For more than 25 years I have used the Head Start program to teach medical students data analysis, how to evaluate the therapeutic efficacy of treatment regimens, and how belief commonly trumps evidence. This process usually followed the recommendation of a course of treatment by a consultant to a patient who was on a general medicine service that was under my direction. I sometimes expressed doubt as to the efficacy of the recommendation whereupon a student or resident would explain that the treatment at issue was standard medical practice. The following brief exchange would then take place:

NK: Is Head Start a good program?
Student: Yes
NK: Why?
Student: Because it helps children.
NK: How do you know that it helps children?
Student: [Silence]
NK: Go to the library (later on Google) and find the evidence that it helps children.

After a short while the student would tell me that that after a interval of a few years – say by the third grade – that there was no discernible difference between children who had been in Head Start and those who were otherwise the same but who had not been in the program.

NK: So does Head Start work?
Student: No
NK: Should we cancel it?
Student: [Silence]


Here’s the 2010 Head Start Impact Study prepared by HHS. Its Executive Summary is 39 pages long. You’ll need a larger hard drive to download the whole thing. Here’s a key passage:

However, the benefits of access to Head Start at age four are largely absent by 1st grade for the program population as a whole. For 3-year-olds, there are few sustained benefits, although access to the program may lead to improved parent-child relationships through 1st grade…

Here’s a passage from the 2012 report:

The lasting effects of Head Start and early childhood education in general on children’s outcomes have been the focus of much study. Considering only outcomes through early elementary school and middle childhood, results for the HSIS cognitive outcomes are in line with other experimental and non-experimental early education studies. Non-experimental Head Start studies showed initial positive impacts of a roughly similar magnitude to those found in the HSIS that dissipated as the children entered early elementary school (Currie & Thomas 1995; Garces, et al., 2002; Ludwig & Phillips 2008; Deming 2009). Moreover, recent longitudinal data from the experimental evaluation of Early Head Start (Vogel, et al., 2010) showed a similar pattern of early positive impacts that were not sustained into elementary school. Experimental results from the HighScope Preschool Curriculum Comparison study found negligible differences between study groups in cognitive and academic outcomes in the first decade of study (Schweinhart & Weikart, 1997). Similar conclusions about the size and lack of persistence of early impacts were reported in a recent broader meta-analysis of early childhood interventions (Leak et al., 2010). However, as we discuss later, some studies, including those that did not show differences in elementary school, reported finding positive effects later in adulthood. Although the underlying cause of the rapid attenuation of early impacts is an area of frequent speculation, we don’t have a good understanding of this observed pattern. All we can say is after the initially realized cognitive benefits for the Head Start children, these gains were quickly made up by children in the non-Head Start group.

In other words, it’s very hard to show that Head Start has any lasting effect on children who went through the program. It’s always been virtually impossible to show a beneficial effect, but the government can’t bring itself to say so and the public is impervious to argument and uncritically loves the program so:

In time the Rockies may crumble, Gibraltar may tumble
They’re only made of clay
But Head Start is here to stay

But what does this have to do with medicine? Here as in Head Start intentions and outcomes are often felt synonymous. Consider the study Endovascular Treatment for Acute Ischemic Stroke in the March 7, 2013 New England Journal of Medicine and the editorial Endovascular Treatment for Acute Ischemic Stroke — Still Unproven which accompanied the study. Intravenous tissue plasminogen activator (t-PA) is the only proven reperfusion therapy for acute ischemic stroke, and its clinical effectiveness is critically time-dependent. Other techniques that might open clogged arteries were developed. Thus thrombectomy devices were used in combination with t-PA and the FDA approved these devices under 510(k) clearance, which does not require proof of clinical efficacy. Later, Medicare provided reimbursement for these procedures, leading to widespread use of the devices despite the absence of evidence establishing their efficacy. “The trial [linked above] showed similar safety outcomes and no significant difference in functional independence with endovascular therapy after intravenous t-PA, as compared with intravenous t-PA alone.”

Hundreds of Stroke Centers have opened in the United States which provide endovascular treatment for strokes despite no evidence for the effectiveness of such treatment. Will this study in the NEJM change what they offer? I doubt it as there is always a rationale for believing that an unproven therapy may yet be found effective. Remember this unproven treatment is paid for by medical insurance and Medicare. If we stopped paying for unproven treatments we could probably shrink the medical workforce in half and solve all our problems paying for medical care. But the emotional concession required for such a move is beyond consideration. It would be far more painful than bankruptcy. While money makes this uncertain therapy possible it’s not what drives it. The motive force is that bluntest of instruments – the desire to do good. This urge is fine and laudable, but only if married to intellectual rigor. Without this rigor it can crush and mutilate those it intends to help.

Finally, another unintended consequence. The NEJM of March 14, 2012 presents Risk of Ischemic Heart Disease in Women after Radiotherapy for Breast Cancer. As if getting breast cancer were not bad enough the study concludes: Exposure of the heart to ionizing radiation during radiotherapy for breast cancer increases the subsequent rate of ischemic heart disease. The increase is proportional to the mean dose to the heart, begins within a few years after exposure, and continues for at least 20 years. Women with preexisting cardiac risk factors have greater absolute increases in risk from radiotherapy than other women. It’s not surprising that radiating the heart incidental to radiation treatment of breast cancer causes cardiac injury and that the risk of this injury injury is directly proportional to the amount of radiation applied and, furthermore, that women with risk factors for heart disease are more impacted by this treatment. The Journal also published an editorial in this issue, The Cardiovascular Perils of Cancer Survivorship, which points out that other forms of cancer therapy have cardiotoxic effects.

None of this should be interpreted as a call for therapeutic nihilism, but only as a plea for caution and thought. The practice of medicine is not as easy as it looks.