Predictors of quality of life of TB/HIV co-infected patients in the Northern region of Ghana

We conducted an analytic cross-sectional study among TB/HIV co-infected patients. Interviews were done using a standardized questionnaire and WHOQOL-BREF instrument to gather patients’ socio-demographic information and to evaluate their QOL respectively.

The study was carried out in the Northern Region of Ghana with 2,310,939 people [19]. The capital city of the area is Tamale, and there are 16 administrative districts. The region is bordered to the north by the North East Region, to the east by the Eastern Ghana-Togo International Border, to the south by the Oti Region, and to the west by the Savannah Region. More than 30 hospitals in the area identify and treat more than 200 TB patients who also co-infected with HIV.

All patients/participants who were 18 years or older and had received a TB/HIV co-infection diagnosis at a clinic at least one month prior to enrolment, and were willing to participate in the study were enrolled. The study excluded patients who were admitted to wards.

The sample size was calculated based on the QOL of TB patients in Nigeria [20]. The region recorded 452 TB/HIV co-infected patients. Using Yamane formula for sample size calculation at 95% confidence level and P = 0.5 [21];  \(\mathrm n=\;\frac N{1+N\left(e\right)^2}\), Where; n = sample size; N = population size = 452 e = precision level = 0.05 Therefore; \(n=\;\frac{452}{1+452\left(0.05\right)^2}\;=\;213\). Therefore, the minimum sample size was 213.

All 16 districts in the Northern Region with 32 hospitals that provide services to TB/HIV co-infected patients were selected in the study using total enumerative sampling method. The number of patients sampled was proportionately determined based on size.

At each health facility, simple random sampling method was used to select patients based on proportionate allocation (Table 1).

List of all the patients at each facility was obtained from patients’ registers. All the selected patients at each facility were contacted by healthcare workers through their telephone numbers. Patients were told to report to facilities on a particular day and interviewed by research assistants at a private place after consenting.

WHOQOL-BREF assessment tool was used to assess the QOL of study participants. It is an abbreviated form of the WHOQOL-100 assessment tool, made up of 26 items on QOL in physical health, psychological well-being, social relationships and environmental health domains. See Supplementary 1, Additional File 1.

Kessler Psychological Distress Scale (K10) was used to measure depression of study participants. The scale is made up of 10 questions about the emotional states of participants each with a 5-level response scale from 1 ‘none of the time’ to 5 ‘all of the time’. Questions 3 and 6 are not asked provided the response to the preceding question was ‘none of the time’, in which case they will be automatically scored 1. Scores of all 10 items are added giving a minimum score of 10 and a maximum score of 50 [22]. See Supplementary 2, Additional File 2.

Perceived stigma was measured with a 6-item Internalized AIDS-Related Stigma Scale (IA-RSS) as described in Supplementary 3, Additional File 3. Scale scores were summed to obtain composite stigma scores and those who scored below the mean were labelled as ‘not stigmatized’ and those above the mean as ‘stigmatized’ [23].

Adherence to medications was assessed using the 5-item AIDS Clinical Trials Group (ACTG) Adherence Self Report Tool. The questionnaire queries the patient about the number of doses missed of medication during each of the 4 days before a clinic visit. Because patients usually take more than 1 drug, the adherence percentage for each of the 4 days before visiting the clinic was calculated as [1-(number of missed dosage units/number of dosage units prescribed] ×100. The adherence percentage was also dichotomized into “adherent” (≥ 95%) or “non-adherent” (˂ 95%) [24, 25]. See Supplementary 4, Additional File 4.

The WHOQOL-BREF, K10, IA-RSS and the ACTG instruments used in the study are valid, widely used and internationally accepted. Also, pre-testing with respondents using the instruments was performed. The WHOQOL-BREF reliability was determined using the Cronbach’s alpha (α). Internal consistency of items in the various domains was considered when the α ≥ 0.7 [26].

Three (3) field assistants with a minimum of higher national diploma (HND) degree were recruited and trained by the principal investigator (PI) on data collection tools and process for 4 days. Debriefing meetings were held with the interviewers and research team to identify and address any problems identified during the pretesting. The research team led by the PI conducted the pretest of data collection tools. Private spaces at various facilities were used to collect respondents’ relevant information using all the data collection tools. The interviews lasted 30 minutes on average and the data were collected over a period of 4 months.

Data collected using KoBoCollect questionnaire were extracted into Microsoft Excel, cleaned and validated. Data were analysed using STATA version 15.1. Descriptive statistics were performed by estimating frequencies and percentages for categorical data. Means and Standard deviations were computed for normally distributed continuous data.

In assessing QOL using WHOQOL-BREF assessment tool, domain scores were calculated using Excel. Mean scores were obtained by adding all the items within each domain and dividing by number of items in that domain. Domain score was obtained by multiplying the mean score by 4 [27]. To ensure that the domain scores are comparable with the scores used in the WHOQOL-100, the scores were transformed to a 0-100 scale by employing the formula:

\(\mathrm{Transformed}\;\mathrm{score}\;=\;(\mathrm{score}-4)\;\mathrm x\;(100/16)\) [27]. QOL was dichotomized as 'poor' and 'good'. This was done by using the mean for each domain as a cut-off point. Individuals with scores greater than the mean were considered as having  good QOL whilst individuals with a score less than the mean were considered as having  poor QOL.

Multiple logistic regression was used to determine predictors of QOL at a 5% significant level.

A total of 213 respondents were included in the study. The mean age of the respondents was (38.99 ± 14.00) years with most, 33.3% (71) aged 30-39 years. More than half, 54.9% (117) of the respondents were males. More over two-thirds of respondents, or 72.8% (155), were married, and 61.0% (130) identified as Muslims. Less than a quarter, or 21.6% (46) of the respondents, were working (Table 2).

Seventy-six percent 76% (161) of the 213 respondents had depression. Nearly half of those, 47.8% (77) had mild depression, and 18.6% (30) had severe depression. The majority of respondents, or 81.7% (174), stated that they have experienced stigma because of their medical issues. Nearly two-thirds of respondents, or 67.3% (142), reportedly exercised, while more than three-fourths, or 78.9% (168), indicated they took their medications as prescribed. (Table 3).

The mean scores and standard deviations for QOL domains from highest to lowest among respondents were environmental health (59.40 ± 16.70), physical health (55.70 ± 13.64), psychological health (54.89 ± 10.86), and social relationships (47.46 ± 17.17) (Table 4).

According to the WHOQOL-BREF domains, the bulk of respondents who self-rated their QOL to be poor, 64.3% (137) were found in the environmental health domain. Social relationship domain however, recorded highest, 53.5% (114) number of respondents with good self-rated QOL. Generally, most respondents had poor self-rated QOL in all domains except in social domain (Table 5).

Of the 213 respondents studied, 70.0% (149) had their overall QOL to be poor across all domains of the WHOQOL-BREF assessment tool. Of the 117 respondents who self-rated their QOL to be poor with respect to physical health domain, almost all, 88.0% (103) of them also had their overall QOL to be poor (Table 6).

At the crude stage of the multivariate logistic regression analysis, variables that were suggestive to be significantly associated with the overall QOL were religion, level of education, sex, marital status, current employment status, the main mode of transport to and from clinic, depression, stigma, exercise, adherence to medications and out-pocket-purchase of drugs. Respondents who were single had 5.2 times increased odds of good QOL compared to those who were married (cOR = 5.19, 95% CI: 2.11 – 12.82, p < 0.000). Respondents who had tertiary education had 5.7 times increased odds of good QOL as against those who did not have any formal education (cOR = 5.71, 95% CI: 1.72 – 18.95, p < 0.004). Similarly, compared with those who were unemployed, employed respondents had 5.7 times increased odds of good QOL (cOR = 5.66, 95% CI: 2.82 – 11.37, p < 0.000). Respondents who had mild depressive conditions had 71% reduced odds of having good QOL in comparison to those who were not depressed (cOR = 0.29, 95% CI: 0.14 – 0.62, p < 0.001). Similarly, compared to those who were not depressed, respondents who had severe depressive conditions experienced a reduction of 97.5% less in the odds of having good QOL (cOR = 0.03, 95% CI: 0.03 – 0.20, p < 0.000). Also, respondents who reportedly adhere to their medications had about 8 times increased odds of good QOL compared to their counterparts (cOR = 7.98, 95% CI: 2.37 – 26.84, p < 0.001).

After adjusting for other variables, religion, education level, sex, marital status, current employment status, primary mode of transportation to and from the clinic, depression, stigma, exercise, and adherence to medications remained associated with overall QOL. Respondents who had tertiary education had 6.3 times increased odds of good QOL compared to their counterparts (aOR = 6.32, 95% CI: 1.64 – 24.38, p < 0.007). Respondents who were gainfully employed had 5.2 times increased odds of good QOL compared to their unemployed counterparts (aOR = 5.23, 95% CI: 1.87 – 14.60, p < 0.002). Also, compared to those who were not depressed, mildly depressed patients had a reduction of 80% in the odds of having good QOL) (aOR = 0.20, 95% CI: 0.08 – 0.52, p < 0.001), whilst those who were severely depressed experienced a reduction of 97% in the odds of having good QOL (aOR = 0.03, 95% CI:0.03 – 0.29, p < 0.003). Similarly, respondents who had perceived stigma had a reduction of 69% in the odds of having good QOL compared to those who were stigmatized (aOR = 0.31, 95% CI: 0.11 – 0.85, p < 0.022). Furthermore, respondents who reportedly adhere to their medications had 6.4 times increased odds of good QOL compared to their counterparts (aOR = 6.36, 95% CI: 1.52 – 26.65, p < 0.011)(Table 7).