Concordance between medical records and interview data in correctional facilities

Area of Science:

• Correctional healthcare research
• Public health in incarcerated populations
• Data reliability and validity

Background:

• Self-administered questionnaires, interviews, and medical records are common research data sources.
• Evaluating the concordance and reliability of these data sources is essential.
• This study assessed data concordance between medical records and interviews in correctional facilities.

Purpose of the Study:

• To evaluate the concordance between medical and behavioral data from medical records and interview questionnaires.
• To compare data reliability between male and female inmates.
• To inform data source selection for research in correctional settings.

Main Methods:

• Compared medical record and interview data for 679 inmates (male and female maximum security prisons).
• Conducted gender-stratified and non-stratified analyses using SPSS to calculate prevalence and kappa coefficient scores (κ).
• Calculated sensitivity and specificity between medical records and interviews.

Main Results:

• Strong concordance (κ=0.66-0.89) observed for HIV, hepatitis C, diabetes, asthma, and tattoos.
• Moderate concordance (κ=0.49-0.57) for hypertension, renal disease, smoking, recent antibiotic use, and cocaine use.
• Poor concordance (κ=0.23-0.36) for illicit drug use and marijuana use.
• Females generally showed higher concordance and prevalence rates.
• Medical conditions were more frequent in records; behavioral conditions more frequent in interviews.

Conclusions:

• Agreement between medical records and self-reports differs significantly by data type.
• Medical conditions show higher concordance with records, while behavioral factors align better with interviews.
• Data source selection should be tailored to the specific information required for research.