Getting the most out of digital healthcare systems: Data repurposing to improve inpatient diabetes care

ABSTRACT

Aims: At a North London hospital which lacked remote blood glucose monitoring, we sought to introduce a system that automatically alerted the Diabetes Team of inpatients with deranged blood glucose (BG) measurements. We hypothesised that this would streamline the team's workflow, and improve glycaemic control in inpatients. Methods: We developed a feature that used data from recently deployed e-vitals software to generate daily reports containing lists of inpatients who had experienced hypo-or substantial hyper-glycaemia ( < 4 or >20mmol/L) within the previous 24h. These reports were automatically sent each morning to the Diabetes Team. Results: Feedback from specialist nurses suggested improvement in the efficiency of their patient identification workflow, which previously had principally involved taking phone referrals and manually searching ward lists. Initial post-intervention data did not suggest improvement in hospital-wide deranged BG rates, however this was confounded by a sharp rise in covid-19 admissions shortly after deployment, with the majority receiving corticosteroids. After several months of use, the feature unexpectedly failed for approximately six weeks, during which time on average significantly more daily hypoglycaemic episodes occurred vs the preceding six-week period (two-sided rank sum, p < 0.001), with rates returning to baseline after it was reintroduced. Conclusions: Our intervention aided staff workflows and possibly improved inpatient glycaemic control, although worsening glycaemic control outcomes upon intervention withdrawal cannot reliably be extrapolated to infer overall benefit of the feature vs pre-intervention standard-of- care. Financial barriers often preclude deployment of gold-standard digital systems in healthcare;innovative exploitation of data generated by more affordable systems can improve productivity and patient care without additional cost.