Counselling and depressive symptoms in older adults with HIV/AIDS in mbarara, Uganda | BMC Psychology

Research design

265 older people 60 years of age and over who were living with HIV/AIDS and receiving medical care at certain medical facilities in Mbarara, southwest Uganda, participated in this cross-sectional study. A cross-sectional research design involves collecting data from a population or sample at a single point in time [35].

Study setting

The study was carried out in Mbarara City at the Nyakayojo and Kakoba H/C IIIs, Mbarara City Council H/C IV, and the Mbarara branch of The AIDS Support Organization (TASO). They are one of the medical facilities in the city with the greatest number of HIV patients and offer HIV/AIDS care services. According to 2022 clinical records, TASO Mbarara and Mbarara City Council H/C IV Nyakayojo health centre III and Kakoba health centre III, the 4 combined Health facilities have a combined total of 603 clients accessing HIV/AIDS aged 60 years [36]. For instance, Mbarara City Council HC IV = 156, Nyakayojo Health Center = 11, Kakoba Health Center = 10, and TASO Mbarara = 426. The aforementioned health centers provide medical care, family care, and counselling to senior citizens. The study took a period of six months, including data collection preparing the required paper work, data analysis and presentation of findings to the relevant authorities.

Study population

The study population consisted of older adults (male and female) with HIV/AIDS who were 60 years of age or older and seeking medical care in Mbarara City.

Criteria for selection

Criteria for inclusion

For a duration of four months, we recruited older adults with HIV/AIDS who were 60 years of age or older and receiving care at TASO Mbarara, Mbarara City Council H/C IV, Nyakayojo, and Kakoba H/C IIIs.

Criteria for exclusion

The study population did not include people with serious cognitive impairments or those who were critically ill.

Sampling and sample size

The Yamane [37] formula of sample size determination, which reads as follows, was used to calculate the sample size:

Where e is the marginal significance level at 0.05, N is the entire population, and n is the sample size. When computed n = 241 participants. 10% marginal error was added to make a sample size of 265 participants. Based on the number of possible participants per institution, the sample was divided as follows: TASO Mbarara, Mbarara Municipal Council Health Center IV, Nyakayojo HC III, and Kakoba Health Center III each had 185, 70, 5, and 5 participants, respectively. The participants were enlisted one after the other [38]. The effect size for the power calculation was informed by prior studies on counselling interventions in similar populations, which reported an odds ratio (OR) of ≥ 2.0 for the association between counselling frequency and reduced depression [17, 24]. Using G*Power 3.1, a post-hoc power analysis was conducted for logistic regression with a two-tailed test, α = 0.05, and a sample size of 265. The analysis assumed a medium effect size (Cohen’s f² = 0.15), consistent with psychosocial intervention studies in PLWH [42]. This yielded a power of 87%, which exceeds the conventional 80% threshold, ensuring sufficient capacity to detect clinically meaningful associations. The rationale for accepting 87% power was its alignment with resource constraints and feasibility, while maintaining robustness against Type II errors.

Purposive sampling was used to select the health facilities because they had already been identified and selected for the study. They were selected because they have the highest number of people living with HIV/AIDS in Mbarara City. However, stratified sampling was used by dividing health units into different strata i.e. government Health centres and NGOs and in each stratum, appropriate number of patients was selected according to the total population. This was done to enable balanced coverage of the targeted respondents. There was some selection bias since the health facilities did not have equal numbers of clients but this was minimized by ensuring that appropriate proportion of respondents were selected from each health facility.

Within each stratum, participants were selected proportionally based on the total number of clients at each facility. However, patients were enrolled consecutively (one after another) as they accessed services, rather than through random selection. This sequential enrollment method was likely chosen for practicality, ensuring the target sample size per stratum was met.

Data collection procedure

The PHQ-9 and structured questionnaires were used to gather information on depression and counselling among older adults living with HIV in Mbarara, southwest Uganda. The clinical records of the four medical institutions, which included demographic data, were used by the facility staff to identify possible research participants. The responders were given informed consent forms to read and, if they had any questions, to ask. Consent forms and surveys were translated into Runyankore for illiterate respondents. Each customer who signed the consent form and agreed to participate in the study was assigned a unique identification number so that they could be easily identified when the data was being collected [38].

Study tools

Questionnaire

A questionnaire was developed by the principal investigator and administered by trained Research Assistants to respondents. The questionnaire was made in English but was also translated in the local language– Runyankore. All relevant data was collected using the instruments mentioned above. A well-structured Questionnaire was used for demographic characteristics and determined associated factors for depression like how often one seeks counselling services, pill burden, ART adherence, family support, having a drug companion, stigma and discrimination, presence of HIV comorbidities among others.

Patient health questionnaire (PHQ-9)

Uganda has validated and implemented the PHQ-9 [39]. HIV-positive individuals in Uganda have previously used it [40, 41]. The cut-off values for the PHQ-9 are as follows: 1–4 for little depression, 5–9 for mild depression, and 10–14 for major depression. moderately severe depression in 15–19 and severe depression in 20–27. When a person scores 10 or higher, depression is diagnosed. The frequency with which older people spoke with their counselors was used to measure counselling, while clients’ experiences with decreased stigma, beginning income-generating activities, feeling better about themselves, coping with HIV/AIDS, and improving drug adherence were used to gauge the role of counselling [38].

Quality control

Validity of research instruments

Content validity refers to the degree to which the research instrument measures what it should measure [42]. Content Validity Index (C.V.I) by Kothari [42] was used to test validity of research instruments using the formula;

The obtained CVI was 0.92 which was greater than 0.7 hence the tools were considered valid to yield results for the study.

The study operationalized counselling using single-item, researcher-developed measures: frequency of interaction (e.g., “How often do you talk to your counsellor?”), home visits (yes/no), family involvement (yes/no), and perceived benefits (checklist of outcomes like reduced stigma). No validated scales were used for counselling dimensions, though content validity (CVI = 0.92).

Reliability of research instruments

The reliability of the instruments was tested using Cronbach’s Alpha, where the Researcher entered the data of the total number of items in the questionnaire into SPSS program and was able to obtain proper values for reliability analysis. All alpha (α) coefficient values found to be above 0.81 (81%); that is α > 0.7 were sufficient enough for the tools/instruments to be regarded reliable according to Amin [43]. While depression was assessed via the validated PHQ-9, counselling variables relied on non-standardized items, limiting psychometric rigor and generalizability.

Statistical analysis

Using Stata version 17, continuous data were summarized using means and standard deviations, and categorical variables were summarized using percentages and frequencies. Chi-square tests were used to identify differences between depression and independent variables. Binary logistic regression examined the relationship between counselling factors and depression, with a significance level of p ≤ 0.05. Variables significant at p < 0.1 in bivariate analysis were considered for multivariable analysis [38].

Model selection process

Stepwise logistic regression was employed to select the final model, with a p-value threshold of 0.05 for variable inclusion and 0.10 for retention. While stepwise methods carry risks of overfitting and exclusion of theoretically relevant variables, they were justified here due to the exploratory nature of the study and the need to identify key predictors in a resource-constrained setting with limited prior local evidence. To mitigate these limitations, variable selection was guided not only by statistical significance but also by theoretical relevance such as frequency of counselling and home visits were prioritized based on literature linking social support and depression in PLWH.

Multicollinearity assessment

Variance Inflation Factor (VIF) values were calculated for all predictors in the final model. All VIFs were < 2.0, indicating no significant multicollinearity (VIF > 5 is typically considered problematic).

Model fit

The Hosmer-Lemeshow test confirmed adequate fit (χ² = 6.38, p = 0.2402). The final model explained 26% of variance (Pseudo R² = 0.26), suggesting additional unmeasured factors like socioeconomic status, viral load, may influence depression.

Assumptions

Individual counselling was provided to all clients who came at the facilities.

All clients were adhering well on the drugs.

All clients honoured their appointment dates.