Implementation of an Automated Primary Care Acute Kidney Injury Warning System: A Quantitative and Qualitative Review of 2 Years of Experience.

BACKGROUND: Acute kidney injury (AKI) is often detected late, leading to worse clinical outcomes. In 2012, we pioneered an AKI-alerting system for primary care clinicians (PCCs). We retrospectively analysed the alerts and evaluated PCC satisfaction to assess the feasibility of the system.
METHODS: The study used a 2-pronged approach. AKI alerts, generated by an algorithm designed by University College London Hospital biochemistry department between June 2012 and June 2014, were analysed to reveal the demographics and outcomes of each patient generating an alert. Second, a survey was sent to all PCCs assessing awareness and satisfaction with the service. Simple statistical methods were applied (mean, median, SD and interquartile range).
RESULTS: One hundred forty-two alerts were generated, of which 101 were genuine. Generally, the patient demographics, AKI stratification and aetiology were in keeping with the inpatient AKI population. Forty-eight percent of cases were referred to the hospital with a median length of stay of 9.9 days. Three-month mortality was 12%. Among PCCs, there was good awareness of the system with most finding it valuable. The key complaints around the system were to do with lack of knowledge of its existence.
CONCLUSIONS: Our evaluation has demonstrated that the implementation of AKI alerts in the community is technically feasible, does not result in excessive demand on hospital services, appears to influence PCC behaviour and was perceived overwhelmingly as a useful service by these clinicians. This experience should inform further developments including behavioural interventions (such as clinician alerts) to improve community AKI care.