ABSTRACT Aim To explore whether a delay from referral to first contact with nurse‐led community health services is associated with the likelihood of subsequent emergency department attendance. Design We use individual linked administrative data on use of community health and hospital services. We identify a cohort of 343,721 individuals referred to community health services in England by their primary care provider in 2019. We then track their subsequent community healthcare contacts and emergency department attendances. Methods We exploit variation in the time to contact caused by weekend delays, which create longer times to first contact for people referred later in the working week. The main analysis compares patients referred on Thursday with those referred on Tuesday. Results We show that 6.7% of patients referred on Thursday wait an extra two days for their first community contact relative to those referred on Tuesday. Despite this delay, we find no evidence that people referred on Thursday are more likely to have a subsequent emergency department attendance compared to those referred on Tuesday. Conclusions We do not find delayed community health services contact to be associated with an increased risk of emergency attendance amongst patients referred to community services by their primary care provider. This suggests that short delays in contact time are not detrimental for this group. Impact Shifting care from hospital to community settings is a key priority for health systems internationally. In England, community health services face significant staffing shortages, limiting the extent to which services can be responsive and support the desired strategic shift. Our findings suggest that these constrained community providers could use their limited capacity to prioritise responding quickly to other patients without harming those referred via primary care. Reporting Method STROBE guidelines. Patient or Public Contribution This study did not include patient or public involvement in its design, conduct, or reporting.
Cameron et al. (Fri,) studied this question.