Cost-effectiveness of Hypertension Treatment by Pharmacists in Black Barbershops

Abstract

Background

In the Los Angeles Barbershop Blood Pressure Study (LABBPS), pharmacist-led hypertension care in Los Angeles County Black-owned barbershops significantly improved blood pressure control in non-Hispanic Black men with uncontrolled hypertension at baseline. In this analysis, 10-year health outcomes and healthcare costs of one year of the LABBPS intervention versus control are projected.

Methods

A discrete event simulation of hypertension care processes projected blood pressure, medication-related adverse events, fatal and non-fatal cardiovascular disease events, and non-cardiovascular disease death in LABBPS participants. Program costs, total direct healthcare costs (2019 USD), and quality-adjusted life years (QALYs) were estimated for the LABBPS intervention and control arms from a healthcare sector perspective over a 10-year horizon. Future costs and QALYs were discounted 3% annually. High and intermediate cost-effectiveness thresholds were defined as <$50,000 and <$150,000 per QALY gained, respectively.

Results

At 10 years, the intervention was projected to cost an average of $2,356 (95% uncertainty interval [UI] −$264-$4,611) more per participant than the control arm and gain 0.06 (95% UI 0.01-0.10) QALYs. The LABBPS intervention was highly cost-effective, with a mean cost of $42,717 per QALY gained (58% probability of being highly and 96% of being at least intermediately cost-effective). Exclusive use of generic drugs improved the cost-effectiveness to $17,162 per QALY gained. The LABBPS intervention would be only intermediately cost effective if pharmacists were less likely to intensify antihypertensive medications when systolic blood pressure was ≥150 mmHg or if pharmacist weekly time driving to barbershops increased.

Conclusions

Hypertension care delivered by clinical pharmacists in Black barbershops is a highly cost-effective way to improve blood pressure control in Black men.

INTRODUCTION

Hypertension prevalence remains higher among non-Hispanic Black men than in any other racial or ethnic group in the U.S.^{1, 2} Hypertension awareness and treatment have plateaued in the U.S. since 2010, and Black men continue to have worse blood pressure control and higher hypertension-related cardiovascular disease mortality rates compared with other groups.^{3, 4} Improving hypertension control among Black men is a national priority.

Healthcare interventions delivered in community settings outside of the traditional medical clinic that leverage established relationships may overcome mistrust due to racism in the healthcare system and improve health outcomes in Black men.^5-11 The Los Angeles Barbershop Blood Pressure Study (LABBPS) built upon barber-patron relationships and teamwork between barbers and clinical pharmacists to deliver hypertension care in Los Angeles County Black barbershops.^{12, 13} The LABBPS randomized barbershops and their Black patrons with uncontrolled hypertension to either clinical pharmacist blood pressure management in the barbershop (intervention) or barber-delivered education alone (control).^{12, 13} After one year, the intervention reduced mean systolic blood pressure by 20.8 mmHg (95% confidence interval [CI] 13.9-27.7 mmHg) relative to the control arm. Blood pressure control <130/80 mmHg was attained by 68% of participants in the intervention arm and 11% of participants in the control arm.

Despite the clinical success of the LABBPS intervention in lowering blood pressure, concerns about the sustainability and scalability of the intervention remain due in part to its complexity and potentially high costs.^{14, 15} This analysis projected the 10-year total healthcare costs and health gains associated with the LABBPS intervention in hypertensive Black men using individual participant data from the one-year trial, and explored how exclusive generic medication use or more efficient intervention design might improve cost-effectiveness.

METHODS

The simulation model and key inputs used to conduct this research are available to interested researchers upon reasonable request and approval by the model team. Interested researchers can submit a 1- to 2-page research proposal and collaboration plan to Dr. Bellows and will be required to sign a Creative Commons agreement.

The LABBPS intervention and results have been described in detail elsewhere.^{12, 13} Briefly, the LABBPS recruited Black male patrons aged 35-79 years with systolic blood pressure ≥140 mmHg from Black-owned barbershops; women and individuals requiring dialysis or chemotherapy were excluded. In the intervention arm, participants received hypertension management for one year from clinical pharmacists combined with barber education. Pharmacists received specialized training and certification in hypertension management and obtained collaborative practice agreements with providers to manage participants’ antihypertensive medications. Pharmacists followed a treatment algorithm that included initiation of two half-standard dose antihypertensive medications when participants were not on any medications at baseline (amlodipine plus a long-acting angiotensin converting enzyme inhibitor or angiotensin receptor blocker, such as, telmisartan) and favored generic medications when covered by insurance. Participants received a $25 incentive per visit with the pharmacist to help cover antihypertensive medication costs and travel expenses. In the control arm, participants received hypertension education from their barber only. A simulation model was created to determine the cost-effectiveness of this intervention.

Simulation Overview

The Blood Pressure Control-Cardiovascular Disease Policy Model (BP-CVDPM) is a hybrid, discrete event simulation model that combines the independently validated Blood Pressure Control Model and the Cardiovascular Disease Policy Model (Figure 1, Figure I in the Supplement).^16-19 The Blood Pressure Control Model predicts long-term blood pressure outcomes by simulating modifiable hypertension care management processes (i.e., visit frequency, blood pressure measurement accuracy, probability of regimen intensification, and medication adherence).^{16, 18} The Cardiovascular Disease Policy Model predicts cardiovascular disease events, survival, quality-adjusted survival, and direct healthcare costs, conditioned on blood pressure and other clinical characteristics.^{17, 19}

Figure 1. Effects of the LABBPS barber-pharmacist intervention on components of the Blood Pressure Control-Cardiovascular Disease Policy Model.

BMI – body mass index, BP – blood pressure, CVD – cardiovascular disease, DBP – diastolic blood pressure, eGFR – estimated glomerular filtration rate, HDL-C – high-density lipoprotein cholesterol, LABBPS – Los Angeles Barbershop Blood Pressure Study, LDL-C – low-density lipoprotein cholesterol, QALYs – quality-adjusted life years, systolic blood pressure.

Notes: The figure shows the components of the Blood Pressure Control-Cardiovascular Disease Policy Model, the modifiable inputs to each component, and the projected outcomes of the model. Simulated blood pressures in Blood Pressure Control Model component are used in the Cardiovascular Disease Policy Model as risk factors that impact clinical and economic outcomes.

The BP-CVDPM simulated the complex healthcare delivery systems and participant-healthcare provider interactions contributing to the LABBPS intervention (Supplemental Methods). Events included in the simulation are: (1) physician office visits, (2) pharmacist visits in the barbershop, (3) medication-related adverse events, (4) medication discontinuation, (5) fatal or non-fatal cardiovascular disease events (heart failure; stroke; and coronary heart disease – including myocardial infarction, angina, and cardiac arrest), and (6) non-cardiovascular disease death. When any of these events occurs, the model stores the associated clinical and healthcare cost outcomes, updates patient characteristics, and determines the time to the next event. In the absence of any events within a given year, the model updates the patient’s characteristics, cardiovascular disease risk factors, and, subsequently, the predicted time to a cardiovascular disease event given these changes.

Simulated Population

Individual participant data from the one-year LABBPS were used in BP-CVDPM simulations. To create lifetime cardiovascular disease risk factor trajectories for each LABBPS participant, participants were matched 1:1 to non-Hispanic Black men in the National Heart, Lung, and Blood Institute (NHLBI) Pooled Cohort Study, for whom lifetime cardiovascular disease risk factor trajectories were estimated previously.^{20, 21} In the primary analysis, 263 LABBPS participants were matched to NHLBI Pooled Cohorts Study participants on baseline characteristics: age (+/− 5 years), systolic blood pressure (+/− 5 mmHg), diastolic blood pressure (+/− 5 mmHg), use of antihypertensive medications (yes or no), body mass index (+/− 3 kg/m²), diabetes status, and current cigarette smoking status (current, former, or never). In a scenario analysis, participants were only matched on systolic blood pressure (+/− 5 mmHg) and the use of antihypertensive medications (yes or no) in order to simulate all 319 LABBPS participants (Supplemental Methods). To generate stable cost-effectiveness estimates, cohorts of 10,000 LABBPS participants were simulated in all analyses by sampling with replacement from the matched population, assuming equal probability of being sampled (Supplemental Methods).

Modeled Interventions

The one-year intervention and control arms from the LABBPS were compared. In the intervention arm, participants had the same mean frequency of blood pressure management visits with a pharmacist in the barbershop as in the LABBPS, while continuing usual care with their primary care physician, which provided more opportunities for medication titration. Additionally, the intervention arm had a higher probability of medication titration during visits with the pharmacist and improved medication adherence (see Calibration and Validation section below). In the control arm, participants received encouragement and hypertension education from their barber, in addition to usual care with their regular primary care physician. In the primary analysis, participants were assumed to return to usual care alone after the year-long intervention and were subject to the same processes of hypertension care (number of physician visits, probability of intensification, adherence) for the remainder of the 10-year time horizon.

Model Inputs

Inputs regarding time, resources used, and other hypertension care processes in the LABBPS intervention arm (e.g., time spent performing clinical care, administrative tasks, patient communication, and travel between barbershops) were derived from semi-structured interviews with the trial pharmacists (Supplemental Methods). All model inputs related to the LABBPS intervention are reported in Table 1. Other model inputs (e.g., usual care medication adherence, probability of adverse drug reactions, time between physician office visits, probability of cardiovascular disease events, medication costs) were derived from the published literature and analyses of publicly available national data sources (Tables I -V in the Supplement). Total hospitalization costs for cardiovascular disease and serious adverse events were calculated by multiplying age-stratified, survey-weighted means from the National Inpatient Sample by a published professional fee ratio.^{22, 23} Data inputs were specific to U.S. non-Hispanic Black men whenever feasible; when these were not available, inputs for the general U.S. population were used.

Table 1.

Los Angeles Barbershop Blood Pressure Study (LABBPS) Intervention Model Inputs.

LABBPS Intervention Component	Base-case	Lower Range	Upper Range	Source
Weeks Between Pharmacist Visits when Measured Blood Pressure was:*
<130/80 mmHg	4.50	3.00	6.50	LABBPS Individual Participant Data12, 13
≥130/80 mmHg	2.00	1.00	4.00
Probability of Antihypertensive Regimen Intensification by Pharmacist when Measured Systolic Blood Pressure was
130-134 mmHg	0.16	0.13	0.65	LABBPS Pharmacist Interview, Calibrated
135-139 mmHg	0.29	0.13	0.90
140-149 mmHg	0.31	0.21	1.00
≥150 mmHg	0.95	0.33	1.00
Relative Reduction in Non-Adherence with LABBPS Intervention	0.80	0.50	1.00	Calibrated, Assumption
Intervention Staff Wages
Specialty training hours for each pharmacist	26.00	0.00	100.00	LABBPS Pharmacist Interviews
Hours driven per week for each pharmacist	8.75	7.00	17.50
Administrative task hours per week for each pharmacist	19.09	5.00	25.00
Telephone follow up hours per week for each pharmacist	2.13	0.54	3.25
Hours per visit with pharmacist when measured blood pressure was
<130/80 mmHg	0.25	0.21	0.75
≥130/80 mmHg	0.50	0.25	0.88
Hours per initial visit with pharmacist	0.71	0.42	1.00
Administrative assistant hours per week (one full-time assistant per two pharmacists)	40.00	20.00	80.00
Pharmacist hourly wage†	$60.64	$42.21	$77.52	Bureau of Labor Statistics
Administrative assistant hourly wage†	$18.69	$11.87	$30.88
Use of barbershop station (one per Intervention barbershop)	$1,000.00	$500.00	$2,000.00	LABBPS Pharmacist Interviews
Employee benefit rate	30.00%	-	-	Assumption
Driving Reimbursement
Miles driven per week for each pharmacist	262.00	209.60	524.00	LABBPS Pharmacist Interviews
Reimbursement rate per mile driven	$0.58	$0.20	$0.58	Internal Revenue Service
Hypertension Certification Test for Each Pharmacist	$600.00	$0.00	$1,000.00	American Society for Hypertension
Participant Financial Incentive Per Pharmacist Visit	$25.00	$0.00	$50.00	Victor et al. (2018),12, Victor et al. (2019),13 Range Assumption
Cost Per Mobile Blood Pressure Machine (one per Intervention barbershop)	$995.00	$746.25	$1,243.75	Soma Technology§, Range +/−25%
Point-of Care Mobile Laboratory Testing
Laboratory testing machine (one per pharmacist)	$10,162.59	$7,621.94	$12,703.24	Abbott¶, Range +/−25%
Laboratory test (per test)	$11.24	$8.43	$14.05
Laboratory test frequency (number of visits between tests)‡	2.00	0.00	5.00	LABBPS Individual Participant Data12, 13

LABBPS – Los Angeles Barber Blood Pressure Study.

^*The time to the next visit was randomly sampled from a conditional gamma distribution after each LABBPS visit depending on blood pressure control at the last visit. The SD for the gamma distributions was equal to the mean.

^†Base-case is equal to median hourly wage, range derived from the 10^th and 90^th percentiles (2019 USD).

^‡Number of LABBPS visits that occur between laboratory tests (2 visits between laboratory tests means tested every third visit).

^§Soma Technology: manufacturer of BP machine used in trial, depreciated assuming replacement every 3 years

^¶Abbott: manufacturer of point of care laboratory devices, depreciated assuming replacement every 3 years

Notes: The table shows the model inputs for the LABBPS intervention.

Outcomes

Primary outcomes were mean direct medical costs (2019 USD), mean quality-adjusted life years (QALYs), and incremental cost-effectiveness ratios (ICERs) over 10 years. Costs and QALYs were also disaggregated into their individual components (Supplemental Methods). Chronic quality of life estimates were calculated using a published formula, including adjustments for age, sociodemographic characteristics, comorbidities, and cardiovascular disease history.²⁴ Short-term quality of life decrements were included for acute cardiovascular disease and serious adverse events. A healthcare sector perspective was adopted, including all direct medical costs regardless of payer type, and future costs and QALYs were discounted 3% annually.²⁵ Cost-effectiveness thresholds of <$50,000, $50,000 to <$150,000, and ≥$150,000 per QALY gained defined highly, intermediately, or not cost-effective, respectively, as recommended by the American College of Cardiology and American Heart Association.²⁶

Additional projected clinical outcomes included: proportion with controlled blood pressure (<130/80 mmHg), cardiovascular disease (including heart failure, stroke, and coronary heart disease) incidence rate, serious medication-related adverse events, and all-cause and cause-specific mortality over 10 years.

Calibration and Validation

The BP-CVDPM was calibrated to reproduce contemporary cardiovascular disease incidence and total event rates with usual care for non-Hispanic Black men from the NHLBI Pooled Cohorts Study, the National Inpatient Sample, and the original, population-dynamic version of the U.S. Cardiovascular Disease Policy Model (Figures II -III in the Supplement). The calibrated model reproduced all-cause, non-cardiovascular disease, and cardiovascular disease mortality rates reported by the Centers for Disease Control and Prevention Wide-ranging ONline Data for Epidemiologic Research (CDC WONDER) for U.S. non-Hispanic Black men (Figure IV in the Supplement).

Two LABBPS hypertension processes of care were calibrated to reproduce both the blood pressure outcomes observed and number of antihypertensive medications used in the trial: (1) probability of antihypertensive medication intensification when blood pressure was uncontrolled and (2) patient medication adherence, which was defined as taking medications exactly as prescribed and was not measured in the trial. Medication adherence in the control arm was derived from published meta-analyses and varied by number of antihypertensive medications (Supplemental Methods).^{27, 28} Calibrated medication adherence was similar to other published estimates of pharmacist interventions.^{29, 30} Finally, the reduction in cardiovascular disease events and mortality based on the achieved systolic blood pressures with the LABBPS intervention and control arms were validated against a published meta-analysis (Supplemental Methods).³¹

Statistical Analysis

Main Analysis

The primary analysis used a 10-year time horizon to capture meaningful differences in cardiovascular disease outcomes and costs relevant to healthcare payers in the U.S. However, outcomes over time horizons varying from one year to a lifetime (age 100 or death) were also examined in sensitivity analyses. For the primary analysis, the mean outcomes and 95% uncertainty intervals (UI) were estimated by running 500 randomly sampled cohorts and probabilistically sampling model inputs from pre-specified distributions. The uncertainty around the projected ICER was estimated by determining the proportion of the 500 probabilistic iterations that resulted in an ICER <$50,000 or <$150,000 per QALY gained.²⁶

Three primary scenario analyses examined the impact on cost-effectiveness of the intervention with: (1) exclusive use of generic antihypertensive medications as the primary analysis included both brand and generic medications to represent current national utilization of antihypertensive medications; (2) an “optimized intervention” that assumed a synergistic effect of efficiency gains resulting in less time spent on intervention tasks (i.e., pharmacist administrative and driving time, administrative assistant time), lower equipment costs (blood pressure and point-of-care laboratory machines and supplies), exclusive use of generic medications, and no participant incentive costs; and (3) shortening the one-year LABBPS intervention to 26 weeks, as a substantial proportion of LABBPS participants achieved blood pressure control by 6 months and treatment effects between 6 and 12 months were similar (Table VI in the Supplement).^{12, 13}

Sensitivity and Other Scenario Analyses

In one-way sensitivity analyses, the LABBPS intervention parameters to which the model was most sensitive were identified by independently varying all of these parameters across plausible ranges while all others were held constant at their base-case value. Plausible ranges were defined as the maximum and minimum values from pharmacist interviews, 95% confidence interval from the LABBPS individual participant data, reported 10^th and 90^th percentiles for salaries, or assumed percentage or value changes (Table 1). The following scenario analyses were also performed: (1) no improvement in medication adherence with the LABBPS intervention, (2) extending improved medication adherence to five years, and (3) increasing the duration of the intervention to 10-years (Table VI in the Supplement).

This analysis followed best practice recommendations from the Second Panel on Cost-Effectiveness in Health and Medicine (Tables VII-VIII in the Supplement).²⁵ The BP-CVDPM was constructed in TreeAge Pro 2019 (TreeAge Software Inc, Williamstown, Massachusetts) with additional analyses performed using R version 3.4.4 (Vienna, Austria). The study protocols for the LABBPS and the NHLBI Pooled Cohorts Study were approved by the Institutional Review Boards at each participating institution and all participants provided written informed consent. This study was reviewed and approved by the Institutional Review Board at Columbia University Irving Medical Center.

RESULTS

Calibration and Validation

At baseline, mean age was 55.2 (95%UI: 55.0-55.4) years, mean systolic blood pressure was 150.7 (150.5-150.9) mmHg, and mean diastolic blood pressure was 88.8 (88.6-89.0) mmHg (Table IX and Table X in the Supplement). At one year, the simulated reduction in systolic blood pressure for the intervention arm relative to the control arm was 22.5 (95%UI 22.0-23.0) mmHg, compared with 22.2 (95%CI 18.9-25.6) mmHg observed in matched LABBPS participants (Table X in the Supplement). Simulated hazard ratios for mortality and cardiovascular disease for the intervention vs. control were similar to a published meta-analysis (Table XI in the Supplement).³¹

Main Analysis

After the simulated one-year intervention, the projected difference in blood pressure control between the intervention and control arms was somewhat attenuated by returning to usual care alone in subsequent years. However, at 10 years, the projected blood pressure control rate remained higher in the intervention arm (60.4% [95%UI 51.5%-69.0%]) than in the control arm (24.7% [95%UI 18.4%-32.4%]) (Table XII and Figure V in the Supplement). The intervention arm was projected to experience fewer cardiovascular disease events (17.4% [95%UI 13.9%-21.8%]) than the control arm (24.8% [95%UI 20.1-31.2%]), with an incidence rate ratio of 0.67 (95%UI 0.61-0.73) (Figure VI and VII in the Supplement).

Compared with the control arm, the LABBPS intervention was projected to cost $2,356 (95%UI −$264-$4,611) more per participant over 10 years and gain a mean of 0.06 QALYs (95%UI 0.01-0.10) (Table 2). The resulting ICER was $42,717 per QALY gained (58.2% probability of being highly and 95.6% probability of being at least intermediately cost-effective) (Table 2, Figure 2, Figure VIII in the Supplement). First-year programmatic costs of the LABBPS intervention ($3,713 per participant) were accompanied by higher medication, adverse event, and background healthcare costs over 10 years, which were partially offset by reduced cardiovascular disease costs (Figure 3, Figure IX and Table XIII in the Supplement). The largest programmatic LABBPS intervention cost was pharmacist time ($2,623 per participant), of which, 45.0% was driving to barbershops and providing clinical care (barbershop visits and telephone follow up) (Figure 3).

Table 2.

Costs and Effectiveness of Los Angeles Barbershop Blood Pressure Study (LABBPS) Intervention at 10 Years.

Scenario	Mean Costs (2019 USD)			Mean QALYs			ICER	Probability cost- effective
	Intervention	Control	Incremental (95%UI)	Intervention	Control	Incremental (95%UI)		At $50k per QALY gained	At $150k per QALY gained
Primary Analysis	$89,154	$86,797	$2,356 (−$264 to $4,611)	6.387	6.332	0.055 (0.008 to 0.097)	$42,717	58.2%	95.6%
Generic Drugs	$86,455	$85,505	$950 (−$1,705 to $3,375)	6.372	6.317	0.055 (0.007 to 0.099)	$17,162	86.4%	96.6%
Optimized Intervention	$84,619	$85,410	−$791 (−$3,704 to $1,612)	6.375	6.318	0.057 (0.008 to 0.101)	Dominant*	98.2%	99.4%
26-Week Intervention	$87,574	$86,658	$916 (−$1,500 to $3,480)	6.369	6.319	0.050 (0.002 to 0.089)	$18,300	84.6%	95.0%

ICER – incremental cost-effectiveness ratio, LABBPS – Los Angeles Barbershop Blood Pressure Study, QALY – quality-adjusted life year, 95%UI – 95% uncertainty interval (i.e., 2.5^th to 97.5^th percentiles).

^*Dominant – the intervention costs less and is more effective than control in this scenario; ICERs are not presented when a strategy is dominant.

Notes: The table shows the mean total direct healthcare costs, mean QALYs, ICER, probability of the LABBPS intervention being cost-effective. The four scenarios vary key components of the simulation model: (1) the primary analysis, (2) shortening the LABBPS intervention to 26 weeks, (3) using generic drug costs, and (4) an “optimized intervention” that uses both generic drug costs, more efficient delivery of the LABBPS intervention (i.e., lower limit for pharmacist administrative and driving time, administrative assistant time, lower equipment costs for blood pressure and lab machine), and no participant incentive costs.

Figure 2. Probability LABBPS intervention is cost-effective vs. control at 10 years.

LABBPS – Los Angeles Barbershop Blood Pressure Study, QALY – quality-adjusted life years.

Notes: The figure shows the probability of the LABBPS intervention being cost-effective vs. the control arm across increasing cost-effectiveness thresholds. The four lines show different scenarios varying key components of the simulation model: (1) the primary analysis, (2) shortening the LABBPS intervention to 26 weeks, (3) using generic drug costs, and (4) an “optimized intervention” that uses both generic drug costs, more efficient delivery of the LABBPS intervention (i.e., lower limit for pharmacist administrative and driving time, administrative assistant time, lower equipment costs for blood pressure and lab machine), and no participant incentive costs. The probabilities were derived from 500 probabilistic iterations of the model for each scenario.

Figure 3. Intervention costs and cumulative cost outcomes over time in the primary analysis.

A) One-year LABBPS intervention costs by type B) Cumulative incremental costs of the intervention vs. control arms over time

BP – blood pressure, CVD – cardiovascular disease, LABBPS – Los Angeles Barbershop Blood Pressure Study, QALYs – quality-adjusted life years.

Notes: The figure shows the projected cumulative incremental cost and quality-adjusted life year outcomes over 10 years for the Los Angeles Barbershop Blood Pressure Study intervention vs. control from the Blood Pressure Control-Cardiovascular Disease Policy Model. (A) LABBPS mean intervention costs per participant by cost component; (B) cumulative mean incremental costs per participant by type of cost (≥$0 increased costs with intervention, <$0 decreased costs). Costs are in 2019 USD.

When using generic drugs exclusively, the ICER decreased to $17,162 per QALY gained (86.4% probability of being highly cost-effective) (Figure X in the Supplement). The optimized intervention scenario resulted in the LABBPS intervention being dominant, that is, it was both less costly and more effective (98.2% probability of being highly cost-effective). Compared with the one-year intervention arm in the primary analysis, the 26-week intervention had lower programmatic costs, but resulted in worse blood pressure control and more cardiovascular disease events at 10 years (Table XII in the Supplement). The resulting ICER for the 26-week intervention compared with the control arm was $18,300 per QALY gained (84.6% probability of being highly cost-effective).

Sensitivity and Other Scenario Analyses

Cost-effectiveness results were most sensitive to probability of medication intensification when systolic blood pressure was ≥150 mmHg and pharmacist administrative and driving time, though even when each of these inputs was less favorable, the intervention remained at least intermediately cost-effective (Figure XI in the Supplement). The LABBPS intervention was highly or intermediately cost-effective in nearly all scenarios, including when extending the projected time horizon following the one-year intervention (Table XIV in the Supplement). Over a lifetime, the ICER for the one-year intervention arm was $7,884 per QALY gained relative to the control arm. However, when the duration of the intervention was extended to 10 years (i.e., a pharmacist managed each participant’s blood pressure in a barbershop for 10 years), the intervention arm was no longer cost-effective (ICER $361,004 per QALY gained).

DISCUSSION

Compared with barber education-only, the LABBPS barber-pharmacist hypertension management intervention is a highly cost-effective way to control hypertension and reduce cardiovascular disease morbidity and mortality in Black men. The projected high upfront costs of the LABBPS intervention were partially offset by preventing cardiovascular disease events and would cost $42,717 per QALY gained over 10 years. Exclusively using generic drugs improved cost-effectiveness to $17,162 per QALY gained. An “optimized intervention” with improved efficiency of care delivery and a shorter 26-week intervention could further improve cost-effectiveness.

There are a limited number of recent economic analyses of pharmacist-delivered and team-based hypertension management interventions.^32-34 In one analysis, one year of team-based hypertension care provided by a nurse or pharmacist reduced systolic blood pressure 8.1 mmHg more than usual care, and the intervention cost $974 per patient (inflated to 2019 USD) for office and telephone visits and patient time. Over 10 years, it was projected to increase total healthcare costs nationally by $2.6 billion (inflated to 2019 USD) and gain 922.6 thousand QALYs in U.S. adults.³² Another analysis of a pharmacist-managed home blood pressure telemonitoring intervention reduced systolic blood pressure by 9.7 mmHg compared with usual care over one year, and increased healthcare costs by $1,173 per patient (inflated to 2019 USD), including pharmacist visits, telemonitoring, and savings in medical care.^{33, 34} This study did not include long-term projections or an estimated impact on quality of life. The LABBPS intervention differs from these in that it was more intensive and delivered hypertension care to Black men exclusively in a trusted community setting. These factors may explain both the higher costs and why the observed blood pressure decreases were substantially larger in the barber-pharmacist intervention arm of the LABBPS (>20 mmHg systolic blood pressure reduction) than in other studies (<10 mmHg reduction). The lifetime projected QALY gains with the intervention in the current analysis are similar to those in an economic analysis of the Systolic Blood Pressure Intervention Trial (SPRINT), which compared an intensive systolic blood pressure goal of <120 mmHg vs. a standard goal of <140 mmHg in high cardiovascular disease risk patients.³⁵ The current LABBPS analysis projected a gain of 0.35 QALYs over a lifetime (Table XIV in the Supplement) compared with a projected gain of 0.27 QALYs in SPRINT. The projected lifetime QALY gains in SPRINT were similar when restricting to a population aged <75 years, more closely resembling the age of the LABBPS population.³⁵

Black-owned barbershops may be a particularly effective setting to deliver health promotion, improve hypertension control, and reduce cardiovascular disease disparities. The barbershop represents social capital that cannot be leveraged by home- or clinic-based hypertension management programs.^36-39 This cost-effectiveness analysis accounted for individual health gains and healthcare costs, but it did not account for the value of improved health equity in society as a whole when an intervention gains health for Black men specifically. Non-monetary societal benefits of interventions that reduce health disparities may be hard to quantify, but future health economic evaluations should consider setting higher willingness-to-pay thresholds for interventions that decrease health disparities, thereby making the added societal value of these interventions explicit.^{26, 40, 41} The current study projected future costs and health outcomes of the LABBPS intervention using a validated computer simulation model, which may also be used to refine and improve the efficiency of the LABBPS intervention when it is replicated and scaled up broadly.^{14, 15} One concern raised as a potential barrier to widespread LABBPS implementation is the specialty training of clinical pharmacists.¹⁵ In this analysis, the cost of specialty training and certification was included; the results suggest that long-term health benefits and avoided healthcare costs of the LABBPS offset these upfront training costs.

Interventions like the LABBPS have the potential to reduce cardiovascular disease disparities.⁴² The disproportionate burden of Coronavirus Disease 2019 (COVID-19) in Black communities provided a stark reminder of the devastating impact of healthcare disparities on minority communities.⁴³ Higher burden and worse control of chronic diseases likely contributed to the two-fold higher COVID-19 death rate in Black, compared with White communities.⁴⁴ COVID-19 has highlighted the need for public health programs to partner with trusted community partners outside of traditional healthcare settings.⁴⁵ Community partnerships, such as with barbershops, may improve access to care, disease screening and education, and control of other underlying chronic diseases that build resilient communities, and may be an effective step toward narrowing Black-White health disparities that result in part from centuries of systemic racism in medicine.^{10, 37, 39}

The current analysis benefited from access to the complete LABBPS trial data and research team. Individual participant data was directly simulated and trial-related inputs were obtained (e.g., pharmacist time) directly from the two LABBPS clinical pharmacists. However, outcomes after the one-year trial and the effect of transitioning back to usual care on blood pressure are uncertain. Therefore, a conservative approach was adopted and it was assumed that after the one-year trial period, processes of hypertension care management returned to usual care over the remaining years of the 10-year primary time horizon. Additionally, antihypertensive adherence information was not directly collected in the trial and the long-term impact of the intervention on pill-taking adherence is uncertain. Another limitation is that, although most BP-CVDPM inputs were specific to Black men, some were based on the general population. However, the model was calibrated to reproduce observed contemporary cardiovascular disease event and mortality rates in Black men. These findings may also be somewhat limited in scope as a healthcare sector perspective was used, which only considers direct healthcare costs, rather than a societal perspective, which may include indirect costs such as improvements in productivity. Finally, cost-effectiveness estimated for the LABBPS may not be generalizable to other U.S. communities, as it was specific to Los Angeles County and was driven in part by the high underlying risk of cardiovascular disease in Black men. It is unknown how an intervention like LABBPS might impact Black men in different locations across the U.S. However, the main results and exploratory analyses provide a useful framework to guide implementation of the LABBPS intervention and other community-based hypertension control interventions on a larger scale. National scale health impact projections and specific considerations regarding the national scale-up of the LABBPS intervention are described in another report published in this issue.⁴⁶

CONCLUSION

A Black-owned barbershop-led, pharmacist-based hypertension control intervention is a cost-effective strategy to reduce cardiovascular disease morbidity and mortality in U.S. non-Hispanic Black men. Exclusive use of generic drugs could further improve the cost-effectiveness of LABBPS and the feasibility for widespread implementation in the U.S.

Supplementary Material

Supplemental Publication Material

Table I. Blood Pressure Control-Cardiovascular Disease Model Inputs.

Table II. Total Annual Background Healthcare Costs Compared with US Department of Health and Human Service Estimates (2019 USD).

Table III. Codes Used to Identify Cardiovascular Disease Events and Selected Comorbidities.

Table IV. Codes Used to Identify Serious Adverse Events.

Table V. Key Assumptions and Sources for Primary Analysis.

Table VI. Key Assumptions and Sources for Alternate Scenarios.

Table VII. Reporting Checklist for Cost-Effectiveness Analyses from the Second Panel on Cost-Effectiveness in Health and Medicine.

Table VIII. Formal Health Care Sector Impact Inventory Assessment.

Table IX. LABBPS Validation of Baseline Characteristics.

Table X. LABBPS Validation of 1-Year Observed Outcomes.

Table XI. LABBPS CVD and Mortality Risk Reduction External Validation.

Table XII. Blood Pressure and CVD Event Outcomes at 10 years Across Scenarios.

Table XIII. Detailed Mean Cost and Effectiveness Outcomes by Type at 10 Years for the Primary Analysis.

Table XIV. Alternate Scenario Analyses.

Figure I. Detailed Blood Pressure Control-Cardiovascular Disease Policy Model Structure.

Figure II. Calibration and Internal Validation of CVD Incidence Rates.

Figure III. Calibration and Internal Validation of CVD Total Event Rates.

Figure IV. Calibration and Internal Validation of Mortality Incidence Rates.

Figure V. Blood Pressure Control for the LABBPS Intervention vs. Control Over 10 Years.

Figure VI. Cumulative Incidence of CVD Events and Mortality for the LABBPS Intervention vs. Control Over 10 Years.

Figure VII. Mortality Rates for the LABBPS Intervention vs. Control Over 10 Years.

Figure VIII. Cost-Effectiveness Scatter Plots Across Scenarios.

Figure IX. Cumulative QALY Outcomes Over Time and Incremental Cost-Effectiveness Ratio the Primary Analysis.

Figure X. Incremental Costs Over 10 Years by Cost Type Across Scenarios.

Figure XI. One-Way Sensitivity Analysis of LABBPS Intervention Components in the Primary Analysis.

CLINICAL PERSPECTIVE.

1. What is new?

• Hypertension care provided by pharmacists in Black-owned barbershops is a highly cost-effective way to manage blood pressure in Black men and reduce cardiovascular risk over ten years, costing an average of $42,717 per quality-adjusted life-year (QALY) gained.

• The exclusive use of generic antihypertensive medications improved the cost-effectiveness of the intervention, resulting in an average cost of $17,162 per QALY gained.

• This cost-effectiveness analysis identified ways to design efficient and scalable community-based hypertension control programs.

2. What are the clinical implications?

• Black-owned barbershops may be a particularly effective and cost-effective setting to improve hypertension control and reduce cardiovascular disease disparities in Black men.

• Delivery of healthcare outside of traditional clinical settings and leveraging trusted community partnerships may improve health outcomes when traditional medical care fails to reach populations at risk.

Non-standard Abbreviations and Acronyms

LABBPS

Los Angeles Barbershop Blood Pressure Study

BP-CVPDM

Blood Pressure – Cardiovascular Disease Policy Model

ICER

Incremental Cost-Effectiveness Ratio

QALY

Quality-Adjusted Life Year

NHLBI Pooled Cohorts

National Heart Lung and Blood Institute

SPRINT

Systolic Blood Pressure Intervention Trial

CDC WONDER

Centers for Disease Control and Prevention Wide-ranging ONline Data for Epidemiologic Research