Abstract

Background Recent studies have found conflicting results on the effect of external oversight on clinical outcomes: a cross-sectional study found that fully-accredited hospitals have lower risk-adjusted mortality rates compared to nonaccredited ones (Falstie-Jensen et al., 2015); while in a cohort study no difference was found in measures of process of care (Bogh et al., 2015). The aim of this study was to determine the effect of external inspections of acute trusts performed by the Care Quality Commission (CQC) on rates of adverse events in the English National Health Service (NHS). Methods Interrupted time series study using a mixed-effects random-slopes multilevel model. Data on rates of falls with harm and pressure ulcers were retrieved from the NHS thermometer for all acute NHS trusts (n = 155) from April 2012 until January 2016. From CQC’s website, dates of on-site inspection were obtained, choosing one inspection for each trust for the period between September 2013 and October 2014, or the closest date to it. In case of several inspections, the one assessing more standards was selected. Median (P25−P75) are presented for rate of falls and pressure ulcers for pre and post intervention periods. Logarithmic transformations were used to model the effect of external inspections. 5% significance level was used and all analyses were performed in Stata/SE 14. Results 150 (97%) acute trusts reported information to NHS thermometer during the study period. Mean (standard deviation) number of data points were 18.72 (5.8) before and 24.7 (6.1) after intervention. Before inspection, median (P25−P75) rate of falls with harm was 5.8 (2.3−10.7) and pressure ulcers was 49.3 (36.5–63.2) per 1,000 patients/month. After inspection, rate of falls was 3.75 (1.6−6.7) and pressure ulcers was 41.8 (30−53.9) per 1,000 patients/month. A significant downward trend was found before inspection (−1.95 falls per 10,000 patients/month, p < 0.0001 and −6.3 pressure ulcers per 10,000 patients/month, p < 0.0001). No significant change in level was observed after inspection, although there was a change in trend (1.06 falls per 1,000 patients/month, p = 0.001 and −2.1 pressure ulcers per 10,000 patients/month, p = 0.01). Conclusion External inspection has no clinically relevant effect on adverse event rates, moreover, downward trend flattens after inspection. This might reflect the diversion of managerial and clinical staff's attention away from patient care during on-­site visits or simply a floor effect due to continuous improvement over time. Future research should focus on the impact on other measures of processes of care and outcomes such as mortality.