What value reassurance?

Colonoscopies are usually performed to exclude serious pathology, such as colonic neoplasm or colitis. Young patients presenting with symptoms of functional gastrointestinal disturbances, such as those fulfillling diagnostic criteria for irritable bowel syndrome, are very unlikely to harbor serious pathology, if no warning features are present. Colonoscopy in this setting has extremely low yield.¹ Therefore, the main reasons for performing colonoscopy are (i) the value of reassurance to the patient; and (ii) the value of reassurance to the treating physician (principally, medicolegal concerns). How valid is this idea of patient reassurance? Can this be evaluated, and either established or debunked as a valid indication?

This is not an abstract idea. In most countries, availability of colonoscopy is limited by resources and often rationed by procedural payments. In many of these countries, colon cancer screening with colonoscopy, or faecal occult blood testing followed by colonoscopy, is expanding, stretching these limited resources. Physicians who perform colonoscopy for functional bowel disorders may expect to be challenged on the point.

Reassurance though is an imprecise term, and measuring it is difficult. Investigators usually adopt surrogate measures such as subsequent consultation rates, symptoms, and measures of psychological and physical function.² Most frequently, self-rated scales that assess health-related anxiety (HRA), quality of life, and physical symptoms are employed before and after a health intervention. This approach is supported by evidence that functional gastrointestinal disorders are associated with adverse scores for these self-rating scales, and improvement in symptoms is associated with improvement in scores.^3,4

Health-related anxiety has been reported to improve in dyspeptic patients after endoscopy;⁵ on the other hand, health-related quality of life (HRQOL) did not improve in patients with irritable bowel syndrome after colonoscopy.⁶ In this issue of The Journal of Gastroenterology and Hepatology, Esfandyari and Harewood⁷ report improvement in health anxiety scores after colonoscopy, in a considerably smaller sample than that reported by Spiegel et al.⁶ What could explain these different findings?

The use of HRA in the ‘positive’ studies and HRQOL in the ‘negative’ study might suggest the former instrument is more sensitive to reassurance. On the other hand, HRA may also be more sensitive to situational anxiety. The timing of the first questionnaire was not specified in the Quadri and Spiegel studies; in the Esfandyari study the first health-related anxiety questionnaire was administered immediately before the colonoscopy. As a result, it is reasonable to assume that the high health-related anxiety scores may have had a component of (quite normal) pre-procedure anxiety. The second health-anxiety questionnaire was completed after recovery from sedation, and after reassurance. The significant improvement in overall anxiety scores immediately after colonoscopy therefore may have reflected relief that the procedure (and not coincidentally, the period of fasting) was over, rather than reassurance about the findings. The anxiolytic effects of sedation likely also played a role.

On the other hand, the questions in the Health Anxiety Questionnaire address global anxiety rather than situational (pre-procedure) anxiety. Use of an instrument such as the State-Trait Anxiety Inventory questionnaire may have been useful to address this point;^8,9 state but not trait anxiety has been shown to increase immediately prior to endoscopy,¹⁰ and state anxiety falls after completion of colonoscopy.¹¹

Therefore, in both the Quadri and Esfandyari studies, the fact that reduction in anxiety scores was sustained at 6 months may merely reflect the fact that the baseline was abnormally high. Nor can this weakness be addressed by the inclusion of a control population (no colonoscopy). Another potentially important difference between the Esfandyari study and the Spiegel study is the heterogeneous population of the former, while the latter study assessed patients fulfillling strict diagnostic criteria for the irritable bowel syndrome.

Nevertheless, it might be premature to dismiss the study by Esfandyari and Harewood on methodological grounds; some clinicians may be comforted by the findings and the issue has important resource implications (see the excellent editorial by Lin¹²). In these imprecise areas of medicine, there is no such thing as a ‘definitive’ study; we can, however, expect prospective studies of ‘reassurance’ in defined study populations, where the impact of situational anxiety is deliberately measured and kept separate from global anxiety measures.