Portuguese Primary Care physicians response rate in surveys: A systematic review

• Autor(a) principal: Basílio,Nuno
• Data de Publicação: 2018
• Outros Autores: Cardoso,Sara, Nunes,José Mendes, Laranjo,Liliana, Antunes,Maria da Luz, Heleno,Bruno
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Revista da Associação Médica Brasileira (Online)
• Texto Completo: http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0104-42302018000300272
• topic: Physicians; Family Practice; Primary Health Care; Surveys and Questionnaires; Portugal

Summary Introduction: Surveys are a useful tool in primary care. However, low response rates can introduce selection bias, impairing both external and internal validity. The aim of this study was to assess the average response rate in surveys with Portuguese general practitioners (GPs). Method: We searched the Medline, Web of Science, Scopus, Embase, PsychInfo, SciELO, IndexRMP, RCAAP, Revista Portuguesa de Medicina Geral e Familiar, Acta Médica Portuguesa and the proceedings of conferences of general practice from incepton to December 2016. We included all postal, e-mail, telephone and personal surveys to primary care physicians without language restrictions. We did not assess risk of bias of included studies, since the main outcome was survey response rate. We performed planned subgroup analyses of the use of monetary incentives, the use of non-monetary incentives, survey delivery modes and prior contact with participants. Results: A total of 1,094 papers were identified and 37 studies were included in this review. The response rate in surveys done to Portuguese GPs was 56% (95CI 47-64%). There was substantial heterogeneity among included studies (I2=99%), but subgroup analysis did not explain this heterogeneity. Conclusion: Consistent with other published studies, the average response rate in surveys done with Portuguese GPs was 56%, with substantial variation among studies. Use of monetary incentives, one of the most effective strategies to increase response rates, was not present in any of the included studies.