Impact of case management by advanced practice nurses in primary care on unplanned hospital admissions: a controlled intervention study

BACKGROUND Increasing unplanned hospital admissions disrupt planned health care, lead to additional morbidity and are expensive. A recent review found only weak evidence for case management preventing unplanned admissions, yet case management of older people is being implemented widely in the UK. We aimed to study the effect of advanced practice nurse case management on unplanned medical and geriatric hospital admission rates in patients 50 years and over, and on admission risk in a 'higher risk' sub-group of patients in the UK. METHODS Case management by advanced practice nurses in NHS primary care practices in the Swansea Local Health Board area, Wales, UK. We conducted a prospective non-randomized controlled intervention study comparing unplanned medical and geriatric patient admissions between five intervention and thirty non-intervention practices during a pre-intervention year and an intervention year. RESULTS For all lengths of stay, comparing intervention (n = 5) with non-intervention practices (n = 30) from pre-intervention to intervention year, we found that the unplanned medical and geriatric admission rate was significantly lower in the intervention group - adjusted relative risk of 0.909; relative risk reduction 9.1% (95% credible limit 0.840 to 0.984, p = 0.018); absolute risk reduction 0.99 admissions per 100 patients (95% credible limit 0.17 to 1.86, p = 0.018). For lengths of stay of one night or more we observed a stronger effect - adjusted relative risk 0.896; relative risk reduction 10.41% (95%, credible limit 0.820 to 0.979, p = 0.015). Most of the rate reduction was due to a reduction in the number of new admissions but much less so for admissions of lengths of stay of at least one night, compared to all lengths of stay. We did not find a statistically significant effect on re-admission or multiple re-admission rates in 'higher risk' patients previously admitted one or more times - adjusted relative risk of further multiple admissions per previously admitted patient 0.908 (95% credible limit 0.765, 1.077); relative risk reduction 9.3%; adjusted relative risk of total admissions per multiple admitter 0.995 (95% credible limit 0.940, 1.053) relative risk reduction 0.6%. CONCLUSION Although this study reports a reduction in unplanned admission rates in the intervention practices, this appears to be only in part directly due to nurse case management: most of the reduction did not occur in multipe admitters whom were case managed. Further research is needed to explain this finding, to elucidate how best to target the attention of case managers and to examine the complexity of potential outcomes in terms of the nature and necessity of admissions and most suitable lengths-of-stay in terms of acute care or rehabilittion need.