Validation of Outcomes Through the Analysis of the Process-Outcome Relation: Limitations
International Journal for Quality in Health Care
Validation of outcomes of different levels of neonatal care through the analysts of the relationship between process and outcome of care was attempted as part of an effectiveness stndy and quality assurance programme. Adequacy rates and outcome measures were calculated for second-and third-level units and centres according to the babies' level of risk. Differences in adequacy rates were apparently influenced by the differential distribution of babies by care level sectors among maternity
... ng maternity centres. Hie comparison of mortality rates for "treated" and "untreated" subgroups did not show higher mortality rates for the "untreated" subgroups. This finding could be related to the phenomenon of "confounding by indication". Consistency of differences in average rates of adequacy with differences in general outcomes was observed for most subgroups analysed; lack of consistency could be related to low effectiveness. Indications of low effectiveness related to secondary and more so to tertiary level technologies were observed. The study approach was found more useful for comparisons made at the level of maternity services as a whole, where "confounding by Indication" does not operate than at the level of sectors.