Objective: To compare the scores on the Relative Stress Scale (RSS) with those on the General Health Questionnaire (GHQ) and the Geriatric Depression Scale (GDS), and to establish a cut-off score for RSS in order to distinguish carers with symptoms of psychiatric disorders from those without.
Methods: One hundred and ninety-four carers of 194 patients suffering from dementia according to ICD-10 were included in the study. Burden of care was assessed by the 15-items RSS, and psychiatric symptoms by means of the GHQ-30 and the 30-items GDS. A case score above 5 on GHQ and above 13 on GDS were used to define carers with probable psychiatric morbidity. Sensitivity (SS), specificity (SP), accuracy and likelihood ratio for a positive test (LR+) were calculated for different cut-points of the RSS.
Results: Fifty-six percent of the carers had a GHQ score above 5, and 22% had a GDS score above 13. A two-step cluster analysis using 192 of the 194 carers, identified three groups of carers; a low risk group for psychiatric morbidity (LRG), 82 carers with GHQ ≤ 5 and GDS ≤ 13; a medium risk group (MRG), 69 carers with GHQ > 5 and GDS ≤ 13; and a high-risk group (HRG), 40 carers with GHQ > 5 and GDS > 13. The optimal RSS cut-off to distinguish between the LRG and the others was > 23 (SS 0.72, SP 0.82, accuracy 0.76, LR + 4.0), whereas the optimal cut-off to separate the HRG from the others was >30 (SS 0.74, SP 0.87, accuracy 0.84, LR + 5.7).
Conclusion: The RSS is a useful instrument to stratify carers according to their risk of psychiatric morbidity. Copyright © 2006 John Wiley & Sons, Ltd.