Defining the appropriate use of community hospital beds

Patients of GPs who have access to community hospitals (CHs) as well as district general hospitals (DGHs) tend to spend on average more days in hospital each year. Increasing attention is being paid to the efficient management of medical admissions; however, there has been no previous prospective study investigating the appropriateness of CH admissions.

To develop a protocol to assess the clinical appropriateness of admission and length of stay of patients in CHs and to simultaneously compare the appropriateness of admissions to all DGHs and CHs in the county.

A protocol named Community Hospital Appropriateness Evaluation Protocol (CHAEP) was developed to assess CH admissions through a process of consultation and a series of pilot studies. The appropriateness evaluation protocol (AEP) was also reviewed and used to assess DGH admissions.

A prospective cohort of 440 DGH admissions from five DGH sites and 440 CH admissions from nine CHs.

The admissions were assessed and followed for 28 days. If an admission failed to satisfy any of the criteria then the researcher interviewed the clinician to decide whether it was justified to override the protocol and still classify the admission as appropriate. To assess validity, a proportion of these 'clinical overrides' and the researcher's classifications were reviewed retrospectively by a clinical panel. The kappa statistic was used to assess the level of agreement.

Applying the CHAEP, 82% of CH admissions satisfied a criterion for admission and a further 3% were given clinical overrides. A lower intensity of care was required for the majority of the remainder while three admissions required DGH care according to AEP criteria. Sixty-eight per cent of bed days satisfied day-of-care criteria within CHAEP and only a further 2% were given clinical override. These results were similar to those found with the AEP at the DGHs where 75% of admissions (plus 16% given clinical override) and 55% of days-of-care (plus 20% given clinical override) satisfied the AEP criteria. The review panel generally did not agree with the clinician's use of the clinical override at the CHs. Agreement between research nurse and review panel was better for the AEP and DGH (kappa = 0.9, 95% confidence interval (CI) = 0.7-1.0) than for the CHAEP and CH (kappa = 0.37, 95% CI = 0.1-0.8).

The CHAEP could be used to audit the appropriateness of admission and length of stay in CHs. Other health communities would need to review the CHAEP before it could be applied within their context.