Abstract
Background
Boosted atazanavir (ATV/r) is associated with clinical jaundice, which may affect ART adherence. However, the prevalence of clinical jaundice among people on ATV/r and its association with ART adherence is unknown in Uganda.
Methods
We conducted a cross-sectional study among 236 adult patients on ATV/r at the Infectious Disease Institute. Clinical jaundice was assessed by self-report and observation; ART adherence was measured using the 8-item Morisky scale.
Results
Period prevalence of clinical jaundice was 31.8%, while point prevalence was 21.1%. While 8.5% had poor adherence, clinical jaundice was a significant predictor of poor adherence (adjusted odds ratio = 3.03, 95% CI: 1.18–7.76, p = 0.021).
Conclusion
Clinical jaundice is common among patients on ATV/r in Uganda and is associated with poor ART adherence. Strengthening pre-ATV/r (pre-switch) counselling and exploring safer regimen alternatives are required to improve outcomes.
Keywords: HIV, Boosted atazanavir, Clinical jaundice, Adherence, Antiretroviral therapy
Introduction
According to Uganda’s consolidated guidelines for HIV/AIDS prevention and treatment[22], Dolutegravir (DTG) is the preferred anchor drug for adult first- and second-line ART. Boosted atazanavir (ATV/r) is the alternative second-line option, especially for patients who cannot use DTG or who fail non-nucleoside reverse transcriptase inhibitor or DTG-based first-line therapy. However, ATV/r is linked to indirect hyperbilirubinemia, which may progress to jaundice [2, 4].
Jaundice occurs when blood bilirubin levels exceed 3 mg/dL and may present clinically with yellowing of eyes, mucous membranes, and skin [17]. ATV/r-induced jaundice is due to inhibition of Uridine 5’-Diphospho-Glucuronosyl Transferase 1A1 (UGT1A1), an enzyme responsible for bilirubin conjugation [6]. Genetic polymorphisms such as UGT1A1*28 are common across African populations, with particularly high frequencies in equatorial African groups [8], increasing susceptibility to jaundice [18]. Clinical jaundice may lead to non-adherence or discontinuation of ART among people living with HIV (PLHIV) [6].
Previous studies have assessed ART adherence [12, 15, 23], and ATV/r-associated hyperbilirubinemia [4, 11, 16]. Clinical trials conducted in predominantly white populations report a point prevalence of clinical jaundice of 5–8% [11, 20], while a Hispanic cohort reported a period prevalence of 21% [4]. However, no studies from Africa, and none from Uganda, have described the prevalence of clinical jaundice, or its association with ART adherence among people taking ATV/r. This limits the applicability of findings from non-African populations, given the high frequency of UGT1A1 variants in equatorial African populations, which may increase susceptibility to ATV/r-related clinical jaundice. Understanding clinical jaundice in this context is relevant not only for HIV programs in Uganda but also for global treatment decisions where ATV/r remains a key second-line option. Recognizing and addressing visible ART side effects such as clinical jaundice is essential for sustaining adherence, improving counselling, and informing national pharmacovigilance and treatment guidelines. This study aimed to determine the prevalence of clinical jaundice and its association with ART adherence among adults on ATV/r attending an HIV clinic in Kampala, Uganda.
Methods
Study design and setting
A cross-sectional study nested within an existing HIV care cohort was conducted from 8th April to 5th May 2019 at the HIV clinic of Infectious Diseases Institute (IDI), Makerere University, located within Mulago National Referral Hospital, Kampala, Uganda. As of 2019 the clinic cared for over 8,300 PLHIV, including approximately 871 on ATV-based ART regimen, in line with Uganda MoH HIV guidelines and the WHO guidelines. Routine data were captured in both physical records and an electronic database, including information on ART regimens, adverse drug reactions, reasons for switching and stopping ART. A quality control officer routinely checked and validated the data entered.
As part of routine care, patients received counselling at ART initiation on adherence, potential adverse drug effects, and actions to take if symptoms occurred such as promptly returning to the clinic or contacting healthcare providers. Clinical assessments and management, including regimen review, were conducted during follow-up.
Sample size
A sample size of 270 participants was calculated using Cochran’s formula [1], assuming 50% prevalence of clinical jaundice, 5% margin of error, and a 95% confidence interval (CI). This calculation was based on IDI clinical records for 2019, which indicated an estimated population of 871 PLHIV on ATV/r.
Selection criteria and sampling
Participants were selected from the existing cohort of PLHIV receiving care at the IDI clinic and were eligible if they were HIV-positive adults (≥ 18 years) on ATV/r for at least 3 months. Those who were too ill were excluded. Consecutive sampling was used, and 242 participants were enrolled due to the time-bound recruitment period, representing 89.6% of the planned sample size.
Data collection and data variables
Data were obtained from routinely collected data and primary data collected through a semi-structured, interviewer-administered questionnaire. The questionnaire was pre-tested among 10 adults on ATV/r at IDI clinic who were not part of the main study. The pre-test confirmed clarity and relevance of the questions, and no modifications to the tool were made. Collected data were entered into EpiData software and exported for analysis.
The main outcome was ART adherence, assessed using the 8-item Morisky Medication Adherence Scale (8-MMAS)[21], a widely used self-report tool for measuring medication adherence. Responses were coded and summed to generate a total score ranging from 0 to 8, with higher scores indicating better adherence. Adherence was categorized using standard 8-MMAS thresholds as high (8), medium (6–<8), and low (< 6). For the primary analysis, adherence was dichotomized into good (high and medium combined) and poor (low).
Independent variables included current or past clinical jaundice while on ATV/r, duration on ART, comorbidities (hypertension, diabetes and tuberculosis), social support, pill burden, disclosure, drug and substance abuse, stigma score and discrimination and socio-demographic characteristics (age, tribe, gender, religion, education level, employment, economic status, marital status). Clinical jaundice was assessed through self-report and visual examination of the sclera and buccal mucosa by trained study staff under natural light or bright indoor lighting. Distress due to clinical jaundice was graded using a visual analogue scale from 0 to10 [19]. Point 0 indicated that the person didn’t think about the clinical jaundice and 10 indicated that its presence maximally bothered them. Social support was assessed using the 3-item Oslo support scale, and numbers were assigned to each response. Total individual scores were categorized into 3: 3–8, 9–11, and 12–14, representing poor, moderate, and strong social support, respectively [10]. Stigma score and discrimination were measured using the 8-item Stigma Score for Chronic Illnesses (8-SSCI) [13]. Numbers were assigned to each response, and total individual scores were dichotomized; > 5 and < 5 indicated stigma and no stigma, respectively. Drug and substance abuse were assessed through self-report.
Data analysis
Data were analyzed using Stata version 13. Descriptive statistics were used to summarize participant characteristics. The prevalence of clinical jaundice was calculated as point prevalence and period prevalence, expressed as percentages with 95% CI. Point prevalence was defined as the presence of clinical jaundice at the time of the study, while period prevalence was defined as having ever experienced clinical jaundice since initiation of the ATV/r regimen.
Bivariate logistic regression analysis was conducted to identify variables associated with adherence. Variables with p-values < 0.20 at the bivariate level (sex, alcohol use, clinical jaundice, stigma, and having discussed jaundice with a doctor) were considered for inclusion in the multivariable logistic regression model. Additional covariates of epidemiological relevance (age, duration on ATV/r, and social support), were also assessed. All variables were reported with crude odds ratios (cOR) and corresponding 95% CIs. A stepwise elimination approach was used to develop the final model, retaining variables with p < 0.05. Alcohol use was retained in the final model due to its clinical relevance.
Multicollinearity was assessed using variance inflation factors, and model fit was evaluated using the Hosmer–Lemeshow goodness-of-fit test. Effect modification by sex and age was assessed by including interaction terms. Predicted probabilities (marginal effects) of poor adherence by clinical jaundice status were estimated to aid interpretation of findings. Statistical significance was assessed at p < 0.05.
Results
Study participants
Of the 242 study participants enrolled, complete data from 236 respondents were analyzed (97.5% completion rate), and the mean age was 39.9 years (SD 10.8). Participants with clinical jaundice were younger (mean age 37.4 vs. 41.1 years, p = 0.0149) and more likely to be female (80.0% (40/50) vs. 60.3% (112/186), p = 0.003). No significant differences were observed in tribe, religion, education level, marital status, drug and substance abuse.
Prevalence and characteristics of clinical jaundice
The point prevalence of clinical jaundice was 25.4% (60) (95% CI: 20.2–31.4), while period prevalence was 31.8% (75) (95% CI: 26.1–38.1) (Table 1). Clinical jaundice episodes varied in onset (≥ 1week) and duration (≥ 2weeks) after initiation of ATV/r and were often recurrent. Among those with clinical jaundice, 84.7% (47) reported that it minimally or did not bother them at all (visual analog score; 0–3) (Table 1). Only 2 participants (< 1%) discontinued treatment because of jaundice.
Table 1.
Prevalence of clinical jaundice and ART adherence among 236 adults on ATV/r at IDI, Kampala, Uganda
| Characteristic, n = 236 | Frequency (n) | Percentage (%) | 95% CI |
|---|---|---|---|
| Clinical jaundice (point prevalence) | |||
| Yes | 60 | 25.4 | 20.2 − 31.4 68.6–79.7 |
| No | 176 | 74.6 | |
| Clinical jaundice (period prevalence) | |||
| Yes | 75 | 31.8 | 26.1–38.1 |
| No | 161 | 68.2 | 61.9–73.9 |
| Visual analog score*, n = 70 | |||
| 0–3 | 47 | 67.1 | |
| 4–7 | 16 | 16 | |
| 8–10 | 7 | 7.0 | |
| Adherence (8-MMAS) | |||
| Poor < 6 | 20 | 8.5 | 5.5–12.8 |
| Good ≥ 6 | 216 | 91.5 | 87.2–94.5 |
*5 missing values
ART adherence and associated factors
The majority (91.5%, n = 216) of the participants had good ART adherence. At bivariate analysis, sex, alcohol use, clinical jaundice, discussing jaundice with a doctor, and stigma and discrimination were associated with adherence (p < 0.20) (Table 2).
Table 2.
Bivariate and Multivariable logistic analysis of the associations between variables and ART adherence among 236 adults on ATV/r at IDI, Kampala, Uganda
| Variables, n = 236 | Good adherence (n = 216), n (%) | Poor adherence (n = 20), n (%) | cOR (95% CI), p value | aOR (95% CI), p value |
|---|---|---|---|---|
| Sociodemographic characteristics | ||||
| Age in years | 0.99 (0.99–1.03), 0.699 | |||
| Sex | ||||
| Male | 75 (94.9) | 4 (5.1) | 1.00 (ref) | |
| Female | 141 (89.8) | 16 (10.2) | 2.13 (0.69–6.59), 0.191* | |
| Tribe | ||||
| Muganda | 124 (93.2) | 9 (6.8) | 1.00 (ref) | |
| Ankole | 22 (91.7) | 2 (8.3) | 1.25 (0.25–6.19), 0.782 | |
| Others | 70 (88.6) | 9 (11.4) | 1.77 (0.67–4.67), 0.248 | |
| Religion** | ||||
| Anglican | 61 (91.0) | 6 (9.0) | 1.00 (ref) | |
| Catholic | 69 (88.5) | 9 (11.5) | 1.33 (0.44–3.93), 0.611 | |
| Muslim | 37 (94.8) | 2 (5.1) | 0.55 (0.10–2.87), 0.477 | |
| Pentecostal | 45 (95.7) | 2 (4.3) | 0.45 (0.09–2.3), 0.344 | |
| SDA | 4 (100) | 0 | - | |
| Education level | ||||
| None | 22 (95.6) | 1 (4.4) | 1.00 (ref) | |
| Primary | 77 (89.5) | 9 (10.5) | 2.57 (0.31–21.4), 0.382 | |
| Secondary | 88 (92.6) | 7 (7.4) | 1.75 (0.20–14.9), 0.609 | |
| Tertiary | 29 (90.6) | 3 (9.4) | 2.28 (0.22–23.4), 0.449 | |
| Marital status | ||||
| Married/Cohabiting | 114 (91.2) | 11 (8.8) | 1.00 (ref) | |
| Separated/Widowed | 63 (92.6) | 5 (7.4) | 0.82 (0.27–2.47), 0.728 | |
| Single | 39 (90.7) | 4 (9.3) | 1.06 (0.32–3.53), 0.921 | |
| Alcohol | ||||
| No | 184 (92.9) | 14 (7.1) | 1.00 (ref) | 1.00 (ref) |
| Yes | 32 (84.2) | 6 (15.8) | 2.5 (0.88–6.88), 0.085* | 2.65 (0.93–7.65), 0.069 |
| Clinical characteristics | ||||
| Clinical jaundice | ||||
| No | 152 (94.4) | 9 (5.6) | 1.00 (ref) | 1.00 (ref) |
| Yes | 64 (85.3) | 11 (14.7) | 2.9 (1.15–7.34), 0.024* | 3.03 (1.18–7.76), 0.021 |
| Talked to the doctor about jaundice | ||||
| No | 168 (93.3) | 12 (6.7) | 1.00 (ref) | |
| Yes | 48 (85.7) | 8 (14.3) | 2.33 (0.90–6.03), 0.081* | |
| Perception | ||||
| Disclosure of HIV status | ||||
| No | 13 (86.7) | 2 (13.3) | 1.00 (ref) | |
| Yes | 203 (91.9) | 18 (8.1) | 0.58 (0.12–2.75), 0.49 | |
| Stigma and discrimination (8 item score > 5) | ||||
| No | 209 (92.1) | 18 (7.9) | 1.00 (ref) | |
| Yes | 7 (77.8) | 2 (22.2) | 3.32 (0.64–17.16), 0.153* | |
| Pill burden (they take a lot of pills) | ||||
| No | 143 (92.3) | 12 (7.7) | 1.00 (ref) | |
| Yes | 73 (90.1) | 8 (9.9) | 1.31 (0.51–3.34), 0.577 | |
Variables with p < 0.20 at bivariate analysis were considered for inclusion in the multivariable logistic regression model. A stepwise elimination approach was used to obtain the final adjusted model, retaining variables independently associated with poor adherence at p < 0.05
*p value < 0.05
**1 missing value
In multivariable logistic regression, clinical jaundice was independently associated with poor adherence (aOR = 3.03; 95% CI: 1.18–7.76; p = 0.021), while alcohol use showed a borderline association (aOR = 2.65; 95% CI: 0.93–7.65; p = 0.069). There was no evidence of multicollinearity, and the model demonstrated good fit (Hosmer–Lemeshow: χ² = 0.00, p = 0.9999). No statistically significant interaction was observed between clinical jaundice and sex or age (p > 0.05 for interaction terms).
The predicted probability of poor adherence was 5.5% (95% CI: 2.0%–9.0%) among participants without clinical jaundice, compared to 14.9% (95% CI: 6.9%–22.9%) among those with clinical jaundice, corresponding to an absolute increase of 9.3% points.
Discussion
This study found a high prevalence of clinical jaundice among PLHIV on ATV/r at IDI in Kampala, Uganda, with a significant association between clinical jaundice and poor ART adherence. These findings highlight how ART side effects may compromise adherence, leading to virologic failure, opportunistic infections, and increased mortality [3, 5].
The point prevalence of clinical jaundice (21.1%) was higher than in earlier trials, a randomized controlled trial (6–8%) and the CASTLE study (5%) [11, 20]. Similarly, the period prevalence of clinical jaundice (31.8%) exceeded that of a Hispanic cohort (21%) [4]. The difference in results may be attributed to the variation in UGT1A1 enzyme expression between the African, Hispanic and white populations. The UGT1A1*28 allele, which decreases the expression of the UGT1A1 enzyme, resulting in hyperbilirubinemia, is more prevalent in Africans (40%) than in Hispanics (38%) and whites (31%) [9, 18, 21].
Most participants reported good adherence, consistent with findings from a multicenter study conducted among adolescents in Uganda [15]. The good ART adherence may be attributed to good counselling skills, motivation schemes, and ART refill reminders by the IDI staff. However, adherence was likely to be impeded by the presence of clinical jaundice. Additionally, discontinuation of treatment due to clinical jaundice was rare, consistent with a prior study where less than 1% discontinued ART because of scleral icterus or jaundice [11]. This suggests that while clinical jaundice affects adherence, it rarely leads to treatment cessation since, its temporary. Clinical Jaundice episodes occurred more than one week after initiation of ATV/r, lasted more than two weeks, and were often recurrent, but most participants reported minimal concern. Although ART-related side effects are commonly reported as barriers to adherence, counseling and social support have been shown to influence patient perceptions and may influence how individuals respond to such effects [7].
Implications for practice and policy
These findings suggest the need to strengthen counselling at ART initiation by preparing patients for the possibility of benign, reversible jaundice associated with ATV/r and clearly communicating that it does not necessitate treatment discontinuation. Routine assessment for jaundice during follow-up visits may help identify patients requiring additional adherence support. In addition, ART guidelines and patient education materials should incorporate practical guidance on the recognition and management of ATV/r-related jaundice, and healthcare providers should be supported to address patient concerns proactively.
Future directions
Further research is needed to evaluate whether enhanced counselling and follow-up support can mitigate the impact of treatment-related side effects on adherence. Exploration of alternative regimens with a lower risk of jaundice, including unboosted atazanavir in appropriate settings, may also be warranted.
Conclusion
Clinical jaundice is common among PLHIV on ATV/r and is associated with increased odds of poor ART adherence, posing challenges to achieving the 95-95-95 targets. Strengthening pre-ATV/r (pre-switch) counselling and exploring safer regimen alternatives are critical steps toward improving outcomes.
Limitations of the study
Self-reported measures of jaundice and adherence may be subject to recall and reporting bias. Although validated tools were used, the 8-MMAS, Oslo social support scale, and 8-SSCI were not formally validated in this study population. Clinical jaundice was assessed through self-report and visual examination without laboratory confirmation using bilirubin levels, which may have resulted in misclassification, and inter-rater reliability was not assessed. The study did not incorporate objective adherence measures such as antiretroviral drug concentrations or viral load data, which may have provided additional insight into adherence patterns. In addition, some clinically relevant covariates, including HBV/HCV coinfection, hemolysis or anemia, genetic predisposition such as Gilbert syndrome, concomitant hepatotoxic medications, and nutritional status, were not available and may have resulted in residual confounding. The study was conducted at a single high-volume urban clinic, which may limit generalizability, although findings may be relevant to similar settings. Finally, recruitment was time-bound and the achieved sample size was slightly below the target, which may have reduced statistical precision.
Acknowledgements
The authors would like to thank the staff of Infectious Diseases Institute for their support that enabled data collection and the patients for their willingness to participate in this study. The authors also acknowledge Ronald Olum for reviewing the manuscript and updating some of the references.
Abbreviations
- 8-MMAS
8-item Morisky Medication Adherence Scale
- 8-SSCI
8-item Stigma Scale for Chronic Illnesses
- AIDS
Acquired immunodeficiency syndrome
- AOR
Adjusted odds ratio
- CI
Confidence interval
- HIV
Human Immunodeficiency Virus
- IDI
Infectious Disease Institute
- OR
Odds ratio
- PLHIV
People living with HIV
- SD
Standard Deviation
- WHO
World Health Organization
Author contributions
AKN, SKN, EL and FEK conceived the idea and conducted the study. AKN developed the first draft of the manuscript. AKN, SKN, EK and FEK performed analysis and interpreted the results. AKN, SKN, EL, EK, NO, IL, BC, ML and FEK contributed to results interpretation, reviewed the first draft and drafted the subsequent versions of the manuscript. All authors read and approved the final manuscript.
Funding
No funding was obtained for this study.
Data availability
The datasets during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The research and ethics committees at Makerere University School of Health Sciences, Uganda National Council of Science and Technology reviewed and approved the study. Informed consent was obtained from all the research participants.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Bartlett II, Kotrlik JE, J.W. and, Higgins CC. Organizational research: Determing appropriate sample size in survey research. Inform Technol Learn Perform J. 2001;19(1):43. [Google Scholar]
- 2.Bentu-Ferrer D, et al. Clinical pharmacology, efficacy and safety of atazanavir: A review. Expert Opin Drug Metabolism Toxicol. 2009;5(11):1455–68. 10.1517/17425250903321514. [DOI] [PubMed] [Google Scholar]
- 3.Biset Ayalew M. Mortality and its predictors among HIV infected patients taking antiretroviral treatment in Ethiopia: A systematic review. AIDS Res Treat. 2017;2017:1–10. 10.1155/2017/5415298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bissio E, Lopardo GD. Incidence of hyperbilirubinemia and jaundice due to Atazanavir in a cohort of Hispanic patients. AIDS Res Hum Retroviruses. 2013;29(3):415–7. 10.1089/aid.2012.0233. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bukenya D, et al. What causes non-adherence among some individuals on long term antiretroviral therapy? Experiences of individuals with poor viral suppression in Uganda. AIDS Res Therapy. 2019;16(1):2. 10.1186/s12981-018-0214-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gammal RS, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for UGT1A1 and atazanavir prescribing. Clin Pharmacol Ther. 2016;99(4):363–9. 10.1002/cpt.269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hlophe LD, et al. Barriers and facilitators to anti-retroviral therapy adherence among adolescents aged 10 to 19 years living with HIV in sub-Saharan Africa: A mixed-methods systematic review and meta-analysis. PLoS ONE. 2023;18(5):e0276411. 10.1371/journal.pone.0276411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Horsfall LJ, et al. Prevalence of clinically relevant UGT1A alleles and haplotypes in African populations. Ann Hum Genet. 2011;75(2):236–46. 10.1111/j.1469-1809.2010.00638.x. [DOI] [PubMed] [Google Scholar]
- 9.Kanestri V, et al. Clinical significance of the UGT1A1*28 allele detection in HIV-infected patients. J Int AIDS Soc. 2014;17(4 Suppl 3):19579. 10.7448/ias.17.4.19579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kocalevent R-D, et al. Social support in the general population: standardization of the Oslo social support scale (OSSS-3). BMC Psychol. 2018;6(1):1–8. 10.1186/S40359-018-0249-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.McDonald C, et al. Clinical significance of hyperbilirubinemia among HIV-1-infected patients treated with atazanavir/ritonavir through 96 weeks in the CASTLE study. AIDS Patient Care STDs. 2012;26(5):259–64. 10.1089/apc.2011.0092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mills EJ, et al. Adherence to HAART: A systematic review of developed and developing nation patient-reported barriers and facilitators. PLoS Med. 2006;3(11):2039–64. 10.1371/journal.pmed.0030438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Molina Y, et al. The Stigma Scale for Chronic Illnesses 8-Item Version (SSCI-8): Development, validation and use across neurological conditions. Int J Behav Med. 2013;20(3):450–60. 10.1007/s12529-012-9243-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Nabukeera-Barungi N, et al. Adherence to antiretroviral therapy and retention in care for adolescents living with HIV from 10 districts in Uganda. BMC Infect Dis. 2015. 10.1186/s12879-015-1265-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Panagopoulos P, et al. Hyperbilirubinemia in atazanavir treated HIV-infected patients: the impact of the UGT1A1*28 allele. Pharmacogenomics Personalized Med. 2017;10:205–8. 10.2147/PGPM.S107152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ravindran R. Jaundice. Surg (United Kingdom). 2020;38(8):446–52. 10.1016/j.mpsur.2020.06.008. [DOI] [Google Scholar]
- 18.Ribaudo HJ, et al. Impact of UGT1A1 gilbert variant on discontinuation of ritonavir-boosted atazanavir in AIDS clinical trials group study A5202. J Infect Dis. 2013;207(3):420–5. 10.1093/infdis/jis690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Shillingford JP, Aitken CB. Measurement of feelings using visual analogue scales. Proc R Soc Med. 1969;62(10):989. 10.1177/003591576906201005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Squires K, et al. Comparison of once-daily atazanavir with efavirenz, each in combination with fixed-dose zidovudine and lamivudine, as initial therapy for patients infected with HIV. JAIDS J Acquir Immune Defic Syndr. 2004. 10.1097/00126334-200408150-00003. [DOI] [PubMed] [Google Scholar]
- 21.Turatti L, et al. Short communication: UGT1A1*28 variant allele is a predictor of severe hyperbilirubinemia in HIV-infected patients on HAART in Southern Brazil. AIDS Res Hum Retroviruses. 2012;28(9):815–8. 10.1089/aid.2011.0261. [DOI] [PubMed] [Google Scholar]
- 22.Uganda Ministry of Health. Consolidated guidelines for the prevention and treatment of HIV and AIDS in Uganda. Kampala: Ministry of Health, Uganda. 2022.
- 23.Weiser S, et al. Barriers to antiretroviral adherence for patients living with HIV infection and AIDS in Botswana. JAIDS J Acquir Immune Defic Syndr. 2003;34(3):281–8. 10.1097/00126334-200311010-00004. [DOI] [PubMed] [Google Scholar]
- 24.Krousel-Wood M, Islam T, Webber LS, Re RN, Morisky DE, Muntner P. New medication adherence scale versus pharmacy fill rates in seniors with hypertension. Am J Manag Care. 2009 Jan;15(1):59-66. [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets during and/or analyzed during the current study are available from the corresponding author on reasonable request.