Factors affecting survival of elderly nursing home residents

To examine whether the admission characteristics of nursing home residents predict mortality and to look for factors that account for different mortality rates between those homes.

A retrospective case note audit of the admission characteristics of residents in nursing homes registered with the Manchester Health Authority who died within a 12 month period and a group of current residents matched by age, sex and length of stay. Time between admission and death was examined using actuarial estimators of survival and factors predicting time between admission and death were examined using the proportional hazards model of Cox. Standardised mortality ratios (SMR) were calculated on each of the homes from the Office for National Statistics for Greater Manchester Metropolitan County. Logistic regression was used to assess those factors that predicted death at 1, 2 and 4 weeks and 3, 6 and twelve months after admission.

Following admission the mean survival time was 5.9 years. There was a wide variation in the mortality ratio between the homes and five of the 32 studied had SMRs of over 7. Increasing age, male sex, poor appetite, sleep disturbance, build, place of admission from, history of malignancy and respiratory disease and the number of prescribed drugs were all found to be significant predictors of mortality. Malignancy, pressure sores, poor appetite and number of prescribed drugs predicted death with 4 weeks after admission.

Some nursing homes have higher mortality rates than others although this was accounted for by the characteristics of the residents and not by any measurable characteristic of the nursing home environment.