Phone-interview version of barthel index : a study of its validity and utility

Abstract

Independence in self-care activities is one of the primary goals of rehabilitation of people with disabilities. The validity of its assessment has become a common interest for researchers and practitioners in the field. Most of the assessment methods are designed around the expert-observational method which appears to be the most valid. However, such a method was criticized to be more expensive and lack of ecological validity than compared with the self-report method. Its opponents, on the other hand, argue that other alternative methods such as self-report rely too much on the cognitive competence of patients which tends to put negative effect on the validity of the assessment. This study is aimed to test the validity of a telephone-based protocol designed for the Barthel Index (BI) which is to be used in a Chinese community. The instrumentation process involved the evaluation of the content validity and inter-rater reliability when the BI was translated from its English to Chinese versions, and converted from expert-observational to self-report format. Forty-six rehabilitation practitioners participated in different focus groups and panel reviews to evaluate the relevance of the task content and appropriateness of the rating criteria of the original BI. Results of the review suggest that six of the items were, to a different extent, irrelevant to the Chinese elderly people. They were "stair-climbing", "bladder control", "bowel control", "feeding", "bathing" and "toileting". Based on the comments from the panels, either the content or rating criteria was modified. The revised BI was then translated to Chinese and further simplified to a self-report format. The self-report format was then developed into the phone-interview protocol. Field tests were conducted to ensure the inter-rater reliability of each of the versions developed toward the phone-interview protocol. They involved a total of 63 older patients who were admitted to the geriatric wards of a regional general hospital. The mean age of the patients was 77.0 (SD=7.7). The Kappa statistics on the ratings assigned by pairs of rehabilitation therapists ranged from 0.63 to 1.00 (p<0.001) indicating that the reliability of the translated versions was either substantial or excellent. Seventeen patients with comparable demographic characteristics with those of previous field tests were then recruited to establish the equivalence between the expert-observational and self-report BI. The patients were screened for their cognitive and communicative skills using the Mini-Mental State Examination (MMSE) with scores ranged from 18 to 29 (Mean=24.6 SD=3.4). The Kappa statistics on the ratings generated from the two versions of BI were between 0.60 and 0.92 (p<0.001), which suggest an excellent agreement for six items and a substantial agreement for four items. Another 40 patients with similar demographic characteristics were recruited to test on the equivalence between the self-report phone-interview (within the hospital setting) and the expert-observational versions of BI. To control for the age effect, the patients were divided into the young-old (n=20; Mean=69.3, SD=3.1) and old-old groups (n=20; Mean=83.5, SD=7.1). Their mean MMSE score was 24.3. In general, the consistency of the ratings assigned by the patients in the young-old group was higher than that of patients in the old-old group. The Kappa statistics on nine items were 0.73-1.00 and 0.34-0.86 (p<0.02) respectively. A field test was conducted to establish the validity of the phone-interview protocol for use by patients after discharging from hospital to home. Thirty-one post-stroke patients who had similar demographic characteristics participated in this part of the study. Their mean age was 70.4 (SD=3.2) and mean MMSE score was 24.9. The Kappa statistics of the ratings generated from the patients via the phone-interview at the home setting and those obtained from on-site assessment by rehabilitation therapist were comparable with those obtained from the hospital setting (0.71 to 0.90, p<0.001). The findings suggest that the task content and demand of some of the BI self-care tasks is specific to the Chinese culture. Modifications of the original task items do not seem to compromise the psychometric properties of the instrument. After the control of cognitive abilities, the patients' age seems to exert a noticeable effect on hampering the validity of their self-report ratings on the BI items. This prompted us to further control on the age of patients. The results indicate that, for the young-old patients (aged under 75) with competent cognitive abilities, the use of the phone-interview method did not appear to generate additional errors onto the BI ratings when compared with those generated from the self-report format. Our findings further strengthen the utility of self-care assessments such as BI in the community settings via the phone-interview format. This study also demonstrates that, with a good screening system and appropriate training, patients are competent informants who can participate in the monitoring process of their own health at home. The same procedures could be used to demonstrate generalizability for other self-care instruments used in other Chinese communities such as in Taiwan, Mainland China, and eastern countries.