Henry Ford Hospital Medical Journal


Mark W. Shatz


Evaluation of the neuropsychological effects of surgical treatment on cerebrovascular disease is beset by numerous methodological difficulties. These include problems specific to this patient population as well as others inherent in all retrospective studies. Five such problems are described: 1) nonrandomized subject selection; 2) dropout from follow-up; 3) natural history of cerebrovascular disease; 4) effects of hospitalization; and 5) the role of practice effects. This paper examines these methodological problems for their impact on our knowledge and proposes alternative research directions to address their shortcomings.