for men 55-69, the decision to receive PSA-based screening should be between the clinician and the patient and include a complete understanding of all potential harm as well as benefit, and incorporate the patient’s values and preferences (C grade). For men 70 and older, the USPSTF recommends against PSA-based screening because the potential benefits do not outweigh the harm (D grade).¹ Clearly this is an admission by a panel of experts that, based on a systematic review of the evidence, there is, at best, no more than a moderate level of certainty that the net benefit of screening is small. In fact, in certain age groups of men, the harm outweighs the benefit.
So, what do we need for a screening programme to be successful? Well, we need a suitable disease. This may sound odd, but not all diseases lend themselves to screening. Certainly, orphan and ultra-orphan diseases are not suitable for screening programmes. Not because they don’t have serious outcomes but because their prevalence is too low. And, critically, a suitable disease must be one in which treatment before symptoms emerge is better than treatment after symptoms emerge. Otherwise, the costs to the system and the burden to the patient render the screening inconsequential. Good examples of screening programmes where the early detection, diagnosis and treatment have favourable outcomes are glaucoma, hypertension and breast cancer, to name a few. Scenarios in which early screening makes very little difference are diseases like ovarian and pancreatic cancer. It’s not that we don’t want to detect these diseases early. It’s that these diseases are highly fatal and early detection before symptoms develop has little impact on mortality.
What else? We’ve got our suitable disease. Well, now we need a suitable test. A suitable test is characterised by its simplicity and ease of administration, its low cost and the fact that it is well-tolerated by the patient (ie not too invasive). But most of all, a suitable test must have very high validity. In other words, the test must do what it says it will do. It must have high sensitivity and specificity. If an individual tests positive, the individual should truly have the disease and if the individual tests negative, the individual should be disease-free.
OK, so now we’ve got a suitable disease and a suitable test. Now, we put them together to develop a programme and ask ourselves whether the programme is, indeed, feasible and effective. Can we do it? And does it matter if we do it? On the question of feasibility, we need to know if we can identify enough cases and can those cases be reliably moved forward to diagnosis and treatment? On the question of effectiveness, the central point is about actual outcomes. If we have a suitable disease and a suitable test, and the test is feasible but we don’t pick up early disease or have an impact on mortality, then we can say the screening programme is not effective.
So, there you have it. It’s not easy. The USPSTF has a list of 98 published recommendations for primary care practice. Everything from cancer screenings to preventive medications like folic acid for the prevention of neural tube defects. Perhaps the greatest benefit, though, of screening programmes is that we get people ‘in the system’ on a regular and routine basis. People who otherwise would not be seeing their doctors are now doing so. And maybe this unintended consequence of screening is how we ought to be looking at it.
* Some of the content of this article is based in part on the writings and opinions of Julie Buring, ScD at Harvard T.H. Chan School of Public Health.
¹ Draft Recommendation Statement: Screening for Prostate Cancer. US Preventive Services Task Force. April 2017. www.screeningforprostatecancer.org
