Objective: To investigate bias and concurrent validity of patients' alcohol use reports by examining concordance with relatives' reports.
Design: Using a quasiexperimental static group comparison design, alcohol use was examined at 1 year after injury. Chi-square analyses, t tests, and a Fisher exact probability test were conducted to investigate the relationship between patients' and caregivers' reports on measures of alcohol use.
Setting: Medical center outpatient clinic.
Participants: 175 adult patients with traumatic brain injury and 175 family informants. Inclusion criteria were: patient older than 18 years of age and completion of self-report and family member versions of questionnaires. The full range of brain injury severity was represented.
Main outcome measures: Quantity-Frequency-Variability Index (QFVI; Cahalan and Cisin, 1968); Brief Michigan Alcohol Screening Test (B-MAST; Porkorny et al, 1972).
Results: A high rate of concordance, typically greater than 90%, was noted between patients' and relatives' reports on each of the alcohol use measures. Chi-square analyses revealed no significant differences (p > .05) between patients' and relatives' reports for the QFVI or the B-MAST. Inspection of the data found lesser rates of agreement for persons with severe injuries. Chi-square analyses revealed no differences based on family members' relationship to patients or whether they lived together or apart. However, a Fisher exact probability test (p < .03) for the QFVI indicated higher agreement rates for spouses relative to parents.
Conclusions: The results do not support the hypothesis that patients underreport drinking and that their reports are inconsistent with those of caregivers. The high levels of concordance are consistent with those found in similar research studies involving alcoholic populations. Results suggest that patients' reports should be given higher credibility and should not be routinely dismissed in favor of information provided by families. However, more caution is necessary when collecting information from persons with severe injury.