Previous Article in Journal
Impact of Incretin Mimetics on Thyroid Cancer Among Patients with Type 2 Diabetes: A Retrospective Cohort Time-to-Event Analysis
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Pharmacoepidemiology
Volume 4
Issue 2
10.3390/pharma4020010
pharmacoepidemiology-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Evaluation of Screening Tool of Older People’s Prescriptions (Stopp) Criteria in an Urban Cohort of Older People with HIV

by
Lauren F. O’Connor
1,
Jenna B. Resnik
1,
Sam Simmens
2,
Vinay Bhandaru
2,
Debra Benator
3,
La’Marcus Wingate
4,
Amanda D. Castel
1 and
Anne K. Monroe
1,*,† on behalf of the DC Cohort Executive Committee
1
Department of Epidemiology, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA
2
Department of Biostatistics, Milken Institute School of Public Health, George Washington University, Washington, DC 20052, USA
3
Washington DC VA Medical Center, Washington, DC 20422, USA
4
Department of Clinical & Administrative Pharmacy Sciences, Howard University College of Pharmacy, Washington, DC 20059, USA
*
Author to whom correspondence should be addressed.
†
Collaborators of the DC Cohort Executive Committee are indicated in Acknowledgments.
Pharmacoepidemiology 2025, 4(2), 10; https://doi.org/10.3390/pharma4020010
Submission received: 8 April 2025 / Revised: 29 April 2025 / Accepted: 2 May 2025 / Published: 12 May 2025
Browse Figures
Versions Notes

Abstract

:
Background: The validated Screening Tool of Older People’s Prescriptions (STOPP) identifies potentially inappropriate prescribing (PIP)—treatments where potential risk outweighs potential benefit. STOPP is particularly important for people aging with HIV and comorbidities, since PIP may exacerbate symptoms and decrease adherence. Methods: We analyzed data from the DC Cohort, a longitudinal cohort of people with HIV (PWH). We applied STOPP criteria to identify PIP among DC Cohort participants aged ≥ 50 years who completed a Patient Reported Outcomes (PROs) survey. All medications prescribed in the 2 years prior to PROs survey completion were considered. Negative binomial models were used to evaluate factors associated with PIP and structural equation modeling was used to evaluate whether symptom burden mediates the relationship between PIP and quality of life. Results: Of 1048 eligible DC Cohort participants, 486 (46%) had at least one PIP. The most common systems implicated were musculoskeletal (23%), analgesic drugs (16%), and the central nervous system (13%). Age, race/ethnicity, HIV transmission factor, social determinants of health, and type of HIV care site were significantly associated with number of PIP in the crude models. In the multivariable model with just demographic variables, the association between age (aIRR: 1.03 (95% CI: 1.02, 1.04)), intravenous drug use (aIRR: 1.68 (95% CI: 1.20, 2.35)), White, non-Hispanic race (aIRR: 0.67 (95% CI: 0.50, 0.92)), site type (aIRR: 0.75 (95% CI: 0.62, 0.92)), and the expected number of PIPs remained significant. In the fully adjusted multivariable model with demographics and SDOH, the association between age, intravenous drug use, White, non-Hispanic race, and expected number of PIPs remained significant. Statistical evidence that symptom burden mediates the relationship between PIP and each of the QOL dimensions was present. Conclusions: Future interventions should work to decrease PIP among these high-risk groups, especially for PIP associated with increased symptom burden.

1. Introduction

People with HIV (PWH) have increased susceptibility to multiple comorbid conditions, especially as they age [1,2,3]. As the number of comorbidities increases, so do the number of medications that may be prescribed. With an extremely high comorbidity burden and increasing prescriptions needs, older PWH are at a high risk for polypharmacy and potentially inappropriate or contraindicated prescribing (PIP) [4,5]. PIP is when the potential risks of a medication outweigh the potential benefits [6,7]. There are various methods for evaluating PIP. One commonly used tool is the “screening tool of older people’s prescriptions” (STOPP) [8]. The goal of STOPP is to provide evidence-based rules of avoidance of commonly encountered instances of PIP [9]. In clinical settings, subsequent to identifying PIP with a tool like STOPP, a clinical team then evaluates if the medication is inappropriate for a particular patient and discontinues it, or if, for that patient, the benefits of the medication actually outweigh the risks and continues it.
Multiple studies in various clinical settings around the world have found polypharmacy and PIP are highly prevalent in older populations [10,11,12,13], and the nature of this polypharmacy can differ by age [14,15,16]. One study specifically evaluating PWH estimated that two thirds of the prescriptions were inappropriate, with reasons including incorrect dosage, lack of explanation for prescribing, interactions between prescriptions, and medication not being appropriate for older individuals [4,17]. In addition, these PIP are overwhelmingly related to comorbid conditions, rather than HIV treatment itself [4,17]. For example, one study found that older PWH were taking an average of 13 different medications at one time, the majority of which were not HIV-related [17]. Some of the most common comorbidities that require additional prescriptions include dyslipidemia, hypertension, depression, diabetes, and chronic obstructive pulmonary disease, although this is not an exhaustive list [17,18].
The high prevalence of polypharmacy and PIP is concerning given that both can lead to poorer physical health outcomes and a reduced quality of life (QOL), especially in an older population [6]. It is important to better understand how PIP may play a role in health outcomes among older PWH, but this area remains largely unexplored. Previous studies have also found disparities that exist in which vulnerable populations have higher rates of PIP. For example, it has been previously found that women are more likely to have PIP than men [4,19,20]. Differences have also been found across racial groups and race/ethnicity has been found to serve as an effect modifier of the relationship between PIP and socioeconomic status [21,22,23].
Given the high prevalence of polypharmacy and increased likelihood of PIP in older PWH, it is critical to understand the nature of PIP among older PWH and evaluate associated factors. It is also important to better understand the impacts of inappropriate prescribing on HIV health-related outcomes and QOL. Therefore, we evaluated the prevalence of PIP in a sample of older PWH in the DC Cohort, a longitudinal cohort of PWH receiving care in Washington, District of Columbia (DC). The primary objective of this study was to characterize PIP among older PWH and evaluate factors such as clinical characteristics, demographics, and social determinants of health (SDOH), that may be associated with PIP. The secondary objective was to evaluate the relationship between PIP, QOL, and symptom burden, as well as examine whether symptom burden serves as a mediator in this relationship. As this is our initial investigation of this topic in our cohort, the scope of this study is limited to the application of STOPP rules to our cohort. Subsequent research will focus on the potential downstream impact of applying this screening tool: for example, performing clinical reviews based on screening results to definitively identify medications that should be discontinued and examining the impact of those continuations on multiple domains, including adverse events, quality of life, and cost.

2. Methods

2.1. Participant Recruitment and Data Collection

Our analysis used data from a cohort study to characterize PIP among PWH receiving care in the DC area. We used data from the DC Cohort, which is an ongoing longitudinal cohort of consenting HIV patients in the DC area. The cohort began enrollment in 2011 and currently has over 13,000 participants enrolled. The cohort uses data from electronic health record (EHR) extraction to collect information on medications, diagnoses, labs, and procedures recorded at each participant’s site of HIV care. Data is extracted, deidentified, and updated on a quarterly basis throughout the calendar year and all information is stored in one secure database [24].
We administered a self-reported survey of patient-reported outcomes (PROs) among a subset of DC Cohort participants. Participants previously enrolled in the DC Cohort were recruited during visits with their HIV care provider and consenting participants were then asked to complete the survey. All participants providing PROs had to be at least 18 years old or adolescents seeking care independently. Participants had the option to complete the survey in either English or Spanish. The survey consisted of questions surrounding various PROs such as GAD-7 and PHQ-8 measures used to classify mental health, items from the modified Memorial Symptom Assessment Scale to measure social SDOH, the EuroQOL 5-Dimensional scale to evaluate QOL, and symptom burden measured by the Accountable Health Communities Screening Tool [25,26,27,28,29,30,31]. Participant responses were stored in a secure REDCap database and surveys were then linked to their DC Cohort data [32,33]. Surveys containing PROs that could not be linked to the DC Cohort database were excluded from the study.
We collected information on baseline demographics, comorbidities, HIV viral load (copies/mL), prior diagnoses, vital signs, and medications from the DC Cohort database. All DC Cohort data was collected prior to 30 September 2024. Then, we collected data on QOL, symptom burden, and SDOH from the completed PROs surveys. All PRO surveys completed and linked to the DC Cohort database by 30 September 2024 were included in the analysis.
We examined all DC Cohort participants who were enrolled in the cohort and completed a PROs survey by 30 September 2024. To be included in the final analysis, participants had to be 50 years old or older at the time of their PROs survey.

2.2. Measuring Potentially Inappropriate Prescriptions

The “screening tool of older people’s prescriptions” (STOPP) and “screening tool to alert to right treatment” (START) are previously defined tools to examine both PIP and prescriptions that should be initiated among older people [9]. For the purposes of this study, we focused on STOPP criteria alone to examine the prevalence of PIP among our study participants. The tool consists of a list of 77 different criteria involving medications that should not be given after or within a certain amount of time of a diagnosis, lab result, or in combination with another medication. The criteria are also split into subsections for each physiological system. A list of the STOPP criteria has previously been published by O’Mahony in 2015 and the specific definitions used for this criteria can be found in Supplemental Materials [34].
We applied the STOPP criteria to all participants in our study. Only medications prescribed within 2 years prior to PROs completion were considered. The number of individual STOPP criteria each participant met was evaluated, as well as whether participants met at least one of the STOPP criteria within each individual physiological system.

2.3. Measuring Symptom Burden

To measure symptom burden, we used the modified Memorial Symptom Assessment Scale (Short Form) (MSAS-SF). This scale is widely used and has been previously validated [27]. Participants filled out a series of yes/no questions on whether they had recently experienced a symptom. If the participant said they had the symptom, they then filled out a second question regarding how severe, distressing, or frequent the symptom was. This severity/distress/frequency score can range from 1 to 4 and serves as the score for each symptom. Then, continuous measures of symptom burden are calculated from these symptom scores. For instance, the physical symptom score is the average of symptom scores for lack of appetite, lack of energy, pain, feeling drowsy, constipation, dry mouth, nausea, vomiting, change in taste, weight loss, feeling bloated, and dizziness. The psychological symptom score is the average of symptom scores for feeling sad, worrying, feeling irritable, feeling nervous, difficulty sleeping, and difficulty concentrating. Finally, the overall global distress index (GDI) is calculated from the average symptom scores for feeling sad, worrying, feeling irritable, feeling nervous, lack of appetite, lack of energy, pain, feeling drowsy, constipation, and dry mouth [27]. We primarily focused on the GDI for the purposes of this study as it encapsulates general symptom burden.

2.4. Measuring Quality of Life

To measure QOL, we used the EuroQol 5-Dimensional Scale (EQ-5D) [30]. This is a widely used and previously validated scale to evaluate problems within various areas of QOL. The scale focuses on five different dimensions of QOL: mobility, conducting usual activities, anxiety/depression, self-care, and pain/discomfort. For each QOL dimension, participants answer whether they have had no problems, moderate problems, or severe problems. Based on these responses, a 5-digit result can be generated for each participant, or each dimension can be evaluated independently as five different discrete outcome variables.

2.5. Measuring Social Determinants of Health

To measure SDOH, we used the Accountable Health Communities Screening (AHC) Tool [31]. This tool was previously developed and tested by the Centers for Medicare and Medicaid Services. It focuses on five different SDOH domains: housing instability, food insecurity, transportation difficulties, utility assistance needs, and interpersonal safety. For each of these domains, participants answer questions that are used to determine whether the participant has a need in that domain. We therefore measured transportation difficulties, utility assistance needs, housing instability, and food insecurity as separate dichotomous yes/no variables. For our PROs survey, we did not include questions about interpersonal safety.

2.6. Statistical Analysis

First, we examined the proportion of study participants who had at least one PIP and compared the distribution of demographic characteristics between those with and without at least one PIP. Chi-squared and Wilcoxon rank-sum tests were used to evaluate differences between those with and without a PIP. We then evaluated the number of people who had at least one PIP within each STOPP physiological system using descriptive statistics. Then, to calculate the number of PIPs each person had, we evaluated the number of individual STOPP criteria each person met at least once and used that sum as the overall outcome variable.
We then evaluated the association between each covariate and the number of PIPs each person met. We evaluated a series of unadjusted Poisson models to check for overdispersion. If overdispersion was found, we evaluated a negative binomial model, zero-inflated model, and zero-inflated negative binomial model. Each model was compared to the respective nested model using a likelihood ratio test. If the likelihood ratio test was non-significant (p > 0.05), then the simpler model was selected to calculate the incidence rate ratios (IRR). Once a model with adequate fit was chosen, we examined unadjusted associations between demographics (race/ethnicity, gender, HIV transmission factor, age), type of HIV care site, viral load lab at the time of PROs completion, and number of PIPs to determine baseline characteristics that are associated with PIP. We then ran a multivariable regression model that adjusted for each of the above baseline characteristics and SDOH. Finally, we used a likelihood ratio test to compare the reduced model to the fully adjusted model, using only participants who were included in both models, for goodness of fit. All regression analyses were conducted using SAS 9.4 (Cary, NC, USA).

2.7. Mediation Analysis

After evaluating baseline characteristics associated with PIP, we examined the impact of PIP on symptom burden and QOL. It has previously been established that PIP is associated with increased symptom burden and lower QOL [35,36]. Given that all our prescriptions were prior to the date the PROs survey was completed, we evaluated whether symptom burden mediates the relationship between PIP and QOL. An overview of the hypothesized relationship can be found in Figure 1. We evaluated whether the association between inappropriate prescribing and each QOL factor was mediated by symptom burden, as measured by the GDI. To do so, we first used Baron and Kenny’s approach to check the following assumptions: (1) The unadjusted association between inappropriate prescribing and symptom burden is significant, (2) the unadjusted association between inappropriate prescribing and the QOL factor is significant, (3) symptom burden is associated with QOL, after adjusting for PIP, and (4) PIP is not significantly associated with QOL, after adjusting for GDI. If each of the above assumptions were met, we evaluated the indirect and direct effect of PIP on each QOL dimension through symptom burden, as measured by GDI. We then calculated a 95% confidence interval for the indirect and direct effect using bias-corrected bootstrapping methods (iterations: 2000). Each of the above models were run after adjusting for age, gender, and race/ethnicity. Effects with a p-value < 0.05 were considered statistically significant evidence of mediation by symptom burden. Mediation analyses were conducted using MPlus Eighth Edition (Los Angeles, CA, USA) [37].

3. Results

As of 30 September 2024, there were over 13,000 participants across 14 different sites in the DC Cohort. Of these, there were 1048 cohort participants who met the inclusion criteria for this study and were included in the analysis. The average age of participants was 61 years (±7). A majority of participants were Black, non-Hispanic (N = 830, 79%), male (750, 72%), and men who have sex with men (N = 383, 37%) (Table 1). When compared to the total DC Cohort participants who were eligible to take the PROs survey, but did not enroll, our subset of study participants did not differ significantly with regards to race/ethnicity (p = 0.1505), gender (p = 0.6165), or HIV transmission factor (p = 0.4239).
After applying the list of STOPP criteria, which is divided by physiological body systems, to all 1048 participants, 486 (46%) had at least one PIP. Those with PIP significantly differed from those without PIP with regards to age, race/ethnicity, gender, HIV transmission factor, site type, viral suppression, SDOH, QOL, and symptom burden (Table 1). The most common body system for which participants had PIP was the musculoskeletal system (N = 245, 23%), followed by analgesic drugs (N = 172, 16%), and the central nervous system (N = 140, 13%) (Table 2). For the musculoskeletal system, the most commonly met STOPP criteria were H7 (N = 133, 13%) and H8 (N = 151, 14%)), which both refer to the inappropriate prescribing of non-steroidal anti-inflammatory drugs (NSAIDs). For analgesic drugs, L2 was the most commonly met STOPP criteria (N = 172, 16%), which refers to the use of opioids without a concomitant laxative. For the central nervous system, D14 (N = 99, 9%), which refers to the use of first-generation antihistamines, was the most commonly met criteria. See Supplemental Table S1 for the full distribution of individual STOPP criteria.

3.1. Characteristics Associated with Potentially Inappropriate Prescribing

We used Poisson models to evaluate the simple associations between each covariate and PIP but found evidence of overdispersion. Therefore, we evaluated a negative binomial model for each covariate and compared the fit to the Poisson model. The resulting likelihood ratio test was significant, suggesting the negative binomial model had better fit. We then compared the negative binomial model to the zero-inflated negative binomial model and found that the likelihood ratio test was non-significant, suggesting the negative binomial model was adequate. Therefore, we used a negative binomial model for all analyses. An overview of fit statistics for an example model between age and PIP can be found in Supplemental Table S2.
In the unadjusted model, age, race/ethnicity, HIV transmission factor, site type, and food insecurity were significantly associated with the expected number of PIPs. Specifically, a one-year increase in age was associated with a 3% increase in the expected number of PIPs (IRR: 1.03 (95% CI: 1.02, 1.04)). Those with an HIV transmission factor of heterosexual contact (IRR: 1.39 (95% CI: 1.12, 1.73)) or intravenous drug use (IDU) (IRR: 2.18 (95% CI: 1.58, 3.02)) had a higher expected number of PIPs than men who have sex with men. Having food insecurity was also associated with an increased expected number of PIPs (IRR: 1.28 (95% CI: 1.04, 1.57) while White, non-Hispanic race (IRR: 0.61 (95% CI: 0.45, 0.81)), and receiving care at a hospital site were associated with a decreased expected number of PIPs (IRR: 0.77 (95% CI: 0.64, 0.93)) (Table 3).
In the multivariable model including demographic variables only, the association between age (aIRR: 1.03 (95% CI: 1.02, 1.04)), IDU (aIRR: 1.68 (95% CI: 1.20, 2.35)), White, non-Hispanic race (aIRR: 0.67 (95% CI: 0.50, 0.92)), site type (aIRR: 0.75 (95% CI: 0.62, 0.92)) and expected number of PIPs remained significant (Table 3, Model I). In the fully adjusted multivariable model with demographics and SDOH, the association between age, IDU, White, non-Hispanic race and expected number of PIPs remained significant (Table 3, Model II).

3.2. Mediation Analysis Results

The proposed relationship between PIP and other covariates is displayed in Figure 1. We evaluated whether the relationship between each QOL dimension and PIP is mediated by symptom burden, measured by GDI. The direct path between PIP and GDI (Path A) was significant, as well as the direct path between GDI and each QOL dimension (Path B). The direct path between PIP and QOL, controlling for GDI, was also significant for each QOL dimension (Path C) (Table 4).
Given that each of these dimensions met the assumptions for the evaluation of mediation, we evaluated the indirect effect of PIP on each QOL dimension, where a value of 0 was considered to have null effect. The indirect effect of PIP on anxiety/depression (β: 0.22 (95% CI: 0.11, 0.33)), usual activities (β: 0.12 (95% CI: 0.06, 0.18)), mobility (β: 0.09 (95% CI: 0.05, 0.14)), self-care (β: 0.11 (95% CI: 0.06, 0.18)), and pain/discomfort (β: 0.12 (95% CI: 0.06, 0.19)) were all significant. Therefore, there is statistical evidence that symptom burden mediates the relationship between PIP and each of the QOL dimensions. A visualization of the mediation analysis can be found in Figure 2 and details of the mediation estimates can be found in Table 4.

4. Discussion

Using the STOPP criteria, we determined that PIP was present in nearly half of patients in an urban cohort of PWH. This high frequency of PIP has been found previously among older PWH [8]. Prescriptions related to the musculoskeletal system, those for analgesic drugs, and prescriptions related to the central nervous system were most commonly identified as potentially inappropriate. It was beyond the scope of this investigation to determine what proportion of PIP would ultimately be deemed appropriate or inappropriate (necessitating medication discontinuation) by the treating clinician. We found that age, White, non-Hispanic race and having an HIV transmission factor of IDU—were associated with PIP in the fully adjusted models. We also importantly found that the relationship between PIP and QOL concerns is mediated by symptom burden.
For the musculoskeletal system, the two STOPP criteria that were most commonly met were related to prescribing long courses of NSAIDs or prescribing NSAIDs to individuals with other comorbidities. NSAIDs can have multiple complications in older adults, including gastrointestinal bleeding, kidney failure, and heart failure [38,39,40]. Meeting NSAID-related STOPP criteria was frequently observed in other studies, such as a prior study conducted among older PWH, and the potential harm of NSAIDs been well documented [13,17,41,42,43,44,45,46]. A systematic review of hospitalizations among adults 60 and older showed a mean prevalence of adverse drug reaction-related hospitalizations of 8.7% (95% CI 7.6–9.8), with NSAIDs being the most commonly implicated class [47]. However, given that many PWH have chronic pain [48], providers may feel that the benefits of NSAIDs, especially compared to other classes of pain medications, outweigh their risks. This is a challenging clinical conundrum, and, in future work, we can explore how many potentially inappropriate NSAID prescriptions are actually discontinued after review by the treating provider.
Another commonly identified PIP was the prescription of an opioid without a concurrent laxative. This can lead to constipation, which is particularly detrimental among older individuals [49]. Opioid use among older individuals on its own is problematic due to adverse effects including impact on cognition and balance [50]. Although opioid prescribing has historically been common among PWH [51], in recent years due to the heightened awareness of the multiple adverse effects of opioids, prescribing of these agents has steeply declined [52]. Illicit use of these substances remains a concern, however. In general, constipation among older adults is a major concern and should be a focus of providers caring for aging PWH. Non-pharmacologic interventions (increasing water intake, increasing consumption of high fiber food), in addition to pharmacologic interventions as appropriate, can alleviate this concern.
PIP related to the central nervous system has been commonly found in older PWH [8]. For our criteria, the most commonly met criteria for PIP referred to the use of first-generation antihistamines. First generation antihistamines such as diphenhydramine and hydroxyzine may commonly be prescribed among aging individuals in an attempt to address symptoms such as anxiety and insomnia. Although these agents may be perceived to be less harmful than other agents for these symptoms, they have significant effects in older populations due to their anticholinergic effects and their potential to raise the risk of dementia [53].
In our fully adjusted model, the only demographic factors associated with PIP were increasing age, White, non-Hispanic race, and an HIV transmission factor of IDU. Given the strong evidence that older individuals are at increased risk for polypharmacy and PIP, it is not surprising that age was an important risk factor for PIP in our study [6,14,15,16,17,54,55,56]. This continuing trend is especially concerning given the risks of adverse drug events, falls, and dangerous prescribing cascades that can make inappropriate prescriptions especially dangerous among older populations [54,57,58,59]. The addition of HIV medications into this risk of polypharmacy puts our population of PWH at especially high risk for these adverse events. As the population of PWH ages, it is increasingly important to develop more targeted interventions to reduce PIP among PWH specifically.
PWH with a history of IDU have an excess of certain comorbidities, specifically end-stage kidney and liver disease, compared to PWH with other HIV transmission factors [60]. Higher comorbidity burden may explain in part why we observed more polypharmacy in individuals with IDU. These individuals may also experience challenges regarding chronic pain and pain management that are not experienced by individuals without a history of IDU [61]. In a polypharmacy analysis from the HIV Outpatient Study (HOPS), individuals with illicit substance use had a higher odds of receiving medications with adverse interactions, although there was not an observed higher odds of interacting or contraindicated medications by HIV transmission status [62].
Evidence of mediation in the relationship between polypharmacy and health-related outcomes has been found previously, though none have looked at PIP specifically. For instance, multimorbidity has been found to mediate the relationship between polypharmacy and self-rated health, as well as depression and obesity [63,64]. Drug interactions have also been found to mediate the relationship between polypharmacy and dementia [65]. Furthermore, polypharmacy has also been found to mediate the relationship between other health-related factors [66,67]. These findings suggest that polypharmacy and PIP are part of a complicated web of health-related outcomes. Future interventions should take these complex relationships into account when evaluating how to reduce inappropriate prescribing and improve patient outcomes.
Our study did not find differences in inappropriate prescribing by gender, after controlling for other covariates. This contrasts with prior studies that found women are more likely to receive PIP than men [4,19,68]. The high proportion of men in our study population may have limited our ability to ascertain an association; however, it is possible that the dynamics of gender inequity among PWH specifically may be leading to alternative findings. Future studies should evaluate why gender inequities may occur with regard to inappropriate prescribing and whether these gender inequities may be explained by other patient characteristics.

4.1. Clinical Relevance and Potential Interventions

As discussed previously, not all STOPP-identified PIP results in a subsequent medication discontinuation, as the individual patient risk/benefit ratio as assessed by the treating clinician varies. STOPP is a tool that can be used to flag patients for further review. Various researchers have examined the clinical significance of PIP detected by STOPP and START, which evaluates medications that have been omitted. One Swedish study of 302 primary care patients demonstrated that the majority of patients (86%) met a STOPP/START criteria, and 15% of those resulted in a change in medication [69]. Another Swedish study involving 200 hip fracture patients determined that half of STOPP/START criteria findings were clinically relevant and necessitated medication changes [70]. Therefore, in observational work, STOPP/START flags have been associated with worse outcomes, although not every flag will result in an intervention on the part of the clinician. However, there is certainly room for improvement, which has led researchers to try to develop interventions to identify PIP and implement changes based on these findings.
It has previously been reported that comprehensive reviews of the medical chart, especially among elderly patients, can effectively lead to the discontinuation of PIP [18,54,71]. Educational interventions from the provider have been found to effectively reduce PIP, especially for tricyclic antidepressants [72,73,74,75]. Computerized systems that automatically flag potential PIP have also been found to lower the incidence [72,76]. While these interventions have been effective at lowering PIP, providers have reported that limited time and resources create barriers to these and similar interventions [77]. More efficient interventions that target high-risk individuals or common PIP may be more effective at lowering the incidence. For example, limited computer flags that only appear for common medications such as NSAIDs, tricyclic antidepressants, and selective alpha-1 blockers may be more efficient at alerting providers. Targeting patients for pharmacist chart reviews, such as those with hypertension or a reported housing need, may also lower the burden on providers while reducing PIP in the most at-risk groups. Additionally, the mediated relationship found here suggests that interventions to lower PIP should especially focus on those prescriptions that have the highest impact on symptom burden. These would theoretically have the greatest impact on QOL and therefore should be emphasized. Implementing such targeted intervention methods would be more effective at improving QOL due to symptom burden for older PWH. Future studies should evaluate which prescriptions have the greatest impact on symptom burden and develop targeted interventions towards them.

4.2. Strengths and Limitations

Our study was strengthened by our large sample size and available longitudinal data on PWH. We were able to ascertain exact dates of prescriptions, diagnoses, labs, and vitals in order to characterize whether a prescription was potentially inappropriate. We also only included prescriptions that were given within the 2 years before the date of the PROs survey, therefore ensuring all PIPs occurred before the QOL and symptom burden reports and could theoretically be a causal factor. While prior studies have evaluated medication counts, we were able to apply a standard list of criteria to classify whether medications are potentially inappropriate. This ensures we are not just measuring who receives more medications, which could be a by-product of comorbidity burden, but rather, who was given medications that were contraindicated. These strengths allowed us to better characterize the extent of the problem and ascertain the underlying factors that may be causing it.
Our study also had several limitations. First, the STOPP criteria used to characterize PIP were originally designed to be implemented on the individual patient-level and not applied to a large cohort. Therefore, we were limited in the precision with which we could determine whether an individual participant received PIP. It is possible that certain criteria may have been over- or under-inflated due to our inability to individually review the medication and diagnosis history of each participant. Similarly, we used version 2 of the STOPP/START criteria as that was the most recent version available when this analysis began. However, at the time of this publication, O’Mahony et al. has released version 3 of the criteria [34]. Future analyses should incorporate the more recent version of this criteria. Furthermore, all of our data came from EHRs exported from each participant’s HIV care provider. Therefore, medications or diagnoses that were not entered into the chart or were entered at a different provider site would not have been included in the analysis. This may have introduced misclassification into our outcome measure, although it is unlikely that this misclassification would have differed by any factor other than site of HIV care.
Although we found evidence of potential mediation by symptom burden, we did not collect information on other potential mediating variables, such as financial burden, that may also be contributing to the effect of PIP. Therefore, we are limited in our ability to conclude whether symptom burden fully mediates the relationship between PIP and QOL. Future studies should evaluate other covariates that may be contributing to this relationship.
Another limitation is that there is variation by clinical site with regards to EHR systems and data availability. With regard to EHR systems, some clinics’ EHR systems include alerts to clinicians to avoid certain drug combinations. With regards to data availability, from some sites, the DC Cohort receives prescription information while from others, we receive drug dispensing information, which is potentially more accurate. Also, insurance types among DC Cohort participants vary and this may impact the prescribing choices of physicians and the medication use patterns of participants. Finally, all participants in our study were aged 50 or older. Therefore, our results may not be generalizable to younger populations or a subset of older populations, such as only those over 65.

5. Conclusions

We found that close to half of the participants in our cohort of PWH had PIP, and that PIP was associated with reporting a lower QOL, a relationship which was mediated by symptom burden. Some of the medications determined to be potentially inappropriate, although seemingly benign, can have significant consequences in aging PWH. Future interventions should target those identified here as high-risk for PIP and focus on medications that most commonly lead to increased symptom burden.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/pharma4020010/s1, Table S1: Overview of the number of participants with at least one inappropriate prescription for each STOPP criteria; Table S2: Comparison of fit statistics for the Poisson, negative binomial, zero-inflated, and zero-inflated negative binomial models comparing the association between age and PIP; Table S3: Detailed definitions of defining STOPP criteria in the study sample. Please see original list of STOPP criteria for full references [9]; Table S4: Detailed definitions of drug classifications in the study sample; Table S5: Detailed overview of diagnosis and other comorbidity definitions.

Author Contributions

Conceptualization, L.W. and A.K.M.; Methodology, L.F.O., S.S. and A.K.M.; Software, L.F.O., S.S. and V.B.; Validation, J.B.R., S.S. and V.B.; Formal Analysis, L.F.O. and S.S.; Investigation, A.K.M. and L.W.; Resources, S.S., D.B., L.W., A.D.C. and A.K.M.; Data Curation, D.B., A.D.C. and A.K.M.; Writing—Original Draft Preparation, L.F.O., J.B.R., S.S. and A.K.M.; Writing—Review and Editing, L.F.O., J.B.R., S.S., V.B., D.B., L.W., A.D.C. and A.K.M.; Supervision, S.S., D.B., L.W., A.D.C. and A.K.M.; Project Administration, D.B., A.D.C. and A.K.M.; Funding Acquisition, A.D.C. and A.K.M. All authors have read and agreed to the published version of the manuscript.

Funding

The DC Cohort is funded by the National Institute of Allergy and Infectious Diseases, 1R24AI152598-01. The Milken Institute School of Public Health Research Innovation Award provided funding support for the Patient Reported Outcomes Survey.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of George Washington University (Approval code: 071029 and Approval date: 14 September 2010).

Informed Consent Statement

Informed consent was obtained from all participants involved in the study.

Data Availability Statement

Per DC Cohort protocols, data are available upon request and approval of the DC Cohort Executive Committee. Interested parties should email the Principal Investigator at acastel@gwu.eu.

Acknowledgments

Data in this manuscript were collected by the DC Cohort Study Group with investigators and research staff located at the Children’s National Hospital Pediatric clinic (Natella Rakhmanina); the Senior Deputy Director of the DC Department of Health HAHSTA (Clover Barnes); Family and Medical Counseling Service (Rita Aidoo); Georgetown University (Princy Kumar); The George Washington University Biostatistics Center (Tsedenia Bezabeh, Vinay Bhandaru, Asare Buahin, Nisha Grover, Lisa Mele, Alla Sapozhnikova, Greg Strylewicz, and Marinella Temprosa); The George Washington University Department of Epidemiology (Shannon Barth, Morgan Byrne, Amanda Castel, Alan Greenberg, Shannon Hammerlund, Paige Kulie, Anne Monroe, Megan O’Brien, Lauren O’Connor, James Peterson, Jonathon Rendina, Su Yadana, and Martin Zavala) and Department of Biostatistics and Bioinformatics; The George Washington University Medical Faculty Associates (Jose Lucar); Howard University Adult Infectious Disease Clinic (Jhansi L. Gajjala) and Pediatric Clinic (Sohail Rana); Kaiser Permanente Mid-Atlantic States (Michael Horberg); La Clinica Del Pueblo (Suyanna Barker); Washington Health Institute, formerly Providence Hospital (Jose Bordon); Unity Health Care (Gebeyehu Teferi); Us Helping Us, People Into Living, Inc. (DeMarc Hickson); Veterans Affairs Medical Center (Rachel Denyer); Washington Hospital Center (Adam Klein); and Whitman-Walker Institute (Stephen Abbott).

Conflicts of Interest

The authors declare no conflicts of interest. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.

References

  1. Cuzin, L.; Katlama, C.; Cotte, L.; Pugliese, P.; Cheret, A.; Bernaud, C.; Rey, D.; Poizot-Martin, I.; Chirouze, C.; Bani-Sadr, F.; et al. Ageing with HIV: Do comorbidities and polymedication drive treatment optimization? HIV Med. 2017, 18, 395–401. [Google Scholar] [CrossRef]
  2. Mata-Marín, J.A.; Martínez-Osio, M.H.; Arroyo-Anduiza, C.I.; Berrospe-Silva, M.d.l.Á.; Chaparro-Sánchez, A.; Cruz-Grajales, I.; Cruz-Herrera, J.E.; Uribe-Noguez, L.A.; Gaytán-Martínez, J.E.; Jerónimo-Morales, M. Comorbidities and polypharmacy among HIV-positive patients aged 50 years and over: A case–control study. BMC Res. Notes 2019, 12, 556. [Google Scholar] [CrossRef] [PubMed]
  3. Levy, M.E.; Greenberg, A.E.; Hart, R.; Powers Happ, L.; Hadigan, C.; Castel, A. DC Cohort Executive Committee High burden of metabolic comorbidities in a citywide cohort of HIV outpatients: Evolving health care needs of people aging with HIV in Washington, DC. HIV Med. 2017, 18, 724–735. [Google Scholar] [CrossRef]
  4. Livio, F.; Deutschmann, E.; Moffa, G.; Rrustemi, F.; Stader, F.; Elzi, L.; Braun, D.L.; Calmy, A.; Hachfeld, A.; Cavassini, M.; et al. Analysis of inappropriate prescribing in elderly patients of the Swiss HIV Cohort Study reveals gender inequity. J. Antimicrob. Chemother. 2021, 76, 758–764. [Google Scholar] [CrossRef] [PubMed]
  5. Morillo-Verdugo, R.; Blanco Ramos, J.R.; Abdel-Kader Martín, L.; Álvarez de Sotomayor, M. The challenge of aging and pharmacoterapeutic complexity in the HIV + patient. Farm Hosp. 2018, 42, 120–127. [Google Scholar] [CrossRef]
  6. Back, D.; Marzolini, C. The challenge of HIV treatment in an era of polypharmacy. J. Int. AIDS Soc. 2020, 23, e25449. [Google Scholar] [CrossRef] [PubMed]
  7. Dimitrow, M.S.; Airaksinen, M.S.A.; Kivelä, S.-L.; Lyles, A.; Leikola, S.N.S. Comparison of prescribing criteria to evaluate the appropriateness of drug treatment in individuals aged 65 and older: A systematic review. J. Am. Geriatr. Soc. 2011, 59, 1521–1530. [Google Scholar] [CrossRef]
  8. Vélez-Díaz-Pallarés, M.; Silveira, E.D.; Fradejas, J.F.; Llorente, B.M.; Fernández, C.P.; Errasquín, B.M.; Cruz-Jentoft, A.J.; Álvarez Díaz, A.M. Potentially inappropriate prescribing in older people living with HIV: A scoping review. JAIDS J. Acquir. Immune Defic. Syndr. 2023, 94, 445–460. [Google Scholar] [CrossRef]
  9. Mahony, D.O.; Sullivan, D.O.; Byrne, S.; Connor, M.N.O.; Ryan, C.; Gallagher, P. Corrigendum: STOPP/START criteria for potentially inappropriate prescribing in older people: Version 2. Age Ageing 2018, 47, 489. [Google Scholar] [CrossRef]
  10. Mekonnen, A.B.; Redley, B.; Courten, B.; Manias, E. Potentially inappropriate prescribing and its associations with health-related and system-related outcomes in hospitalised older adults: A systematic review and meta-analysis. Br. J. Clin. Pharmacol. 2021, 87, 4150–4172. [Google Scholar] [CrossRef]
  11. Vinuesa-Hernando, J.M.; Gimeno-Gracia, M.; Malo, S.; Sanjoaquin-Conde, I.; Crusells-Canales, M.J.; Letona-Carbajo, S.; Gracia-Piquer, R. Potentially inappropriate prescriptions and therapeutic complexity in older HIV patients with comorbidities. Int. J. Clin. Pharm. 2021, 43, 1245–1250. [Google Scholar] [CrossRef] [PubMed]
  12. Loste, C.; Moltó, J.; Pérez-Álvarez, N.; Puig, J.; Echeverría, P.; Bonjoch, A.; Fumaz, C.R.; Lemos, B.; Estany, C.; Clotet, B.; et al. Potential prescribing issues among older HIV-infected subjects in a Mediterranean cohort: Does the current prevalence give cause for concern? Br. J. Clin. Pharmacol. 2021, 87, 1310–1317. [Google Scholar] [CrossRef] [PubMed]
  13. Deardorff, W.J.; Jing, B.; Growdon, M.E.; Yaffe, K.; Boscardin, W.J.; Boockvar, K.S.; Steinman, M.A. Medication misuse and overuse in community-dwelling persons with dementia. J. Am. Geriatr. Soc. 2023, 71, 3086–3098. [Google Scholar] [CrossRef] [PubMed]
  14. Bjerrum, L.; Søgaard, J.; Hallas, J.; Kragstrup, J. Polypharmacy: Correlations with sex, age and drug regimenA prescription database study. Eur. J. Clin. Pharmacol. 1998, 54, 197–202. [Google Scholar] [CrossRef]
  15. Golchin, N.; Frank, S.H.; Vince, A.; Isham, L.; Meropol, S.B. Polypharmacy in the elderly. J. Res. Pharm. Pract. 2015, 4, 85–88. [Google Scholar] [CrossRef]
  16. Gokce Kutsal, Y.; Barak, A.; Atalay, A.; Baydar, T.; Kucukoglu, S.; Tuncer, T.; Hizmetli, S.; Dursun, N.; Eyigor, S.; Sarıdogan, M.; et al. Polypharmacy in the Elderly: A Multicenter Study. J. Am. Med. Dir. Assoc. 2009, 10, 486–490. [Google Scholar] [CrossRef]
  17. Greene, M.; Steinman, M.A.; McNicholl, I.R.; Valcour, V. Polypharmacy, drug-drug interactions, and potentially inappropriate medications in older adults with human immunodeficiency virus infection. J. Am. Geriatr. Soc. 2014, 62, 447–453. [Google Scholar] [CrossRef]
  18. A Pharmacist-Led Program to Evaluate and Reduce Polypharmacy and Potentially Inappropriate Prescribing in Older HIV-Positive Patients—McNicholl—2017—Pharmacotherapy: The Journal of Human Pharmacology and Drug Therapy—Wiley Online Library. Available online: https://accpjournals.onlinelibrary.wiley.com/doi/10.1002/phar.2043 (accessed on 12 May 2023).
  19. Liu, G.G.; Christensen, D.B. The Continuing Challenge of Inappropriate Prescribing in the Elderly: An Update of the Evidence. J. Am. Pharm. Assoc. 2002, 42, 847–857. [Google Scholar] [CrossRef]
  20. Fulone, I.; Lopes, L.C. Potentially inappropriate prescriptions for elderly people taking antidepressant: Comparative tools. BMC Geriatr. 2017, 17, 278. [Google Scholar] [CrossRef]
  21. Assari, S.; Bazargan, M. Race/Ethnicity, Socioeconomic Status, and Polypharmacy among Older Americans. Pharmacy 2019, 7, 41. [Google Scholar] [CrossRef]
  22. Cashion, W.; McClellan, W.; Howard, G.; Goyal, A.; Kleinbaum, D.; Goodman, M.; Prince, V.; Muntner, P.; McClure, L.A.; McClellan, A.; et al. Geographic region and racial variations in polypharmacy in the United States. Ann. Epidemiol. 2015, 25, 433–438.e1. [Google Scholar] [CrossRef]
  23. Wang, X.; Liu, K.; Shirai, K.; Tang, C.; Hu, Y.; Wang, Y.; Hao, Y.; Dong, J.-Y. Prevalence and trends of polypharmacy in U.S. adults, 1999–2018. Glob. Health Res. Policy 2023, 8, 25. [Google Scholar] [CrossRef] [PubMed]
  24. Greenberg, A.E.; Hays, H.; Castel, A.D.; Subramanian, T.; Happ, L.P.; Jaurretche, M.; Binkley, J.; Kalmin, M.M.; Wood, K.; Hart, R.; et al. Development of a large urban longitudinal HIV clinical cohort using a web-based platform to merge electronically and manually abstracted data from disparate medical record systems: Technical challenges and innovative solutions. J. Am. Med. Inform. Assoc. 2016, 23, 635–643. [Google Scholar] [CrossRef]
  25. Kroenke, K.; Strine, T.W.; Spitzer, R.L.; Williams, J.B.W.; Berry, J.T.; Mokdad, A.H. The PHQ-8 as a measure of current depression in the general population. J. Affect. Disord. 2009, 114, 163–173. [Google Scholar] [CrossRef] [PubMed]
  26. Spitzer, R.L.; Kroenke, K.; Williams, J.B.W.; Löwe, B. A brief measure for assessing generalized anxiety disorder: The GAD-7. Arch. Intern. Med. 2006, 166, 1092–1097. [Google Scholar] [CrossRef] [PubMed]
  27. Sharp, J.L.; Gough, K.; Pascoe, M.C.; Drosdowsky, A.; Chang, V.T.; Schofield, P. The modified Memorial Symptom Assessment Scale Short Form: A modified response format and rational scoring rules. Qual. Life Res. 2018, 27, 1903–1910. [Google Scholar] [CrossRef]
  28. Earnshaw, V.A.; Smith, L.R.; Chaudoir, S.R.; Amico, K.R.; Copenhaver, M.M. HIV stigma mechanisms and well-being among PLWH: A test of the HIV stigma framework. AIDS Behav. 2013, 17, 1785–1795. [Google Scholar] [CrossRef]
  29. Christopoulos, K.A.; Neilands, T.B.; Hartogensis, W.; Geng, E.H.; Sauceda, J.; Mugavero, M.J.; Crane, H.M.; Fredericksen, R.J.; Moore, R.D.; Mathews, W.C.; et al. Internalized HIV Stigma Is Associated with Concurrent Viremia and Poor Retention in a Cohort of US Patients in HIV Care. J. Acquir. Immune Defic. Syndr. 2019, 82, 116–123. [Google Scholar] [CrossRef]
  30. EuroQol Group EuroQol—A new facility for the measurement of health-related quality of life. Health Policy 1990, 16, 199–208. [CrossRef]
  31. Standardized Screening for Health-Related Social Needs in Clinical Settings: The Accountable Health Communities Screening Tool—National Academy of Medicine. Available online: https://nam.edu/standardized-screening-for-health-related-social-needs-in-clinical-settings-the-accountable-health-communities-screening-tool/ (accessed on 14 September 2022).
  32. Harris, P.A.; Taylor, R.; Thielke, R.; Payne, J.; Gonzalez, N.; Conde, J.G. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. J. Biomed. Inform. 2009, 42, 377–381. [Google Scholar] [CrossRef]
  33. Harris, P.A.; Taylor, R.; Minor, B.L.; Elliott, V.; Fernandez, M.; O’Neal, L.; McLeod, L.; Delacqua, G.; Delacqua, F.; Kirby, J.; et al. The REDCap consortium: Building an international community of software platform partners. J. Biomed. Inform. 2019, 95, 103208. [Google Scholar] [CrossRef]
  34. STOPP-START—CGA Toolkit Plus. Available online: https://www.cgakit.com/m-2-stopp-start (accessed on 22 May 2023).
  35. Barlow, A.; Prusak, E.S.; Barlow, B.; Nightingale, G. Interventions to reduce polypharmacy and optimize medication use in older adults with cancer. J. Geriatr. Oncol. 2021, 12, 863–871. [Google Scholar] [CrossRef] [PubMed]
  36. Schenker, Y.; Park, S.Y.; Jeong, K.; Pruskowski, J.; Kavalieratos, D.; Resick, J.; Abernethy, A.; Kutner, J.S. Associations Between Polypharmacy, Symptom Burden, and Quality of Life in Patients with Advanced, Life-Limiting Illness. J. Gen. Intern. Med. 2019, 34, 559–566. [Google Scholar] [CrossRef]
  37. Muthén, B.O.; Muthén, L.K.; Muthén, B.O. Mplus User’s Guide, 8th ed.; Muthén & Muthén: Los Angeles, CA, USA, 2017. [Google Scholar]
  38. Griffin, M.R.; Ray, W.A.; Schaffner, W. Nonsteroidal anti-inflammatory drug use and death from peptic ulcer in elderly persons. Ann. Intern. Med. 1988, 109, 359–363. [Google Scholar] [CrossRef] [PubMed]
  39. Dixit, M.; Doan, T.; Kirschner, R.; Dixit, N. Significant Acute Kidney Injury Due to Non-steroidal Anti-inflammatory Drugs: Inpatient Setting. Pharmaceuticals 2010, 3, 1279–1285. [Google Scholar] [CrossRef] [PubMed]
  40. Mamdani, M.; Juurlink, D.N.; Lee, D.S.; Rochon, P.A.; Kopp, A.; Naglie, G.; Austin, P.C.; Laupacis, A.; Stukel, T.A. Cyclo-oxygenase-2 inhibitors versus non-selective non-steroidal anti-inflammatory drugs and congestive heart failure outcomes in elderly patients: A population-based cohort study. Lancet 2004, 363, 1751–1756. [Google Scholar] [CrossRef]
  41. Meng, L.-C.; Kojima, T.; Suzuki, Y.; Weng, S.-E.; Chen, H.-M.; Huang, S.-T.; Akishita, M.; Chen, L.-K.; Hsiao, F.-Y. Medication overload: A closer look at polypharmacy and potentially inappropriate medications among older people in Taiwan and Japan. Arch. Gerontol. Geriatr. 2023, 115, 105100. [Google Scholar] [CrossRef]
  42. Nguyen, T.N.M.; Laetsch, D.C.; Chen, L.-J.; Holleczek, B.; Meid, A.D.; Brenner, H.; Schöttker, B. Comparison of Five Lists to Identify Potentially Inappropriate Use of Non-Steroidal Anti-Inflammatory Drugs in Older Adults. Pain Med. 2021, 22, 1962–1969. [Google Scholar] [CrossRef]
  43. Schmidt-Mende, K.; Andersen, M.; Wettermark, B.; Hasselström, J. Drug–disease interactions in Swedish senior primary care patients were dominated by non-steroid anti-inflammatory drugs and hypertension—A population-based registry study. Scand. J. Prim. Health Care 2020, 38, 330–339. [Google Scholar] [CrossRef]
  44. Hudhra, K.; García-Caballos, M.; Casado-Fernandez, E.; Jucja, B.; Shabani, D.; Bueno-Cavanillas, A. Polypharmacy and potentially inappropriate prescriptions identified by Beers and STOPP criteria in co-morbid older patients at hospital discharge. J. Eval. Clin. Pract. 2016, 22, 189–193. [Google Scholar] [CrossRef]
  45. Cahir, C.; Fahey, T.; Teeling, M.; Teljeur, C.; Feely, J.; Bennett, K. Potentially inappropriate prescribing and cost outcomes for older people: A national population study. Br. J. Clin. Pharmacol. 2010, 69, 543–552. [Google Scholar] [CrossRef] [PubMed]
  46. Davidoff, A.J.; Miller, G.E.; Sarpong, E.M.; Yang, E.; Brandt, N.; Fick, D.M. Prevalence of Potentially Inappropriate Medication Use in Older Adults Using the 2012 Beers Criteria. J. Am. Geriatr. Soc. 2015, 63, 486–500. [Google Scholar] [CrossRef]
  47. Oscanoa, T.J.; Lizaraso, F.; Carvajal, A. Hospital admissions due to adverse drug reactions in the elderly. A meta-analysis. Eur. J. Clin. Pharmacol. 2017, 73, 759–770. [Google Scholar] [CrossRef] [PubMed]
  48. Addis, D.R.; DeBerry, J.J.; Aggarwal, S. Chronic P—Ain in HIV. Mol. Pain. 2020, 16, 1744806920927276. [Google Scholar] [CrossRef]
  49. Gallagher, P.; O’Mahony, D. Constipation in old age. Best Pract. Res. Clin. Gastroenterol. 2009, 23, 875–887. [Google Scholar] [CrossRef]
  50. Chau, D.L.; Walker, V.; Pai, L.; Cho, L.M. Opiates and elderly: Use and side effects. Clin. Interv. Aging 2008, 3, 273–278. [Google Scholar] [CrossRef] [PubMed]
  51. Edelman, E.J.; Gordon, K.; Becker, W.C.; Goulet, J.L.; Skanderson, M.; Gaither, J.R.; Brennan Braden, J.; Gordon, A.J.; Kerns, R.D.; Justice, A.C.; et al. Receipt of Opioid Analgesics by HIV-Infected and Uninfected Patients. J. Gen. Intern. Med. 2013, 28, 82–90. [Google Scholar] [CrossRef]
  52. Prescription Opioid Trends in the United States. Available online: https://www.iqvia.com/insights/the-iqvia-institute/reports-and-publications/reports/prescription-opioid-trends-in-the-united-states (accessed on 4 April 2025).
  53. Campbell, N.; Boustani, M.; Limbil, T.; Ott, C.; Fox, C.; Maidment, I.; Schubert, C.C.; Munger, S.; Fick, D.; Miller, D.; et al. The cognitive impact of anticholinergics: A clinical review. CIA 2009, 4, 225–233. [Google Scholar] [CrossRef]
  54. Dagli, R.J.; Sharma, A. Polypharmacy: A Global Risk Factor for Elderly People. J. Int. Oral Health 2014, 6, i–ii. [Google Scholar]
  55. Tseng, A.; Szadkowski, L.; Walmsley, S.; Salit, I.; Raboud, J. Association of Age With Polypharmacy and Risk of Drug Interactions With Antiretroviral Medications in HIV-Positive Patients. Ann. Pharmacother. 2013, 47, 1429–1439. [Google Scholar] [CrossRef]
  56. Wastesson, J.W.; Morin, L.; Laroche, M.-L.; Johnell, K. How Chronic Is Polypharmacy in Old Age? A Longitudinal Nationwide Cohort Study. J. Am. Geriatr. Soc. 2019, 67, 455–462. [Google Scholar] [CrossRef] [PubMed]
  57. Lai, S.-W.; Liao, K.-F.; Liao, C.-C.; Muo, C.-H.; Liu, C.-S.; Sung, F.-C. Polypharmacy Correlates With Increased Risk for Hip Fracture in the Elderly: A Population-Based Study. Medicine 2010, 89, 295. [Google Scholar] [CrossRef] [PubMed]
  58. Rochon, P.A.; Gurwitz, J.H. Optimising drug treatment for elderly people: The prescribing cascade. BMJ 1997, 315, 1096–1099. [Google Scholar] [CrossRef] [PubMed]
  59. Ziere, G.; Dieleman, J.P.; Hofman, A.; Pols, H.A.P.; Van Der Cammen, T.J.M.; Stricker, B.H.C. Polypharmacy and falls in the middle age and elderly population. Br. J. Clin. Pharmacol. 2006, 61, 218–223. [Google Scholar] [CrossRef]
  60. Lesko, C.R.; Moore, R.D.; Tong, W.; Lau, B. Association of injection drug use with incidence of HIV-associated non-AIDS-related morbidity by age, 1995–2014. AIDS 2016, 30, 1447–1455. [Google Scholar] [CrossRef]
  61. Tsao, J.C.I.; Plankey, M.W.; Young, M.A. Pain, psychological symptoms and prescription drug misuse in HIV: A literature review. J. Pain. Manag. 2012, 5, 111–118. [Google Scholar]
  62. Holtzman, C.; Armon, C.; Tedaldi, E.; Chmiel, J.S.; Buchacz, K.; Wood, K.; Brooks, J.T. Polypharmacy and Risk of Antiretroviral Drug Interactions Among the Aging HIV-Infected Population. J. Gen. Intern. Med. 2013, 28, 1302–1310. [Google Scholar] [CrossRef]
  63. Assari, S.; Wisseh, C.; Bazargan, M. Obesity and Polypharmacy among African American Older Adults: Gender as the Moderator and Multimorbidity as the Mediator. Int. J. Environ. Res. Public Health 2019, 16, 2181. [Google Scholar] [CrossRef]
  64. Bazargan, M.; Smith, J.; Saqib, M.; Helmi, H.; Assari, S. Associations between Polypharmacy, Self-Rated Health, and Depression in African American Older Adults; Mediators and Moderators. Int. J. Environ. Res. Public Health 2019, 16, 1574. [Google Scholar] [CrossRef]
  65. Trevisan, C.; Limongi, F.; Siviero, P.; Noale, M.; Cignarella, A.; Manzato, E.; Sergi, G.; Maggi, S. Mild polypharmacy and MCI progression in older adults: The mediation effect of drug–drug interactions. Aging Clin. Exp. Res. 2021, 33, 49–56. [Google Scholar] [CrossRef]
  66. Bose, S.; Kelly, L.; Shahn, Z.; Novack, L.; Banner-Goodspeed, V.; Subramaniam, B. Sedative polypharmacy mediates the effect of mechanical ventilation on delirium in critically ill COVID-19 patients: A retrospective cohort study. Acta Anaesthesiol. Scand. 2022, 66, 1099–1106. [Google Scholar] [CrossRef]
  67. Weng, Y.-A.; Deng, C.-Y.; Pu, C. Targeting continuity of care and polypharmacy to reduce drug–drug interaction. Sci. Rep. 2020, 10, 21279. [Google Scholar] [CrossRef]
  68. López-Centeno, B.; Badenes-Olmedo, C.; Mataix-Sanjuan, A.; Bellón, J.; Pérez-Latorre, L.; López, J.; Benedí, J.; Khoo, S.; Marzolini, C.; Calvo-Alcántara, M.; et al. Potentially inappropriate medications in older adults living with HIV. HIV Med. 2020, 21, 541–546. [Google Scholar] [CrossRef]
  69. Parodi López, N.; Svensson, S.A.; Wallerstedt, S.M. Clinical relevance of potentially inappropriate medications and potential prescribing omissions according to explicit criteria—A validation study. Eur. J. Clin. Pharmacol. 2022, 78, 1331–1339. [Google Scholar] [CrossRef]
  70. Lönnbro, J.; Wallerstedt, S.M. Clinical relevance of the STOPP/START criteria in hip fracture patients. Eur. J. Clin. Pharmacol. 2017, 73, 499–505. [Google Scholar] [CrossRef]
  71. Sergi, G.; Rui, M.D.; Sarti, S.; Manzato, E. Polypharmacy in the Elderly. Drugs Aging 2011, 28, 509–518. [Google Scholar] [CrossRef]
  72. Kaur, S.; Mitchell, G.; Vitetta, L.; Roberts, M.S. Interventions that can Reduce Inappropriate Prescribing in the Elderly. Drugs Aging 2009, 26, 1013–1028. [Google Scholar] [CrossRef]
  73. Rognstad, S.; Brekke, M.; Fetveit, A.; Dalen, I.; Straand, J. Prescription peer academic detailing to reduce inappropriate prescribing for older patients: A cluster randomised controlled trial. Br. J. Gen. Pract. 2013, 63, e554–e562. [Google Scholar] [CrossRef]
  74. Clyne, B.; Fitzgerald, C.; Quinlan, A.; Hardy, C.; Galvin, R.; Fahey, T.; Smith, S.M. Interventions to Address Potentially Inappropriate Prescribing in Community-Dwelling Older Adults: A Systematic Review of Randomized Controlled Trials. J. Am. Geriatr. Soc. 2016, 64, 1210–1222. [Google Scholar] [CrossRef]
  75. McDonagh, M.S.; Peterson, K.; Winthrop, K.; Cantor, A.; Lazur, B.H.; Buckley, D.I. Interventions to reduce inappropriate prescribing of antibiotics for acute respiratory tract infections: Summary and update of a systematic review. J. Int. Med. Res. 2018, 46, 3337–3357. [Google Scholar] [CrossRef]
  76. Tamblyn, R.; Huang, A.; Perreault, R.; Jacques, A.; Roy, D.; Hanley, J.; McLeod, P.; Laprise, R. The medical office of the 21st century (MOXXI): Effectiveness of computerized decision-making support in reducing inappropriate prescribing in primary care. CMAJ 2003, 169, 549–556. [Google Scholar]
  77. Clyne, B.; Bradley, M.C.; Hughes, C.M.; Clear, D.; McDonnell, R.; Williams, D.; Fahey, T.; Smith, S.M. Addressing potentially inappropriate prescribing in older patients: Development and pilot study of an intervention in primary care (the OPTI-SCRIPT study). BMC Health Serv. Res. 2013, 13, 307. [Google Scholar] [CrossRef]
Pharmacoepidemiology 04 00010 g001
Figure 1. Proposed relationship between participant characteristics, PIP, symptom burden, and quality of life. Separate models were evaluated for each quality-of-life factor: anxiety/depression, problems with usual activities, problems with mobility, problems with self-care, and pain/discomfort.
Figure 1. Proposed relationship between participant characteristics, PIP, symptom burden, and quality of life. Separate models were evaluated for each quality-of-life factor: anxiety/depression, problems with usual activities, problems with mobility, problems with self-care, and pain/discomfort.
Pharmacoepidemiology 04 00010 g001
Pharmacoepidemiology 04 00010 g002
Figure 2. Overview of the statistical relationship between potentially inappropriate prescriptions, symptom burden, and quality of life. All models were adjusted for age, gender, and race/ethnicity. Each quality-of-life factor model was shown to have the general relationship with PIP and symptom burden. See Table 4 for details on regression estimates.
Figure 2. Overview of the statistical relationship between potentially inappropriate prescriptions, symptom burden, and quality of life. All models were adjusted for age, gender, and race/ethnicity. Each quality-of-life factor model was shown to have the general relationship with PIP and symptom burden. See Table 4 for details on regression estimates.
Pharmacoepidemiology 04 00010 g002
Table 1. Overview of participant demographics, stratified by those who met at least one STOPP criterion and those who did not. Differences in the distribution of categorical and continuous variables were evaluated using chi-square tests and Wilcoxon rank-sum tests, respectively.
Table 1. Overview of participant demographics, stratified by those who met at least one STOPP criterion and those who did not. Differences in the distribution of categorical and continuous variables were evaluated using chi-square tests and Wilcoxon rank-sum tests, respectively.
DemographicTotal
(N = 1048)		Had a PIP
(N = 486)		Did Not Have a PIP
(N = 562)		p-Value
Age [years] (Mean [St.Dev.])61.26 (7.00)62.03 (6.98)60.60 (6.95)0.0006
Race/Ethnicity <0.0001
Black, non-Hispanic830 (79.20)418 (86.01)412 (73.31)
White, non-Hispanic142 (13.55)37 (7.61)105 (18.68)
Hispanic44 (4.20)16 (3.29)28 (4.98)
Other13 (1.24)2 (0.41)11 (1.96)
Unknown19 (1.81)13 (2.67)6 (1.07)
Gender <0.0001
Male750 (71.56)317 (65.23)433 (77.05)
Female281 (26.81)154 (31.69)127 (22.60)
Transgender: male-to-female16 (1.53)14 (2.88)2 (0.36)
Transgender: female-to-male1 (0.10)1 (0.21)0 (0.00)
HIV Transmission Factor <0.0001
Men who have sex with Men (MSM)383 (36.55)142 (29.22)241 (42.88)
Intravenous Drug Use (IDU)87 (8.30)59 (12.14)28 (4.98)
MSM and IDU11 (1.05)3 (0.62)8 (1.42)
High Risk Heterosexual (HRH)358 (34.16)184 (37.86)174 (30.96)
Perinatal1 (0.10)1 (0.21)0 (0.00)
Other207 (19.75)96 (19.75)111 (19.75)
Unknown1 (0.10)1 (0.21)0 (0.00)
Viral Suppression (<200 copies/mL) a <0.0001
Virally Suppressed961 (91.70)455 (93.62)506 (90.04)
Not Virally Suppressed50 (4.77)28 (5.76)22 (3.91)
Unknown37 (3.53)3 (0.62)34 (6.05)
Type of HIV Care Site <0.0001
Community483 (46.09)261 (53.70)222 (39.50)
Hospital565 (53.91)225 (46.30)340 (60.50)
Social Determinants of Health
Housing Need291 (27.77)148 (30.45)143 (25.44)0.0710
Transportation Need164 (15.65)93 (19.14)71 (12.63)0.0039
Utility Need103 (9.83)49 (10.08)54 (9.61)0.7972
Food Insecurity360 (34.35)191 (39.30)169 (30.07)0.0017
Quality of Life
Pain/Discomfort Problems469 (44.75)254 (52.26)215 (38.26)<0.0001
Mobility Problems265 (25.29)151 (31.07)114 (20.28)<0.0001
Self-Care Problems74 (7.06)46 (9.47)28 (4.98)0.0047
Usual Activities Problems213 (20.32)127 (26.13)86 (15.30)<0.0001
Anxiety/Depression Problems356 (33.97)187 (38.48)169 (30.07)0.0042
Symptom Burden Score (Mean [St. Dev.])
Psychological Symptoms0.79 (0.94)0.86 (0.95)0.73 (0.92)0.0094
Physical Symptoms0.38 (0.48)0.48 (0.51)0.30 (0.43)<0.0001
General Distress Index0.66 (0.72)0.76 (0.75)0.57 (0.68)<0.0001
PIP = Potentially inappropriate prescription; a Based on the most recent viral load lab prior to completion of the PROs survey.
Table 2. Number of participants who had PIP within each category (N = 1048).
Table 2. Number of participants who had PIP within each category (N = 1048).
SystemN (%)
Musculoskeletal System245 (23.38)
Analgesic Drugs172 (16.41)
Central Nervous System140 (13.36)
Coagulation System88 (8.40)
Antimuscarinic/Anticholinergic Drugs71 (6.77)
Cardiovascular System57 (5.44)
Endocrine System43 (4.10)
Fall Risk Increase35 (3.34)
Respiratory System11 (1.05)
Urogenital System6 (0.57)
Gastrointestinal System6 (0.57)
Renal System0 (0.00)
Any STOPP Criteria486 (46.37)
Table 3. Unadjusted and adjusted incidence rate ratios evaluating patient characteristics associated with number of PIPs.
Table 3. Unadjusted and adjusted incidence rate ratios evaluating patient characteristics associated with number of PIPs.
DemographicCrude ModelModel I bModel II c
Unadjusted IRR
(95% CI)		Adjusted IRR
(95% CI)		Adjusted IRR
(95% CI)
Age1.03 (1.02, 1.04)1.03 (1.02, 1.04)1.03 (1.02, 1.05)
Race/Ethnicity
Black, non-HispanicREFREFREF
Hispanic0.67 (0.41, 1.09)0.77 (0.47, 1.25)0.70 (0.40, 1.21)
White, non-Hispanic0.61 (0.45, 0.81)0.67 (0.50, 0.92)0.62 (0.44, 0.88)
Other/Unknown0.78 (0.45, 1.35)0.78 (0.45, 1.33)0.79 (0.42, 1.48)
Gender Identity
MaleREFREFREF
Female1.20 (0.98, 1.48)1.05 (0.83, 1.33)1.07 (0.82, 1.41)
Transgender1.68 (0.86, 3.31)1.66 (0.86, 3.20)1.59 (0.68, 3.76)
HIV Transmission Factor
Men who have sex with Men (MSM) aREFREFREF
High-Risk Heterosexual (HRH)1.39 (1.12 1.73)1.12 (0.86, 1.45)1.03 (0.77, 1.38)
Intravenous Drug Use (IDU)2.18 (1.58, 3.02)1.68 (1.20, 2.35)1.67 (1.15, 2.41)
Other/Unknown1.18 (0.91, 1.53)0.99 (0.76, 1.31)0.96 (0.71, 1.31)
Viral Suppression (<200 copies/mL)
Not Virally SuppressedREFREFREF
Virally Suppressed0.94 (0.62, 1.43)0.96 (0.63, 1.44)1.16 (0.71, 1.88)
Unknown0.12 (0.04, 0.33)0.15 (0.05, 0.42)0.19 (0.07, 0.56)
Site Type
CommunityREFREFREF
Hospital0.77 (0.64, 0.93)0.75 (0.62, 0.92)0.85 (0.68, 1.07)
Social Determinants of Health
Housing Need1.20 (0.97, 1.49)-1.19 (0.95, 1.48)
Food Insecurity1.28 (1.04, 1.57)-1.16 (0.91, 1.48)
Transportation Need1.26 (0.97, 1.63)-1.19 (0.90, 1.57)
Utility Need0.97 (0.70, 1.33)-0.84 (0.61, 1.16)
IRR = Incidence Rate Ratio; a Includes participants who are MSM and IDU; b Adjusted for age, race/ethnicity, gender, HIV transmission factor, viral suppression, and site type; c Adjusted for age, race/ethnicity, gender, HIV transmission factor, viral suppression, site type, and social determinants of health
Table 4. Evaluating whether GDI mediates the relationship between PIP and QOL using the bootstrapping method (iterations: 2000). Each model was adjusted for age, gender, and race/ethnicity. Model is just identified so fit statistics cannot be reported.
Table 4. Evaluating whether GDI mediates the relationship between PIP and QOL using the bootstrapping method (iterations: 2000). Each model was adjusted for age, gender, and race/ethnicity. Model is just identified so fit statistics cannot be reported.
Mediation of the relationship between PIP and having problems with anxiety/depression.
PathModel VariablesRegression Estimates (95% CI)
Path APIP → GDI0.08 (0.04, 0.11) A
Path BGDI → Anxiety/Depression18.19 (12.16, 27.19) B,C
Path CPIP → Anxiety/Depression1.16 (1.06, 1.28) B,D
Mediated PathPIP → GDI → Anxiety/Depression0.22 (0.11, 0.33) E
Mediation of the relationship between PIP and having problems with usual activities.
PathModel VariablesRegression Estimates (95% CI)
Path APIP → GDI0.08 (0.04, 0.11) A
Path BGDI → Usual Activities4.64 (3.56, 6.05) B,C
Path CPIP → Usual Activities1.28 (1.16, 1.42) B,D
Mediated PathPIP → GDI → Usual Activities0.12 (0.06, 0.18) E
Mediation of the relationship between PIP and mobility problems.
PathModel VariablesRegression Estimates (95% CI)
Path APIP → GDI0.08 (0.04, 0.11) A
Path BGDI → Mobility3.45 (2.72, 4.37) B,C
Path CPIP → Mobility3.31 (2.61, 4.20) B,D
Mediated PathPIP → GDI → Mobility0.09 (0.05, 0.14) E
Mediation of the relationship between PIP and self-care problems.
PathModel VariablesRegression Estimates (95% CI)
Path APIP → GDI0.08 (0.04, 0.11) A
Path BGDI → Self-Care4.21 (3.00, 5.91) B,C
Path CPIP → Self-Care1.44 (1.09, 1.45) B,D
Mediated PathPIP → GDI → Self-Care0.11 (0.06, 0.18) E
Mediation of the relationship between PIP and pain and discomfort.
PathModel VariablesRegression Estimates (95% CI)
Path APIP → GDI0.08 (0.04, 0.11) A
Path BGDI → Pain/Discomfort5.15 (3.83, 6.92) B,C
Path CPIP → Pain/Discomfort1.26 (1.13, 1.40) B,D
Mediated PathPIP → GDI → Pain/Discomfort0.12 (0.06, 0.19) E
A Regression estimate is based on a linear regression; B Regression estimate represents an odds ratio from logistic regression; C Adjusted for potentially inappropriate prescriptions; D Adjusted for GDI; E Estimate is the indirect effect based on a combination of linear and logistic regression.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

O’Connor, L.F.; Resnik, J.B.; Simmens, S.; Bhandaru, V.; Benator, D.; Wingate, L.; Castel, A.D.; Monroe, A.K., on behalf of the DC Cohort Executive Committee. Evaluation of Screening Tool of Older People’s Prescriptions (Stopp) Criteria in an Urban Cohort of Older People with HIV. Pharmacoepidemiology 2025, 4, 10. https://doi.org/10.3390/pharma4020010

AMA Style

O’Connor LF, Resnik JB, Simmens S, Bhandaru V, Benator D, Wingate L, Castel AD, Monroe AK on behalf of the DC Cohort Executive Committee. Evaluation of Screening Tool of Older People’s Prescriptions (Stopp) Criteria in an Urban Cohort of Older People with HIV. Pharmacoepidemiology. 2025; 4(2):10. https://doi.org/10.3390/pharma4020010

Chicago/Turabian Style

O’Connor, Lauren F., Jenna B. Resnik, Sam Simmens, Vinay Bhandaru, Debra Benator, La’Marcus Wingate, Amanda D. Castel, and Anne K. Monroe on behalf of the DC Cohort Executive Committee. 2025. "Evaluation of Screening Tool of Older People’s Prescriptions (Stopp) Criteria in an Urban Cohort of Older People with HIV" Pharmacoepidemiology 4, no. 2: 10. https://doi.org/10.3390/pharma4020010

APA Style

O’Connor, L. F., Resnik, J. B., Simmens, S., Bhandaru, V., Benator, D., Wingate, L., Castel, A. D., & Monroe, A. K., on behalf of the DC Cohort Executive Committee. (2025). Evaluation of Screening Tool of Older People’s Prescriptions (Stopp) Criteria in an Urban Cohort of Older People with HIV. Pharmacoepidemiology, 4(2), 10. https://doi.org/10.3390/pharma4020010

Article Metrics

Back to TopTop