Antihypertensive Medication Adherence and Medical Costs, Health Care Use, and Labor Productivity Among People With Hypertension Authors: Jun Soo Lee, Raul Segura Escano, Nicole L. Therrien, Ashutosh Kumar, Ami Bhatt, Lisa M. Pollack, Sandra L. Jackson, Feijun Luo Published: 2024 ABSTRACT BackgroundHypertension affects nearly half of US adults yet remains inadequately controlled in over three‐quarters of these cases. This study aimed to assess the association between adherence to antihypertensive medications and total medical costs, health care use, and productivity‐related outcomes.Methods and ResultsWe conducted cross‐sectional analyses using MarketScan databases, which included individuals aged 18 to 64 years with noncapitated health insurance plans in 2019. Adherence was defined as ≥80% medication possession ratio for prescribed antihypertensive medications. We used a generalized linear model to estimate total medical costs, a negative binomial model to estimate health care use (emergency department visits and inpatient admissions), an exponential hurdle model to estimate productivity‐related outcomes (number of sick absences, short‐term disability, long‐term disability), and a 2‐part model to estimate productivity‐related costs in 2019 US dollars. All models were adjusted for age, sex, urbanicity, census region, and comorbidities. We reported average marginal effects for outcomes related to antihypertensive medication adherence. Among 379 503 individuals with hypertension in 2019, 54.4% adhered to antihypertensives. Per person, antihypertensive medication adherence was associated with $1441 lower total medical costs, $11 lower sick absence costs, $291 lower short‐term disability costs, and $69 lower long‐term disability costs. Per 1000 individuals, medication adherence was associated with lower health care use, including 200 fewer emergency department visits and 90 fewer inpatient admissions, and productivity‐related outcomes, including 20 fewer sick absence days and 442 fewer short‐term disability days.ConclusionsAdherence to antihypertensives was consistently associated with lower total medical costs, reduced health care use, and improved productivity‐related outcomes. Nonstandard Abbreviations and Acronyms Commercial Claims and Encounters long‐term disability medication possession ratio short‐term disability Clinical PerspectiveWhat Is New? Adherence to antihypertensive medications is associated with significant reductions in health care use, including emergency department visits and inpatient admissions, and is also associated with lower overall medical costs.Adherence to antihypertensive medications is further associated with improvements in work productivity, measured by reductions in sick absences and both short‐term and long‐term disability days. What Are the Clinical Implications? Enhanced adherence to antihypertensive medications among working‐age adults may provide triple benefits of improving clinical outcomes and reducing health care costs, and increased work productivity, offering additional economic advantages to both patients and employers.Interventions aimed at improving antihypertensive medication adherence could have a broad impact, from better controlling hypertension and reducing cardiovascular risks to yielding substantial health care savings and reducing productivity losses in the workforce. Clinical Perspective What Is New? Adherence to antihypertensive medications is associated with significant reductions in health care use, including emergency department visits and inpatient admissions, and is also associated with lower overall medical costs.Adherence to antihypertensive medications is further associated with improvements in work productivity, measured by reductions in sick absences and both short‐term and long‐term disability days. Adherence to antihypertensive medications is associated with significant reductions in health care use, including emergency department visits and inpatient admissions, and is also associated with lower overall medical costs. Adherence to antihypertensive medications is associated with significant reductions in health care use, including emergency department visits and inpatient admissions, and is also associated with lower overall medical costs. Adherence to antihypertensive medications is further associated with improvements in work productivity, measured by reductions in sick absences and both short‐term and long‐term disability days. Adherence to antihypertensive medications is further associated with improvements in work productivity, measured by reductions in sick absences and both short‐term and long‐term disability days. What Are the Clinical Implications? Enhanced adherence to antihypertensive medications among working‐age adults may provide triple benefits of improving clinical outcomes and reducing health care costs, and increased work productivity, offering additional economic advantages to both patients and employers.Interventions aimed at improving antihypertensive medication adherence could have a broad impact, from better controlling hypertension and reducing cardiovascular risks to yielding substantial health care savings and reducing productivity losses in the workforce. Enhanced adherence to antihypertensive medications among working‐age adults may provide triple benefits of improving clinical outcomes and reducing health care costs, and increased work productivity, offering additional economic advantages to both patients and employers. Enhanced adherence to antihypertensive medications among working‐age adults may provide triple benefits of improving clinical outcomes and reducing health care costs, and increased work productivity, offering additional economic advantages to both patients and employers. Interventions aimed at improving antihypertensive medication adherence could have a broad impact, from better controlling hypertension and reducing cardiovascular risks to yielding substantial health care savings and reducing productivity losses in the workforce. Interventions aimed at improving antihypertensive medication adherence could have a broad impact, from better controlling hypertension and reducing cardiovascular risks to yielding substantial health care savings and reducing productivity losses in the workforce. Hypertension elevates the risk for heart disease and stroke, 2 leading causes of death for people in the United States. 1 Almost half of US adults have hypertension, and over three‐quarters of these individuals do not have their blood pressure controlled to <130/80 mm Hg. 2 Prescription medication is recommended, along with lifestyle modifications, for nearly 80% of those with hypertension to achieve blood pressure control. 2 Adherence to antihypertensive medications (ie, consistent use as prescribed) plays a critical role in hypertension control. 3 , 4 , 5 , 6 Medication adherence is multifactorial and may be influenced by socioeconomic and demographic, health care system, therapeutic, and patient‐related factors. 4 Improving adherence to antihypertensive medications among people with hypertension is crucial for improving national hypertension control rates. 5 , 7 In addition to improved hypertension control, adherence to antihypertensive medications is associated with reduced risk of stroke, fewer hospitalizations, and decreased cardiovascular morbidity and mortality. 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 Despite its importance, rates of adherence to antihypertensive medications are suboptimal, though estimates vary by method of measurement and population characteristics. 4 Chang et al 16 found that nearly two‐thirds of US adults with hypertension were considered adherent to their antihypertensive medications, with lower adherence rates observed among Medicare beneficiaries, women, and younger adults. Improving medication adherence may impact economic outcomes, including total medical costs, 17 , 18 , 19 hospitalization and emergency department (ED) visits, 12 , 20 , 21 , 22 and other indirect outcomes such as work productivity 23 , 24 and mortality rates. 25 , 26 Lost work productivity, attributed to chronic conditions and resulting in absenteeism or disability leave, imposes an economic burden on the patient, employers, and health care systems. 27 , 28 Enhanced medication adherence among employees with hypertension may lead to fewer days of absence and health care use savings. 28 , 29 , 30 Although some studies have documented costs and health care use associated with adherence to antihypertensive medications, 12 , 17 , 18 , 19 , 20 none have investigated productivity‐related outcomes, which are indirect benefits associated with such adherence. This study aimed to investigate the economic burden associated with adherence to antihypertensive medications, including total medical costs, health care use, and labor productivity. Administrative commercial claims linked with employer‐provided payroll databases from 2019 were used to examine both medical costs and productivity‐related costs associated with medication adherence. METHODS Data We used the 2019 MarketScan Commercial Claims and Encounters (CCAE) and Health and Productivity Management databases. 31 The CCAE database contains claims for inpatient, ED, outpatient, and prescription drugs from enrollees and their dependents covered by employer‐sponsored commercial health insurance plans. This database aggregates data from >300 employers, >30 health plans, and >500 hospitals across the United States. The Health and Productivity Management database is derived from a subset of employers' payroll databases, including information on recreational, sick, and other absences, as well as absences related to short‐term disability (STD) and long‐term disability (LTD). Linking the Health and Productivity Management database to commercial claims information in the CCAE database enables researchers to analyze associations between medical conditions, medication adherence, total medical costs, health care use, and labor productivity outcomes. This study involved secondary data analysis using deidentified information and was categorized as nonresearch and thus exempt from institutional review board review. Due to data use agreement restrictions, the MarketScan data cannot be made publicly available. However, the program codes used for this study can be requested from the corresponding author. Identification of Individuals With Hypertension We included individuals aged 18 to 64 years, continuously enrolled in the 2019 MarketScan CCAE and Health and Productivity Management databases. Within this group, individuals with hypertension were defined if they had at least 1 hypertension diagnosis (International Classification of Diseases, Tenth Revision, Clinical Modification [ICD‐10‐CM] codes I10–I15) and at least 1 antihypertensive drug claim in 2019 (Table S1). We excluded individuals with a history of pregnancy (Table S2) or those covered under capitated insurance. Identification of Adherence to Antihypertensive Medications Antihypertensive medications (antihypertensives) were identified using generic names and categorized by therapeutic class (angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, β‐blockers, calcium channel blockers, diuretics, other antihypertensives, and renin‐angiotensin system antagonists [angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, and direct renin inhibitors]) (Table S1). The average medication possession ratio (MPR) for antihypertensives by therapeutic class in 2019 was calculated as the ratio of the sum of the days' supply for all drug fills to the number of days in the study period. 32 For individuals taking multiple antihypertensives, the MPR was calculated as an average of the MPR for each therapeutic class. Adherence was defined as an average MPR ≥80%, whereas nonadherence was indicated by an average MPR <80%. 32 A dummy indicator of adherence to antihypertensives was created, with a value of 1 for adherent individuals and 0 for nonadherent individuals. Outcome Variables Dependent variables included (1) total all‐cause medical costs (the sum of individuals' out‐of‐pocket costs and insurance payments), (2) health care use measured by the number of ED visits and inpatient admissions, and (3) productivity‐related outcomes quantified by the number of sick absences and absences related to STD and LTD. The total all‐cause medical costs encompass all medical expenses, including outpatient visits, ED visits, inpatient admissions, and pharmacy claims. The economic analysis was conducted from the perspective of both the health care system and the employer, capturing the broader economic impact of antihypertensive medication adherence. Our focus was on sick absences rather than other types, such as recreational, jury duty, military leave, or plant shutdowns, which may not be directly affected by medication adherence. Productivity‐related costs were calculated by multiplying absence hours by average hourly wages in 2019, 33 with adjustments: 100% for absences, 70% for STD‐related absences, and 60% for LTD‐related absences. 33 , 34 Average hourly wages in 2019 US dollars ($27.99) were sourced from the US Bureau of Labor Statistics for all employees on private nonfarm payrolls. Statistical Analysis We categorized individuals into age groups (18–34 [reference], 35–44, 45–54, and 55–64 years), sex (men [reference] and women), urbanicity (rural [reference] and urban), census regions (Northeast [reference], Midwest, South, and West), and comorbidities, specifically the Quan 17 Charlson Comorbidities. 35 Comorbidities were defined if individuals had at least 1 inpatient admission or 2 outpatient visits with a 30‐day interval. Summary statistics were documented for age, sex, urbanicity and census region of residence, and comorbidities. Differences in average values by medication adherence status were tested using the Wilcoxon rank sum test for continuous variables and Pearson χ2 test for categorical variables. For total medical costs, a generalized linear model with a log link and γ distribution was used. For costs associated with productivity‐related outcomes, we used a 2‐part model. The first part was a logit model, and the second part was a generalized linear model with a log link and γ distribution. For health care use (number of ED visits and inpatient admissions), a negative binomial model was used. For productivity‐related outcomes, where they contained excess 0s, we used an exponential hurdle model. All models were adjusted for individuals' age, sex, urbanicity, census regions, and dummy indicators for the 17 Charlson comorbidities. Average marginal effects of outcomes associated with adherence to antihypertensives were reported. For the sensitivity analysis, we used the overlap weighting method so measured confounders were equally distributed between the adherence and nonadherence groups. 36 , 37 , 38 , 39 The overlap weighting method mimics random assignment by creating a pseudopopulation through weighting so that measured confounders are equally distributed between adherence and nonadherence groups. 37 , 38 We calculated overlap weights that were proportional to the probability of individuals belonging to the opposite group. Weights were obtained from propensity scores, estimated by logistic regression, with medication adherence as the outcome. We used the iterative method to estimate propensity scores proposed by Imbens and Rubin. 40 , 41 This method involved using baseline covariates, as reported in the main model, to perform higher‐order interactions for covariate selection in logistic regression, maximizing the likelihood function. Estimates of logistic regression, assessment of covariate balance, and probability distributions for weighted samples for overall and by sex, age group, and urbanicity are presented in Tables S3 through S5 and Figures S1 through S4. Lastly, as a sensitivity analysis, we further adjusted for the number of antihypertensives in the model. A P value <0.05 was considered statistically significant. All analyses were conducted by using Stata SE version 17 (StataCorp, College Station, TX) in 2023. RESULTS A total of 379 503 individuals with hypertension were identified who had availability of absence, STD, or LTD information (Figure). All 379 503 individuals had information on medical costs, health care use, and drug claims. Among these individuals, 54 608 (14%) had absence information, 328 073 (86%) had STD information, and 323 445 (85%) had LTD information. Study sample selection of individuals diagnosed with hypertension, MarketScan Commercial and Health and Product Management Database, 2019. HPM indicates Health and Productivity Management; ICD‐10‐CM, International Classification of Diseases, Tenth Revision, Clinical Modification; LTD, long‐term disability; and STD, short‐term disability. Out of the 379 503 individuals, 206 310 (54.4%) individuals adhered to antihypertensives (Table 1). Those who adhered (mean±SD age, 52.5±7.7 years) were older than those who did not (mean±SD age, 49.4±8.9; P<0.001). Adherent individuals were less likely to be women (31.4% versus 36.7%; P<0.001). Medication adherence rates were also lower for women than men (50.4% versus 56.4%; P<0.001) and for individuals living in urban areas compared with those in rural areas (54.2% versus 55.8%; P<0.001) (Table S6). Adjusted results showed that women were 5.83% (95% CI, 6.16–5.50; P<0.001) less likely to adhere to medications than men, and urban residents were 1.38% (95% CI, −1.89 to −0.86; P<0.001) less likely to adhere to medications than those residing in rural areas (Table S7). Table 1Summary Statistics of Individuals With Hypertension by Medication Adherence Status, MarketScan Commercial Insurance, 2019* Characteristics and outcomesAllMedication adherence† Medication nonadherence P valueN=379 503 (100%)N=206 310 (54.4%)N=173 193 (45.6%)Age, y, mean (SD)51.1±8.452.5±7.749.4±8.9<0.001Age groups, y, n (%)18–3418 480 (4.9%)5808 (2.8%)12 672 (7.3%)<0.00135–4461 732 (16.3%)26 358 (12.8%)35 374 (20.4%)<0.00145–54142 361 (37.5%)76 387 (37.0%)65 974 (38.1%)<0.00155–64156 930 (41.4%)97 757 (47.4%)59 173 (34.2%)<0.001Women, n (%)128 245 (33.8%)64 679 (31.4%)63 566 (36.7%)<0.001Urban residency, n (%)341 292 (89.9%)184 978 (89.7%)156 314 (90.3%)<0.001Census region, n (%)Northeast52 646 (13.9%)31 090 (15.1%)21 556 (12.5%)<0.001Midwest96 314 (25.4%)54 545 (26.4%)41 769 (24.1%)<0.001South182 220 (48.0%)94 087 (45.6%)88 133 (50.9%)<0.001West47 611 (12.6%)26 286 (12.7%)21 325 (12.3%)<0.001Comorbidities, n (%)Myocardial infarction6951 (1.8%)3468 (1.7%)3483 (2.0%)<0.001Congestive heart failure7549 (2.0%)3547 (1.7%)4002 (2.3%)<0.001Peripheral vascular disease5493 (1.5%)2980 (1.4%)2513 (1.5%)0.87Cerebrovascular disease1775 (0.5%)749 (0.4%)1026 (0.6%)<0.001Dementia49 (0.0%)19 (0.0%)30 (0.0%)0.028Chronic pulmonary disease16 189 (4.3%)8117 (3.9%)8072 (4.7%)<0.001Rheumatic disease3981 (1.1%)2116 (1.0%)1865 (1.1%)0.12Peptic ulcer disease719 (0.2%)324 (0.2%)395 (0.2%)<0.001Mild liver disease6196 (1.6%)3079 (1.5%)3117 (1.8%)<0.001Diabetes without chronic complication69 362 (18.3%)40 754 (19.8%)28 608 (16.5%)<0.001Diabetes with chronic complication15 073 (4.0%)8584 (4.2%)6489 (3.8%)<0.001Hemiplegia or paraplegia802 (0.2%)270 (0.1%)532 (0.3%)<0.001Renal disease9818 (2.6%)5210 (2.5%)4608 (2.7%)0.009Any malignancy10 373 (2.7%)5972 (2.9%)4401 (2.5%)<0.001Moderate or severe liver disease468 (0.1%)182 (0.1%)286 (0.2%)<0.001Metastatic solid tumor1379 (0.4%)657 (0.3%)722 (0.4%)<0.001AIDS/HIV1256 (0.3%)694 (0.3%)562 (0.3%)0.53Outcomes, $, mean±SDTotal medical costs14 081.8±40 871.812 882.2±32 497.415 510.7±48 976.3<0.001Total insurance medical payments12 425.0±40 176.811 250.1±31 776.913 824.5±48 275.8<0.001Total individuals' out‐of‐pocket medical costs16 56.8±19 32.01632.1±1756.21686.3±2122.2<0.001Sick absence costs76.6±807.970.6±754.983.9±866.9<0.001STD costs686.5±3676.0533.8±3143.7868.5±4216.3<0.001LTD costs159.9±4232.6120.3±2766.5207.1±5489.5<0.001No. of emergency department visits0.34±0.990.25±0.780.46±1.18<0.001No. of inpatient admissions0.09±0.410.06±0.320.12±0.49<0.001No. of outpatient visits12.90±16.0512.73±15.2913.11±16.90<0.001No. of pharmacy prescriptions25.21±20.0528.30±20.7821.53±18.49<0.001No. of sick absences2.38±9.252.22±8.702.57±9.85<0.001No. of STD5.07±25.153.93±21.496.43±28.88<0.001No. of LTD1.40±34.121.06±22.361.80±44.12<0.001No. of antihypertensive prescriptions7.93±6.429.43±6.806.14±5.42<0.001ED indicates emergency department; LTD, long‐term disability; and STD, short‐term disability.*The Wilcoxon nonparametric rank‐sum test was used to test the differences in means for continuous variables, and the Pearson χ2 test was used to test differences in proportions for categorical variables by medication nonadherence to antihypertensives status. † Adherence to antihypertensives if the average medication possession ratio of the 7 therapeutic classes was ≥80%, otherwise medication nonadherence. Summary Statistics of Individuals With Hypertension by Medication Adherence Status, MarketScan Commercial Insurance, 2019* ED indicates emergency department; LTD, long‐term disability; and STD, short‐term disability. The Wilcoxon nonparametric rank‐sum test was used to test the differences in means for continuous variables, and the Pearson χ2 test was used to test differences in proportions for categorical variables by medication nonadherence to antihypertensives status. Adherence to antihypertensives if the average medication possession ratio of the 7 therapeutic classes was ≥80%, otherwise medication nonadherence. Table 2 presents the total medical costs associated with adherence to antihypertensives. The predicted total medical cost for individuals who adhered to antihypertensives was $17 770 (95% CI, $17 238–$18 301), whereas for those who did not adhere, the predicted total medical cost was $19 210 (95% CI, $18 643–$19 778). Thus, compared with individuals who did not adhere, those who adhered had a $1441 per individual lower total medical costs (95% CI, −$1709 to −$1172; P<0.001). The association of medication adherence with total medical costs was more pronounced for individuals residing in urban areas, showing an additional decrease of $855 (95% CI, −$1633 to −$77; P<0.05) compared with those living in rural areas. Table 2Total Medical Costs (Per Individual) Associated With Adherence to Antihypertensives in 2019* OutcomesAllWomenMenWomen vs men† UrbanRuralUrban vs rural† Total medical costs, $Medication nonadherence19 21020 48818 5581930‡ 19 49816 6412857‡ 18 643 to 19 77819 870 to 21 10717 957 to 19 1601474 to 2386‡ 18 907 to 20 08915 977 to 17 3062221 to 3493‡ Medication adherence17 77018 83717 2251611‡ 17 97115 9702002‡ 17 238 to 18 30118 265 to 19 40816 663 to 17 7871205 to 2017‡ 17 419 to 18 52315 373 to 16 5661454 to 2550‡ Difference–1441‡ –1652‡ –1333‡ −319−1527‡ −671.7−855‡ −1709 to –1172‡ −2131 to –1172‡ −1650 to –1016‡ −886 to 249−1813 to –1241‡ −1399 to 55.76−1633 to –77‡ Observation379 503128 245251 258379 503341 29238 211379 503*A generalized linear model with a γ distribution and log link was used. The average marginal effects with 95% CIs were reported. All models were adjusted for the patient's age, sex, urbanicity of residence, Census regions, and comorbidities. The differences and the 95% CIs of the predicted and average marginal effects by sex and urbanicity were calculated. All the outcomes are per individual. † The differences with 95% CIs in the predicted values and average marginal effects by sex (women–men) and urbanicity (urban–rural) were reported. ‡ Statistical significance at the P<0.05 level. Total Medical Costs (Per Individual) Associated With Adherence to Antihypertensives in 2019* A generalized linear model with a γ distribution and log link was used. The average marginal effects with 95% CIs were reported. All models were adjusted for the patient's age, sex, urbanicity of residence, Census regions, and comorbidities. The differences and the 95% CIs of the predicted and average marginal effects by sex and urbanicity were calculated. All the outcomes are per individual. The differences with 95% CIs in the predicted values and average marginal effects by sex (women–men) and urbanicity (urban–rural) were reported. Statistical significance at the P<0.05 level. Table 3 documents the number of ED visits and inpatient admissions per 1000 individuals associated with adherence, both overall and stratified by sex and urbanicity of residence. Adherence to antihypertensives was associated with 200 fewer ED visits per 1000 individuals (95% CI, −206 to −194; P<0.001) and 89.6 fewer inpatient admissions per 1000 individuals (95% CI, −95.5 to −83.7; P<0.001). Women experienced an additional 113 fewer ED visits (95% CI, −126 to −99.5; P<0.001) associated with medication adherence compared with men, whereas men had an additional 26.6 fewer inpatient admissions (95% CI, 17.9 to 25.3; P<0.001) associated with medication adherence compared with women. Individuals living in urban areas had fewer inpatient admissions associated with medication adherence than those living in rural areas by 35.5 (95% CI, −46.8 to −24.2; P<0.001). Table 3Health Care Use (per 1000 Individuals) Associated With Adherence to Antihypertensives in 2019* OutcomesAllWomenMenWomen vs men† UrbanRuralUrban vs rural† Emergency department visitsMedication non‐adherence456580393187‡ 4574497.6451 to 462570 to 591387 to 399176 to 199‡ 451 to 463433 to 465−9.2 to 24.4Medication adherence25630623174.4‡ 255266−10.9‡ 253 to 260299 to 312228 to 23567.4 to 81.3‡ 252 to 259257 to 276−21.0 to –0.1‡ Difference−200‡ −275‡ −162‡ −113‡ −202‡ −183‡ −18.5−206 to –194‡ −287 to –263‡ −168 to –155‡ −126 to –99.5‡ −208 to –195‡ −202 to –165‡ −38.1 to 1.1Observation379 503128 245251 258379 503‡ 341 29238 211379 503Inpatient admissionsMedication nonadherence210177226−49.5‡ 21614967.6‡ 199 to 220169 to 185214 to 239−58.8 to –40.3‡ 205 to 227140 to 15856.4 to 78.7‡ Medication adherence120105128−22.9‡ 12391.132.0‡ 114 to 12699.6 to 110120 to 135−28.7 to –17.1‡ 117 to 12984.9 to 97.225.0 to 39.1‡ Difference−89.6‡ −72.0‡ −98.6‡ 26.6‡ −93.2‡ −57.7‡ −35.5‡ −95.5 to –83.7‡ −78.5 to –65.5‡ −106 to –91.1‡ 17.9 to 35.3‡ −99.6 to –86.9‡ −67.7 to –47.7‡ −46.8 to –24.2‡ Observation379 503128 245251 258379 503341 29238 211379 503*A negative binomial regression was used. All models were adjusted for the patient's age, sex, urbanicity of residence, Census regions, and comorbidities. We reported the average marginal effects with a 95% CI. The differences and the 95% CIs of the predicted and average marginal effects by sex and urbanicity were calculated. All the outcomes are per 1000 individuals. † The differences with 95% CIs in the predicted values and average marginal effects by sex (women–men) and urbanicity (urban–rural) were reported. ‡ Statistical significance at the P<0.05 level. Health Care Use (per 1000 Individuals) Associated With Adherence to Antihypertensives in 2019* A negative binomial regression was used. All models were adjusted for the patient's age, sex, urbanicity of residence, Census regions, and comorbidities. We reported the average marginal effects with a 95% CI. The differences and the 95% CIs of the predicted and average marginal effects by sex and urbanicity were calculated. All the outcomes are per 1000 individuals. The differences with 95% CIs in the predicted values and average marginal effects by sex (women–men) and urbanicity (urban–rural) were reported. Statistical significance at the P<0.05 level. Table 4 presents productivity losses per 1000 individuals associated with medication adherence. Individuals adhering to antihypertensives had fewer sick absences (−19.8 days per 1000 individuals [95% CI, −28.0 to −11.6]; P<0.001) and STD days (−442 days [95% CI, −548 to −335]; P<0.001). The association between medication adherence and the number of sick absences was more pronounced for individuals living in rural areas, showing a stronger association in sick absences by 46.6 days (95% CI, 11.3–81.9; P<0.001) than those living in urban areas. Table 4Productivity Losses (per 1000 Individuals) Associated With Adherence to Antihypertensives in 2019* OutcomesAllWomenMenWomen vs men† UrbanRuralUrban vs rural† No. of sick absencesMedication nonadherence308329297−31.6‡ 299385−86.0‡ 300 to 316316 to 341288 to 306−44.8 to –18.3‡ 291 to 307357 to 413−114 to –57.9‡ Medication adherence288306279−26.8‡ 284323−39.4‡ 281 to 295294 to 318271 to 286−39.2 to –14.4‡ 277 to 291302 to 345−61 to –17.8‡ Difference−19.8‡ −22.9‡ −18.2‡ 4.75−15.1‡ −61.7‡ 46.6‡ −28.0 to –11.6‡ −38.7 to –7.2‡ −27.4 to –8.9‡ −13.3 to 22.8−23.3 to –6.9‡ −96.1 to –27.3‡ 11.3 to 81.9‡ Observation54 60813 34641 26254 60849 125548354 608No. of STDMedication nonadherence451443544596241‡ 45484211337‡ 4414 to 46154230 to 44794475 to 471696.7 to 386‡ 4444 to 46523976 to 444696.8 to 577‡ Medication adherence407239254147222‡ 40863.9501373978 to 41663799 to 40524036 to 425975 to 369‡ 3988 to 41843731 to 4168−88.7 to 362Difference−442‡ −429‡ −448‡ −19.2−462‡ −262−200−548 to –335‡ −585 to –273‡ −586 to –311‡ −224 to 186−575 to –349‡ −572 to 48.5−529 to 128Observation328 073111 160216 913328 073295 10432 969328 073No. of LTDMedication nonadherence150812331648415‡ 14791761−2811319 to 16971045 to 14211412 to 1884173 to 657‡ 1289 to 16691243 to 2278−794 to 232Medication adherence16241445171526915961870−2741414 to 18331187 to 17041460 to 1970−43.4 to 5821382 to 18101314 to 2427−836 to 287Difference11621266.8−1461171106.8−94.2 to 326−64.2 to 489−215 to 348−537 to 245−102 to 335−618 to 838−752 to 766Observation323 445110 894212 551323 445290 85332 592323 445LTD indicates long‐term disability; and STD, short‐term disability.*An exponential hurdle model was used. All models were adjusted for the patient's age, sex, urbanicity of residence, Census regions, and comorbidities. We reported the average marginal effects with a 95% CI. The differences and the 95% CIs of the predicted and average marginal effects by sex and urbanicity were calculated. All the outcomes are per 1000 individuals. † The differences with 95% CIs in the predicted values and average marginal effects by sex (women–men) and urbanicity (urban–rural) were reported. ‡ Statistical significance at the P<0.05 level. Productivity Losses (per 1000 Individuals) Associated With Adherence to Antihypertensives in 2019* LTD indicates long‐term disability; and STD, short‐term disability. An exponential hurdle model was used. All models were adjusted for the patient's age, sex, urbanicity of residence, Census regions, and comorbidities. We reported the average marginal effects with a 95% CI. The differences and the 95% CIs of the predicted and average marginal effects by sex and urbanicity were calculated. All the outcomes are per 1000 individuals. The differences with 95% CIs in the predicted values and average marginal effects by sex (women–men) and urbanicity (urban–rural) were reported. Statistical significance at the P<0.05 level. Table S8 presents the productivity costs per individual associated with medication adherence. Individuals adhering to antihypertensives had $11 lower costs associated with sick absences (95% CI, −16 to −6; P<0.001), $291 lower costs associated with STD (95% CI, −315 to −268; P<0.001), and $69 lower costs associated with LTD (95% CI, −97 to −41; P<0.001) compared with those who did not adhere. Women experienced a stronger association between STD costs and medication adherence ($72 [95% CI, −121 to −22]; P<0.01) than men, and individuals living in urban areas had a stronger association in costs associated with STD (−$173 [95% CI, −249 to −97]; P<0.001) than those living in rural areas. Sensitivity analyses using an overlap‐weighted model showed overall results similar to the main results (Tables S9 through S12). Specifically, compared with those who did not adhere, individuals adhering to antihypertensives had a $1378 lower total medical costs per individual (95% CI, −1628 to −1128; P<0.001) and 82 fewer inpatient admissions per 1000 individuals (95% CI, −87 to −77; P<0.001) (Tables S9 and S10). Tables S11 and S12 present productivity‐related outcomes, demonstrating consistent results. Tables S13 through S15 presented the results after adjusting for the number of antihypertensives. The findings indicate even higher reductions for those who adhered, including significantly lower total medical costs (−$1988 per individual [95% CI, −2262 to −1715]; P<0.001), fewer ED visits (−191.7 per 1000 individuals [95% CI, −208.2 to −175.1]; P<0.001), inpatient admissions (−290.7 per 1000 individuals [95% CI, −451.4 to −129.9]; P<0.001), sick absences (−13.5 per 1000 individuals [95% CI, −22.0 to −5.0]; P<0.01), and STD instances (−512.9 per 1000 individuals [95% CI, −622.9 to −403.0]; P<0.001). DISCUSSION In our study, we used administrative commercial claims linked with employer‐provided payroll databases, a novel approach that allowed us to examine both medical costs and productivity‐related costs associated with medication adherence. This innovative linkage provided insights into the broader economic burden of hypertension management, encompassing not only health care use but also the indirect benefits of adherence from an employer's perspective. Individuals with higher adherence rates had significantly lower total medical costs ($1441 per individual), along with fewer ED visits (200 fewer visits per 1000 individuals per year) and fewer inpatient admissions (89.6 fewer admissions per 1000 individuals per year) compared with their nonadherent counterparts. For productivity‐related outcomes, individuals who adhered to antihypertensives had fewer sick absences, STD, and LTD, as well as lower related costs, compared with those who did not adhere. When analyzing the association of medication adherence among population subgroups, we observed differences by sex and urbanicity of residence. Our findings align with existing literature on the association between medication adherence and reduced preventable health care use, decreased total medical costs, and improved productivity‐related outcomes. 23 , 24 , 30 In a 2018 systematic review, Cutler et al 17 identified 12 studies focused on cardiovascular disease and total health care costs, revealing that adjusted annual economic costs of nonadherence for cardiovascular disease ranged from $3347 to $19 472 per person. Baker‐Goering et al in 2019 19 found predicted expenditures to be $610 lower among individuals who were adherent to antihypertensives compared with those who were not adherent. Although our findings indicate more modest savings, variations in the formula used to calculate medication adherence and the medication nonadherence cutoff points, as well as our focus on hypertension, likely contribute to these differences. With regard to productivity, Gifford et al's systematic review in 2018 42 on adherence to antihypertensives demonstrated that employee populations with chronic conditions can generate productivity and health care use savings at the organization level as adherence increases. Our analysis aligns with these findings in productivity and health care use savings. However, differences in indicators, outcome measures, and chronic conditions may have influenced the magnitude of estimated savings. Describing the economic costs of uncontrolled hypertension has been recognized as a strategic approach to elevate hypertension control to a national priority and improve hypertension control across the United States. 5 The association between medication adherence and total medical costs, health care use, and productivity‐related outcomes can be explained through various potential mechanisms. Individuals who adhere to medications typically exhibit better control of chronic disease and health outcomes, 8 , 9 , 10 , 11 , 13 , 14 , 15 , 43 leading to lower number of ED visits and inpatient admissions related to costly acute events, subsequently lowering total medical costs. Moreover, individuals with improved health outcomes tend to experience fewer sick absences 42 and rely less on disability‐related insurance, contributing to increased productivity and reduced costs to employers. However, it is crucial to acknowledge the potential influence of unmeasured social determinants of health, especially economic factors such as poverty, unstable housing, food insecurity, lack of transportation, and lack of social support. These factors may have affected both lower medication adherence and medical costs, ED visits, and sick absences. 44 , 45 , 46 Although our study controlled for many confounding variables, the potential impact of these unmeasured social determinants of health indicators cannot be ignored and may represent an important area for future research. Prior studies have highlighted variations in medication adherence based on sex. 47 Women, although more likely to use medication annually than their male counterparts, were less likely to adhere to prescribed medication. 48 This difference might be attributable to women more frequently experiencing side effects 49 , 50 and managing complex regimens, 51 despite their higher likelihood of seeking preventive care. 52 A separate study found that, although men demonstrated higher rates of adherence than women in the age range of 20 to 40 years, this trend reversed later in life (50–70 years). 53 Further research could help elucidate the reasons behind sex‐based differences. The Surgeon General's Call to Action to Control Hypertension recognizes the prioritization of medication adherence as an evidence‐based strategy to achieve improvements in hypertension control. 5 Moreover, the Call to Action highlights several interventions designed to empower and equip patients for better medication adherence. These interventions include self‐measured blood pressure monitoring, tailored interventions by pharmacists, shared decision‐making and motivational interviewing, reducing or eliminating cost sharing for medications, simplifying medication regimens, including the use of fixed‐dose combination antihypertensives, and synchronizing refills for multiple medications to a single date. 4 , 5 , 54 Clinicians can implement or expand these interventions to support medication adherence in hypertension care. 4 , 5 Employers can play a pivotal role by supporting coverage of these interventions in employer‐sponsored health plans or implementing medication adherence support services in the workplace. Employers may find that the reductions in costs due to averted health care use, and productivity losses may exceed program costs. 5 , 55 This study has several limitations. First, the MarketScan CCAE database is not a random sample, and the results may not be fully generalizable to all individuals with commercial insurance. Second, our results excluded individuals without insurance or those publicly insured, such as Medicare and Medicaid beneficiaries. If individuals with public insurance or without insurance exhibit different patterns in medication adherence, our results may not be universally applicable. Additionally, our study was restricted to individuals aged 18 to 64 years, limiting generalizability to older (≥65 years) or younger (≤17 years) individuals. Third, our inclusion criteria required continuous enrollment, and individuals with gaps in insurance coverage or job disruptions during the period were not captured in our analysis. Fourth, although the MPR is a widely used method for measuring medication adherence, it may overestimate adherence for individuals who routinely refill medications early. MPR, based on drug claims data, captures prescription filling behavior but does not directly measure individuals' medication‐taking behavior. Additionally, the MPR calculation used in this study implemented total days of drug supply in 2019 without capturing the different fill durations (eg, 30, 60, 90, 120+ days) for antihypertensives. 32 This limitation may influence the adherence rates observed. Fifth, our study focused exclusively on adherence to antihypertensives, and did not capture adherence to medications for other conditions that may impact health, and subsequent health care use and labor productivity. Sixth, although we conducted a sensitivity analysis using the propensity score overlap weighting method, with all results remaining consistent, due to the limited information available in the claims database, we were unable to include other potential confounding factors such as individuals' education level, income, race, and poverty level. Seventh, the cross‐sectional nature of our study may not capture longer‐term effects of medication adherence. The observed higher associations on STD compared with LTD may be attributed to cross‐sectional data; examining longitudinal data might reveal different associations with LTD. Eighth, STD costs were calculated using the same average hourly wages for working adults in the United States. The decrease in the STD and LTD costs associated with medication adherence was notably higher for men and those living in urban areas. Given the higher hourly wages for men and urban residents, our estimates of productivity‐related costs might be underestimated. Because our estimates are derived from interactions of medication adherence, sex, and urban indicators, we could not account for different hourly wages by sex and urbanicity of residence. Our study focused on outcomes available from administrative claims data and employer‐provided payroll system; consequently, we could not look at individuals' health outcomes. Future research endeavors could delve into the association of medication adherence with health outcomes, productivity‐related outcomes, and other nonmedical costs for individuals who are uninsured or publicly insured, extending the analysis over a longer time horizon. Additionally, our study focused solely on pre‐COVID‐19 pandemic periods, whereas future studies might broaden the scope to include pandemic periods when medication adherence may have been impacted by interruptions and delays in health care services. 56 Future studies may explore the associations of medication adherence across different racial and ethnic groups to further enhance our understanding of disparities in health outcomes. Lastly, our results showed the lowest (highest) medication adherence rates in the South (Northeast) Census regions. Given the stroke belt in the southeastern region, future studies may explore the potential differences in medication adherence and their impacts on economic burdens and stroke prevalence. CONCLUSIONS Our study identified the associations of antihypertensive medication adherence with total medical costs, health care use, and labor productivity, including number of sick absences, STD, and LTD. Medication adherence was associated with fewer ED visits and inpatient admissions, lower total medical costs, and fewer sick absences, STD, and LTD. These findings highlight not only the potential benefits of medication adherence in reducing medical costs but also improving productivity‐related outcomes. The results emphasize the potential for implementing and expanding programs to support medication adherence among individuals with hypertension covered by commercial insurance. Sources of Funding None. Disclosures None. Supporting information Tables S1–S15 Figures S1–S4