Participant-proxy reliability in traumatic brain injury outcome research

Area of Science:

• Neuroscience
• Rehabilitation Medicine
• Psychometrics

Background:

• Traumatic Brain Injury (TBI) can significantly impact an individual's ability to self-report outcomes.
• Proxies are often used in TBI research to gather outcome data, but their reliability needs careful assessment.
• Self-selected proxies may offer a more accurate perspective compared to assigned proxies.

Purpose of the Study:

• To evaluate the reliability of participant-proxy agreement for outcome measures in individuals with moderate to severe TBI.
• To determine the extent to which self-selected proxies can provide dependable data for TBI research.

Main Methods:

• Intraclass Correlation Coefficients (ICC) were employed to quantify participant-proxy reliability.
• Data were collected using the Craig Handicap Assessment and Reporting Technique (CHART), Community Integration Questionnaire (CIQ), and Functional Independence Measure (FIM).
• The study included 204 individuals with TBI and their self-selected proxies, who had been discharged from inpatient rehabilitation.

Main Results:

• Eighty-seven percent of items across the three instruments showed moderate to high ICC, indicating good reliability.
• Participant-proxy agreement was highest for concrete, observable information.
• Agreement was lowest for items assessing cognitive abilities, money management, and out-of-home activities.

Conclusions:

• Participant-proxy reliability is adequate for many outcome measures in TBI research when participants choose their own proxies.
• The findings support the use of self-selected proxies in TBI outcome studies, particularly for observable aspects.
• Further investigation may be needed for cognitive and financial domains where agreement is weaker.