The doctrine of early intervention

When assessing interventions, context is more important than doctrine

The principle of early intervention in a disease process is intuitively appealing. Clearly, where early intervention is synonymous with primary prevention, the principle is sound. You need only to look at the fall in recorded cases of poliomyelitis or meningitis caused by Haemophilus influenzae type B after the introduction of their respective vaccines for confirmation of the value of appropriate early intervention.1 But when early intervention means secondary prevention the issue is less clear. We know now that a combination of exercise, diet, and intensive treatment with insulin can delay the start of microvascular complications in patients with insulin dependent diabetes mellitus.2 On the other hand, considerable debate surrounds the treatment and surveillance of patients diagnosed with ductal carcinoma of the breast,3 prostate cancer,4 or malignant melanoma.5

A recent editorial in the New England Journal of Medicine appraising a paper by Weinberger et al6 presented what its author called the “heretical view” that the doctrine of early intervention might be wrong.7 The aim of Weinberger et al's multicentre study, based on data from Veterans' Affairs centres across the United States, was to assess whether early intervention in patients who had recently been discharged from hospital with diabetes, chronic obstructive pulmonary disease, or congestive cardiac failure could prevent their readmission to hospital. The intervention, which comprised close follow up by a nurse and primary care physician at the local clinic and in the patient's home, produced the opposite effect: patients who received the early intervention had higher monthly readmission rates and spent longer in hospital than patients in the control group receiving usual care. The author of the editorial observed that “closer scrutiny of the patients simply led to more medical care and perhaps to harm”—although there was no evidence that the patients were harmed by the care they received. Is the conclusion of the editorial—that the doctrine of early intervention may be wrong—correct?

Weinberger et al's study was really an exercise in tertiary prevention, and their use of the term “early intervention” in this context is misleading. The patients in the study were in the end stages of their diseases—half of those with congestive cardiac failure were classified as grade III or IV, a third of the patients with diabetes had end organ damage, and a quarter of those with chronic obstructive pulmonary disease required domiciliary oxygen. Two thirds of the intervention group were classified at the beginning of the study as at medium or high risk of readmission, and nearly 8% of the whole study population died during the six month study. Accordingly, the failure of the intervention to prevent readmission may reflect more the severity of the study population's illness than the inadequacy of the strategy.

The extensive burden of illness in the patients in this study, coupled with the large amount of direct medical care and proactive telephone advice they received, created ample opportunities for the initiation of what have been described as clinical cascades.8 Here, a single seemingly innocuous clinical characteristic, like a patient's or even health professional's anxiety, initiates an unstoppable series of events resulting in more and more medical care being offered. Even so, the analysis offered by the authors of this study says nothing about the appropriateness of the interventions and admissions, which may have been entirely justifiable.

The exclusion criteria used in this study pose some questions about its generalisability. Firstly, the Veterans' Affairs system offers medical care to a predominantly low income population, including those who are homeless and chronically mentally ill.9 Less than half of those initially eligible to take part in the study actually participated, roughly 10% of the screened population. Nearly a third of those who were eligible to take part declined to do so because they did not wish to be assigned to a new doctor whom they did not know. This makes the study less relevant to doctors in Britain and other countries where patients are discharged back into the care of their own general practitioner.10

Should doctors support the doctrine of early intervention, or any other doctrine for that matter? A doctrine is defined as “the principles or dogmas of a scientific school.”11 when an intervention is being assessed, is it not more important to assess whether the activity is appropriate and effective? Any intervention should be based on evidence that is relevant, robust, and reliable; it should meet the patient's needs, and be tailored to the context in which those needs arise. Weinberger et al's paper shows that this vulnerable patient group consumed a large amount of medical resources: what it does not show is whether these resources were justified and effectively consumed.