Thousands of patients around the world suffer harm from unsafe health care. The place of hospital care has been widely discussed in this context. The same is true for factors contributing to the distribution of risks for incidents in hospitals. However, the same cannot be said for unsafe primary care. Despite the great potential for incidents in primary care (where the majority of health care is provided), there are persistent gaps in knowledge on patient safety at this level.

In this issue of Cadernos de Saúde Pública, the article by Marchon & Mendes Junior (p. 1815-35) offers an instigating critical review that expands the debate on advances and impasses in the evaluation of patient safety in primary care. Contrasting with the research output on the theme, the authors examine factors that affect comparability between findings from different approaches, for example: differences in the operationalization of concepts, particularly in the definition of events (i.e., what does a given study aim to measure?); availability of valid and accurate instruments (how to measure?); and the classification of incidents among extremely heterogeneous groups. This synthesis allows reflecting on the theme’s implications for quality of care and the need for more robust research methods. In addition, the identification and understanding of patterns in primary care incidents and contributing factors makes a relevant contribution to the effectiveness of preventive methods. Identifying the processes that generate errors is both necessary and useful.

The theme’s relevance urges us to tackle methodological issues that are not always trivial, involving complexity of care in terms of both the group’s composition and the specific context. Which incidents are most common? What proportion of incidents involving harm can be explained by differences in risk distribution and the severity of the patient’s disease (composition)? What is the contribution of the institutional context (availability of supplies, workload, staff shortage, infrastructure, treatment protocols)? We also need to deepen our understanding of such events based on their frequency over time and their geographic scope. Can such events occur more than once in the same individual? Can the same risk factor produce different events? How does one analyze different events arising from the same risk situation? Distinct answers and findings can be obtained for each question based on the same variables, depending on the research methods employed. Likewise, the scale of the effect of determinants in the occurrence of events does not always coincide with the scale for which the data were captured (patients, health professionals, family members, institutions). Finding the best fit between scales is a common objective. In this sense, analyses limited to voluntary reporting systems can present major limitations.

There is certainly much to learn from various approaches applied to different primary care scenarios. Marchon & Mendes Junior have identified the first questions and paths for such research.

Enirtes Caetano Prates Melo

Associate Editor

Publication Dates