SHBC1455

H.GUO¹, Z.J.HILDON², V.W.K.LOH³, M.SUNDRAM⁴, M.A.B.IBRAHIM¹, W.E.TANG⁵, A.CHOW¹

Tan Tock Seng Hospital¹, Saw Swee Hock School of Public Health², Yong Loo Lin School of Medicine³, National University Polyclinics⁴, National Healthcare Group Polyclinics⁵

Singapore’s healthcare system presents an ideal context to explore processes underpinning decision-making for antibiotic prescribing by primary care doctors in different primary care settings in Singapore.

Thirty semi-structured interviews were conducted with 17 polyclinic doctors and 13 private general practitioners (GPs). Data were analysed using applied thematic analysis following realist principles, synthesised into a theoretical model, informing solutions to appropriate antibiotic prescribing.

Antibiotic prescribing practices are non-standardised due to lack of national guidelines in primary care. Themes contributing to optimal prescribing related to personal valuing of reduction in antimicrobial resistance (AMR) which was enabled further by organisational culture creating and sustaining such values. Trusting patient-doctor relationships, supported by reasonable patient loads, allow shared decision-making enabling optimal prescribing. Transparency and applying data to inform practice underpins all levels of optimal care delivery. 

The VALUE model reinforces intrapersonal Values consistent with prioritising AMR reduction, and Aligning organisational culture to these by leveraging standardised guidelines and interpersonal intervention tools. A focus on Liaison between patient and doctor is crucial, by building inadequate consultation time and props as discussion aids, or quick turnover communication tools in time-constrained settings. Message consistency will ultimately improve trust, helping to enable shared decision-making. Lastly, Use of monitoring data to track and Evaluate antibiotic prescribing using meaning indicators, that account for the role of shared decision-making can also be leveraged for change.

These VALUE dimensions are recommended as potentially transferable to diverse contexts, and as an implementation tool to be tested empirically and updated accordingly.