Table 1.
Characteristics of included studies
| First author, year (country) | Participant characteristics and study setting | Methods | Intervention reported or tools used for medicine optimisation | Key findings | Limitations |
|---|---|---|---|---|---|
| Ahmed et al., 2021 (Pakistan) [65] |
n = 602 32.9% > 50 years Hospital |
Participants were given a validated generic HRQOL questionnaire | EuroQol quality of life scale EQ-5D-3L and Visual Analogue Scale | 59.5% of participants reported no impairment in self-care, however, 63.1% were extremely anxious/depressed. Overall, the mean EQ-5D utility and Visual Analogue Scale scores were 0.388 (SD = 0.41) and 66.20 (SD = 17.22), respectively. Multiple regression analysis has shown that age over 50, the female gender, primary or secondary education, less than a year since HIV diagnosis, having a detectable viral load, and a longer time to ART were all factors significantly associated with HRQOL | The results cannot be generalised to non-adherent PLWH as participants who failed to show up regularly according to their dispensing records were excluded from the study. Moreover, the cross-sectional study design does not allow for causality analysis and there is a possibility of social desirability bias as participants are likely to under-report socially undesirable behaviours |
| Ventuneac et al., 2020 (USA) [53] |
n = 406 M = 50.7 Community |
Intervention study: This study consisted of a single-arm prospective study design with assessments at baseline and 6 months | Rango (mobile health application) | 95% of participants returned for a follow-up visit, with 59% (38/65) of those who were unsuppressed at baseline achieving viral suppression. Viral suppression among Rango participants and those receiving usual care were similar (p = 0.84) and increased in both groups at six months (p < 0.001). Significant difference in the number of unsuppressed participants (n = 65) at baseline who were suppressed (n = 38) at 6 months (p = 0.006) | The findings are limited by the short study duration and a lack of useability data. Usage changes were not analysed over the 6 months |
| Nguyen et al., 2018 (USA) [62] |
n = 176 M = 58.7 years (SD = 5.4) Community |
Data pertaining to HIV-positive participants was obtained from the 2012 baseline cohort from the Research Core of the Rush Centre of Excellence on Disparities in HIV and Aging | The CESD-10, the 30-item MMSE, the 5-item De Jong-Gierveld Loneliness Scale, and the Everyday Discrimination Scale | Participants with good/excellent health showed greater purpose in life, fewer depressive symptoms, more education, and less loneliness than those with poor fair health. Less depressive symptoms, disabilities, adverse life events, and loneliness were associated with higher healthy days index scores. Health-related quality of life was linked to disabilities, smoking status, depression, race/ethnicity, and purpose in life | The cross-sectional study design does not allow for causality interpretations, and self-reported data may include response bias. The participants in this study had good virologic control and thus generalisability to those with less virological control may be reduced |
| DeFulio et al., 2021 (USA) [98] |
n = 50 M = 52.4 years (SD = 10.7) Community |
Intervention study: Medication Event Monitoring System (MEMS) cap were given to participants attached to their primary ART medication bottle to measure adherence. Participants were placed into either the intervention or control group. Participants were given smart phones that had the intervention app and had to submit videos of medication consumption. At the beginning and every month after surveys were given to participants | Medication Event Monitoring System (MEMS) cap and smartphone-based intervention app ("SteadyRX") | Participants in the intervention group submitted 75% of required videos, of which 81% met validity criteria, thus indicating a high usability level. Over the study duration the percentage of adherent participants decreased in the control group (p = 0.031). The control group self-reported an average adherence of 91.10% adherence and 94.34% was reported by the intervention group, but this was not a significant difference | A limitation of this study is the small sample size and single site recruitment, reducing generalisability of the findings. Another study limitation could be the lack of requiring a detectable viral load. Finally, the biometric data collecting procedure was flawed, creating another study limitation |
| Lopez-Centeno et al., 2020 (Madrid) [20] |
n = 1292 Median = 69 years (67–73) Community and Hospital pharmacies |
Dispensation registries of community and hospital pharmacies from the Madrid Regional Health Service was analysed between January to June 2017. The Beers criteria was used to identify potentially inappropriate medications among older PLWH | The 2019 American Geriatrics Society Beers criteria | Polypharmacy was observed in 65.9% of older PLWH. Cardiovascular (69.7%), gastrointestinal and metabolism (68.2%), and nervous system (61.0%) drugs were among the most prescribed co-medications among participants. At least one potentially inappropriate medication was identified in 37.3% (482) participants. 667 potentially inappropriate medications were identified in 482 participants, 60.8% (293) involved benzodiazepines, and 27.2% (131) involved nonsteroidal anti-inflammatory drugs | One limitation to the study is that over-the-counter medicines were not included, this may have led to an under-estimation of potentially inappropriate medications. Moreover, the study is limited by a lack of information on participants co-morbidities, medical managements, such as potential dosage adjustments, and the absence of information on the clinical outcomes of patients with potentially inappropriate medication |
| Jakeman et al., 2021 (Switzerland) [74] |
n = 1019 Median = 70 years Nationwide |
Prescriptions of eligible participants from the Swiss HIV Cohort Study (SHCS) were reviewed to assess anticholinergic (ACH) medication | An average of 5 (± 3.6) non-HIV drugs were taken by participants. 20% of participants were on one ACH medication, reporting average ACH scores of 1.7 (± 1.3). Self-reported neurocognitive impairment was associated with depression and being on one ACH medication | A limitation to this study includes that adherence to ACH medications were not assessed so it cannot be determined if participants were taking their prescribed medications. Dose or duration of ACH medication and history of CNS infection was not evaluated, which all impacts neurocognitive impairment. Self-reported depression rating scales were not available; thus, the diagnosis of depression may have been missed in some of the cohort | |
| Fischetti et al., 2022 (USA) [51] |
n = 1144 Median = 52 years Community |
Data were collected retrospectively via medical chart reviews | Most participants (48%) had one or two co-morbidities, with two participants having 5 co-morbidities. 80% of participants had an undetectable viral load. Higher viral suppression was seen in participants with more co-morbidities (p = 0.009). It was reported that participants with psychiatric disorders had the lowest viral suppression compared to other co-morbidities | Generalisability of the study findings is reduced due to recruitment being from a single site. The number of medications taken per disease and data on disease control was not taken | |
| Hartzler et al., 2019 (USA) [60] |
n = 44 M = 52.3 years Hospital and Community |
Data were collected through focus groups and a survey for demographic information | The study found an emerging theme across staff focus groups of wanting therapy to be patient centred, adaptable, and mission-congruent. Patients reported desiring therapy to have patient autonomy in illness management and fairness among service users. Staff perceived higher compatibility for motivation interviewing than cognitive behavioural therapy or contingency management, this was similar among patients albeit a less robust or reliable pattern | A limitation of the study includes potential bias in participant responses within the focus groups due to social desirability. The generalisability of the study findings is reduced due to the small sample size | |
| Furlotte et al., 2017 (Canada) [92] |
n = 11 Age: 52–67 years Community |
Semi-structured interviews, a checklist of health and social services, and a demographic questionnaire were used to collect data | Three main themes emerged from the interviews: uncertainty, stigma, and resilience. Uncertainty impacting on mental health was reported due to unexpected survival, medical uncertainty, and perception of one's symptoms. Stigma experiences were caused by discrimination in health care interactions, being stigmatised due to physical appearance, anticipated stigma, misinformation, and compounded stigma. Individual approaches to resilience helped participants cope with these experiences, examples include decreasing the space that HIV consumes in their life, making lifestyle changes around the condition, and using social supports | Limitations to the study include that recruitment was of participants attending their local clinic service and the results cannot be generalised to those not attending care services. Full mental health histories were not taken; thus, it is not possible to determine whether participants had mental health concerns before or after their HIV diagnosis | |
| Lee, 2019 (USA) [82] |
n = 97 M ≥ 50 years for all groups Community |
Self-report measures were used to collect data on demographics, ART medication information and medication adherence. Pharmacy and medical records were also used for data collection | Medication adherence, both self-reported and pharmacy-based, or executive functioning was not significantly associated with cigarette smoking. Poor self-reported medication adherence was associated with symptoms of clinical and subclinical levels of anxiety and depression | Generalisability of the study findings are reduced due to recruitment being from a single site. Overall status of health was not measured, which may have impacted the findings. Some participants found the self-reported adherence questionnaire to be difficult to understand, this may have affected the findings, also pharmacy-based adherence data was based on pharmacy refill data, which may not always mean participants are taking the medications they collect | |
| Jacomet et al., 2020 (France) [59] |
n = 1137 M = 50.2 years Hospital and community |
Data were collected through surveys | The presence of co-morbidities was reported in 64.2% of participants and 90% had a undetectable viral load. 58% of participants knew of the medication file, however, only 40% of pharmacists reported to offering it systematically. 32% of participants would like to use the medication file programme, particularly those with shorter ARV duration, a less often undetectable viral load and those who experience anxiety more often | Generalisability of the study results are reduced due to the sample population being predominantly male | |
| Zheng et al., 2022 (China) [61] |
n = 185 M = 58 years (SD = 7.5) Hospital and community |
Data were collected using questionnaires | The Chinese version of the Living with Medicines Questionnaire version 3 (LMQ-3) and the Centre for Adherence Support Evaluation (CASE) Adherence Index | Polypharmacy was reported in 40% of participants. A higher level of medicine-related burden was reported in females, who took more drugs and had a lower monthly income. ART adherence was negatively associated with medicine-related burden (p = 0.001) | Limitations include that data were collected via self-report and was not verified using medical records, prevalence of PDDI in this study may have been underestimated as the Liverpool interaction database does not include information on traditional Chinese medicines or herbal drugs, and as recruitment occurred only at two sites in the Hunan province, findings may not be generalisable elsewhere |
| Uphold et al., 2004 (USA) [46] |
n = 19 (≥ 50 years) and n = 18 (< 40 years) Older group: M = 58 years (SD = 7) Hospital |
Data were collected using electronic medical records | Adverse side effects from highly active antiretroviral therapy (HAART) were uncommon in the two age groups, with only 4 participants stopping HAART due to adverse effects. Viral load improved significantly for both groups on HAART (p = 0.0001) | Generalisability of the study results are reduced due to the sample population being predominantly male | |
| McInnes et al., 2013 (USA) [81] |
n = 1871 36% = 45–54 years, 43% = 55–64 years and 11.7% ≥ 65 years Community clinics |
Data collected via the Veterans Aging Cohort Study was used to investigate an association between patient electronic personal health record use and ARV adherence. Pharmacy-refill data was used to assess adherence | Personal health record: My HealtheVet | The study found that younger participants (under 45 years) were less adherent than older participants (over 55 years). Participants ≥ 65 years old were less likely to use the personal health record than those < 45 years old. Personal health record use was linked to better adherence | The cross-sectional study design limits the ability to establish a causal relationship between personal health records and adherence. Moreover, there may be unmeasured confounding factors. Pharmacy refill data may overestimate adherence as patients may be collecting their medications without administering them |
| Rosenfeld et al., 2020 (UK) [41] |
n = 100 Median = 56 years (50–87) Hospital and community |
Focus groups, life-history interviews, and surveys were used to collect data | Distinguishing support from HIV-negative (Goffman's 'the wise') people and support based on experiences of PLWH themselves (Goffman's 'the wise'), participants viewed the former as requiring supplementation by the later. Experientially based support varied across groups | A limitation to this study is that focus groups may have led to social desirability bias or moderator bias | |
| Gardenier et al., 2010 (USA) [80] |
n = 56 M = 50.5 years (SD = 8.5) Community (AIDS day health care program) |
Information was extracted from medical records then reviewed and corrected by participants. Participants also completed the Social Provision Scale and the AIDS Clinical Trial Group adherence follow-up instrument | Social Provision Scale (SPS) and the AIDS Clinical Trial Group (ACTG) adherence follow-up instrument | There was a statistical significance between adherence and social support (p = 0.02). Out of the 51 participants who were prescribed ART, 28 (55%) were adherent. There was a statistically significant difference between CD4 T-cell counts between the adherent and non-adherent group, with the latter being lower (p = 0.004) | The study is limited due to the sample size, recruitment not being randomised, and the participants being recruited at two New York city locations from the AIDS Day Health programme |
| Bosire, 2021 (South Africa) [57] |
n = 15 40–70 years Hospital and community |
Data were collected using an ethnographic approach and through 90-min interviews | Participants access to care to manage their co-morbidities were impeded by fragmentation of care, having multiple clinic appointments, conflicting information, and poor patient-provider communication | The small sample size and single site for recruitment reduces the generalisability of the study findings. Language barriers and the use of an interpreter may have changed the meaning of statements reported by participants | |
| Frazier et al., 2018 (USA) [79] |
n = 3672 over 50 years old Hospital and community |
Matched interview and medical record abstraction data from a surveillance system, the Medical Monitoring Project (MMP), was analysed | Women living with HIV over 50 years old were more likely to be prescribed antiretroviral therapy, be virally supressed, be dose adherent, and are less likely to have received sexually transmitted infection prevention information from a healthcare provider, have condomless sex with a negative or unknown partner and report depression compared to women under 50 years old | The limitations to this study are that generalisability of findings are reduced due data being collected from HIV positive women who are in care and not those who are not receiving medical care and the cross-sectional study design does not allow for causality analysis of findings | |
| Schatz et al., 2021(South Africa) [48] |
n = 23 Age ≥ 50 years Community |
In-depth semi-structured interviews | Perceived shame of sexuality and disrespect by clinical staff, disclosing serostatus to others, affording transport to clinics and co-morbidities were key age-related barriers to ART access. Age-related facilitators were financial and moral support from families and access to social grants | Limitations to this study include the small sample size and recruitment of participants who were already tested and linked to care, reducing the generalisability of the findings. Moreover, there may be other barriers or difficulties relating to not testing, late testing, and failing to attend care services that the study was unlikely to identify due to the sample population | |
| Schatz et al., 2022 (South Africa) [84] |
n = 161, PLWH > 40 years Community |
Focus group discussions | Participants reported fewer negative consequences of discloser in 2018 compared to 2013. Participants reported positive outcomes such as building trust, and greater support with adherence and medication collection | A limitation of the study is that participants in the 2013 focus group were different to those in 2018, which may have influenced the findings | |
| Contreras-Macias et al., 2020 (Spain) [73] |
n = 19 M = 69.4 years Outpatient |
Data were obtained from medical records, the electronic prescription programme, and the outpatient dispensing programme | Medication Regimen Complexity Index (MRCI), The European AIDS Clinical Society (EACS) guideline of “Selected Top 10 Drug Classes To Avoid (Top-10-A) in elderly PLWHIV”, and the STOPP-Beers criteria | Polypharmacy was reported in 84.2% (16) of participants and a Top-10-A potentially inappropriate prescription was evident in 47.4% (9) of participants. The most prevalent group of prescribed drugs were benzodiazepines, reported in 30% (6) of participants. 57.9% of participants were complex patients with a MRCI index above 11.25. A higher sum of STOPP-Beers criteria was identified in older patients. Analysis using the t-student test showed a statistically significant relationship between MRCI score and the sum of the STOPP-Beers criteria with increasing age (p < 0.05) | The generalisability of the study findings is reduced by the single-centre study design and small sample size. Moreover, the STOPP-Beers criteria were validated to non-HIV patients 65 years or older |
| Contreras-Macias et al., 2021 (Spain) [70] |
n = 428 M = 50 years (SD = 10.9) Outpatient |
Data were collected using the Capacity-Motivation-Opportunity pharmaceutical care model at routine follow-up appointments | Medication Regimen Complexity Index (MRCI) and the EQ-5D-5L questionnaire | Polypharmacy was identified in 25.9% (111) of participants, with 5.4% (23) being on 11 or more medications. A negative correlation between ED-5D and MRCI scores was identified (p = 0.0002). The relationship between co-morbidity and quality of life was statistically significant in the thyroid-mechanic (p = 0.002) and geriatric-depressive (p = 0.003) patterns | Although the study had a large sample size, the generalisability of the findings is limited due to the study being based at a single urban safety net hospital. Moreover, participants included in the study were those engaged in the care and cannot be generalised to patients who fail to attend care |
| Engelhard et al., 2018 (The Netherlands) [26] |
n = 331 (HIV positive) M = 51 years (SD = 11.2) Outpatient |
HRQOL was measured using a survey in a nationwide sample of PLWH. Data from studies in diabetes and rheumatoid arthritis were added | Medical Outcomes Study Short Form 36-item Health Survey | The HIV sample had the lowest mental health score, with the odds of poor mental HRQOL being higher in HIV patients than other groups. The chances of poor physical HRQOL were similar in both the HIV and diabetes groups, but lower in the rheumatoid arthritis group. Poor physical HRQOL among PLWH was linked to a history of AIDS, longer time on combination ART and severe co-morbidity. Being of Sub-Saharan African descent and having CD4+ counts of less than 350 was linked to poor mental HRQOL | A limitation of this study is that comparing HRQOL between different datasets may lead to findings resulting from other unmeasured factors and not the diversity of the diseases. Socioeconomic status, substance use, and sexual orientation were potential confounders that the study was unable to adjust for, due to inconsistent recording across data sets. Other limitations include that disease severity was not corrected for and participant samples of the other diseases were not a national cross-section of the patient populations |
| Moitra et al., 2011 (USA) [87] |
n = 16 M = 52.5 years (SD = 5) Community clinic |
Intervention study: 3–5 weekly 60-min acceptance-based behaviour sessions were conducted in groups of 3–5 participants. Discussions included overall acceptance-based principles, with each session being a stand-alone intervention | Acceptance-based behaviour therapy | 37.5% (6) participants found the groups to be very helpful and another 37.5% (6) found them to be moderately helpful, whilst 25% (4) found them minimally helpful. The study reported that qualitative observations suggested that the acceptance-based intervention strategies were well suited in the target population. A significant point made in every session was that avoiding the realities of living with HIV can lead to worsened health | The small sample size reduces the generalisability of the study. Group formats for each session may have impeded recruitment or undermined treatment acceptability for potential participants |
| Farahat et al., 2020 (Saudi Arabia) [47] |
n = 13 M = 50.1 years Hospital |
Data were collected retrospectively via medical chart reviews | Out of 130 participants that were included, 48.5% had one or more co-morbidities. Diabetes (15.4%), dyslipidaemia (10.8%), hypertension (10.8%) and lymphoma (10.0%) being the most common co-morbidities. An increase in co-morbidities was seen with an increase in age, with 40.7% of participants aged 60 years or older having three or more co-morbidities. Logistic regression analysis showed that only patients aged 50 years and older were more likely to have at least one co-morbidity | The generalisability of the study is limited by the small sample size. Also, the cross-sectional study design does not allow for causality analysis. Moreover, the study did not look at adherence and medical records were reviewed over an 18-year period, but several ART medication doses and durations were missing | |
| Guaraldi et al., 2017 (Italy) [28] |
n = 482 M = 53.9 years (SD = 6.9) Multidisciplinary clinic |
Patients were evaluated using two frailty tools as part of routine protocol | The Frailty Index and frailty phenotype | The frailty phenotype categories were: 51.9% pre-frail, 3.1% frail, and 45% robust. The mean Frailty Index score was 0.28±0.1. Falls and disability were linked to the Frailty Index but not the frailty phenotype | Due to the cross-sectional study design, the Frailty Index could not be assessed over time for the prediction of adverse outcomes. The lack of a HIV-negative control group and more objective tools to assess disability are further limitations |
| Owen et al., 2012 (UK) [71] |
n = 10 Median = 57 (52–78) years Hospital and community |
A biographical narrative approach was used to collect data | Findings showed that some participants were positive about ageing regarding it as progressing towards valued life goals, whereas others were more conflicted with future prospects. The individual's biographic relationship with the HIV epidemic history rather than their age influenced the differences in views of the future. Participants who were involved with HIV for longer were more likely to have interrupted careers due to illness, depend on state benefits, and have damaged social networks | The study findings were limited by the small sample size and recruitment location | |
| Date et al., 2022 (UK) [101] |
n = 164 Age: MOR: median = 59.5 (SD = 50–78) years; standard care: median = 60 (SD = 50–82) years Hospital |
Intervention study: Participants were randomized to either receive standard care or a Medicines Management Optimisation Review (MOR) | Medicines Management Optimisation Review toolkit, the University of Liverpool and Toronto General Hospital HIV drug interaction references, EuroQol five-dimension five-level (EQ-5D-5L) questionnaire and visual analogue scale (EQ-VAS) | Seventy participants were in the intervention group and ninety-four in the standard care group. Significantly more medicine-related problems were identified in the intervention group at baseline (p = 0.001) and 6 months (p = 0.001). There was a significant reduction in new medicine-related problems at 6 months in the intervention group compared to baseline (p = 0.001), with 44% being resolved at baseline and 51% at 6 months. There were no changes in HRQOL identified between groups or after the intervention. Participants and healthcare professionals found the MOR highly acceptable | The limitations to this study include the sample size, MORs required extra attendance to the clinic which may have precluded some participants, and as this was a feasibility study, it was not powered to measure the effectiveness of the intervention |
| Cho et al., 2018 (USA) [100] |
n = 36 Control group: M = 52 years (SD = 6.6) Intervention group: M = 51 years (SD = 13) Community |
Intervention study: Follow-up focus groups lasting 60–90 min were conducted using semi-structured discussion guides to allow participants to discuss their experiences and any issues with using the mVIP app, after the clinical trial had ended | mVIP (a web-app) | Focus groups revealed the five following themes related to predisposing factors: ease of using the app, being user-friendly, self-efficacy for management of symptoms, design preference of illustrated strategies with videos, and user-control. The four themes identified relating to enabling factors included: information requirements of symptom management, tracking symptoms, fit in lifestyle/living/schedule conditions, and more languages. The five themes reported relating to reinforcing factors included: communication with healthcare providers, information visualisation for each user, social networking, improvement in quality of life, and individual-tailored information quality | The generalisability of the findings may be limited due to the small sample size and the study sample being predominantly female. The app was only in English, therefore PLWH who are primary-Spanish speakers (an underserved population in the USA) were not included in the study |
| McNicholl et al., 2017 (USA) [38] |
n = 248 M = 57.8 years (SD = 5.1) Community |
Intervention study: Electronic medical records were used for medication reconciliations conducted by pharmacists | Pharmacist medication reconciliation, Beers and STOPP criteria, Patient Health Questionnaire-9, the Montreal Cognitive Assessment, and the Veterans Ageing Cohort Study (VACS) Index | Hypertension (56%), depression (52%), COPD/asthma (48%), dyslipidaemia (39%), coronary artery disease (27%), and diabetes (22%) were the most common co-morbidities found. 35% of participants were taking 16 or more medications and 16% were taking more than 20. Beers and STOPP criteria were present in 156 and 134 participants, respectively. 25 contraindicated drug interactions were identified in 20 participants. A mean of 2.2 medications were stopped after medication reconciliation | A limitation to this study is that only a subsample of the clinic’s population was included due to limited resources. Underestimations of the pharmacist’s role in correcting potentially inappropriate prescribing may have occurred as only medications that can be corrected without collaboration and conducted during the clinic visit were measured. Interventions undertaken after the clinic visit was not included |
| Hojilla et al., 2021 (USA) [95] |
n = 584 M = 50.5 years Hospital |
Secondary analysis of data collected from a RCT of behavioural interventions for reducing unhealthy alcohol use in PLWH | Berger HIV stigma scale, the 12-item short form survey (SF-12) | African American participants reported higher personalised stigma scores and disclosure concerns compared to Caucasians. Both Hispanic/Latinx and African American participants were more likely to report having concerns around public attitudes towards PLWH than Caucasians. Woman were more likely to have increased negative self-image scores than men | Sample size limited the studies ability to evaluate correlates of HIV stigma within sex and race/ethnicity subgroups. The generalisability of the results is reduced as participants are an insured cohort with well managed HIV |
| McAllister et al., 2013 (Australia) [88] |
n = 335 M = 52 years Hospital |
Data were collected via anonymous surveys | 19.6% (65) of participants reported meeting pharmacy dispensing costs as difficult or very difficult, 14.6% (49) stated that due to pharmacy dispensing costs they have delayed purchasing medications, and 9% (30) reported stopping medication due to pharmacy costs. Amongst the 19.6% of participants finding difficulty meeting pharmacy costs, 29.2% (19) had stopped medication compared to 4.1% (11) of the remaining 270 patients (p < 0.0001). 5.7% (19) patients found travel to the clinic difficult or very difficult. Difficulty meeting pharmacy and clinic travel costs were independently associated with treatment cessation and interruption. 4.9% of patients reported being asked if they were having difficulty with payments for medication | The study findings are limited due to a single-site recruitment, participants being mostly male, and the lack of viral load data meaning that clinical significance of patient responses could not be determined | |
| Heron et al., 2019 (Australia) [31] |
n = 2406 (HIV-positive) and n = 648,205 (HIV-negative) Age: 45–64 years = 58.9%; 65–74 years = 10.3%; over 75 years = 2.3% Community |
Data were collected directly from MedicineInsight—a national primary care data programme | HIV-positive males were less socioeconomically at a disadvantage than HIV-negative males. The HIV-positive males in this cohort were at an increased risk of cancer, chronic kidney disease, anxiety, depression, and osteoporosis. Younger PLWH are at risk of premature onset of osteoporosis, cancer, and cardiovascular disease. A high prevalence of depression and anxiety was reported among HIV-positive males | A limitation to this study includes the potential of duplicated patient records on MedicineInsight as it is not linked across practices, thus if a patient were to visit more than one practice it would be entered as separate records | |
| McMillan et al., 2019 (Canada) [67] |
n = 716 Age range = 50–92 years and M = 59.2 (SD = 6.5) Outpatient clinic |
Health data routinely collected at the Southern Alberta Clinic that included laboratory, self-reported and clinician-reported results | The 29-item Frailty Index developed for Southern Alberta Clinic | The mean Frailty Index, 0.303 (± 0.128), did not differ between genders. It was not linked to current CD4 counts or nadirs. Frailty Index increased with age, ART duration, and duration since HIV diagnosis. Higher Frailty Indexes was seen in those who died compared to the survivors | The cross-sectional design of the study limits the ability to draw conclusions of causality and directionality of the associations found. Another limitation is that the data used was collected for purposes other than calculating frailty, raising questions about accuracy and comprehensiveness |
| Kiplagat et al., 2019 (Kenya) [34] |
n = 57 Age: 50–59 years = 33.3%; 60–69 years = 49.2%; 70–79 = 17.5% Hospital and community |
In-depth interviews and four focus groups were used to collect data | Participants reported that co-morbidities and visiting multiple healthcare providers to manage their HIV as factors that impact their adherence to medication and clinic attendance. Other challenges included poor quality of facilities and patient-provider communications. Matched gender and older age for healthcare providers were reported as preferential by participants | The generalisability of the results is reduced as patients that had been lost to follow up or disengaged from care were not included, also participants included had been in care for at least a year, thus their views and experiences may have improved within that time | |
| Drewes et al., 2021 (Germany) [50] |
n = 897 M = 57 years (SD = 6.7) Community |
Data were collected using a self-administered questionnaire | An Adapted version of the negative self-image subscale of the HIV Stigma Scale, the Silver Lining Questionnaire (SLQ), the OSLO 3 Social Support Scale (OSSS-3), and the UCLA Loneliness Scale | 18% (165) of participants reported having one or more falls in the 12 months prior to the study. A higher risk of falling was significantly associated with a lower economic status, living alone and being single. Having one or more co-morbidity increased the risk of falls by 2.5 times. Diseases of the central nervous system, heart disease, rheumatism, osteoporosis, and chronic pain were strongly associated with fall risk. In addition, internalised and experienced HIV stigma, social support, and loneliness were significantly related to a fall risk | The cross-sectional study design does not allow for causality analysis. The self-administered questionnaires may have led to bias on recall and social desirability. The study was not able to fulfil a probability sample of people ageing with HIV in Germany, thus reducing the generalisability of the results. Moreover, several potential risk factors for falls were not included in the analysis of this study, for example, specific medication, problems with balance or gait, or mobility |
| Vinuesa-Hernando et al., 2021 (Spain) [49] |
n = 30 Median = 71 Hospital |
Observational study using data from patient hospital medical records | The Medication Regimen Complexity Index (MRCI), the Simplified Medication Adherence Questionnaire (SMAQ), the Screening Tool of Older People’s Prescriptions (STOPP) and List of Evidence-baSed depreScriing for CHRONic patients (LESS-CHRON) crtieria | The most common co-morbidities were dyslipidaemia (70%), hypertension (66.7%), diabetes (43.4%), mental health disorders (26.7%). 30% of participants were taking 10 or more medications and 70% were taking more than five. 66.7% of participants were adherent to their medications. The MRCI score of concomitant medications was higher than the score of ART at 18.3 points and 5.1 points, respectively. Potentially inappropriate prescribing was present in 70% of participants according to the STOPP or LESS-CHRON criteria. Polypharmacy was significantly associated with meeting deprescribing criteria (p = 0.008) | Limitations to this study include the use of dispensing records as some information may be missing and the reliability of the information depends on the inputting physician. Furthermore, the small sample size limits the generalisability of the data |
| Rozanova et al., 2020 (Ukraine) [97] |
n = 123 Age = 55–81 years Community |
Data were collected via telephone surveys | Older PLWH with substance misuse disorders maintained their HIV and substance use disorder therapies over the Covid-19 lockdown, however, social support was highlighted to be critical to avoid treatment interruptions | A limitation to this study is that telephone surveys may have led to social desirability bias in participant answers | |
| Chayama et al., 2021 (Canada) [55] |
n = 42 Age = 50+ years Community |
Semi-structured in-depth interviews were used to collect data | Participants viewed co-morbidities as more urgent and prioritised them over HIV. Access to care for co-morbidities were hindered by stigma and discrimination. Participants reported difficulty concurrently managing their co-morbidities and HIV due to poorly managed co-morbidities. Concerns and frustrations were stated regarding the potential impact of ART on the development of co-morbidities. Treatment approaches that integrated services aided engagement with care | The study was only conducted with English speaking participants, thus the opinions and experiences of marginalised individuals such as immigrants and refugees were not collected | |
| Siefried et al., 2017 (Australia) [85] |
n = 522 M = 50.8 years (SD = 12.3) Hospital and community clinics |
A study-provided laptop was used by participants to complete a 204-item questionnaire | University of Liverpool HIV drug interaction database, and the Charlson co-morbidity Index | 77 (14.8%) of participants reported being linked to one or more HIV community organisations or peer support groups. The median duration on ART was 11 years. 78 (14.9%) of participants missed an average of one or more ART medications per month in the 3 previous months | The self-reported nature of the questionnaire for adherence may overestimate true adherence levels. Due to recruitment strategies, there is a risk of selection bias. The study findings are less generalisable to patients without subsidised healthcare, community supports, those with virological failure, females, and heterosexual males |
| Siefried et al., 2018 (Australia) [44] |
n = 522 M = 50.8 years (SD = 12.3) Hospital and community clinics |
Participants were given a 204-item questionnaire on dedicated laptops | 204-question questionnaire incorporating other existing or pre-validated instruments | 292 (55.9%) participants reported having co-morbidities. 392 (75.1%) participants took at least one concomitant drug. The daily pill burden for concomitant drugs was 6 and the ART daily pill burden was 1.2. Cardiovascular, antidepressants, over the counter, endocrine agents and anti-effectives were the most common classes of concomitant medication. 122 (23.4%) participants were taking at least 5 concomitant medications. The concomitant medication taken in 17 participants were contraindicated with their ART. Overall, 730 ART-concomitant combinations were identified as being a potential drug–drug interaction. 178 participants reported adverse drug reactions | A majority of the participants were male and in a country with subsidised healthcare systems, thus reducing generalisability to females and those in countries without subsidisation. The cross-sectional study design does not allow for causality analysis. The study did not collect data on concomitant medication dosage, therefore it cannot report whether dose adjustments would mitigate potential drug–drug interactions |
| Bogart et al., 2021 (USA) [94] |
n = 76 Median = 52.9 (SD = 12.9) Community |
Intervention study: Individually randomised group-treatment trial using cognitive behavioural therapy. Participants were clustered into groups. Semi-structured interviews were conducted post intervention | Medication Event Monitoring System (MEMS) bottle cap | The intervention group showed improved adherence compared to the control group (electronically monitored: p = 0.06; Self-reported: p = 0.02). There was significantly lower medical mistrust amongst the intervention group compared to the control group (p = 0.02) | Generalisability of the study findings are reduced due to convenience sampling at one community site and the small sample size |
| Harris et al., 2020 (USA) [64] |
n = 35 M = 58.3 years (SD = 5.4) Community and hospital |
Data were collected through surveys, interviews, and focus groups | The Berger HIV Stigma Scale, Perceived Stress Scale, Engagement with Health Care Providers Scale, and the Composite of Engagement in HIV care | 54.3% of participants were reported to be moderately engaged in care. The overall stigma in participants were high and participants were reported to be moderately stressed. There was a significant correlation between engagement in care and the stigma subscales, including negative self-image stigma (p = 0.03). Perceived stress was also associated with overall stigma, disclosure stigma, personalised stigma, negative self-image stigma, and public attitudes stigma. Race was highlighted as an additional cause of stigmatisation among African Americans | Limitations include data collected from surveys being self-reported, and items related to engagement in care were not based on reviews of medical charts. Self-reports may have led to bias on recall and social desirability. The small sample size limits generalisability |
| Mao et al., 2018 (Australia) [37] |
n = 98 Median = 51.5 years (26–65) Community |
Intervention study: A 6-week randomised SMS reminder intervention for ART adherence was conducted, followed by a mixed-method evaluation consisting of one-to-one interviews and a self-completed online survey | SMS reminders to mobile phones | The most common reasons for previous ART interruption were experiencing side effects and attending to other life priorities. There were no statistically significant differences between the intervention and control arms at the end of the SMS campaign. A common suggestion for improvement of the intervention was that it should be tailored to each individual's needs and synchronised with their dosing regimens. The SMS campaign had several positive responses, describing impacts beyond ART adherence | The small sample size may reduce generalisability of the findings and limits the ability to find differences between the experiment groups. The cross-sectional mixed methods evaluation does not allow causal conclusions to be drawn. Moreover, evaluation consisted of self-reported data |
| Jiménez-Guerrero et al., 2018 (Spain) [77] |
n = 242 Median = 57.5 years (54–62) Hospital |
Data from electronic clinical records were used with a computer system (Diraya) to identify home treatment and an application for outpatient dispensing (Farmatools). Interactions were identified using four independent prescribers, product specification and an online database (http://www.drugs.com) | http://www.drugs.com | 61% (148) of participants were receiving concomitant treatment and 243 potential interactions were detected among 110 participants. 46 of the interactions were considered severe, whilst 197 were moderate. 76% (35) of the severe interactions were associated with boosted protease inhibitors. Statins and inhaled corticosteroids caused most severe interactions | The retrospective nature of the study could be a limitation, however, being merely descriptive, it may be considered irrelevant. Another limitation could be that over the counter and herbal medicines were not included |
| Knight et al., 2018 (South Africa) [35] |
n = 23 PLWH ≥ 50 years Community |
In-depth semi-structured interviews | Participants received care for both HIV and other conditions provided by different healthcare professionals and at different health facilities. Older PLWH and non-communicable diseases experience several physical and structural barriers to accessing care. These difficulties can worsen health outcomes | A limitation to this study is that recruitment was conducted via referrals from the HIV service, thus this may have led to participants with more or less barriers being missed | |
| Halloran et al., 2019 (UK/Ireland) [14] |
n = 698 PLWH ≥ 50 years and 374 PLWH ≤ 50 years 304 HIV-negative participants≥ 50 years old Hospital |
Potential drug–drug interactions were analysed using two interaction checking tools. The Pharmacokinetic and Clinical Observations in People Over 50 (POPPY) study is a prospective, observational, multicentre study that collected data over a three-year period from 2013 to 2016 | The Lexicomp database and the Liverpool drug interaction database (wwwhiv-druginteractionsorg) | Polypharmacy was prevalent in 65.8, 48.1 and 13.2% in older PLWH, younger PLWH and the HIV-negative group, respectively. This reduced to 29.8% of the older group and 14.2% of the younger group when ARVs were excluded. 36.1% of older PLWH, 20.3% of younger PLWH and 16.4% of the HIV-negative group had a prevalence of ≥ 1 PDDI involving non-ARV medications. The prevalence of ≥ 1 PDDI between ARV and non-ARV medication was 57.3% in older PLWH and 32.4% in younger PLWH | One limitation of this study was a lack of data on dosing information of most medications, leading to some interactions being overestimated. Another limitation is that the medication lists were self-reported by participants, which may have been under-reported. Lexicomp is a sensitive interaction checker and some of the interactions flagged may not be clinically relevant in daily practice |
| John et al., 2016 (USA) [32] |
n = 359 Median = 57 years (50–80) Hospital |
The following four domains were evaluated using a questionnaire: social support, physical health and function, mental health, and behavioural and general health | The Lubben Social Network Scale-6, the Social Provisions Scale, the UCLA 8-item Loneliness Scale, the Patient Health Questionnaire-9), the Generalised Anxiety Disorder-7, the Breslau 7-item PTSD screen, and the Montreal Cognitive Assessment | Participants experienced the burden of ageing related conditions over the domains evaluated. Nearly 60% reported mild symptoms of loneliness, 50% showed low social support, 41% had a fall in the last year and 34% met the criteria for possible mild cognitive impairment. Participants 60 years old or over had higher frequencies of balance issues compared to the group aged 50–59 years old. Fewer participants reported "very good" or "excellent" HRQOL in the 50–59 years old age group compared to the older group | The cross-sectional study design meant that changes were not measured over time. The findings may not be generalisable as the participants were largely male and from one city. Patients in long-term care facilities were not included. Most of the participants were diagnosed over 10 years prior, thus results may not be generalisable to newly diagnosed older PLWH |
| Townsend et al., 2007 (USA) [63] |
n = 58 M = 51.5 years (SD = 8.8) Hospital |
Data were collected using electronic medical records | Findings showed a non-significant correlation between viral loads and 6-month pharmacy medication refill-based adherence (r = 0.1). Adherence rates lower or equal to 70% led to CD4+ levels progressively declining. AIDS-related events incidence or past ARV experience did not significant affect the distribution of participant CD4+ levels or adherence | Limitations to this study include that pharmacy refill data was used to measure adherence as patients may have collected their medications but not taken them, the small sample size, and short study duration | |
| Levy et al., 2017 (USA) [36] |
n = 7018 Median = 50 years (39–57) Hospital and community |
Electronic medical records were used for data collection | Half of the participants reported having hypertension, 48% had dyslipidaemia, and 35% had obesity. A higher prevalence of co-morbidities was seen in older PLWH (p < 0.001). Hypertension was reported more in black patients, diabetes and obesity were reported more in female and black patients, and dyslipidaemia was reported more in male and white patients (all p < 0.001). Metabolic co-morbidities were associated with controlled immunovirological factors, longer time since HIV diagnosis, and a greater duration of ART | Limitations to this study include the cross-sectional study design does not allow for temporality or causality analysis, electronic medical records may have missing information, thus this may affect the results. Information on ARV adherence was not collected | |
| Gimeno-Gracia et al., 2014 (Spain) [76] |
n = 130 M = 56.7 years (SD = 6.2) Hospital |
Data were collected from out-patient pharmacy records at a University Hospital in Spain | At the end of the study, 90% of participants had an undetectable viral load and 58% had a CD4 count over 500 cells/mm3. Treatment that was based on protease inhibitors were used by 51.5% of older patients and 54.4% of the younger patients, whilst nonnucleoside reverse transcriptase inhibitors were used by 43.8 and 39.8%, respectively. The older group used treatments with abacavir more frequently (p = 0.054) and with tenofovir less frequently (0.105) compared to the younger group | Due to the retrospective nature of this study, some data was incomplete. Also, the varying number of years since diagnosis among patients could have influenced their degree of adherence, the ART received, etc. | |
| Gimeno-Gracia et al., 2016 (Spain) [72] |
n = 119 Median = 52 years (51–56) Hospital |
Data were collected retrospectively to compare polypharmacy, percentage of patients that collected each therapeutic drug class, and the median duration of each drug class between older PLWH and the general population | A higher percentage of HIV-positive males had polypharmacy than males from the general population (8.9 vs 4.4%, p = 0.01). This was also true for females from each group, with older and younger groups having 11.3 and 3.4% of polypharmacy, respectively (p = 0.002). HIV-positive participants recieved more gastrointestinal drugs, analgesics, anti-infectives, central nervous system (CNS) agents, and respiratory drugs than the general population. No differences was observed between both groups for cardiovascular drugs. HIV-positive participants had a higher estimated number of treatment days than the males in the general population for CNS agents (p = 0.02), anti-infectives (p < 0.001) and more were receiving sulphonamides (p < 0.001), quinolones (p = 0.009) and macrolides (p = 0.02) | Generalisability of the study is reduced due to single-site recruitment and only a small number of female participants being included | |
| Greene et al., 2014 (USA) [13] |
n = 89 (HIV positive) Median = 64 years (60–82) Community |
Structured interviews were used to obtain demographic data, HIV history and co-morbidity data. Medication lists were reviewed during interviews to obtain medication usage information. Participants were asked to complete a questionnaire before the interview with all their prescribed and over-the-counter medication | Lexi-Interact drug interaction software, Beers criteria 2012, and the Anticholinergic Risk Scale | Common co-morbidities amongst the participants were identified as hypertension, hyperlipidaemia, and depression. An average of 13 medications (2–38) were taken by PLWH, with only an average of four being ARVs, whereas non-HIV participants took an average of six (3–10) medications. There was at least one potentially inappropriate medicine prescribed in 46 (52%) of PLWH. Ten (11%) of the HIV participants had a Category X (avoid combination) interaction and 62 (70%) had at least one Category D (consider modification), with a third of these interactions being between two non-ARV medications. Fifteen (17%) PLWH were identified as having an anticholinergic risk scale ≥ 3 | The study population consisted of highly educated, Caucasian, men who have sex with men, therefore the findings may be less generalisable. The participants were also diagnosed for an average of 20 years, so the data may not be generalisable to older patients diagnosed more recently. Information about dosing was not obtained, therefore drug–drug interactions may have been overestimated, although the findings are consistent with other studies |
| Greene et al., 2018 (USA) [68] |
n = 356 Median = 56 years (53–62) Hospital |
Survey data were collected at two clinics to assess participants social, physical, mental, and cognitive health | The UCLA eight item loneliness scale, the Montreal Cognitive Assessment, the Lubben Social Network Scale, and the Patient Health Questionnaire | Symptoms of loneliness was reported in 58% of participants, with the majority having mild loneliness. Lonely participants were more likely to have depressive symptoms, poor or fair HRQOL, have fewer physical supports, be current smokers or at-risk drinkers and/or drug users | The study participants were predominantly male, urban, were 57% white, and had long durations since diagnosis, therefore generalisability is limited. As the study was cross-sectional, temporal relationships between loneliness and depression with HRQOL and functional impairment was not examined. The HRQOL measure used was due to ease of administration but did not have a mental health measure. Also, a measure of stigma was not included, previous research has shown links between loneliness, stigma, and depression |
| Greene et al., 2018 (USA) [27] |
n = 77 surveyed and n = 31 focus groups Median = 58 years (50–77) Hospital |
Data were collected through focus groups and a survey | Findings highlighted the need for greater focus on the following: (1) the need for knowledge expertise in HIV and ageing, (2) a focus on determinants of health (e.g., marginal housing) and on medical conditions, (3) locating speciality services together, and (4) social isolation. These findings informed the creation and design of a multidisciplinary care model for PLWH (the Golden Compass programme) | A limitation of this study includes single-site recruitment and that generalisability of the findings is reduced as non-English-speaking participants were excluded | |
| Kuteesa et al., 2012 (Uganda) [86] |
n = 40 Median = 65 years (50–80) Hospital |
Individual in-depth interviews and focus groups were used to collect data. Observations of clinic interactions were also recorded | Key themes that emerged from the qualitative interviews highlighting distinctive healthcare needs in older PLWH were: difficulty disclosing (8%), stigma (43%), access to care (80%), delayed diagnosis and care-seeking (55%), quality of patient = provider relationship (75%), serodiscordance (20%), adherence support (25%), continuity of care (14%), end-of-life issues (13%) and other issues (20%). Participants reported experiencing stigma due to HIV and ageism. Concerns and anxiety regarding securing future healthcare and the lack of social services was expressed by participants. Problems with transport and food compromised adherence to ART for many participants | Limitations to this study affecting generalisability of the findings include recruitment from a limited geographic area and purposive sampling. Duration since diagnosis may have affected individual perspectives reducing generalisability of the findings as interviews did not indicate any participant to be newly diagnosed | |
| Philbin et al., 2021 (USA) [91] |
n = 59 M = 51 years Hospital and community |
In-depth interviews were used to collect data | Four main groups emerged from those interviewed, firstly, those with few long-acting injection related worries who received episodic injections, secondly, those who had regular injections and did not want anymore, thirdly, those with a history of injection drug use that were worried long-acting injections would trigger a reoccurrence, and lastly, those who currently inject drugs and have few worried around long-acting injections. Most participants who have a history of using injectable medication would prefer long-acting injectable ART, but participants with regular injections already and a history of injection drug use may not | The generalisability of the study findings are limited as they are based on the individual women's experiences and concerns, which may not be relatable to other subpopulations | |
| Katende-Kyenda et al., 2008 (South Africa) [75] |
n = 8999 (HIV-positive) 59.42% = 40–60 years and 1.58% ≥ 60 years National medicine claims database |
Data were collected directly from Interpharm Data systems and analysed. Prescriptions were used to determine if combinations of ARVs could cause possible drug–drug interactions | A clinical significance rating of potential drug–drug interactions as described by Tatro [Drug Interaction Facts 2005 St Louis, MO: Facts and Comparisons (2005)] | Participants received a mean of 2.36 ARVs per prescription. 960 drug–drug interactions were identified. Patients aged 40–60 years old had the highest number of ARV prescriptions and the highest number of drug–drug interactions. The most drug–drug interactions were seen between Indinavir and ritonavir, efavirenz and indinavir, efavirenz and lopinavir/ritonavir | A limitation to the study was that demographic and clinical information was not available on the database. Moreover, dosage information was also not supplied |
| Schreiner et al., 2019 (USA) [42] |
n = 103 M = 53.16 years (SD = 7.2) Community |
Data from a parent study examining physical activity patterns in PLWH was used for secondary analysis to evaluate treatment burden. The parent study used one-to-one interviews and entered responses directly into Research Electronic Data Capture | The Treatment burden Questionnaire-13, the Bullen and Onyx (2007) Social Capital Measurement Tool | Overall, a low level of treatment burden was reported among participants, however, 16% of participants reported experiencing high treatment burden. Treatment burden was significantly associated with number of chronic conditions (p ≤ 0.01) and social capital (p = 0.03). The most prevalent co-morbidities were hypertension, asthma, arthritis, diabetes, hepatitis B/C, and hyperlipidaemia. Remembering to take medications at certain times during the day, paperwork, the limitations linked to taking medications and maintaining a prescribed exercise regimen were items causing the highest treatment burden | Due to treatment burden not being the primary focus of the parent study, the reanalysis could not collect data on other variables of interest. The sampling technique used may have allowed for potential sampling bias, effecting the generalisability of the study findings. PLWH who are not insured or able to afford regular medical care were not represented in this study |
| Hastain et al., 2020 (USA) [29] |
n = 99 Median = 54 years (49–61) Community clinic |
Electronic medical records were used to evaluate ART simplification. Drug–drug interaction scores pre- and post- simplification were calculated. Concomitant medications were identified and evaluated for drug–drug interactions with pre- and post- simplification ART regimens | A drug–drug interaction incidence and severity score was developed and validated and the University of Liverpool's HIV Drug Interaction Checker | A median of 3 ART pills were taken a day. After simplification, the median number of ART pills taken was 2. Discontinuing protease inhibitors and nonnucleoside reverse transcriptase inhibitors occurred frequently and ART changes to integrase strand transfer inhibitor-based regimens were common. Average interaction scores reduced from 3 (1–6) to 1 (0–2) from pre- to post- simplification. The median number concomitant medication taken was 4.5 | The results may not be generalisable due to the sample size and recruitment from only one urban site. Also, not all ART simplification strategies and HIV medications could be thoroughly evaluated. As the study relied on the completeness and accuracy of electronic medical records, it is possible that over the counter, herbal and non-prescriptions items were not included. Lastly, the scoring system does not reflect the clinical significance of drug–drug interactions, for example, some interactions may need dose adjustments or close monitoring |
| Foster et al., 2009 (USA) [93] |
n = 24 Age range = 50–76 years (M = 57) Community |
Four focus groups were conducted, and supplementary data was obtained using two stigma instruments | The Stigma Impact of HIV scale and the Self-Perceptions of HIV Stigma Scale | Participants reported rarely or not experiencing stigma as they had not disclosed their HIV diagnosis to others. Stigma was reported by participants on the Internalised Shame scale. Qualitative data found four themes associated with stigma: (1) disclosure; (2) stigma experiences; (3) need for HIV/AIDS education; and (4) acceptance of the disease | Limitations to this study include that the focus group setting may have led to response-bias and participants who have been stigmatised not wanting to participate in group discussions, data on sexual orientation and mode of sexual transmission was not collected, and the time since diagnosis ranged broadly between participants, which may have affected perceptions on stigma and disclosure |
| Halkitis et al., 2014 (USA) [66] |
n = 180 M = 55.4 years (SD = 4.6) Community |
Data were drawn from Project Gold, a study of ageing HIV-positive men who have sex with men in New York City. Self-report data was collected for sociodemographic characteristics and clinical markers | The AIDS Clinical Trials Group (ACTG) Adherence Questionnaire, the 21-item Beck Depression Inventory-II, the HIV Stigma Scale | 109 (57.2%) participants reported at least one suboptimal adherence behaviour, with 36 (20%) missing doses in the 4 days prior to the assessment; 97 (48.3%) failing to take medication on schedule; 40 (24.1%) failing to follow instructions; and 33 (18.3%) missing doses in the weekend prior. Participants who missed a dose in the four days prior had experienced higher levels of HIV-related stigma compared to those who did not miss doses in those days. Depression and HIV-related stigma was associated with failing to take medication on schedule in the four days before (p = 0.05 and p = 0.01, respectively). These factors were also associated with failing to follow instructions on how to take ARVs | Data on adherence and psychosocial factors were collected using self-reports, which could have resulted in under-reporting. The cross-sectional study design limits the studies ability to find causal conclusions and evidence of a history of adherence behaviours and psychosocial behaviours |
| Shippy et al., 2005 (USA) [43] |
n = 160 Age: 50–59 = 85%; 60+ = 15% Community |
A survey was used to collect variables of interest | The increasing group of ageing PLWH are facing isolation from informal networks due to HIV stigma and ageism. 71% of participants lived alone. Family members and partners are critical for informal support, however, only a third of participants had a partner. 86% of participants used Medicaid | The cross-sectional study design does not allow for causality analysis among mental health indicators, social networks, and support needs of older PLWH | |
| Zepf et al., 2020 (USA) [54] |
n = 281 M = 57.9 Hospital |
Self-reported questionnaires were used to collect data | Symptoms of depression (p = 0.008), stimulant use (p < 0.0001), and PTSD (p = 0.002) were significantly associated with reduced medication adherence. Intimate partner violence, binge drinking and past physical or sexual abuse were not significantly associated with lower medication adherence | Due to the study cross-sectional study design, causality or determination of the sequence of events could not be determined. Self-administered questionnaire answers are subject and may have led to desirability bias in findings | |
| Morillo-Verdugo et al., 2019 (Spain) [39] |
n = 223 Median = 53 years (52–57) Hospital |
Data were collected from an electronic medical record. Questionnaires and hospital dispensing records were used to measure medication adherence | The SMAQ questionnaire, the Morisky-Green questionnaire, and the Medication regimen complexity index | The median number of concomitant medications prescribed per participant was 3 (1–5) and the median number of co-morbidities was also 3 (2–4). Polypharmacy was noted in 56.1% of participants and 9.4% had higher polypharmacy. The polypharmacy was 60% cardiovascular, 27.1% was depression-anxiety, 7.1% was mixed and 5.8% was COPD. Participants with high pharmacotherapy complexity indices had a 50 times greater chance of polypharmacy. It was demonstrated that 11.25 was the best cut-off point to predict polypharmacy in older PLWH | A minor limitation is that private prescription and alternative medicines were not included. Another limitation is that the Medication regimen complexity index tool used is imperfect and has trade-offs between sensitivity and specificity. Also, some dosing options were missed, such as monthly doses |
| Schnall et al., 2018 (Australia) [99] |
n = 80 Control group: M = 51 years (SD = 9) Intervention group: M = 50 years (SD = 11.7) Community |
Intervention study: Symptom burden when using a mobile app with self-care strategies to manage symptoms was compared to a mobile app without the self-care strategies between 2016 and 2017. The study used mVIP, a web-app, that was created to help users self-manage their symptoms. Surveys were given weekly on the app. The study lasted 12 weeks | mVIP (a web-app) | Participants in the intervention group had a significantly higher improvement than the control group in the following 5 symptoms: depression (p = 0.001), anxiety (p = 0.001), neuropathy (p = 0.002), weight loss/wasting (p = 0.020), and fever/chills/sweat (p = 0.037). The intervention group also showed a higher improvement in adherence rates to ART medications than the control group (p = 0.017). Over the 12-week trial, mVIP was linked to improved symptom burden and adherence rates in PLWH | The study had limitations to the generalisability of study findings with a small sample size, and as participants with access to a smart-phone or tablet were required, PLWH who do not have or are unable to use this technology were marginalised |
| Gimenez et al., 2020 (Spain) [90] |
n = 42 Median = 70 years Hospital |
Observational, prospective study where data from hospital patient records were analysed using various tools | The Medication Regimen Complexity Index (MRCI), the Simplified Medication Adherence Questionnaire (SMAQ), the Morisky-Green questionnaire (MMAS) and the Revised Patients’ Attitudes Towards Deprescribing (rPATD) Questionnaire | A majority of participants agreed with the following three statements in relation to deprescribing, firstly, 91.9% agreed to “I have a good understanding of the reasons I was prescribed each of my medicines”, secondly, 89.2% reported that “Overall, I am satisfied with my current medicines” and finally, 89.2% agreed that “I like to be involved in making decisions about my medicines with my doctors” | The small sample size limits the generalisability of the data. Another limitation is the loss of information intrinsically in relation to the dichotomisation of the indices used for analysis |
| Patel et al., 2016 (UK) [40] |
n = 299 M = 58 years (SD = 2) Outpatient clinic |
Data were collected using self-reported questionnaires | One or more co-morbidities was present in 84% of participants and 61% had two or more. High cholesterol, hypertension, depression and sexual dysfunction were commonly reported. Multivariate analysis when adjusted for lifestyle factors, showed that number of years since HIV diagnosis, age, and duration of ART were significant predictors of co-morbidity. Multiple co-morbidities, polypharmacy and higher use of non-HIV services were reported in older PLWH | The self-reported nature of the questionnaire may have led to recall and self-presentation bias. As data was collected at a single timepoint, it was not possible to determine the direction of associations between co-morbidity and risk factors. Generalisability of findings are reduced to due single-site recruitment with a high prevalence of White men who have sex with men | |
| Kteily-Hawa et al., 2019 (Canada) [69] |
n = 1422 M = 55.81 years (SD = 5.25) in older women group Community |
Survey data from the Canadian HIV Women's Sexual and Reproductive Health Cohort Study was used. The questionnaire used in the study included the following nine sections: demographics and socioeconomic status, healthcare utilisation, medical and HIV disease information, stigma, and discrimination, HRQOL, and substance misuse | The 12-item Short Form Survey (SF-12), the Medical Outcome Trust scoring method, the 10-item Resiliency Scale, the Center for Epidemiologic Studies Depression 10-item scale, and the 4-item Abbreviated Social Support Instrument | Older participants had poorer physical HRQOL and less social support than younger participants. Mental HRQOL was similar between the groups. Better physical HRQOL was linked to resilience. Poorer physical health was linked to poorer mental HRQOL, food insecurity and depressive symptoms | The cross-sectional study design limits the ability to determine a causal relationship between correlates and causations of HRQOL. The self-report nature of the questionnaire may have been affected by social desirability biases. Recruitment may have been biased to include women already accessing HIV care, and thus experiencing higher physical functioning and resilience, lowering generalisability of results to those not accessing care |
| Sutton et al., 2016 (USA) [83] |
n = 15602 M = 52.1 years (SD = 9.5) Hospital and community |
Data were collected from the Veterans Health Administration electronic health record system from 2006 to 2012. Data included demographic details, vital signs, mortality, administrative claims, laboratory results, and pharmacy dispensing records | The Charlson co-morbidity index score | A significantly lower proportion of single-tablet regimen compared to multiple-tablet regimen participants had undetectable viral loads at baseline (p < 0.001). It was found that the single-tablet regimen participants had double the odds of being adherent, less chances of hospitalisation, and more chances of an undetectable viral load during follow-up, after controlling for baseline covariates | A limitation to the study was that pharmacy refill data was used as a measure of adherence, however, this may have overestimated adherence. The study was unable to exclude confounding factors that could have influenced the findings, as patients were not randomised to the two treatments. Despite attempting to control for some variables through multivariable models, residual confounding may have remained |
| McKetchnie et al., 2021 (USA) [58] |
n = 15 Age ≥ 50 years Community |
Data were collected using one-on-one interviews | The main themes highlighted from interview transcripts were firstly, the impact of chronic pain and its treatment on HIV care engagement, secondly, the impact of substance use on chronic pain, and lastly, interventions to target substance use and chronic pain. The results showcase the need for interventions addressing structural, physical, and psychological barriers to care engagement | The small sample size limits the generalisability of the data | |
| Heckman et al., 2002 (USA) [78] |
n = 83 M = 55.2 (SD = 5.1) Community clinics in two cities |
Self-administered assessment instruments were completed by participants | The Symptom Checklist-90-Revised, the Beck Depression Inventory, the Barriers to Care Scale, the Provision of Social Relations Scale, the HIV-related Life Stressor Burden Scale, and the Limitations Related to HIV symptoms question | 4% of participants had severe depression, 21% had moderate depression, 33% had mild depression, 42% had minimal or no depression. Participants reporting more psychological symptoms had higher HIV-related life stressor burden, less social support from friends, and experienced more barriers to services due to HIV-related stigma. Elevation of symptoms characteristic of somatisation was seen in participants | Due to recruitment being from only two sites, this may reduce the generalisability of the study findings. Also, the study used a self-selected sample. Participants were already connected to the AIDS service organisation; therefore they might already have their life-care needs met; study findings may underestimate the life chaos among participants |
| Haruna et al., 2021 (Tanzania) [56] |
n = 27 M = 54 years Hospital |
Data were collected through in-depth semi-structured interviews | Participants and healthcare workers held positive attitudes towards integration of services, they reported perceived benefits with this. Inconsistent and limited supplies, for example with screening equipment or medications, low awareness of non-communicable diseases among PLWH, insufficient training of healthcare workers on how to manage non-communicable diseases in PLWH, treatment costs and payment systems were factors hindering the integrations of services | Interviews being conducted within the clinic setting that participants usually received their care may have led to socially desirable answers. Recruitment from five large clinics in Dar es Salaam may not be representative of patients from other smaller clinics | |
| Fritsch, 2005 (Canada) [96] |
n = 22 under 50 years, n = 12 aged 50–65 years Hospital and community |
Data were collected via in-depth individual interviews | The majority of participants had positive experiences accessing medical services. PLWH over 50 years old reported being able to access the social and medical services they required to cope with HIV. Both the younger and older groups accessed a variety of medical services, although the later did slightly less. The older group utilised social and emotional supports and organisations to a similar or higher degree than the younger group | The generalisability of the study findings is limited due to the geographic location, sample size, non-random sampling and snowballing technique used | |
| Kim et al., 2017 (USA) [33] |
n = 250 Median = 50 years (44–56) Community |
Data were extracted from electronic medical records | Sedating medication and opioids were prescribed in 80 and 50% of participants, respectively. 51% of participants exceeded risky drinking limits. Illicit opioid use and illicit opioid sedative use was reported in 23% and 9% of participants, respectively, with 37% reporting lifetime non-fatal overdose and 7% past year non-fatal overdose. Participants were taking a median of 10 (7–14) medications and a median of 2 (1–3) sedating medications. The odds of a lifetime non-fatal overdose were increased with each additional sedation medication taken and any opioid medication. The prevalence of past year non-fatal overdose was higher with each additional sedating medication (p = 0.049) and non-ARV medication (p = 0.048) | Limitations to this study include the absence of a validated method to assess overdose history and the examination of only non-fatal overdose | |
| Shamu et al., 2021 (Zimbabwe) [52] |
n = 420 Median = 55 years Outpatient clinic |
Data were collected from electronic medical records | World Health Organization stage 3 or 4 conditions were present in 40% of participants at baseline. 15% of participants had hypertension at baseline and a further 27% reported incident hypertension at follow-up. 61.7% of participants had one or more chronic non-communicable disease. Results from the follow-up reported the most common incident co-morbidities to be hypertension, arthritis, and chronic kidney disease | The generalisability of the results is reduced due to the nature of the study being single-sited | |
| Hoang et al., 2009 (USA) [30] |
n = 1,018 M = 50.6 (SD = 9.4) Hospital and community |
Data were obtained from electronic medical records and by interviewing the Infectious Disease clinics chiefs at the five Veterans Affairs facilities | 93% of participants had co-morbidities with a mean of 3.2 co-morbidities (SD = 2). Viral suppression was achieved in a median of 231 days (SD = 411.6) by 52% of participants. Participants attending integrated clinics offering hepatitis, psychological, psychiatric, and social services were 3.1 times more likely to achieve viral suppression compared to those visiting clinics that only offered HIV care (Hazard ratio = 3.1, p < 0.001) | A limitation of this study includes that findings may be affected by the unmeasured differences between participants attending different clinics and healthcare providers skills varying, rather than the comprehensiveness of the care. Also, the potency of ART or its complexity was not measured. Duration of combined ART was not measured; therefore, it is not possible to determine whether this affected viral suppression | |
| Njie-Carr et al., 2021 (USA) [102] |
n = 20 completed baseline activities, n = 6 completed all activities, n = 7 included in focus group Age ≥ 50 years Community |
Intervention study: Participants were given The CARE intervention, a three-module intervention for three weeks followed by post intervention measures to test its utility | World Health Organization QOL BREF, Perceived Social Support Scale, Mood tracker, Geriatric Depression Scale, and HIV Self-Care for Symptom Management Survey | Participants found the CARE intervention useful and felt empowered utilising the app. Neuropathic pain (70%), anxiety (55%), fatigue (50%), and depressive symptoms (35%) were reported to cause high-intensity distress and negative impact on life to participants. It was reported that self-regulation and self-awareness improved, and there were positive trends in outcome measures | A small sample size limits the generalisability of the study results. Another limitation is the short intervention duration of three weeks. Also, the participants retained in the study may be healthier and more acquainted with technology |
| Singo et al., 2015 (South Africa) [45] |
n = 12 Age = 50–83 years Community |
Data were collected through interviews using an interview guide | The findings showed that HIV and AIDS can have negative effects on the lives of older PLWH. Socio-economic, household, psychological and emotional health burdens are critical areas in relation to older PLWH | The small sample size limits the generalisability of the data and the population recruited in may not be representative of other geographical areas |
ACH anticholinergic medication, ART antiretroviral therapy, ARV antiretroviral, CNS central nervous system, HAART highly active antiretroviral therapy, HRQOL health related quality of life, MOR medicines management optimisation review, PDDI potential drug-drug interaction, PLWH people living with HIV, PTSD post-traumatic stress disorder