What is the relationship between primary care and unscheduled hospital activity within 30 days of discharge from an emergency admission?

Problem

Unscheduled hospital activity within 30 days of discharge includes readmissions and A&E attendance and admission. Much of this activity is considered potentially avoidable for ambulatory care sensitive conditions where optimum management would be community-based. Our analysis identifies key aspects of primary care associated with overall readmissions, A&E attendance and A&E admission within 30 days of a discharge from an index emergency admission for heart failure (HF) or COPD.

Approach

National hospital administrative data for England were used to identify adult patients with a first emergency admission for HF and COPD. Patients were tracked for 30 days after live discharge. A&E activity - attendance and admission - and readmission via any route were modelled as three outcomes. In addition to patient, hospital and A&E factors aspects of primary care including appointment availability and quality of clinical care from the GP patients’ survey and the Qualities and Outcomes Framework were included as potential explanatory variables in logistic models.

Findings

Between April 2009 and March 2011 66219 HF patients and 90351 COPD patients were discharged alive from emergency admissions, contributing nearly 27,000 A&E attendances and over 28,000 readmissions within 30 days. Conversion (admission) rates from A&E were 76.9% for HF patients and 74.0% for COPD patients; overall 30-day readmission rates were 19.8% and 16.5% respectively.Age, deprivation, various comorbidities, ethnicity and length of stay were all significant predictors of unscheduled hospital activity. There was limited evidence that trust factors were associated with unscheduled hospital activity; A&E attendance during the evening or night shift had increased odds of admission, as did selected measures of busyness.Associations between primary care and unscheduled hospital activity were limited and effect sizes were small. For COPD patients, one extra GP per 1000 patients was associated with lower odds of A&E attendance (OR: 0.86, 95%CI 0.79-0.94) and overall readmission (OR: 0.89, 95%CI 0.82-0.96). Increased practice size was associated with increased odds of readmission for HF patients, but an increase of 1000 patients was only associated with an increase of 0-0.7%. HF patients attending practices with higher HF prevalence had lower odds of A&E attendance (OR: 0.84, 95%CI 0.79-0.90). However, HF prevalence is low and an increase of 0.4% - the interquartile range - would equate to a decrease in odds of 0.9%.

Consequences

Although there is increasing pressure on primary care to improve services to reduce the pressure on A&E we found little evidence that measures of access and quality of care were associated with lower odds of A&E activity or readmissions, although there is some evidence of a minor volume-outcome relationship. The increased odds of admission at night may be associated with severity, but may also be linked to lack of availability of other services during this time.