Population.

Eligible studies included adults (≥18 years) living with HIV and HTN or DM or both comorbidities, irrespective of whether HTN or DM onset preceded or followed HIV infection. We focused on PLHIV because they represent a growing population experiencing chronic multimorbidity and they have established care platforms which can influence self-care. Adults were included because self-care responsibilities and treatment decision-making differ substantially from those of children and adolescents. We did not restrict inclusion by duration of diagnosis or treatment status (e.g., receiving treatment or not), as these characteristics were inconsistently reported and were not central to the review objective.

Phenomenon of interest/outcome.

We included studies reporting self-care practices or behaviors undertaken by individuals to manage HIV and comorbid HTN and/or DM. The primary outcome of this review was self-care practices for managing HIV and HTN or DM multimorbidity in LMICs. Self-care practices were defined as health-related actions or measures undertaken by individuals to manage illness, maintain health, or restore well-being, encompassing physical, psychological, spiritual and social strategies [34]. This review did not include studies reporting clinical or disease outcomes among PLHIV and HTN or DM comorbidities.

Type of studies/ study design.

We included primary studies employing qualitative, quantitative or mixed methods designs. Eligible quantitative studies comprised observational studies (cross-sectional, cohort and case-control studies), interventional studies and surveys. Qualitative studies were eligible regardless of design or data collection method. We excluded secondary research papers (including systematic reviews and protocols), policy analyses, editorials, commentaries, abstracts, conceptual papers, and studies involving non-human participants.

Context.

Only studies conducted in LMICs were included. We focused on LMICs because distinct health system constraints, resource limitations, and the evolving integration of HIV and NCD care shape self-care practices in these settings. Therefore, patient experiences and challenges in LMICs differ markedly from those in high-income counties, influencing how self-care is understood and practiced. A summary of inclusion and exclusion criteria is provided in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Review inclusion and exclusion criteria.

https://doi.org/10.1371/journal.pone.0360076.t001

Data sources.

We searched the following databases: Ovid Medline, Ovid EMBASE, Ovid Global Health, Ovid PsycINFO, Web of Science (including Social Science Citation Index), Scopus, CINAHL via EBSCO, WHO Global Index Medicus, Google Scholar and African Journals Online (AJOL). The AJOL was included to capture African research not indexed in major databases.

Search strategy.

Search terms were developed in consultation with an information specialist at the London School of Hygiene and Tropical Medicine (LSHTM). The search strategy combined three key concepts and their synonyms: “self-care”, “human immunodeficiency virus,” and “hypertension” or “diabetes”, using Boolean operators. We modified search terms as required by each database and used both keywords and medical subject headings (MeSH). Some of the MeSH terms were: “Self-Care”, “Self-manage”, “self-monitor”, “self-medicate”, “adherence”, “PLWHIV”, “PLHIV”, “HIV”, “HIVAIDS”, “HIV Infection”, “acquired immune deficiency syndrome” “SIDA” “Hypertension”, “Blood Pressure”, “Diabetes”, “Diabetes Mellitus”, “high blood sugar”, “hyperglycemia”. Reference lists of included studies were also searched for possible papers. Given AJOL’s limited search and export functionality, a manual search was done and retrieved titles and abstracts. The literature search covered studies published between January 2013 and March 2026 in LMICs. The review involved a two-stage search process. An initial data search was conducted in December 2023. To ensure that the review incorporated the most recent evidence prior to submission, we conducted a final update of the literature search in March 2026 by re-running the same search strategies across all databases. Newly identified records were screened using the same eligibility criteria and incorporated into the final synthesis. The period from 2013 was chosen because it captures both the expansion of self-care concepts in literature and a period during which HIV/NCD integration has evolved [35,36]. There were no language restrictions to maximize the scope of relevant articles from LMICs. Full search strategies, including MeSH terms, are provided in S3 Table.

Study screening and selection.

Following the search, all citations were imported into EndNote X20 software (Clarivate Analytics) [37], and duplicates removed. Screening was managed using Rayyan software [38]. Two reviewers (AKT and ASS) independently screened titles, abstracts, and full texts. We resolved discrepancies through discussion and by involving a third reviewer (SM).

Data extraction.

Two reviewers (AKT and ASS) independently extracted the data using an extraction form guided by Cochrane [39,40], with discrepancies resolved through discussions with another reviewer (SM). Extracted data included authors and year of publication, country, study aim, study design, participant characteristics (including age and sex), participant NCD comorbidity, sample size, data collection and analysis methods, and key findings. We also extracted texts, quotations and relevant data on self-care practices.

Inter-rater reliability.

Study selection and data extraction were performed independently by two reviewers. Inter-rater reliability was assessed using Cohen’s kappa. For full-text study selection, (36 articles assessed; 13 included) the kappa value was 0.84, and for data extraction across double-extracted items, kappa value was 0.80, indicating substantial agreement between reviewers.

Quality assessment.

Since only qualitative studies were included, we used the critical appraisal skills program (CASP) appraisal checklist for qualitative studies to assess the quality of the studies [41]. CASP checklist for qualitative research is a widely used critical assessment tool for evaluating the quality and relevance of qualitative research by assessing their methodological rigor (including assessing key aspects like recruitment, data collection, and analysis) and ensuring that ethical standards were met. Two reviewers independently assessed each article and assigned an overall quality ranking of “low”, “medium”, or “high” quality. We followed the score and classification system suggested by Butler et al [42]. Hence, every “Yes” scored 1 point, every ‘Can’t tell or Not sure’ scored 0.5 points and every ‘No’ scored 0. Articles were then classified into three categories following their total score: 1) high-quality paper (scored between 9.5 and 10), moderate-quality paper (scored between 7.5 and 9), low-quality paper (scored less than 7.5) and those under 6 points were to be excluded. Consistent with qualitative evidence synthesis methodology, no studies were excluded based on quality assessment scores, as excluding studies solely on methodological quality grounds may remove contextually rich evidence and introduce selection bias [43]. Instead, quality appraisal was used to assess methodological rigor and transparency and to inform the interpretation of the review findings through the Grades of Recommendation, Assessment, Development and Evaluation – Confidence in Evidence from Reviews of Qualitative Research (GRADE-CERQual) approach [44,45]. Notably, after the appraisal, all included studies were rated as either high or moderate quality, and no low-quality studies were identified.

Data analysis.

Since only qualitative studies were included, we conducted thematic synthesis following Thomas and Harden’s three stages: 1) line-by-line coding; 2) development of descriptive themes; and 3) generation of analytical themes [43]. The coding framework was developed iteratively. Initial line-by-line coding was conducted independently by two reviewers (AKT and ASS) on a subset of included studies to develop and pilot the preliminary coding framework. The two reviewers compared their coding, discussed differences in interpretation, and refined the coding framework through consensus. This piloting phase allowed reviewers to test and calibrate the coding approach, identify areas of ambiguity, and establish a shared understanding of the coding categories before proceeding to the full data set.

The revised framework was then applied to all included studies and further refined iteratively as new codes and patterns concepts emerged from the data. Regular discussions between reviewers were held throughout the synthesis process to ensure consistency, challenge assumptions, and maintain reflexivity in interpreting the findings. Discrepancies in coding or thematic interpretation were resolved through discussion and consensus, with consultation from a third reviewer (SM) when agreement could not be reached. To address reflexivity and minimize reviewer bias, the team (AT, AS, SM and AK) maintained ongoing critical dialogue throughout the synthesis process, reflecting on how prior assumptions and disciplinary perspectives might influence interpretation of the data. This reflexive approach was embedded throughout the analysis rather than a single discrete step.

This inductive approach enabled reviewers to move beyond descriptive findings to generate interpretative and analytical insights. Descriptive themes of self-care practices were then generated to summarize the meaning of the initial codes and code groups. These descriptive themes were further interpreted and organized into analytical sub-themes and overarching themes, providing a deeper understanding of the data.

Although the review protocol planned to include both qualitative and quantitative studies, our search yielded only qualitative evidence. Consequently, the synthesis focused exclusively on qualitative findings using thematic synthesis, rather than narrative synthesis, or meta-analysis. All eligible qualitative findings were included to ensure comprehensive and unbiased synthesis.

Contributions of studies to the final thematic findings.

We assessed each study’s relative contribution to the final synthesized results: (a process proposed by Thomas and Harden [43]). We used the subthemes of our thematic analysis to determine each study’s contribution. These sub-themes represented granular analytical insights, allowing us to evaluate how each study informed the overall synthesis. Since our review had 17 sub-themes, a score of one meant that the study provided information to one sub-theme, while a score of 17 meant that that study contributed to all the sub-themes of this review. (This approach has been used in prior review studies to score the contribution of individual papers to review findings [46,47]).

Confidence in the findings and identifying the treatment burden.

We used the GRADE-CERQual approach to assess the level of confidence in each review finding under its respective theme [44,45]. We first created a summary of review findings, then evaluated each review finding across the four GRADE-CERQual components: 1) methodological limitations, 2) coherence, 3) adequacy of data, and 4) relevance [44,45]. For each component, we judged the level of concern (in terms of no/very minor, minor, moderate, or serious concern), and used these judgements to assign an overall confidence rating (high, moderate, low or very low), starting with high confidence and downgrading as concerns increased [44,45]. Confidence judgments were reached through discussion and consensus between two reviewers (AKT and ASS).

To identify treatment burdens, findings with high or moderate confidence were onto the NPT four domains: coherence, cognitive participation, collective action, and reflexive monitoring [26,27]. Finding that did not align with these domains were categorized as “unallocated” following Chikumbu et al. [28]. We identified specific treatment burdens under each domain by capturing demands and tasks described to take patients’ time and efforts in each review finding.

Grading and classifying the severity of illnesses.

Patients commonly assessed and ranked their condition based on perceived severity. Across several studies, HIV was described as a less severe, and more manageable illness compared with HTN or DM, which were viewed as unpredictable and potentially fatal [48–51,56,58]. This perception shaped treatment priorities and adherence behaviors. In contrast, HTN and DM were associated with sudden death, reinforcing fear and urgency in their management [56,58–60]:

“The dangerous disease nowadays is blood pressure, because even sitting here, … I can die. While HIV, you stay longer “ (Female, 70 years old; HTN [56]).

Greater confidence in antiretroviral therapy than in NCD medications.

Patients consistently expressed strong trust in ART, describing ART as highly effective, reliable, and lifesaving [56,57,59]. This confidence fostered strong adherence to ART and optimism towards HIV care [50,52,56,57,59]. In contrast, medications for HTN and DM were seen as less effective, prompting doubts in efficacy and, use of herbal remedies to manage their NCD comorbidities [56,59,60]:

“I went to the herbal clinic in search of an alternative, hoping to get treated. Because all these medications (hypertension and diabetes medication) are just for pain relief and not curative. The herbal clinic has medications that could cure diabetes completely” (Male, 60 years; HTN).

Prioritizing care for HTN and DM comorbidities.

Grading illnesses influenced how patients allocated their time, emotional effort, and limited financial resources towards care [48,51,56,58,59]. Because HIV care routines were well-established, participants prioritized HTN and DM management due to their perceived immediacy and danger [51,58,59]. Some patients turned to dietary changes and traditional remedies, but they still indicated frustration due to poor health outcomes for the NCDs [50,51]:

“With pressure, it does not go down …, I try to eat things I am told to eat but it is always high.” (Female, 48 years old; HTN [51]).

Consistent ART adherence versus intentional or selective non-adherence to NCD medication.

Patients demonstrated strong adherence to ART, which was widely perceived as lifesaving and easy to manage through routine pill-taking [49,50,52,56,57,59]. The free availability and reliable access to ART further reinforced consistent use [50,52,56,58]. In contrast, adherence to HTN and DM medications was consistently lower. Although none of the studies quantified this discrepancy, the pattern of non-adherence to NCD medications was observed, even where NCD medicines were accessible [48,52,57–59]. Non-adherence to NCD medicines was linked to the absence of symptoms, inconsistent advice from health workers, medicine stock-outs and side effects [48,52,57–60]:

“I… do not use insulin so much because it affects my legs. They get [so] swollen that I can’t walk. So, I use it some days and skips other days” (Female, 50 years; HTN&DM [58]).

A few patients with reliable access expressed strong commitment to adherence, often motivated by personal or family health histories [48,56]:

“My father died of hypertension, and uncle had a stroke because of it, so I don’t miss the hypertension drugs” (Male, 52 years; HTN [56]).

Use of alternative and complementary treatments.

Low confidence in NCD medications led many patients to supplement or replace treatment with herbal remedies included consuming bitter herbs and vegetables (such as ginger, garlic, mululuza (Vernonia amygdalina), prayer, or consulting with traditional healers [52,55,57,59,60]. The belief that these remedies would cure NCDs and positive experiences of symptom relief often reinforced trust in these remedies:

“I take the mushroom when I see that my heart rate is high. … it helps with blood pressure. When I don’t have it; I know I’m putting my life at risk. (Female, 60 years; HTN [57]).

Stress reduction techniques to manage HTN.

Some patients attributed HTN to emotional stress and adopted avoidance or withdrew strategies including avoiding conflicts, noisy environments, and stressful discussions to manage HTN [48,57,59,60].

“I should not think a lot, …neither should someone speak to me in a loud frightening way” (Female, 49 years; HTN [57]).

Managing multiple medications (Polypharmacy).

Patients adopted practical strategies to manage complex treatment regimens, including synchronizing medication schedules, storing medications together in “drug banks,” and involving family members in supporting treatment adherence [50,56,59]. However, concerns about pill burden and possible drug toxicity due to the many medicines led some to substitute them with physical exercises, herbal remedies, or to discontinue NCD medications [48,59]:

“I refused taking medication for the virus and for pressure … that is too many drugs in your body” (Female, 42 years; HTN [48]).

Dietary modifications and challenges.

Many patients attempted to change their diet by reducing salt, sugar, and fatty foods, eating smaller portions more frequently, drinking a lot of water to “flush out” medicines, and eating food before taking medicines to prevent side-effects [48,51,53,57,60]. However, food insecurity, alongside poverty, limited food options and social norms on food consumption undermined planned dietary changes and medicine adherence, particularly when medicines had to be taken with food [49–52]:

“This is why I have to skip taking my insulin because my body gets weak, I shake and feel like going mad. … because I don’t have food” (woman, 46-years, DM [52]).

Lifestyle adjustments and condition-specific practices.

In addition to dieting, patients engaged in physical activity, reduced alcohol intake, and avoiding smoking [49,51,53,60]. Some patients also took immune boosters to enhance their immunity. We also observed condition-specific practices in this population. For HIV, patients emphasized safe sex and wound care [53]. For HTN and DM, self-monitoring of blood pressure and glucose was highlighted, though glucose monitoring was often limited by cost [50,59]. For DM only, additional practice was foot care [53]:

“… for type two diabetes, … proper foot care, included washing and drying feet properly; and wearing the right shoes “to be safe” [53].

Acceptance as a coping mechanism.

Acceptance of illness, particularly HIV, emerged as a key psychological adaptation that enabled patients to rebuild self-worth and engage in self-care [50,53,54]. Patients described reaching acceptance through counseling and gradual emotional adjustment [51,53,54]:

“I could not accept living with HIV, but as time went by, I accepted … and that’s what I am now” (P1, Female, HTN [53]).

Managing stigma.

HIV-related stigma drove many patients to conceal their status and ART use, mostly through maintaining secrecy of status, hiding HIV medicines and social withdrawal [51,53,59]. In severe cases, some patients would physically withdraw from stigmatizing environments:

“My brothers stigmatize me a lot. …, they insult me the whole night... So sometimes I don’t even stay at my own home.” (Female, 44 years [59]).

Maladaptive coping.

Some patients resorted to maladaptive coping strategies [51,60], particularly alcohol use to cope with depression associated with multimorbidity [51]. While drinking alcohol offered temporary relief from the feelings of hopelessness, it risked worsening their health outcomes [51,60]:

“I do drink alcohol, … Yes, I sometimes have that problem [depression] I drink alcohol day before going to clinic” (Female, 48 years old; HTN [51]).

Navigating limited medicine availability.

The most recurrent challenge was the inconsistent availability of essential medicines for HTN and DM. Many patients were told to purchase medications out-of-pocket, which was financially unsustainable [50,52,53,55,56,58–60]. Consequently, some patients resorted to visiting multiple clinics to find medicines, borrowing from peers, substituting prescribed drugs with painkillers and in more constrained situations, patients discontinued treatment altogether until they could afford refills [50,52,56,60]:

“They give you one pack and tell you to buy the rest. You may not have money for buying drugs at that time.” (Woman, 42 years, DM [55]).

These shortages were reported across both fragmented and integrated HIV/NCD service models, underscoring persistent inequities between HIV and NCD care systems [49,50,52,55,56,58].