SP228
ELECTRONIC CLINICAL DECISION SUPPORT FOR THE EARLY RECOGNITION AND MANAGEMENT OF ACUTE KIDNEY INJURY: QUANTITATIVE EVALUATION OF OUTCOMES Free

Introduction and Aims: Electronic clinical decision support (eCDS) systems have been promoted to help address deficiencies in AKI care and have been widely adopted despite limited published evidence. However, eCDS implementation in other settings has been hampered by alert fatigue, end-user disengagement and unintended clinical consequences. These impediments to integration may hinder progress against the clinical outcomes which these tools were intended to target. With clinical efficacy, therefore, still uncertain, we undertook a single-centre, quantitative evaluation of outcomes after implementation of an AKI eCDS system within our organisation.

Methods: AKI eCDS was configured within the electronic patient record (EPR) system and targeted medical end-users. Alerts were triggered to appear on accessing the affected patient’s EPR at a threshold serum creatinine (Cr) rise of 25 umol/L/day and linked to clinical guidance and automated ordering of relevant investigations. Pilot ward experience was used to refine the tool before activation across all 61, adult, non-critical care wards.

The study was conducted across three, 4 week phases, commencing: (1) 4 months before, (2) immediately after and (3) 13 weeks after go-live. Consecutive patients up to a maximum of 280 were retrospectively identified in each phase, excluding those receiving acute or chronic dialysis or who were on the ‘care of the dying’ pathway at the time of the ‘virtual’ (invisible to end-users in phase 1) or actual trigger (phases 2 and 3).

Irregular time series of Cr measurements were created for each subject using the ‘R’ package ‘zoo’. The maximum proportional rise in Cr was calculated for each subject from the time series as (Cr[max] - Cr[index]) / Cr[index], where Cr[index] was the 1st Cr on the day of the trigger and Cr[max] the peak value over the following 7 days.

Results: Baseline patient characteristics are shown in Table 1. The proportional rise in Cr in the 7 days post-trigger is shown in table 2.

Conclusions: This preliminary analysis shows no clear impact of the AKI eCDS on hospital length of stay or hospital mortality but the suggestion that fewer patients progress to more severe dysfunction after the system has become established requires further evaluation.