Distortions in Risk Assessment
Failure to adjust correctly for severity will invariably distort the assessment of risk in pharmacoepidemiologic studies. Drugs are not prescribed to patients at random. The decision to prescribe a given agent to a given patient is based on assessment of a complex of factors which the prescriber may not articulate. Drugs with certain real or perceived characteristics will be prescribed to patients with certain real or perceived characteristics. This process has been described by Urquhart as channelling. Patients who are perceived as sicker will be prescribed drugs perceived as stronger; patients with previous drug intolerance will be prescribed agents believed to be better tolerated. Moreover, the outcome of the patient’s illness is determined primarily by the underlying disease and only secondarily by treatment.
KeywordsPharmacoepidemiologic Study Healthy Work Effect Poor Working Condition Respiratory Mortality Asthma Death
Unable to display preview. Download preview PDF.
- 1.Crane J, Pearce NE, Flatt A et al. (1989) Prescribed fenoterol and death from asthma in New Zealand 1981–1983: case control study. Lancet I: 917–922Google Scholar