Making health care coverage depend on the existence of valid, applicable re
search data and positive cost-effectiveness analyses, as managed care contr
acts are beginning to do, is particularly problematic for children. Because
of research challenges specific to children, there are relatively few pedi
atric data and analyses required under such evidence-based coverage standar
ds. It is too soon to expect major increases from federal efforts to stimul
ate pediatric health care research. But absence of requisite evidence would
entitle a managed care organization or other decision maker to deny covera
ge on the basis of unproven, negative assumptions about an intervention. In
general, population-based evidence is an incomplete basis for decisions on
coverage for individual patients. Cost-effectiveness analyses are not stan
dardized and may be biased. Purchasers of managed care and policy makers sh
ould understand the limits of evidence-based coverage standards. Other uses
of evidence may contribute more to systemic improvements of health care.