Self-assessment and subsequent external review as an effective model for improving drug use for elderly in primary care: A descriptive study of SÄKLÄK2

Rationale, Aims, and Objectives: Primary health care often has overall responsibility for elderly patients and their medication treatment. This is a challenging task due to the issue of multiple caregivers, different systems for documentation and multimorbidity among the elderly. The multiprofessional project SÄKLÄK2 was developed to raise drug safety in Swedish primary health care, and this study aimed to assess whether the action agreements that emerged from the model were effective enough to potentially improve drug use in elderly patients. Method: The SÄKLÄK2 project was conducted during 2016. A total of 12 primary health care centres (PHCs) in three counties participated in the project. The intervention method concerned the management of the PHCs and comprised self-assessment, peer review, feedback, and written agreements for change, which were concluded between the reviewers and the manager of the PHC. The action agreements were analysed using summative content analysis (sorted under predesigned categories) and were also assessed as fulfilled, initiated, or not fulfilled within the follow-up time. Results: The importance of securing an accurate medication list was reflected in the number of action agreements in this area. Other prominent improvement areas were follow-up of prescriptions, pharmacogeriatric further education, and cooperation between caregivers. Action agreements to facilitate for the patient to be able to handle his/her drugs were also common. The great majority (88%) of the action agreements were implemented or initiated within the follow-up time. Conclusions: The SÄKLÄK2 intervention model is considered effective in setting up a variety of relevant measures to improve drug safety in primary health care, which are possible to implement in the near future. Hence, the model is regarded as effective and should therefore be offered in a wider context.