ABSTRACT
This chapter attempts to provide rules for determining whether two-phase designs are likely to be worth their considerable effort. The relative efficiency of the two-phase design compared to a simple diagnostic survey depends heavily on the cost ratio of the screen to the diagnostic assessment. In an effort to facilitate research both on psychiatric morbidity in primary-care populations and on methodology for identifying undetected psychiatric cases in such populations, the chapter examines statistical issues arising from the design and analysis of two-phase assessment procedures. If the purpose of the screening procedure is to identify a sample of patients from a primary-care population who have a specific disorder, then the optimal two-phase prevalence design will indeed be unappealing. P. Williams reviewed screening in the context of mental illness and primary care. He expressed optimism, however, that psychiatric screening measures might be useful for case-finding in primary-care settings.
