Objectives In many healthcare systems, policymakers wish to transform everyday practices and ensure higher quality through digitisation and datafication. Digital health technologies, such as digital administrative systems and health apps, are increasingly introduced to support general practitioners (GPs) and empower patients. However, such initiative can result in unintended outcomes, such as overdiagnosis, errors and wasted resources. We investigated a particular example of the political ambition to make healthcare ‘data driven’, namely a new Danish national quality development program for GPs where doctors are placed in so-called ‘clusters’. In these clusters, GPs are obliged to assess their own and colleagues’ clinical quality with data derived from their own clinics – using comparisons, averages, and benchmarks. We investigated how the GPs use self-produced data for quality assurance about themselves and colleagues. Here, we present the results and their implications in relation to overdiagnosis.

Method As a part of a recently finished Ph.D. called Datafying Diagnosis: Data Work in General Practice, we conducted in-depth, semi-structured, one-to-one interviews with 23 Danish GPs. We analysed the interviews by drawing on Science and Technology Studies (STS) literature, including a concept of data developed by philosopher of science, Sabina Leonelli. In the perspective of overdiagnosis, we explore how GPs understand these data, and what makes them trust – or question – a data analysis. The GPs describe how they do change clinical practices based on these discussions of data. So, when and how do data for quality assurance come to influence their perceptions of quality?.

Results The GPs sometimes considered their immediate interpretation of data as so obvious that they would let the data overrule their own clinical, critical judgements. This can create errors and overdiagnosis if data are invalid or incorrectly interpreted. Mostly, the GPs used data in more sophisticated ways in which the data functioned as support for claims that did not require certainty of the validity of the data. The three main claims were the need to discuss with colleagues, to install a change without having a defined goal, and to articulate reflections on individual practices.

Conclusions The findings show how difficult it can be for the GPs to understand their own data and to use it to evaluate clinical quality. This raises questions of the feasibility of the current great and global ambitions of creating, sharing, and reusing health data between very different contexts. The findings have implications beyond the clinical quality of GPs and cluster collaborations as they show how ‘data driven’ becomes enacted in healthcare practice. The ambitions to be ‘data driven’ involve the risk of being driven without sufficient reflection, which can lead to overdiagnosis as well as other unintended consequences.