Conclusions and Recommendations
The issue of screening for thyroid dysfunction has been a difficult and contentious one for several years and, as discussed in previous sections, has been debated in the literature and among those groups who promulgate clinical guidelines. Much of the debate continues, as described in the preceding chapters, because of the absence of critical evidence on (1) the definition and physiological concomitants of early thyroid dysfunction; (2) the effects of early dysfunction on target tissues and organs; (3) the presence or absence of clinical manifestations in these early stages; (4) the natural history of early dysfunction; and (5) the net personal and population benefits and harms of long-term treatment. However, it should be pointed out that the absence of evidence does not mean that important benefits do not exist.
Based on existing data and analyses of Medicare data performed under contract, the Committee first turned to the question about whether there were obvious deficits in access to thyroid function tests and thyroid care in the absence of a policy allowing periodic screening. Based on this evidence, the Committee concludes that current Medicare coverage of thyroid function testing does not impede the timely diagnosis of thyroid disease. At least 90 percent of all Medicare beneficiaries are currently eligible for thyroid stimulating hormone (TSH) testing under current coverage due to the high prevalence of a wide range of symptoms and conditions that are compatible with common manifestations of thyroid dysfunction but are not specific for such dysfunction.
As noted previously, even when excluding the substantial number of Medicare beneficiaries who carry a diagnosis of thyroid disease or are receiving thyroid hormone therapy, there is a nearly universal prevalence of conditions or symp
toms associated with thyroid disease. These include diabetes, hypertension, elevated serum cholesterol or general systemic symptoms such as fatigue, skin problems, undetected cardiac arrhythmias, weight gain or loss, altered bowel habits, and a host of other clinical complaints common in the population, and particularly among persons aged 65 and older. This high prevalence allows TSH testing as a part of nearly all comprehensive clinical assessments under current Medicare coverage policy. Medicare considers TSH testing under these circumstances to be a standard part of clinical evaluations of symptoms and conditions rather than the screening of individuals who are asymptomatic or who have unrecognized clinical symptoms or other abnormalities.
To address the specific question as to whether routine TSH testing is beneficial in persons 65 years and older who have no recognized clinical symptoms or abnormalities related to thyroid dysfunction, the Committee adopted a process for reviewing the evidence similar to the methodology of the United States Preventive Services Task Force, an evidence-based approach that places a premium on published, peer-reviewed, and systematically scrutinized scientific information. In fact, the review materials being prepared for the Task Force were available to the Committee and formed part of the decision-making process. Based on this method for examining the available evidence, the Committee concluded:
Conclusion #1: There is insufficient evidence to recommend periodic, routine screening for thyroid dysfunction among asymptomatic persons using serum TSH levels.
While the evidence is considered in detail in previous chapters, the basic reasons for this conclusion stemmed from several general considerations:
It is uncertain whether asymptomatic persons with abnormal TSH levels but normal thyroid hormone levels actually have some degree of physiologically meaningful abnormalities that would benefit from early treatment in the absence of clinical manifestations. While a few general, small studies suggest that some “asymptomatic” persons with altered TSH levels have detectable physiological or anatomic abnormalities, questions concerning the factors selecting these individuals for study, how these persons relate to defined populations, and whether these intermediate outcomes presage worse overall morbidity and mortality rates compared to appropriate contrast groups leave doubt as to the significance of these findings.
Some of the potentially important consequences of clinical, if not subclinical, thyroid disease, such as altered blood cholesterol and lipid levels and bone density levels, are themselves the subject of recommended, routine clinical screening procedures, and these should be performed as part of a general program of preventive care regardless of a potential relation to possible thyroid dys
function. Conversely, if these abnormalities are discovered as part of this program, then Medicare will cover the thyroid function evaluation.
While some individuals with unrecognized clinical or physiological abnormalities associated with thyroid dysfunction do progress to overt thyroid disease over several years, the rates, timing, and risk factors for this progression are only partly understood. Furthermore, the Committee could find no evidence that would inform the decision about how often to screen for abnormal TSH levels to optimally detect those who might be progressing to overt disease. Undetected progression of subclinical thyroid dysfunction to overt, life-threatening thyroid disease is rare.
Routine TSH screening of asymptomatic persons over 65 years of age may lead to large numbers of persons receiving thyroid hormone therapy, but no randomized clinical trials have been performed that assess the long-term benefits or adverse effects of early treatment of subclinical thyroid dysfunction. Thus, it is difficult to evaluate the net outcomes of long-term hormonal interventions. Population-based samples of patients receiving thyroid hormone therapy consistently show important proportions of these patients to be either undertreated (elevated levels of TSH) or overtreated, in the range of iatrogenic subclinical hyperthyroidism.
Screening with modern, “third-generation” serum TSH assays will detect thyroid dysfunction among older persons with unrecognized symptoms and clinical abnormalities with reasonable efficiency. However, in the absence of evidence on the occurrence of benefits and harms and the incumbent medical care costs during long-term treatment of these patients, it is not possible to determine the long-term costs and the cost/benefit ratio of such a screening program. In Chapter 5, the cost elements for such an initial screening activity and diagnostic evaluation were estimated, as well as subsequent yearly management. Even using these figures, the cost of adding serum TSH screening as a benefit clearly could be substantial. Predicting subsequent costs for a screening benefit would be hampered further by the unpredictability of whether and to what degree clinicians would adopt the screening activities or how rapidly such adoption would take place. Thus, the Committee concluded
Conclusion #2: Given insufficient evidence about the health benefits of a serum TSH screening program, the cost implications for the Medicare pro gram are uncertain.
Because evidence is lacking on the likely health benefits of screening, there is no reasonable basis for estimating whether a screening program would detect thyroid dysfunction more effectively than usual care and, hence, how the costs of treating thyroid dysfunction under the alternative strategies would compare. We do not have an adequate basis for estimating whether there would be any net savings in the costs of treating future consequences of thyroid dysfunction.
Based on these conclusions, the Committee makes the following recommendation:
The Medicare program at this time should not cover screening for thyroid dysfunction as a preventive services benefit. This recommenda tion is based on the lack of sufficient evidence of either net benefit or harm. Additional evidence is required for a definitive conclusion.
The clinical practice of medicine suffers from the lack of a stronger evidence base. Most decisions made by medical practitioners do not lead to results that clearly and promptly demonstrate the wisdom of those decisions. In the case of screening for thyroid dysfunction, if their actions are repeated millions of times, the cumulative costs can become enormous while the net health effects remain unknown. Because of the large number of older persons who possess biochemical thyroid abnormalities, screening and treating for these abnormalities could generate substantial benefit or harm at considerable financial cost. Randomized, controlled trials of TSH screening, pragmatic in approach, could assess the effectiveness of screening and treatment obtained under usual community conditions and consider both health and economic outcomes.