Predicting length of stay in specialist neurological rehabilitation

A retrospective case series was performed to determine which measures of complexity, dependency and function most accurately predict inpatient neurorehabilitation length of stay for individuals with post-acute neurological disorders.

Sociodemographic, medical and functional variables were extracted from data submitted to the UK Rehabilitation Outcomes Collaborative. Length of stay was calculated as the total number of inpatient days, functional status was measured using Barthel Index, rehabilitation complexity was measured using Extended Rehabilitation Complexity Scale, and nursing dependency was measured using the Northwick Park Dependency Scale.

The mean rehabilitation length of stay was 70.9 days, with length of stay being 35.1 days higher in inpatients with acquired brain injury than inpatients with spinal cord injury. Diagnostic category, Barthel Index scores, Extended Rehabilitation Complexity Scale scores and Northwick Park Dependency Scale scores at admission independently predicted length of stay. Multiple regressions including diagnostic group, Barthel Index, Extended Rehabilitation Complexity Scale and Northwick Park Dependency Scale statistically significantly predicted 37.9% of the variability in length of stay (p < 0.005). Northwick Park Dependency Scale on admission was most closely correlated with inpatient length of stay.

In conclusion, inpatient length of stay is predicted by diagnostic category, Extended Rehabilitation Complexity Scale, Northwick Park Dependency Scale and Barthel Index. The most influential predictor of rehabilitation length of stay was Northwick Park Dependency Scale score at admission. These results may help facilitate rehabilitation resource planning and implementation of effective commissioning plans. Implications for Rehabilitation The most accurate predicting variable for length of stay in inpatient neurological rehabilitation was nursing need as measured by the Northwick Park Dependency Scale score on admission. Service users and commissioners can be provided with more realistic predictions of length of stay derived from admission variables that can be used in planning inpatient rehabilitation. Age and gender do not seem to have an effect on the total length of stay in rehabilitation.