Institutional Review Board Assessment-Balancing Efficiency and Quality.

Satisfactory measurements of the quality of institutional review board (IRB) reviews and services continue to be elusive. For evaluative purposes, the review process can be separated into two parts: the administrative functions that support the review board and the review board's decisions. Administrative performance and board decision-making lend themselves to very different measures of quality. In particular, administrative performance is amenable to measures of process efficiency and correctness, while board decisions require a thoughtful consideration of the meaning of quality in the context of the ethical review of research. In the evaluation of administrative process, simple numbers such as mean or median time from submission to determination allow for easy comparison between IRBs, but their use means foregoing the opportunity for nuanced assessment and continuous improvement. The full distribution of measured values would indicate whether the IRB takes longer with complex studies or with certain categories of studies. An analysis of outliers would give the IRB an opportunity to assess particularly problematic areas. While such measures and analyses are not easily standardized or shared, they would be useful within the IRB, and one measure of quality would be whether an IRB had procedures in place for routinely conducting such analyses. In the evaluation of decision quality, at a minimum, IRBs should be held to a measure of the consistency of their decisions. Consistency should be used as an internal quality measure. A potential indicator of quality would be the existence of tools and processes for routinely monitoring the consistency of decisions and requiring clear rationale for inconsistencies. Ongoing quality assessment and continuous improvement in decision-making are likely to require committed resources, but such a commitment will be necessary to provide balance to the impetus to improve efficiency and administrative performance.