Place of residence, perceived signs and explanations of mental illness were independent correlates of public stigma against family members of people with mental illness.
In multivariate analysis occupational status, area of residence, perceived time spent with health care provider, perceived accessibility, perceived waiting time, perceived professional care and over all patient satisfaction were significantly associated with adherence to TB treatment (P < 0.05).
Long distance to the clinic was the main subjective reason, while low education attainment, rural residence, perceived mild illness and dissatisfaction with the ART service were associated with not using their own clinic in multivariate analyses.
In a second step, linear regressions were successively performed by adding to MIL each sociodemographic variable (gender, age, linguistic region, marital status, education, perceived socioeconomic status, residence, employment) and health status one by one.
We found significant differences between CAM users and non-CAM users in the following parameters: region of residence, perceived health status of the child, self-reported illness, number of doctor visits, surveyed caregiver's relationship to the child, caregiver age and educational level, monthly family income, and CAM use by adult family members (all p < 0.01).
In univariable analyses, lower education level, rural residence, perceived mild illness of the disease episode, and dissatisfaction with the ART service were associated with utilization of health care facilities other than the own ART clinic, and in the multivariate analysis these factors remained significant (table 2). 1. Patients could give more than 1 reason.
Long distance to the clinic and to a lesser degree, poor quality of care and long waiting times were the main subjective reasons given, while lower education level, rural residence, perceived mild illness and dissatisfaction with the services of the own ART clinic were independently associated with using health care facilities other than the own clinic.
To determine whether self-reports of unmet need are biased measures of access to health care, the authors examine the relationship between rural residence and perceived need for physician services.
Odd ratios have been adjusted for CAM group, gender, region of residence, age, perceived health status, self-reported specific illness, doctor visit, relationship to subject, caregiver age, caregiver educational level, family income, and CAM use by adult family members for a GEE for ordinal data.
Odds ratios and p-values have been adjusted for gender, region of residence, age, perceived health status, self-reported specific illness, doctor visit, relationship to subject, caregiver age, caregiver educational level, family income, and CAM use by adult family members in the multivariate logistic regression analysis.
