The use of proxies in community integration research

Objective: To assess the level of agreement between persons with various di sabilities and their proxies in reporting community integration outcomes us ing the Craig Handicap Assessment and Reporting Technique (CHART). Design: Reliability study. Setting: Participants living in the community for a minimum of 6 months aft er onset of disability or completion of inpatient rehabilitation. Participants: Persons (n = 983) with disability resulting from amputation, burn, multiple sclerosis, spinal cord injury (SCI), stroke, or traumatic br ain injury and their self-selected proxies. Interventions: Telephone inter-view of subjects (FIM(TM) instrument, CHART) ; proxies (CHART). FIM instrument assesses the degree of assistance with ph ysical and cognitive subscales; CHART measures community integration in 6 s ubscales: physical, cognitive, and economic independence, and mobility, soc ial integration. and occupation. Main Outcome Measures: Intraclass correlation coefficients (ICCs) were used to assess participant-proxy agreement, and stepwise multiple regressions w ere used to identify patterns of difference in agreement based on disabilit y type and demographic variables. Results: Thirty-seven of the 38 items examined for the entire sample yielde d moderate to strong ICCs. Multiple regression analyses indicated that prox ies overrated participants gr with severe functional cognitive disabilities on the mobility subscale (p <.001), overrated participants with less than a high school education on the total CHART score (p <.01), and underrated p articipants with SCIs on the occupation subscale (p <.01). Differences in a ll cases, however, were less than 6 points out of a possible score of 100 p er subscale. Conclusions: Participant-proxy agreement across the 6 disability groups pro vided evidence in support of the inclusion of proxy data for persons with v arious types of disabilities in community integration research.