Aim 1: Characterize frailty and multimorbidity in older PWH in care at the University Hospital, Kwame Nkrumah University of Science and Technology (KNUST).

We hypothesize that prevalence of phenotypic and indexed frailty will be >15% [13] and associated with older age, worse HIV control, female sex, greater social vulnerability and higher risk of death (VACS Index). For 151 PWH ≥ 50 years on stable antiretroviral therapy, we will assess physical frailty phenotype and frailty index [24,25], cognitive impairment, depression, social vulnerability [26], multimorbidity [27], polypharmacy and VACS Index–a measure of multi-organ function and physiologic frailty predictive of death [28–31].

Aim 2: Assess under-diagnosis and management of discrete cardiometabolic diseases in older PWH.

We hypothesize that at 12 months, the proportion of positive screenings with: a) completed follow-up will be <50%; b) effective management of diabetes, hypertension, dyslipidemia, or renal dysfunction will be <50%. We will screen for hypertension, dyslipidemia, diabetes mellitus and renal dysfunction. The screening results will be compared to medical records and patient self-report to estimate rate of under-diagnosis and under-recognition. After sharing screening findings with participants, they will be referred to see their healthcare providers per the current practice. We will reassess them at month 6 and month 12 for: a) frequency and outcomes of follow-up (referral/treatment) for positive screenings; and b) progression and incidence of cardiometabolic diseases.

Aim 3: Identify barriers and facilitators to providing effective, patient centered healthcare to aging PWH.

We will perform semi-structured qualitative interviews with up to: a) 30 PWH to assess engagement and satisfaction with healthcare for HIV, frailty and discrete diseases of aging; b) 20 HIV healthcare providers and policy makers to assess perceptions about caring for aging PWH; and c) 10 of our research staff about experiences conducting geriatric assessments in this study.

Aim 4: Capacity building in research skills that also have clinical relevance.

We will hire research assistants who are new to or have minimal experience in scientific research and provide them training on how to conduct research. These trainings will be delivered through our weekly research meetings, and special research training seminars. Training areas will include; ethical conduct of research, administering informed consent, understanding of the research protocol, training on assessment procedures, and teaching on how to use project management tools such as Trello, Microsoft Outlook, and Redcap.

A. Chart review.

Medical charts will be reviewed both prior to enrollment and after enrollment. Prior to enrollment, medical chart information about HIV status will be reviewed. We may review medical charts at the end of a participant’s participation in the study to monitor changes in medication regimens, treatment outcomes, as well as collect information from procedures that were part of routine clinical care.

B. Geriatric health assessments.

Participants will undergo a focused geriatric health assessment covering vital signs and organ systems. These assessments will enable us to wholistically evaluate study participants [42]. In addition to the routine clinical examinations, we will pay particular attention to factors such as nutrition, vision, hearing, fecal and urinary continence, and polypharmacy (Table 1) [43].

C. Cognitive and mental health assessment.

We will use the Mini-Cog tool which is composed of a clock drawing and a delayed three-word recall task to assess cognition including subtle impairments [44]. Mini-Cog is a tool designed to assess cognitive dysfunction, and has previously been used among the older people [45,46]. We will screen for depression using a two-question tool, “During the past month, have you been bothered by feelings of sadness, depression, or hopelessness?” and “Have you often been bothered by a lack of interest or pleasure in doing things?” [47] Participants will then complete a more comprehensive depression and anxiety scale (e.g., Geriatric Depression Scale) [48] and the Brief COPE [49] (measures coping strategies). These risk assessment tools have been used in Ghana and other countries in West Africa previously [49,50]. We will also administer the Patient Health Questionnaire-9 (PHQ-9) and General Anxiety Disorder questionnaire. Findings from the Geriatric Depression Scale will contribute to validation work of mental health assessment instruments previously used among Ghanaians; the PHQ-9 will be an additional dimension of mental health assessment in our study and its results will be compared with the former.

D. Social vulnerability.

We will assess deficits in social support and engagement, living situation, self-esteem, sense of mastery or control over one’s life and socioeconomic status using a social vulnerability index assessment questionnaire [26,51]. This will provide information about positive and negative social experiences of aging with HIV in the cultural context of Ghana, and how these experiences facilitate or hinder ones recovery from other life stressors [52].

E. Blood and urine samples.

We will collect 15mls of blood at each visit to assess for dyslipidemia (total cholesterol ≥ 5.2 mmol/l, LDL cholesterol ≥ 3.0 mmol/l, triglycerides ≥ 1.7 mmol/l, HDL cholesterol ≤ 1.45), Diabetes Mellitus (hemoglobin A1C ≥6.5) and renal dysfunction (eGFR < 60 ml/min), as well as CD4+ cell count and viral load, when required. We will also collect 5mls of urine for routine urine analysis.

Sample size and power calculation.

Our study is powered to detect whether the prevalence of phenotypic frailty is different from hypothesized 15% [53] With 126 people, 80% power at a 0.05 significance level, we can detect a minimum difference in prevalence of 10%. Assuming a Short Physical Performance Battery (SPPB; major component of frailty testing) mean score (standard deviation) of 10.2 (1.8) among West African PWH ≥50 years [32], 126 people would enable us detect a minimum difference in SPPB score of 0.25 in our sample compared to a reference population. This power is excellent given that minimum meaningful change estimates for SPPB score ranges from 0.4–1.5 [54]. We factored in 20% loss to follow-up to come to a final sample size of 151. We anticipate reaching theoretical saturation [55] from qualitative interviews with up to 30 patients, 20 healthcare providers and 10 research staff.

Assessment quality assurance.

During the assessment, if the participant provides conflicting answers or answers that did not make logical sense (either within the same section or between sections), the RA will gently try to help the participant arrive at more logical answers. However, the RA will not force the participant to change his or her answers. The RA will review the self-administered section with the participant present. If many “refused” options are selected, the RA will offer the participant the opportunity to complete those sections (the RA will accept the participant’s refusal if he or she does not wish to complete the section).

The study will utilize paper-based case record forms and REDCap for data management. Data stored on computers will be password protected, and accessible only to research associates needing the information for follow-up purposes. Project data analyst will on a weekly basis, compare all paper forms with their associated electronic entry to assess accuracy. The RAs will be engaged weekly regarding any noted inconsistencies.

Aim 1 analytic plan: Characterize frailty and multimorbidity in older PWH.

Frailty phenotype: We will apply the Fried definition for physical frailty phenotype, which assesses components of unintentional weight loss, impaired grip strength and walking speed, fatigue and falling physical activity levels [6]. This definition categorizes people as frail if at least 3 out of the 5 components are positive for frailty. If 1 or 2 are components are positive, the definition categorizes them as pre-frail. If none, are positive, then frailty is considered absent. Prevalence of this frailty phenotype will be calculated as the proportion of participants who are not frail, pre-frail or frail. We will use a one sample proportion test to assess whether the proportion is significantly different from our hypothesized 15% for people who are frail. We will use multinomial logistic regression to identify correlates of the different frailty categories at baseline.

Frailty Index: For frailty assessed as an accumulation of deficits in health, we will create an index derived from 30–40 variables describing symptoms, signs, disabilities and diseases (see Table 1). These variables will be selected to meet five previously reported criteria [56]: 1) association with health status; 2) prevalence increases with age; 3) deficit effect should not saturate too early i.e., large proportion of population has deficit by relatively young age; 4) cover a range of health systems; and 5) same set of variables should be used for longitudinal assessments. To calculate each person’s index score, we will divide the number of deficits that apply to that individual by the total number of deficits in the index. Categorical variables will be coded to fit a scale of 0 (no deficit) to 1 (full expression of deficit) and fractional codes (e.g., a single intermediate response variable of maybe will be coded as 0.5 as it sits between a no deficit and full expression of the deficit). Continuous variables with thresholds related to a clinical hazard will be coded to fit on the 0 to 1 scale. For continuous variables without clinically accepted thresholds, a threshold will be selected that corresponds to a frailty index value of 0.2 when all other deficits are summed into the index score. The value 0.2 on the frailty index is recognized as approaching a frail state [57,58]. We will use linear regression to identify correlates of the frailty index score at baseline.

Multi-morbidity and other analyses: We will count the number of conditions co-morbid with HIV to estimate the prevalence and severity of multimorbidity. Additional analyses will examine the correlations between the two frailty measures and their association with multimorbidity, social vulnerability and 5-year mortality risk assessed by higher VACS Index Scores.

Aim 2 analytic plan: Diagnosis, recognition and management of discrete cardiometabolic diseases.

Under-diagnosis of hypertension, dyslipidemia, diabetes mellitus, and renal dysfunction will be assessed as the proportion of existing conditions newly captured by screening during the current study. Under-recognition will be assessed as the proportion existing conditions that participants were not previously aware of. Positive screening follow-up rate will be assessed as the proportion of people with a positive screening that had a follow-up medical encounter related to that positive screening. We will re-assess blood pressure, lipids, hemoglobin A1c, and estimated glomerular filtration rate to determine disease progression among those with positive screenings and disease incidence among those with negative screenings. We will use a one sample test of proportion to assess whether the positive screening follow-up rate at 12 months is significantly different from our hypothesized 50%. To determine the correlates of disease management over time, we will use generalized linear mixed effects regression models [59]. Subject-specific random intercepts and slopes will be used to account for the correlation due to having repeated observations.

Aim 3: Barriers and facilitators to providing effective patient centered healthcare to aging PWH.

Data analyses for this aim will be descriptive in nature; formal hypothesis testing will not be conducted. Qualitative interview data will be analyzed using the framework approach. This study aim will be guided by Levesque’s framework [60] of access to healthcare as a conceptual framework to identify both supply and demand-side factors that influence access to health care. A priori themes adapted from Levesque’s framework regarding engagement and satisfaction with healthcare for HIV, frailty and discrete diseases of aging and the impact of social vulnerabilities on their health as well as emerging themes from the data will form the basis of the framework analysis.

Aim 4: Capacity building in research skills that also have clinical relevance.

Simple pre- and post-tests will be used to assess if there are any changes in the knowledge and skills of participants as a result of the sensitization and training sessions and capacity building activities. Thematic analysis of qualitative data from research staff interviews will be used to assess knowledge and skills gained and participant’s perception of the benefits and impact of the capacity gained.