Investigating inpatient insulin prescribing practice and intervention use in UK hospitals: a mixed methods study and realist synthesis

Insulin prescription errors are a common, costly problem in NHS hospitals in the United Kingdom (UK), and are a consistent source of unintentional harm to inpatients with diabetes. Prompted by the lack of large-scale quantitative and qualitative research conducted in the UK in this area, this research aimed to investigate inpatient insulin prescribing practice and the current use of interventions designed to prevent insulin prescription errors in UK hospitals.

Following the conduct of a systematic review of interventions designed to reduce inpatient insulin prescribing errors, a questionnaire was developed, validated, and used as part of a cross-sectional survey of chief/diabetes pharmacists representing every NHS acute trust in the UK. Information was captured regarding the use, functionality and perceived effectiveness of insulin prescribing systems and interventions. Follow-up qualitative interviews were conducted with survey respondents to further analyse their experiences and opinions regarding insulin prescribing practice and intervention use. Realist synthesis was then undertaken to further understand how insulin self-administration policy interventions worked in different contexts. A participatory health research approach was taken throughout the research to maximise relevance and impact of the research for end-users, and a combination of middle-range theories were used throughout the research to aid the transferability of findings.

Ninety-five hospital trusts responded to the survey (54%), 18 of whom participated in follow-up interviews. Results indicated that a wide range of prescribing systems with varying functionalities were in use, along with a diverse range and combination of error-prevention interventions. Intervention use was positively associated with the availability of specialist diabetes pharmacists (P=0.002), who worked with diabetes teams to improve insulin safety in their organisations. Although mandatory insulin training was used by only 46% of trusts, it was perceived to be very effective at preventing errors. This was due to the perceived lack of understanding and confidence prescribers have with insulin, but the difficulties associated with accessing staff to deliver training in hospital. The insulin passport was perceived to be ineffective and only used by 31% of trusts on account of faults in its design, incompatibility with existing systems, and unreliable use by patients. Self-administration policy interventions were used by 63% hospitals and were described as salient but complex to implement; The use of realist synthesis generated 10 programme theories to further explain how they work, for whom and in what circumstances. Key contexts, outcomes and mechanisms were identified, including hierarchical and blame cultures, patient empowerment, control, shared decision-making, and clarification of roles.

As the first study to investigate insulin prescribing practice and intervention use at a multi-organisational level in the UK, this research contributes to the literature by describing and explaining how interventions may be used to improve the care received by inpatient with diabetes. Actionable findings are included that may help hospitals and policymakers implement interventions that are most likely to result in successful outcomes.