A service evaluation of subcutaneous insulin administration at a large hospital

Abstract

Abstract Introduction One in six adult patients in hospitals have diabetes1. According to the National Diabetes Inpatient Audit (NaDIA)1, around 40% of these patients are treated with insulin, a high-risk drug used to regulate blood glucose levels. Worryingly, 40% of patients treated with insulin experience an insulin error whilst in hospital1. Although the annual NaDIA shows consistent improvement in patient care since 2011, insulin error rates have not reduced significantly. Local data from a large NHS Trust mirrors the national picture2 and this study was part of the Trust’s 2018 NaDIA Quality Improvement Collaborative work to improve patient care specific to subcutaneous (SC) insulin administration. Preliminary work involved the development of a process map of the use of SC insulin with multidisciplinary teams from two wards at the Trust, reflecting how teams think SC insulin is used (work-as-imagined, WAI). Aims 1) To observe and describe the task of administering SC insulin (work-as-done, WAD), 2) To compare WAI and WAD relating to the task of administering SC insulin. Methods Non-participant observation of SC insulin administration took place over a 2-week period (Nov-Dec 2018), during lunch and dinner times, in the same two wards at the Trust where preliminary work took place. A ward pharmacist identified up to 16 opportunities for observations. Data were collected using a piloted observation template. Data from each observation was coded to identify tasks performed, using MS Word. Tasks were further analysed to develop a Hierarchical Task Analysis (HTA). This HTA was compared with the previously developed process map to identify similarities and differences in tasks. Results A total of 13 observations were made. The HTA top-level goal was to administer SC insulin, from which there were four sub-goals: 1) time to give insulin, 2) collect equipment, 3) enter ward and 4) administer insulin. More than 50 sub-tasks were represented in the HTA, highlighting the complexity of the overall task of administering SC insulin. There were variations in the way nurses prepared and gave insulin such as the use of workarounds and omissions of tasks. The comparison of the HTA (WAD) and process map (WAI) identified broadly similar tasks in sub-goal “administer insulin”. Differences in the boundaries of analysis in the process map and HTA made further comparisons of tasks difficult. For example, when describing the task of administering SC insulin, teams focused primarily on the final stage of insulin administration and not tasks that took place prior to it. Conclusion The administration of SC insulin was a complex process. The observations of SC insulin administration (WAD) largely matched those described by healthcare staff (WAI). Nurses were observed to adapt their work to provide essential patient centred care, The study was limited by the number of observations and differences in the boundary of analysis in the HTA and process map. Future work could investigate and understand the nature and implications of adaptations in SC insulin administration, as a key potential to develop a resilient work system.