Symptoms of possible oncological significance: separating the wheat from the chaff

The solution lies in a collaborative study of symptoms in the community

Symptoms are a common experience for all humans and are rarely synonymous with a particular disease. In the United Kingdom people are often encouraged to consult their general practitioner, and general practitioners are often advised to refer patients to secondary care according to the presence of one or more symptoms. Much of the evidence for this is based on the patterns of symptoms in patients with serious or important illnesses that are encountered by specialist clinicians.1 The discriminant value of a symptom is related not only to the specific properties of that symptom but also to the characteristics of the setting and the population encountered.2 An urgent requirement exists for attendance and referral practices to be based on data generated from epidemiological studies of symptoms in community settings.

Failing to respond to potentially serious illness accurately and efficiently can have adverse consequences for the delivery of high quality health care. Inaccuracy is about not identifying the true disease state resulting in false positive and false negative diagnoses. Inefficiency concerns the inappropriate or excessive use of tests or procedures beyond those needed to make a decision in a timely and cost effective manner.3 For patients with cancer, false negative diagnostic inaccuracies and inefficiencies can have adverse effects on prognosis as well as the nature of the interventions required. Randomised controlled trials of screening for breast cancer and colorectal cancer have shown significant effects on mortality as a result of early detection of cancers. 4 5 In relation to patients with symptomatic breast cancer Richards et al have conducted a systematic review of the available literature in which, independent of the effects of lead time bias, delays of three to six months were shown to be associated with lower survival.6 Furthermore, the EUROCARE high resolution studies have highlighted room for improvement in the pathological stage at which patients arrive at secondary care within the United Kingdom with consequent effects on cancer survival.7 Delayed recognition of cancer as a result of inaccuracy or inefficiency may also lead to increases in distress and disability for the patient, in addition to the eventual need for more radical treatment.8 For example, patients with late stage testicular cancer or cutaneous melanoma need more extensive and aggressive treatment than those with early stage disease.9

In relation to the approach to common symptoms of possible oncological significance in situations where the prevalence of cancer is low, a particular need exists to consider false positive results. Such incorrect decisions may lead to inappropriate or even harmful investigation and treatment in addition to placing a further burden on healthcare resources.

Some attempts have been made to study symptoms of possible oncological significance among patients presenting to general practitioners. In view of the “gatekeeping” role of general practitioners, the primary care setting ought to represent the best denominator for patients seeking help with a particular symptom. However, many of these studies are dogged by problems of selection bias by general practitioners. 10 11 They may recruit only patients whom they would have referred onwards anyway—for example, patients with moderate symptoms of prostatism might be included in a study of such symptoms only if they have also been noted to have an enlarged prostate (selection by test outcome). In other situations the tendency may be to enlist patients with intermediate or greater probabilities of illness (selection by clinical spectrum).12 Patients with comorbidities may be excluded from many diagnostic studies, and the range of differential diagnoses in the study population may consequently not reflect the true situation in primary care. But the over-riding problem of such an approach is that it does not provide any information on the discriminant value of symptoms among people who choose not to consult a doctor or among those who elect to seek help from other primary care professionals such as pharmacists or nurse practitioners.

When patients experience a symptom they need to know what action to take and also to have the confidence that those supplying primary care services will appropriately respond to their concerns. We need to target resources towards generating relevant evidence rather than recycling inadequate or inapplicable evidence. If we continue to expect patients to consult primary care practitioners and for them to initiate onward referral of such patients in accordance with information derived from patients with established disease in hospitals, then we are likely to make little impact on the stage specific spectrum of patients with cancer that are seen in our hospitals.