Skip to main content
NCBI home page
Journal of the American Medical Informatics Association: JAMIA logoLink to Journal of the American Medical Informatics Association: JAMIA
. 2024 Jul 31;31(12):2924–2931. doi: 10.1093/jamia/ocae203

Sex-based disparities with cost-related medication adherence issues in patients with hypertension, ischemic heart disease, and heart failure

Ivann Agapito 1, Tu Hoang 2, Michael Sayer 3, Ali Naqvi 4, Pranav M Patel 5, Aya F Ozaki 6,✉
PMCID: PMC11631144  PMID: 39083847

Abstract

Importance and Objective

Identifying sources of sex-based disparities is the first step in improving clinical outcomes for female patients. Using All of Us data, we examined the association of biological sex with cost-related medication adherence (CRMA) issues in patients with cardiovascular comorbidities.

Materials and Methods

Retrospective data collection identified the following patients: 18 and older, completing personal medical history surveys, having hypertension (HTN), ischemic heart disease (IHD), or heart failure (HF) with medication use history consistent with these diagnoses. Implementing univariable and adjusted logistic regression, we assessed the influence of biological sex on 7 different patient-reported CRMA outcomes within HTN, IHD, and HF patients.

Results

Our study created cohorts of HTN (n = 3891), IHD (n = 5373), and HF (n = 2151) patients having CRMA outcomes data. Within each cohort, females were significantly more likely to report various cost-related medication issues: being unable to afford medications (HTN hazards ratio [HR]: 1.68, confidence interval [CI]: 1.33-2.13; IHD HR: 2.33, CI: 1.72-3.16; HF HR: 1.82, CI: 1.22-2.71), skipping doses (HTN HR: 1.76, CI: 1.30-2.39; IHD HR: 2.37, CI: 1.69-3.64; HF HR: 3.15, CI: 1.87-5.31), taking less medication (HTN HR: 1.86, CI: 1.37-2.45; IHD HR: 2.22, CI: 1.53-3.22; HF HR: 2.99, CI: 1.78-5.02), delaying filling prescriptions (HTN HR: 1.83, CI: 1.43-2.39; IHD HR: 2.02, CI: 1.48-2.77; HF HR: 2.99, CI: 1.79-5.03), and asking for lower cost medications (HTN HR: 1.41, CI: 1.16-1.72; IHD HR: 1.75, CI: 1.37-2.22; HF HR: 1.61, CI: 1.14-2.27).

Discussion and Conclusion

Our results clearly demonstrate CRMA issues disproportionately affect female patients with cardiovascular comorbidities, which may contribute to the larger sex-based disparities in cardiovascular care. These findings call for targeted interventions and strategies to address these disparities and ensure equitable access to cardiovascular medications and care for all patients.

Keywords: cardiovascular health, sex-based disparities, medication adherence, logistic regression modeling, All of Us data

Introduction

Within the United States, one of the critical barriers to patient compliance to medication regimens for chronic cardiovascular conditions is cost. Multiple nationally conducted surveys showed expenses related to medicines being a significant barrier to adherence for at least 10% of patients with atherosclerotic, hypertensive, and heart failure related cardiac comorbidities.1–3 Paradoxically, multiple studies evaluating medication costs to treat cardiac comorbidities suggest strict medication adherence would allow for net savings for the healthcare system, as these regimens significantly reduce costly hospitalizations and procedures.4–6 More importantly, adherence is proven to lead to a significantly improved quality of life and longer life spans for patients.7–9

Medication adherence issues in chronic cardiovascular diseases are particularly challenging. Non-adherent patients with diseases like hypertension (HTN), ischemic heart disease (IHD), and heart failure (HF) often do not feel immediate symptomatic consequences of not taking medications.10,11 Patients also face excessive pill burden in managing chronic cardiovascular disease, often being prescribed several medications creating time and cost nuisances.12–14 In patients facing financial struggles, a lack of immediate health consequences and polypharmacy concerns may justify a lack of compliance and lead to cost-related medication adherence (CRMA) issues. These issues manifest in a variety of ways including skipping or reducing doses without clinician supervision, asking less-costly sub-optimal medications, purchasing medications from a different country, or delaying or not even picking up medications altogether.15–17

Women are potentially more vulnerable to CRMA barriers in cardiovascular disease management. Sex-based disparities have been observed in other elements of cardiovascular care, with underrepresentation of females in relevant clinical trials. Additionally, female patients are significantly less likely to receive appropriate referrals to specialists and appropriate testing, and their unique biological factors lead to uncommon presentations of a variety of conditions.18–21 Multiple studies have demonstrated that female patients face significantly increased cumulative healthcare expenses relative to men, especially among younger populations and ethnic minorities.22–24 It is important to investigate sex-based disparities with cost based medication adherence issues to provide additional understanding of the challenges women face in cardiovascular care.

The All of Us Research Program has recruited thousands of patients representing diverse socioeconomic statuses, many of which filled out surveys relevant to CRMA issues as part of study participation.25 To investigate potential sex-based disparities in CRMA in patients with cardiovascular comorbidities, we collected distinct cohorts of patients with IHD, HTN, and HF within the All of Us researcher workbench. To demonstrate the extent biological sex influences the likelihood patients face different CRMA issues, univariable and multivariable logistic regression models adjusted for clinically relevant confounders were implemented predicting relevant medication adherence outcomes. Our large diverse patient cohorts representing common cardiovascular comorbidities allow our study to provide unique insight into sex-based disparities for CRMA burdens.

Methods

Study design, data sources, and inclusion criteria

This study is a cross-sectional study using data from All of US researcher workbench. Patient data were collected in July of 2022, considering patient records dated from January 2018 to June 30, 2022. Patients were included in our analysis if they meet the following criteria: ≥18 years old, having a relevant diagnosis (HTN, IHD, or HF), having medication use consistent with treatment within 90 days of diagnosis, and having completed CRMA questionnaires. Patients were excluded when missing relevant survey data (Figure 1). We chose to exclude them because they were missing most of the needed information for downstream analysis; imputation or creation of “missing” categories would lead to many entries having mostly missing or imputed data. Electronic health records of potential patients were queried using ICD-9 and ICD-10 codes and relevant past medical history survey responses with the All of Us researcher workbench to identify patients with relevant diagnoses (Table S1).26 Medications relevant to cardiovascular comorbidities were identified based on clinical guidelines and clinician guidance (Table S2). Sources of data, how they were accessed within the All of Us researcher workbench, and which information was utilized are detailed in Table S3.

Figure 1.

This figure represents a flow chart describing how study populations were created from the All of Us research work bench. It starts with all patients enrolled in the study and ends with the Ischemic heart disease, hypertension, and heart failure patient populations utilized.

Open in a new tab

Retrospective collection of ischemic heart disease, hypertension, and heart failure cohorts from All of Us researcher workbench. Bolded black arrows indicate the stepwise process by which patients were identified, and population descriptions for each step are contained within boxes. The red arrow represents patients removed from the initial population. Within the red box, each listed item represents an exclusion criteria along with the number of relevant patients. Population sizes are described following (n=) notations.

Study cohorts

Three distinct study populations were created and utilized in subsequent analysis steps, with each population representing a distinct cardiovascular comorbidity including HTN, IHD, and HF. Patients were included in each population if they had a history of relevant diagnosis codes in addition to medication use within 90 days of a prospective diagnosis consistent with treatment (Tables S1 and S2).26–29 Patients with IHD or HF were excluded from the HTN group and patients with IHD were excluded from the HF group.

CRMA outcomes

Patient responses to the CRMA survey were the outcome of interest in the study (Table 1 and Table S3). The CRMA surveys consisted of 7 questions related to medication adherence, prompting patients to respond yes or no to each question if they met criteria for the prospective question within the previous 12 months.

Table 1.

All of Us Research Program past medical history survey questions related to cost-related medication adherence (CRMA).

During the past 12 months…
1. Was there any time when you needed prescription medicines but did not get it because you could not afford it?
2. Did you skip medication doses to save money?
3. Did you take less medicine to save money?
4. Did you delay filling a prescription to save money?
5. Did you ask your doctor for a lower cost medication to save money?
6. Did you buy prescription drugs from another country to save money?
7. Did you use alternative therapies to save money?
Open in a new tab

These are the 7 CRMA survey questions utilized as outcomes for this study, filled out by patients reflecting their CRMA behaviors over the previous year. Question numbers referenced within the paper are consistent with the numbering given here.

Additional patient data

Patient demographic information utilized included sex at birth, age at time of CRMA survey, race, and ethnicity. Lifestyle factors considered included highest attained level of education, household income, smoking history, and alcohol/drug use. Factors related to patient medical history included comorbid conditions, body mass index (BMI), past medication use, and Charlson Comorbidity Index (Table S3).26

Statistical analysis

Descriptive statistics were utilized to describe our patient populations of interest. Chi-square tests of independence and t-tests were performed to assess for the distribution of patient demographic features across biological sex, using a threshold of significance of P < .05. Utilizing CRMA survey questions as outcomes, univariable logistic regression was performed to evaluate the association of sex with each or the 7 prospective outcomes. Multivariable logistic regression models used CRMA survey responses as outcomes and adjusted for many factors along with biological sex. Adjusted models included income status, race, ethnicity, age range, BMI, Charlson Comorbidity Index, past medication use, comorbidities, education status, smoking history, and drug/alcohol use (detailed in Tables S4 and S5). Effect sizes were presented as odds ratios for univariable models (OR), and adjusted odds ratios (AOR) for multivariable models, with each having a reported 95% confidence interval (CI). Logistic regression models with odds ratios were utilized due to the rarer nature of some of the CRMA response outcomes observed, as opposed to a prevalence ratio approach. Initial significance was assessed using the Wald Test within logistic regression models, with an initial significance threshold P-value of .05 to determine if biological sex was a significant predictor of CRMA outcomes. To account for multiple testing (21 regression models), a Bonferroni corrected P-value threshold of .00238 was also considered for adjusted multivariable models.

In total, descriptive statistics were captured and univariable and multivariable logistic regression analyses predicting the 7 CRMA outcomes were performed within each of the 3 identified patient cohorts of HTN, IHD, and HF patients (21 distinct analyses reporting OR and AOR). Data were accessed with Google BigQuery and analyzed using R v4.1.2 (Boston, MA) in an integrated Jupyter Notebook environment.30–32 All data and results herein are presented in compliance with the All of Us Data and Statistics Dissemination Policy, and no IRB approval was needed for this study.

Results

Baseline characteristics

Initial query of the All of Us researcher workbench identified 123 063 patients with relevant diagnoses of HTN, HF, or IHD. From this population, 9875 patients met inclusion/exclusion criteria featuring HTN 5373 (57.2% women) patients, 3891 (58.9% women) HF patients, and 2151 (44.1% women) IHD patients (Figure 1). The median (IQR) age of men with HTN, HF, and IHD are 60 (50,69), 65 (56,74), and 67 (58,74) years, respectively. The median (IQR) age of women with HTN, HF, and IHD are 58 (47,67), 64 (54,72), and 64 (55,72) years, respectively. For both male and female, the majority individuals were between 40 and 65 years old. The distribution of race/ethnicity, income, education, BMI was similar for all the cohorts. The majority of respondents from all cohorts were non-Hispanic White (>60%), followed by non-Hispanic African American (>15%) and reported low income (>30%). Compared to men, a high prevalence of women in each cohort were African American, had lower income, and higher BMI. Reported Chi-square statistics and t-tests assessing distribution of key demographic features between males and females were significant for all patient demographic features evaluated across the 3 cohorts (Table S4).

CRMA questionnaires

Overall, females with HTN, HF, and IHD reported they could not afford medication, skipped medication to save money, took less medication to save money, delayed filling medication, asked for lower cost medication, and used alternative medication to save money at higher rates compared to males. A higher proportion of males in IHD and HF cohorts reported they bought medication from another country compared to females. Some examples of the largest reported differences include 10.0% of females with HTN responded that they could not afford medication compared to 5.1% of males, 7.2% of females with IHD asked for lower cost medication versus 3.7% of males, and for the same question among HF patients, females reported 6.2% versus 2.2% males (Table 2).

Table 2.

Sex based cost-related medication adherence (CRMA) survey responses, odds ratios, and adjusted odds ratios.

Hypertension Male (n = 2296) Female (n = 3077) OR (95% CI) P-value AOR (95% CI) P-value
Could not afford medication, N (%) 117 (5.1) 307 (10.0) 2.05 (1.64, 2.56) <.0001 1.68 (1.33, 2.13) <.0001
Skipped medication to save money, N (%)  65 (2.8) 177 (5.8) 2.08 (1.56, 2.78) <.0001 1.76 (1.30, 2.39) .0003
Took less medication to save money, N (%)  71 (3.1) 198 (6.4) 2.14 (1.63, 2.83) <.0001 1.86 (1.37, 2.45) <.0001
Delayed filling medication, N (%) 95 (4.1) 258 (8.4) 2.09 (1.64, 2.66) <.0001 1.85 (1.43, 2.39) <.0001
Asked for lower cost medication, N (%)  198 (8.6) 379 (12.3) 1.47 (1.23, 1.76) <.0001 1.41 (1.16, 1.72) .0005
Bought medications from another country, N (%)  39 (1.7) 52 (1.7) 0.94 (0.61, 1.43) .76 1.09 (0.69, 1.73) .7098
Used alternative medication to save money, N (%)  40 (1.7) 142 (4.6) 2.69 (1.88, 3.84) <.0001 2.38 (1.63-3.46) <.0001
Ischemic heart disease Male (n = 2255) Female (n = 1631) OR (95% CI) P-value AOR (95% CI) P-value
Could not afford medication, N (%) 86 (3.8) 138 (8.5) 2.38 (1.80, 3.14) <.0001 2.33 (1.72, 3.16) <.0001
Skipped medication to save money, N (%)  51 (2.3) 91 (5.6) 2.55 (1.80, 3.62) <.0001 2.48 (1.69, 3.64) <.0001
Took less medication to save money, N (%)  55 (2.4) 92 (5.6) 2.42 (1.72, 3.40) <.0001 2.22 (1.53, 3.22) <.0001
Delayed filling medication, N (%) 83 (3.7) 118 (7.2) 2.06 (1.54, 2.75) <.0001 2.02 (1.48, 2.77) <.0001
Asked for lower cost medication, N (%)  174 (7.7) 185 (11.3) 1.53 (1.23, 1.90) .0001 1.75 (1.37, 2.22) <.0001
Bought medications from another country, N (%)  40 (1.8) 24 (1.5) 0.85 (0.51, 1.42) .53 1.03 (0.59, 1.81) .91459
Used alternative medication to save money, N (%)  26 (1.2) 45 (2.8) 2.49 (1.53, 4.04) .0003 2.11 (1.25, 3.57) .0052
Heart failure Male (n = 1202) Female (n = 949) OR (95% CI) P-value AOR (95% CI) P-value
Could not afford medication, N (%) 57 (4.7) 84 (8.8) 2.02 (1.43, 2.88) .0002 1.82(1.22, 2.71) .0031
Skipped medication to save money, N (%)  26 (2.2) 62 (6.5) 3.16 (1.98, 5.04) <.0001 3.15 (1.87, 5.31) <.0001
Took less medication to save money, N (%)  27 (2.2) 59 (6.2) 2.88 (1.81, 4.59) <.0001 2.99 (1.78, 5.02) <.0001
Delayed filling medication, N (%) 49 (4.1) 68 (7.1) 1.86 (1.27, 2.71) .002 2.99 (1.79, 5.03) .0246
Asked for lower cost medication, N (%)  85 (7.1) 97 (10.2) 1.52 (1.12, 2.06) .0077 1.61 (1.14, 2.27) .0073
Bought medications from another country, N (%)  20 (1.7) <20 0.60 (0.27, 1.32) .16 0.81 (0.33, 1.96) .6336
Used alternative medication to save money, N (%)  <20 21 (2.2) 2.70 (1.26, 5.76) .01 2.17 (0.92, 5.12) .078
Open in a new tab

From left to right, the columns in the table represent the CRMA outcome of interest, followed count and percentages (in parenthesis) of the total population experiencing each outcome among Males and Females prospectively. Next, the odds ratio for univariable logistic regression models is reported followed by its accompanying P-value. Lastly, the adjusted odds ratio is reported followed by its accompanying P-value as well. The table has 3 sections for hypertension, ischemic heart disease, and heart failure cohorts distinguished by bolded titles and restated column names with accompanying population sizes for males and females.

All counts less than 20 are reported as “<20” without percentages to comply with All of Us guidelines.

Univariable Model: CRMA Outcome∼ Sex at Birth.

Adjusted Models: CRMA Outcome∼ Sex at Birth + Income + Race + Ethnicity + Age Range + BMI + Charlson Comorbidity Index + Education Status + Medication Use History + Comorbidities + Life Style Factors (detailed in Table S4).

Abbreviations: AOR, adjusted odds ratio; IHD, ischemic heart disease; HF, heart failure; HTN, hypertension.

Regression analyses

In univariable logistic regression models, females were significantly more likely than males to report they could not afford medication, skipped medication to save money, took less medication to save money, delayed filling medication, asked for low-cost medication in HTN and IHD, and HF cohorts (Table 2). These relationships were persistent in multivariable evaluations adjusting for clinically relevant covariates considering a significance threshold of 0.05; however, when considering a Bonferroni multiple testing correction threshold (P < .00238), multiple CRMA responses with the HF cohort including Could not afford medication, Delayed filling medication, and Asked for lower cost medication were not significant. In the HTN and IHD cohorts, females were significantly more likely to use alternative medication to save money compared to males in univariable and adjusted models with P-values less than .05; however, the HTN result was the only one that met the Bonferroni corrected significance threshold (Table 2 and Figure 2). Statistical models showed biological sex being insignificant predictor of purchasing medications in another country to save money in all cohorts, and it further showed that men were not significantly more likely to experience any evaluated CRMA issue compared to women according to univariable and multivariable analysis in all cohorts (Table 2 and Figure 2). For adjusted models, odds ratios for covariates utilized and their significance are detailed in Table S5.

Figure 2.

This figure uses points and horizontally plotted lines to represent adjusted odds ratios and their confidence intervals. Each line describes the extent biological sex is predictive of different medication adherence issues in patients with different cardiovascular comorbidities.

Open in a new tab

Adjusted hazard ratios for cost-related medication adherence (CRMA) questions within hypertension, ischemic heart disease, and heart failure cohorts. The Y-axis labels represent each CRMA question, followed by distinct cohort labels. X-axis values are adjusted hazard ratios ranging from 0 to 5. Values >1 represent increased risk for females, while values <1 suggest the same for males. The vertical dotted line represents 1 for reference. Text above each plotted hazard ratio is the following format: hazard ratio (95% confidence interval), P-value.

Many of the multivariable modeling results demonstrated women being at least 1.5 times more likely face various cost-based medication adherence issues compared to men (Table 2 and Figure 2). For HTN patients, females were significantly more likely to report being unable to afford medication (AOR: 1.68, 95%CI: 1.33-2.13), skip medication (AOR: 1.76, 95%CI: 1.30-2.39), take less medication (AOR: 1.86, 95% CI: 1.37-2.45), delay filling medication (AOR: 1.85, 95% CI: 1.43-2.39), and use an alternative medication to save money (AOR: 2.38, 95% CI: 1.63-3.46) even considering adjusted significance thresholds from multiple testing. For IHD patients, females were significantly more likely to report being unable to afford medication (AOR: 2.33, 95%CI: 1.72-3.16), skip medication (AOR: 2.48, 95%CI: 1.69-3.54), take less medication (AOR: 2.22, 95%, CI: 1.53-3.22), delay filling medication (AOR: 2.02, 95% CI: 1.48-2.77), and ask for a lower cost medication (AOR: 1.75, CI: 1.37-2.22) with significance after accounting for multiple testing. For HF patients, females were significantly more likely to report skipping medication (AOR: 3.15, 95% CI: 1.87-5.31) or take less medication (AOR: 2.99, 95% CI: 1.78-5.02) with significance accounting for multiple testing (Table 2, Table S5A-C, and Figure 2).

Discussion

Utilizing the All of Us researcher platform, implementation of inclusion/exclusion criteria in our study created 3 cohorts featuring thousands of patients with HTN, IHD, and HF. Each of these populations had unique distributions of patient demographic features, including age, race, income, education status, BMI, and Charlson Comorbidity Index. Preliminary evaluation of the percentage of females experiencing cost-based medication adherence issues showed they reported them at a higher rate relative to males with most CRMA questions across all 3 patient cohorts. These reported differences were further demonstrated in univariable and multivariable logistic regression analyses, with mostly significant outcomes showing females significantly more likely to experience CRMA issues even in context with clinically relevant covariates. Our study adds compelling evidence to a growing body of literature suggesting females face significant disparities in cardiovascular care relative to male counterparts.

Our study contributes to a growing body of evidence demonstrating health disparities women face with respect to cost-related barriers. One study utilizing National Financial Capability Study (NFCS) data found that that women reported higher cost-related non-adherence (CRN) rates after controlling other factors such as age, household income, work status, and family status among US adult population.33 Another study using US National Health Interview Survey that female patients with diabetes were more likely to skip the medication doses, delay filling their prescriptions, and take less medication compared to male counterparts as a cost-saving measure.34 Similarly, female individuals with cardiovascular disease were more likely to have any types of cost-related barriers in 2 studies published using National Veterans Affairs (VA) electronic medical record (EMR) data and US National Health Interview Survey (NHIS).35,36 Utilizing the latest data from the All of Us Research Program spanning 4 years (2018-2022), our results demonstrate that sex-based disparities in cost-related barriers in medication adherence persist.

There are many factors that contribute to CRMA in female individuals. One study concluded that gender difference plays a role in preferences and attitudes toward medical spending over other living expenses after control for both insurance coverage and poverty status.36 The reasons for these sex-based disparities are challenging to characterize, but some have suggested different social roles, unique psychological distress, and various social determinants of health may have distinct influences in females leading to these issues.33 While our study data focus on biological sex, gender identity itself has associations with poorer observed outcomes and reduced access to healthcare.18 Even with a global perspective, one study showed that among high-income countries, women from the United States were significantly more likely to experience issues paying medical bills compared to other countries.37 Future studies can explore subpopulations of women to see if distinct age groups, ethnic backgrounds, or regional influences further examine these disparities.

There are many factors that influence the likelihood of patients having CRMA issues, many of which are more likely to impact on women. Multiple studies have demonstrated lower yearly income being a significant predictor of medication non-adherence, and women often report lower annual wages relative to men.38,39 Cumulative out-of-pocket expenses can have a direct impact on cost issues as well. While men and women are insured at similar rates, women are generally more likely to pursue a variety of different health services in general leading to much higher cumulative out-of-pocket expenses.40 These general trends persist in Medicare populations, where women were shown to have significantly lower savings and higher out of pocket expenses.41,42 These cumulative expenses on other needs may lead to reduced willingness to pay for medications treating chronic cardiovascular comorbidities.

These issues may likely be compounded by rising medication costs in general. An evaluation of the pricing common brand-name prescription drugs over a 5-year span demonstrated cumulative median price increase of 76%, with specific out-of-pocket costs for patients more than doubling as well for 78% of medications.43 While generic drugs can be a cost-effective alternative, even their pricing may be inflated and rising due to multiple intermediaries receiving payments with each pharmacy claim.44 These costs also may disproportionately effect women, much like the cumulative higher out-of-pocket expenses experienced with increased participation in different health services. Multiple studies have demonstrated that women were significantly more likely to experience polypharmacy issues and have more active prescriptions relative to men.45,46 While these inequities are challenging to address, awareness of these financial toxicities by providers can lead to implementation of more cost-effective interventions.

Medication non-adherence in cardiovascular care has real consequences. Multiple studies have shown a lack of adherence in hypertension management leads to an increased risk of serious cardiovascular events, and significant reductions in risk based on tiered increases in observed medication adherence rates.47–49 Hospitalizations related to congestive heart failure have similar trends, with medication adherent patients significantly less likely to experience debilitating symptomatic incidences requiring hospitalization.50,51 The impact of adherence with preventative statin therapy in ischemic heart disease is well established as well. Mortality risk and myocardial infarction incidence have strong associations with adherence to prescribed statin regimens.52 Given the obvious implications of medication non-adherence with cardiovascular comorbidities, and the poorer outcomes women repeatedly face in cardiovascular care, addressing cost-related adherence issues may be vital.

Our study has several limitations to address. Utilization of CRMA surveys to assess medication adherence has limitations, as perceived experiences or willingness to admit adherence issues may differ between sexes. CRMA surveys are also not specific to just cardiovascular care, so the results may not just describe adherence to specific HTN, HF, or IHD medications. We also acknowledge that biological sex and gender identities may play distinct roles in medication adherence issues. Future studies can better capture how patients with distinct gender identities relative to biological sex experience medication adherence issues. Implemented exclusion criteria removed a large portion of the initial population with HTN, HF, and IHD. The exclusion of patients not having a relevant medication use represented a large majority of the sample excluded; this was to ensure our sample represented patients receiving medications for these conditions. Additionally, patients were excluded due to missing survey information with the All of Us researcher workbench. Alternative approaches, such as data imputation or assigning values for missing information, would have resulted in numerous entries with mostly missing data limiting the interpretability of our findings. These exclusions can certainly be as source of selection bias; more evaluations of adherence issues with different databases may further validate our findings.

This study does have many strengths, which will make it a much-needed contribution to literature addressing health disparities. We have collected large cohorts of thousands of patients, representing diverse ethnic backgrounds, age ranges, education levels, and income statuses. Our statistical methodologies consider these confounding factors with respect to CRMA issues to demonstrate clear sex-based disparities that persist. We clearly demonstrate inequities with respect to CRMA issues and biological sex. Given that these findings are consistent with other similar studies, more studies are warranted examining why women face these financial stressors as opposed to men. Additional follow-up may also be warranted to demonstrate how this lack of adherence may directly impact women’s health over time. In the future, this study and others like it can increase awareness in providers and lead to health system reform to more effectively address financial disparities in biological sex.

Supplementary Material

ocae203_Supplementary_Data
ocae203_supplementary_data.docx (69.6KB, docx)

Acknowledgments

The All of Us Research Program is supported by the National Institutes of Health, Office of the Director: Regional Medical Centers: 1 OT2 OD026549; 1 OT2 OD026554; 1 OT2 OD026557; 1 OT2 OD026556; 1 OT2 OD026550; 1 OT2 OD 026552; 1 OT2 OD026553; 1 OT2 OD026548; 1 OT2 OD026551; 1 OT2 OD026555; IAA #: AOD 16037; Federally Qualified Health Centers: HHSN 263201600085U; Data and Research Center: 5 U2C OD023196; Biobank: 1 U24 OD023121; The Participant Center: U24 OD023176; Participant Technology Systems Center: 1 U24 OD023163; Communications and Engagement: 3 OT2 OD023205; 3 OT2 OD023206; and Community Partners: 1 OT2 OD025277; 3 OT2 OD025315; 1 OT2 OD025337; 1 OT2 OD025276. In addition, the All of Us Research Program would not be possible without the partnership of its participants.

Contributor Information

Ivann Agapito, School of Pharmacy & Pharmaceutical Sciences, Department of Clinical Pharmacy Practice, University of California Irvine, Irvine, CA 92697, United States.

Tu Hoang, School of Pharmacy & Pharmaceutical Sciences, Department of Clinical Pharmacy Practice, University of California Irvine, Irvine, CA 92697, United States.

Michael Sayer, School of Pharmacy & Pharmaceutical Sciences, Department of Clinical Pharmacy Practice, University of California Irvine, Irvine, CA 92697, United States.

Ali Naqvi, Division of Cardiology, Department of Medicine, University of California Irvine, Irvine, CA 92697, United States.

Pranav M Patel, Division of Cardiology, Department of Medicine, University of California Irvine, Irvine, CA 92697, United States.

Aya F Ozaki, School of Pharmacy & Pharmaceutical Sciences, Department of Clinical Pharmacy Practice, University of California Irvine, Irvine, CA 92697, United States.

Author contributions

Ivann Agapito was involved with conceptualization, designing, data curation, analysis, visualization, and writing draft, review, and editing. Tu Hoang was involved with conceptualization, designing, and writing draft, review, and editing. Michael Sayer was involved with analysis, visualization, interpretation of the results, and writing draft, review, and editing. Ali Naqvi and Pranav M. Patel provided supervision and were involved with the interpretation of the results and provided critical feedback to the manuscript. Aya F. Ozaki supervised and lead the study and was involved with conceptualization, designing, and writing draft, review, and editing.

Supplementary material

Supplementary material is available at Journal of the American Medical Informatics Association online.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflicts of interest

None declared.

Data availability

The data underlying this study were provided by All of Us Research Program by permission. Their policy states “authorized data users will not publish or otherwise distribute any participant-level data from the All of Us Research Program database.” Efforts were made in the manuscript to describe where data was accessed within the platform for interested parties.

References

  • 1. Khera R, Valero-Elizondo J, Das SR, et al.  Cost-related medication nonadherence in adults with atherosclerotic cardiovascular disease in the United States, 2013 to 2017. Circulation. 2019;140(25):2067-2075. 10.1161/circulationaha.119.041974 [DOI] [PubMed] [Google Scholar]
  • 2. Tajeu GS, Muntner P.  Cost-related antihypertensive medication nonadherence: action in the time of COVID-19 and beyond. Am J Hypertens. 2020;33(9):816-818. 10.1093/ajh/hpaa085 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Sukumar S, Wasfy JH, Januzzi JL, Peppercorn J, Chino F, Warraich HJ.  Financial toxicity of medical management of heart failure: JACC review topic of the week. J Am Coll Cardiol. 2023;81(20):2043-2055. 10.1016/j.jacc.2023.03.402 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Bansilal S, Castellano JM, Garrido E, et al.  Assessing the impact of medication adherence on long-term cardiovascular outcomes. J Am Coll Cardiol. 2016;68(8):789-801. 10.1016/j.jacc.2016.06.005 [DOI] [PubMed] [Google Scholar]
  • 5. Jacob V, Reynolds JA, Chattopadhyay SK, Community Preventive Services Task Force (CPSTF), et al.  Pharmacist interventions for medication adherence: community guide economic reviews for cardiovascular disease. Am J Prev Med. 2022;62(3):e202-e222. 10.1016/j.amepre.2021.08.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Axon DR, Vaffis S, Chinthammit C, et al.  Assessing the association between medication adherence, as defined in quality measures, and disease-state control, health care utilization, and costs in a retrospective database analysis of Medicare supplemental beneficiaries using statin medications. J Manag Care Spec Pharm. 2020;26(12):1529-1537. 10.18553/jmcp.2020.26.12.1529 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Shin S, Song H, Oh SK, Choi KE, Kim H, Jang S.  Effect of antihypertensive medication adherence on hospitalization for cardiovascular disease and mortality in hypertensive patients. Hypertens Res. 2013;36(11):1000-1005. 10.1038/hr.2013.85 [DOI] [PubMed] [Google Scholar]
  • 8. Wu JR, Moser DK.  Medication adherence mediates the relationship between heart failure symptoms and cardiac event-free survival in patients with heart failure. J Cardiovasc Nurs. 2018;33(1):40-46. 10.1097/jcn.0000000000000427 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Silavanich V, Nathisuwan S, Phrommintikul A, Permsuwan U.  Relationship of medication adherence and quality of life among heart failure patients. Heart Lung. 2019;48(2):105-110. 10.1016/j.hrtlng.2018.09.009 [DOI] [PubMed] [Google Scholar]
  • 10. Ashoorkhani M, Majdzadeh R, Gholami J, Eftekhar H, Bozorgi A.  Understanding non-adherence to treatment in hypertension: a qualitative study. Int J Community Based Nurs Midwifery. 2018;6(4):314-323. [PMC free article] [PubMed] [Google Scholar]
  • 11. Leslie KH, McCowan C, Pell JP.  Adherence to cardiovascular medication: a review of systematic reviews. J Public Health (Oxf). 2019;41(1):e84-e94. 10.1093/pubmed/fdy088 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Bhatt AS, Choudhry NK.  Evidence-based prescribing and polypharmacy for patients with heart failure. Ann Intern Med. 2021;174(8):1165-1166. 10.7326/m21-1427 [DOI] [PubMed] [Google Scholar]
  • 13. Shaikh F, Pasch LB, Newton PJ, Bajorek BV, Ferguson C.  Addressing multimorbidity and polypharmacy in individuals with atrial fibrillation. Curr Cardiol Rep. 2018;20(5):32. 10.1007/s11886-018-0975-x [DOI] [PubMed] [Google Scholar]
  • 14. Derington CG, Gums TH, Bress AP, SPRINT Research Group, et al.  Association of total medication burden with intensive and standard blood pressure control and clinical outcomes: a secondary analysis of SPRINT. Hypertension. 2019;74(2):267-275. 10.1161/hypertensionaha.119.12907 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Dusetzina SB, Besaw RJ, Whitmore CC, et al.  Cost-related medication nonadherence and desire for medication cost information among adults aged 65 years and older in the US in 2022. JAMA Netw Open. 2023;6(5):e2314211. 10.1001/jamanetworkopen.2023.14211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Van Alsten SC, Harris JK.  Cost-related nonadherence and mortality in patients with chronic disease: a multiyear investigation, National Health Interview Survey, 2000-2014. Prev Chronic Dis. 2020;17:E151. 10.5888/pcd17.200244 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Taha MB, Valero-Elizondo J, Yahya T, et al.  Cost-related medication nonadherence in adults with diabetes in the United States: the National Health Interview Survey 2013-2018. Diabetes Care. 2022;45(3):594-603. 10.2337/dc21-1757 [DOI] [PubMed] [Google Scholar]
  • 18. Corliss J. The heart disease gender gap. Harvard Health. September 1, 2022. Accessed April 1, 2024. https://www.health.harvard.edu/heart-health/the-heart-disease-gender-gap [Google Scholar]
  • 19. Zhao M, Woodward M, Vaartjes I, et al.  Sex differences in cardiovascular medication prescription in primary care: a systematic review and meta-analysis. J Am Heart Assoc. 2020;9(11):e014742. 10.1161/jaha.119.014742 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Holtzman JN, Kaur G, Power JE, et al.  Underrepresentation of women in late-breaking cardiovascular clinical trials. J Womens Health (Larchmt). 2023;32(6):635-640. 10.1089/jwh.2022.0536 [DOI] [PubMed] [Google Scholar]
  • 21. Bosomworth J, Khan Z.  Analysis of gender-based inequality in cardiovascular health: an umbrella review. Cureus. 2023;15(8):e43482. 10.7759/cureus.43482 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Closing the benefit gap to advance women’s health equity. Deloitte United States. September 28, 2023. Accessed April 1, 2024. https://www2.deloitte.com/us/en/pages/life-sciences-and-health-care/articles/womens-health-equity-disparities.html [Google Scholar]
  • 23. Daher M, Al Rifai M, Kherallah RY, et al.  Gender disparities in difficulty accessing healthcare and cost-related medication non-adherence: the CDC Behavioral Risk Factor Surveillance System (BRFSS) survey. Prev Med. 2021;153:106779. 10.1016/j.ypmed.2021.106779 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Williams JS, Bishu K, Dismuke CE, Egede LE.  Sex differences in healthcare expenditures among adults with diabetes: evidence from the medical expenditure panel survey, 2002-2011. BMC Health Serv Res. 2017;17(1):259. 10.1186/s12913-017-2178-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Denny JC, Rutter JL, Goldstein DB, All of Us Research Program Investigators, et al.  The “All of Us” Research Program. N Engl J Med. 2019;381(7):668-676. 10.1056/NEJMsr1809937 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Quan H, Sundararajan V, Halfon P, et al.  Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130-1139. 10.1097/01.mlr.0000182534.19832.83 [DOI] [PubMed] [Google Scholar]
  • 27. Heidenreich PA, Bozkurt B, Aguilar D, et al.  2022 AHA/ACC/HFSA guideline for the management of heart failure: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines [published correction appears in Circulation. 2022 May 3;145(18):e1033. doi: 10.1161/CIR.0000000000001073] [published correction appears in Circulation. 2022 Sep 27;146(13):e185. doi: 10.1161/CIR.0000000000001097] [published correction appears in Circulation. 2023 Apr 4;147(14):e674. doi: 10.1161/CIR.0000000000001142]. Circulation. 2022;145(18):e895-e1032. 10.1161/CIR.0000000000001063 [DOI] [PubMed] [Google Scholar]
  • 28. Unger T, Borghi C, Charchar F, et al.  2020 International Society of Hypertension global hypertension practice guidelines. J Hypertens. 2020;38(6):982-1004. 10.1097/hjh.0000000000002453 [DOI] [PubMed] [Google Scholar]
  • 29. Fihn SD, Blankenship JC, Alexander KP, Society of Thoracic Surgeons, et al.  2014 ACC/AHA/AATS/PCNA/SCAI/STS focused update of the guideline for the diagnosis and management of patients with stable ischemic heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines, and the American Association for Thoracic Surgery, Preventive Cardiovascular Nurses Association, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. J Thorac Cardiovasc Surg. 2015;149(3):e5-e23. 10.1016/j.jtcvs.2014.11.002 [DOI] [PubMed] [Google Scholar]
  • 30. R Core Team. A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2022. [Database] [Google Scholar]
  • 31. Kluyver T.  Jupyter Notebooks—a publishing format for reproducible computational workflows. In: Loizides F, Scmidt B, eds. Positioning and Power in Academic Publishing: Players, Agents and Agendas. IOS Press; 2016:87-90. [Google Scholar]
  • 32.BigQuery documentation. Google; 2024. Accessed April 1, 2024. https://cloud.google.com/bigquery/docs
  • 33. Zhang JX, Crowe JM, Meltzer DO.  The differential rates in cost-related non-adherence to medical care by gender in the US adult population. J Med Econ. 2017;20(7):752-759. 10.1080/13696998.2017.1326383 [DOI] [PubMed] [Google Scholar]
  • 34. Bhuyan SS, Shiyanbola O, Deka P, et al.  The role of gender in cost-related medication nonadherence among patients with diabetes. J Am Board Fam Med. 2018;31(5):743-751. 10.3122/jabfm.2018.05.180039 [DOI] [PubMed] [Google Scholar]
  • 35. Jain V, Al Rifai M, Turpin R, et al.  Evaluation of factors underlying sex-based disparities in cardiovascular care in adults with self-reported premature atherosclerotic cardiovascular disease. JAMA Cardiol. 2022;7(3):341-345. 10.1001/jamacardio.2021.5430 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Bhuyan SS, Shiyanbola O, Kedia S, et al.  Does cost-related medication nonadherence among cardiovascular disease patients vary by gender? Evidence from a nationally representative sample. Womens Health Issues. 2017;27(1):108-115. 10.1016/j.whi.2016.10.004 [DOI] [PubMed] [Google Scholar]
  • 37.Health and health care for women of reproductive age. Commonwealth Fund. April 5, 2022. Accessed April 1, 2024. https://www.commonwealthfund.org/publications/issue-briefs/2022/apr/health-and-health-care-women-reproductive-age
  • 38. Lee H, Park JH, Floyd JS, Park S, Kim HC.  Combined effect of income and medication adherence on mortality in newly treated hypertension: nationwide study of 16 million person-years. J Am Heart Assoc. 2019;8(16):e013148. 10.1161/jaha.119.013148 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Rotman A, Mandel H.  Gender-specific wage structure and the gender wage gap in the U.S. Labor Market. Soc Indic Res. 2023;165(2):585-606. 10.1007/s11205-022-03030-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Gordon D. Women pay $15 billion more than men for medical costs, new report shows. Forbes Magazine. Accessed April 1, 2024. https://www.forbes.com/sites/debgordon/2023/09/26/women-pay-15-billion-more-than-men-for-medical-costs-new-report-shows/?sh=4213a0825a63
  • 41. Cubanski J, Koma W, Damico A, Neuman T. How much do medicare beneficiaries spend out of pocket on health care? KFF; 2019. Accessed April 1, 2024. https://www.kff.org/medicare/issue-brief/how-much-do-medicare-beneficiaries-spend-out-of-pocket-on-health-care/
  • 42. Koma W, Neuman T, Jacobson G, Smith K. Medicare beneficiaries’ financial security before the coronavirus pandemic. KFF; 2020. Accessed April 1, 2024. https://www.kff.org/medicare/issue-brief/medicare-beneficiaries-financial-security-before-the-coronavirus-pandemic/
  • 43. Wineinger NE, Zhang Y, Topol EJ.  Trends in prices of popular brand-name prescription drugs in the United States. JAMA Netw Open. 2019;2(5):e194791. 10.1001/jamanetworkopen.2019.4791 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Trish E, Van Nuys K, Popovian R. U.S. consumers overpay for generic drugs. University of Southern California (USC), Leonard D. Schaeffer Center for Health Policy and Economics; 2024. Accessed April 1, 2024. https://healthpolicy.usc.edu/research/u-s-consumers-overpay-for-generic-drugs/
  • 45. Orlando V, Mucherino S, Guarino I, Guerriero F, Trama U, Menditto E.  Gender differences in medication use: a drug utilization study based on real world data. Int J Environ Res Public Health. 2020;17(11):3926. 10.3390/ijerph17113926 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Wang X, Liu K, Shirai K, et al.  Prevalence and trends of polypharmacy in U.S. adults, 1999-2018. Glob Health Res Policy. 2023;8(1):25. 10.1186/s41256-023-00311-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Hamrahian SM, Maarouf OH, Fülöp T.  A critical review of medication adherence in hypertension: barriers and facilitators clinicians should consider. Patient Prefer Adherence. 2022;16:2749-2757. 10.2147/ppa.s368784 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Hamdidouche I, Jullien V, Boutouyrie P, Billaud E, Azizi M, Laurent S.  Drug adherence in hypertension: from methodological issues to cardiovascular outcomes. J Hypertens. 2017;35(6):1133-1144. 10.1097/hjh.0000000000001299 [DOI] [PubMed] [Google Scholar]
  • 49. Corrao G, Parodi A, Nicotra F, et al.  Better compliance to antihypertensive medications reduces cardiovascular risk. J Hypertens. 2011;29(3):610-618. 10.1097/HJH.0b013e328342ca97 [DOI] [PubMed] [Google Scholar]
  • 50. Fitzgerald AA, Powers JD, Ho PM, et al.  Impact of medication nonadherence on hospitalizations and mortality in heart failure. J Card Fail. 2011;17(8):664-669. 10.1016/j.cardfail.2011.04.011 [DOI] [PubMed] [Google Scholar]
  • 51. Ruppar TM, Cooper PS, Mehr DR, Delgado JM, Dunbar-Jacob JM.  Medication adherence interventions improve heart failure mortality and readmission rates: systematic review and meta-analysis of controlled trials. J Am Heart Assoc. 2016;5(6):e002606. 10.1161/jaha.115.002606 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52. De Vera MA, Bhole V, Burns LC, Lacaille D.  Impact of statin adherence on cardiovascular disease and mortality outcomes: a systematic review. Br J Clin Pharmacol. 2014;78(4):684-698. 10.1111/bcp.12339 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ocae203_Supplementary_Data
ocae203_supplementary_data.docx (69.6KB, docx)

Data Availability Statement

The data underlying this study were provided by All of Us Research Program by permission. Their policy states “authorized data users will not publish or otherwise distribute any participant-level data from the All of Us Research Program database.” Efforts were made in the manuscript to describe where data was accessed within the platform for interested parties.

Articles from Journal of the American Medical Informatics Association : JAMIA are provided here courtesy of Oxford University Press

RESOURCES