This study was a 3 arm randomized single blind clinical trial involving 386 randomly selected and allocated adult HIV patients who were enrolled into Antiretroviral Therapy (ART) care at all 4 comprehensive ART sites in Yola. These comprehensive sites were the; Federal Medical Center (FMC) Yola, State Specialist Hospital Yola (SSHY), St Francis Hospital Jambutu and Adamawa Hospital. The trail took place from January to September 2014. A Clinician Client Centered training module was developed based on the Information Behavior and Motivation (IBM) Model. 9 Clinicians involved in ART care were trained with this module to deliver a 10 to 15 minutes clinic based intervention (Clinician Client Centered (CCC) counseling). Intervention group 1 received 2 counseling sessions; at baseline then at 2 months. Intervention group 2 received 1 counseling session at baseline only and the control group received routine care. An interviewer administered validated and reliable structured questionnaire was used for data collection. The questionnaire consisted of 5 sections; Section 1 consisted of a code number and socio-demographic variables. Section 2 consisted of questions on the knowledge of HIV transmission and prevention. This section has a total of 17 statements and answers that had the options of 'yes', 'no' and 'don't know'. Section 3 consisted of 5 statements to address patient's attitude towards HIV/AIDS. Answer options to these questions were from a 5 point Likert scale ranging from strongly agree, agree, don't know, disagree and strongly disagree. Total score possible for attitude range from 5 to 25. On the basis of summated scores respondents were divided into high or low attitude using the 50 percentile as cut off. Section 4; addressed sexual behavior patterns (condom use and number of sexual partners)questions in this section addressed; recent sexual activity in last 30 days preceding the survey, type of sexual relationship (monogamous or polygamous), condom use, gender of sexual partner and type of sexual practice. Section 5 addressed issues of status disclosure. Questions were directed at number of spouse/sexual partner who were aware of the respondents' HIV status. It also determine number of other persons HIV status had been disclosed and the effects of such disclosure. Data was collected at baseline, 2 months and 6 months.
Questionnaires were manually validated for errors and data analyzed using SPSS (Statistical Package for Social Sciences) version 22. All data with interval and ratio scale were explored using graphical methods (histogram with normal distribution curve and whisker box plot) to exam for normality. Parametric statistical test were used for normally distributed data and non parametric equivalent used were normality assumption was violated. For descriptive statistics measures of central tendency and dispersion were used for continuous data. Percentages were used to describe categorical data. The median and interquartile ranges (IQR) were used when the assumption of normality was violated.
Inferential statistics was used to determine homogeneity of respondent's socio-demographic variable between intervention group 1, intervention group 2 and control group. Data transformation was performed on data which were not normally distributed. Data transformation performed on not normally distributed condom use scores with both log10 and square root transformation was unsuccessful. One way ANOVA was used to compare mean differences of scores of log HIV knowledge, log attitudes scores towards HIV/AIDS and HIV status disclosure scores among the two intervention groups and the control group at baseline, 2 months and at 6 months. Kruskal Wallis ANOVA was used to compare median condom use scores among the 3 study groups at baseline, 2 months and 6 months. Spearman's correlation coefficient was used to investigate relationships between HIV knowledge, attitudes towards HIV and sexual behaviors (condom use, sex with unsteady partners and status disclosure).
McNemar's test was used to asses treatment effects of the intervention on sound knowledge, categorical scores of attitude and categorical sexual behavior practices related to HIV in the present study. It was carried out for changes at 2 months and 6 months.
Mixed design ANOVA was employed to determine the main effects of age, gender, group, time and group - time interaction effects for mean log scores of HIV knowledge, log attitude scores towards HIV/AIDS and log HIV status disclosure scores of the 3 study groups. Friedman's repeated measure ANOVA was used to look at main effects of group, time and group and time interaction effects of the median condom use scores of the 3 study groups. A partial eta square (ἠ2) as a measure of effect size representing the variance in proportions of the dependent variable that can be explained by the independent variable was applied to both mixed design ANOVA and Friedman's repeated measures ANOVA. The interpretation of the strength of eta squared values followed guidelines by Cohen, 1988: small effect (0.01), moderate effect (0.06) and large effect (0.14).
All results of log transformed data considered underlying data transformation during result interpretation. The confidence interval was set at 95% for mean estimations. The level of significance; alpha (α) was set at 0.05. A p value of less than 0.05 pertains to the decision rule. The decision rule used was to reject the null hypothesis when p was less than 0.05.
