Are we allocating limited resources to patients in most need?

Full text

Introduction

We aimed to examine this question by studying the correlation between severity of illness, outcome and the nurse workload (major determinant of cost).

Methods

We did a retrospective analysis of all intensive care patients admitted during 1992-1998 to the 10-bed general ICU. APACHE II [1] was used to determine the hospital mortality risk (MR). Patients were grouped into risk bands, in steps of 20%. Standardised Mortality Ratio (SMR=observed hospital mortality/calculated hospital mortality) was used in each stratum to define clinical efficacy. As a proxy for resource consumption, a modified form [2] of the nursing care recording (NCR) system [3] was used. Workload per patient, per survivor, per non-survivor and `effective' workload (workload all patients/number of survivors) was calculated within each stratum.

Results

4395 patients were admitted. 306 were children and 342 had missing values in scoring. APACHE II for survivors/non-survivors was 12.7/23.4, with estimated MR of 18.0/47.9%. NCR per patient was 2.3 times higher for deceased (385) compared to survivor's (165). In survivors, NCR increased in a linear fashion with increasing MR up to 40<60%. NCR in deceased patients was highest in those patients with lowest MR and lowest in those with highest MR. Effective work load increased gradually in a linear fashion with increasing MR.

Discussion

Resources were allocated to patients in most need. It was difficult to quantify the degree of efficacy. Effective NCR is an indicator of clinical efficacy in relation to severity of illness.