Development of a minimum data set-based depression rating scale for use in nursing homes

Abstract

Background: depression is common but under-diagnosed in nursing-home residents. There is a need for a standardized screening instrument which incorporates daily observations of nursing-home staff.

Aim: to develop and validate a screening instrument for depression using items from the Minimum Data Set of the Resident Assessment Instrument.

Methods: we conducted semi-structured interviews with 108 residents from two nursing homes to obtain depression ratings using the 17-item Hamilton Depression Rating Scale and the Cornell Scale for Depression in Dementia. Nursing staff completed Minimum Data Set assessments. In a randomly assigned derivation sample (n = 81), we identified Minimum Data Set mood items that were correlated (P < 0.05) with Hamilton and Cornell ratings. These items were factored using an oblique rotation to yield five conceptually distinct factors. Using linear regression, each set of factored items was regressed against Hamilton and Cornell ratings to identify a core set of seven Minimum Data Set mood items which comprise the Minimum Data Set Depression Rating Scale. We then tested the performance of the Minimum Data Set Depression Rating Scale against accepted cut-offs and psychiatric diagnoses.

Results: a cutpoint score of 3 on the Minimum Data Set Depression Rating Scale maximized sensitivity (94% for Hamilton, 78% for Cornell) with minimal loss of specificity (72% for Hamilton, 77% for Cornell) when tested against cut-offs for mild to moderate depression in the derivation sample. Results were similar in the validation sample. When tested against diagnoses of major or non-major depression in a subset of 82 subjects, sensitivity was 91% and specificity was 69%. Performance compared favourably with the 15-item Geriatric Depression Scale.

Conclusion: items from the Minimum Data Set can be organized to screen for depression in nursing-home residents. Further testing of the instrument is now needed.