The Case of Distributed Accountability. an Ethnographic Study on the Effects of Performance Measures on Accountability Relationships

Background and purpose: This paper examines the effects of performance measures on accountability relationships in a network of nonprofit health clinics. Performance measurement is routinely used among nonprofit organizations worldwide to demonstrate accountability. From the literature, the effects of performance measures on organizations may be classified in three ways: they facilitate organizational improvement and learning; are decoupled from the organizational processes and practices they are supposed to bear witness to; and/or colonize organizational processes and practices into images of auditable measures, often to the detriment of the broader goals of the organization. Depending on the a priori stance the researcher takes, performance measures may be portrayed as either constructive, ceremonial and/or controlling, each with ensuing implications for accountability relationships. However, the effects of performance measures are ultimately a question of empirics rather than of normative presuppositions.

Methods: Using a single case study and drawing on over 350 hours of fieldwork in a network of community-based health clinics in Midwestern United States, I examine the effects of a single performance measure – whether the patient receives depression screening – that is tracked by a single activity – clicking a box in the electronic medical records that indicates the completion of depression screening, on accountability relationships. To this end, I follow three organizational activities. I participate in monthly quality improvement meetings in which network leaders talk about performance measures; participate in fortnightly meetings among leaders of a single health clinic as they devise and implement initiatives to improve performance in the depression screening measure; and shadow medical assistants who do the frontline work of screening for depression.

Results: Two arguments are drawn. First, in complex, contemporary nonprofit organizations, accountability relationships are distributed. Even for a single measure, the task of measuring performance is not the responsibility of a single group of actors who are held accountable to a single external actor – in this case, the federal government who provides funds based on performance – but is held together by multiple accountability relationships. For example, operations personnel determine why the screening rates are low, process improvement personnel audit the accuracy of the medical records, and primary care physicians’ performance incentives are tied to depression screening rates. Consequently, the effects of performance measures are at least as far-reaching as those accountability relationships, implicating senior leaders, managers, and frontline workers. Second, performance measures are productive – they reorganize human labor by bringing together disparate actors into relation, change organizational processes through reshuffling the layout of workspaces and resequencing frontline workers’ tasks to reduce errors in screening for depression, and produce a slew of paperwork such as standard operating procedures along the way.

Conclusions and Implications: Given the increasingly complex nature of nonprofits worldwide, I argue that the study of performance measurement requires detailed ethnographic scrutiny in order to understand their distributed and potentially far-reaching effects. Only once we have located the distribution of these accountability relationships can we make claims on the effects of performance measures on nonprofit organizations and their human constituents.