Blood Pressure Outcomes Among US Veterans Initiating Hypertension Treatment, 2014–2023

Abstract

Background:

Amid persistently low blood pressure (BP) control rates and pervasive therapeutic inertia, we evaluated trends in pre- and post-treatment BP before and after two major disruptions (the 2017 target-lowering guideline and the COVID-19 pandemic) to assess their impact on early hypertension management.

Methods:

This retrospective cohort study of national Veterans Health Administration (VHA) data included outpatient adults newly diagnosed with hypertension and starting antihypertensive medication between 11/13/2014–5/31/2023 (index date). Patients were stratified into three periods corresponding with pre-guideline/pre-pandemic (Period 1); post-guideline/pre-pandemic (Period 2); and post-guideline/post-pandemic (Period 3) and 3 pre-treatment systolic BP (SBP) groups based on the ≥2-measurement-average in the 90-days pre-index (<140, 140–160, or ≥160 mmHg). Across periods, multivariable analyses evaluated: 1) mean post-treatment SBP (mean of measurements from days 180–365 post-index); and 3) post-treatment BP control (<140/90 or <130/80 mmHg).

Results:

Among 271,496 Veterans (mean 62 years, 92% male, 66% Non-Hispanic White); 39%, 32%, and 29% were in Periods 1, 2, and 3, respectively. Adjusted post-treatment SBP across Periods was 128, 135, and 142 mmHg for the <140, 140–160, and ≥160 groups, respectively. Rates of post-treatment BP control <140/90 mmHg across Periods 1, 2, and 3, respectively, were: 82.3%, 83.3%, and 84.2% (SBP <140 group); 64.1%, 66.1%, 67.3% (140–160 group); and 46.4%, 47.0%, 48.6% (≥160 group). For BP control <130/80, rates were: 38.2%, 40.0%, 39.9% (SBP <140 group); 20.7.%, 22.0%, 22.6% (140–160 group); and 15.1%, 15.4%, 15.2% (≥160 group).

Conclusions:

Despite the target-lowering guideline and the care-disrupting pandemic, BP levels and control among Veterans remained largely unchanged. At one year, only half to two-thirds achieved BP <140/90 mmHg, and few reached <130/80 mmHg, underscoring persistent clinical inertia and the need to improve early hypertension management in the VHA.

Trial registration:

n/a

INTRODUCTION

For decades, the Veterans Health Administration (VHA) has exemplified hypertension care, blending evidence-based treatment with system-wide implementation strategies that pushed blood pressure (BP) control rates to >75% at select facilities in the early 2000s.¹ Yet, after US surveillance data signaled a decrease in population-level BP control starting in 2014,^2,3 there has been scant evidence describing BP control rates among Veterans who newly start antihypertensive therapy, particularly with respect to the stricter targets recommended by contemporary clinical guidelines.

Two major, sequential events may have impacted BP control in the VHA. First, the 2017 American College of Cardiology/American Heart Association (ACC/AHA) BP guideline lowered the threshold for antihypertensive medication initiation and intensification from systolic BP (SBP) ≥140 mm Hg or diastolic BP (DBP) ≥90 mm Hg to SBP/DBP of 130/80 mm Hg for US adults at cardiovascular risk.⁴ Second, the COVID-19 pandemic disrupted clinical care in 2020, reshaping how Americans received healthcare.⁵ While several studies reported modest reductions in BP control in the general population during these times,^6–8 similar VHA studies have included mixed hypertensive populations (treated and untreated alike) or prevalent antihypertensive medication users, and without separately examining the impacts of each event on BP control among new treatment initiators.^9,10

Clinical inertia, the failure to start or intensify treatment despite evidence of uncontrolled disease, occurs in 30–55% of hypertension-related visits and is a central barrier to achieving BP control nationwide.^11–17 Factors contributing to inertia, such as pre-treatment BP level and reliance on monotherapy, may also influence medication adherence, further limiting treatment effectiveness.¹⁸ Understanding BP management in patients newly initiating therapy, who may be especially vulnerable to inertia during periods of change or care disruption, is particularly important because early treatment response strongly influences long-term BP outcomes and cardiovascular risk.^19–23 Prior work has evaluated upstream gaps in rates of hypertension awareness, diagnosis, and treatment in the VA.^9,24,25 To further inform the VA’s future efforts and policies to advance BP control, we sought to evaluate early management among Veterans newly initiating antihypertensive medications using a new-user study design. We evaluated differences in pre- and post-treatment BP levels and one-year BP control and medication adherence in three time periods (pre-guideline/pre-pandemic, post-guideline/pre-pandemic, and post-guideline/post-pandemic). Therefore, we conducted the following analysis to assess differences in pre- and post-treatment BP levels and one-year BP control and medication adherence among Veterans initiating treatment for hypertension in three time periods (pre-guideline/pre-pandemic, post-guideline/pre-pandemic, and post-guideline/post-pandemic). Therapeutic inertia encompasses both failure to initiate and failure to intensify therapy; this study focuses on post-initiation inertia by evaluating early management and outcomes during the first year after treatment start. By providing a current benchmark near the publication of the 2025 AHA/ACC BP Guideline,²⁶ these findings aim to inform the VHA’s future efforts and policies to advance BP control for all Veterans.

METHODS

Data Sources and Oversight

We used data from the VHA Corporate Data Warehouse, which contains information back to 1999 on outpatient and inpatient clinical encounters, patient sociodemographic information, vital signs, laboratory test results, diagnosis codes, and medications filled through VHA pharmacies. The University of Utah institutional review board and Salt Lake City Veterans Affairs Health Care System Research and Development Office approved the current study with a waiver of informed consent. We followed the STROBE reporting guideline for cohort studies (Table S1).²⁷

Study Design and Population

We conducted a new-user cohort study of Veterans who filled a prescription for antihypertensive medication in the VHA outpatient setting from November 13, 2014, through May 31, 2023 (i.e., the index date identification period). The date of each Veteran’s first prescription fill for an oral antihypertensive medication during this period was defined as their index date (N=4,699,410) (Figure S1). We further excluded Veterans who: 1) did not have continuous VHA coverage for at least 2 years prior to the index date (N=876,989); 2) did not have ≥1 inpatient or ≥2 outpatient encounters during the pre-treatment period with a hypertension diagnosis code (401.x, 403.0x, 403.1x, 403.9x; I10, I12.0, I12.9.9) in the 1 year prior to, or 180 days after, the index date (N=911,108)²⁸; 3) filled an antihypertensive medication prior to the index date (N=2,423,299); 4) did not have two valid BP measurements in the 1 year prior to the index date (N=214,618); 5) were <18 years of age or pregnant on the index date (N=39); and 6) were marked as non-Veterans or had a death date prior to the index date (N=1,861). The final sample included 271,496 Veterans.

Data Collection

We collected data on baseline (i.e., pre-index) characteristics from the VHA Corporate Data Warehouse using procedure and diagnosis codes associated with inpatient and outpatient encounters on or before the index date (detailed definitions in Table S2). As treatment recommendations and patterns vary by race and ethnicity, self-reported race and ethnicity was queried from enrollment files and grouped into five mutually exclusive categories: Hispanic, non-Hispanic Black, non-Hispanic White, non-Hispanic Other (inclusive of multi-racial, Asian American, American Indian or Alaskan Native, and Native Hawaiian or Other Pacific Islander), and missing (defined as missing ethnicity OR non-Hispanic, missing race). We also defined a variable indicating the time interval between hypertension diagnosis and the index date, defined as the difference in years between the index date and the first inpatient or outpatient hypertension diagnosis code during the study period.

Exposure

Our research question was whether BP outcomes varied over time, with focus on two key events: the 2017 guideline change and the onset of the COVID-19 pandemic. To characterize the exposure, we first evaluated annualized trends in pre- and post-treatment systolic BP from 2014–2023. This allowed visualization of the acute COVID-19 disruption (2020–2021) and subsequent stabilization (2022–2023). Annualized mean BP values (Figure S2) were unchanged in 2020–2021 followed by re-established stability through 2023, indicating that aggregation of the post-pandemic period would not obscure meaningful within-period differences. Therefore, we selected a 3-level exposure variable to streamline analysis and align with the study aims. Periods were defined by index date (Figure 1, Panel A): 1) pre-guideline/pre-pandemic (November 13, 2014–November 30, 2017; N=104,879), 2) post-guideline/pre-pandemic (December 1, 2017–December 31, 2019; N=86,230), and 3) post-guideline/post-pandemic onset (January 1, 2020–May 31, 2023; N=80,387).

Figure 1: Timeline of the periods and study schema.

^a If two or more blood pressure measurements were recorded up to 90 days prior to the index date, then pre-treatment blood pressure was determined using the mean of these measurements. Otherwise, the number of lookback days was increased up to 365 to obtain two blood pressure measurements.

ACC/AHA: American College of Cardiology/AHA American Heart Association; COVID-19: coronavirus disease 2019; DBP diastolic blood pressure; SBP systolic blood pressure; US United States.

Study Outcomes

The two co-primary outcomes were mean pre-treatment SBP levels and mean one-year post-treatment SBP levels. We secondarily evaluated post-treatment BP control dichotomously at two levels: <140/90 mm Hg (JNC-8 threshold which is still used for Health Effectiveness Data Information Set performance metrics) and <130/80 mm Hg (2017 ACC/AHA threshold). We also evaluated the proportion achieving the individual SBP and diastolic BP (DBP) thresholds separately and evaluated an SBP goal of <150 mm Hg for individuals ≥60 years old (American College of Physicians/American Academy of Family Physicians 2017 threshold²⁹). For SBP and DBP, we selected valid BP measurements taken in outpatient settings that typically manage hypertension (Table S2).^1,30 Pre-treatment BP was defined as the mean of all valid BP measurements, measured on ≥2 separate dates, within 365 days pre-index, and post-treatment BP was defined as the mean of all BP measurements in days 180 through 365 following the index date (Figure 1, Panel B).

Because medication adherence influences BP outcomes and clinical inertia, we secondarily assessed one-year antihypertensive medication adherence using proportion of days covered (PDC) from index to one year post-index.^31,32 We separately defined PDC continuously and categorically at a threshold of 80% as recommended by the Pharmacy Quality Alliance³³ and other studies.^28,34

The administrative end of follow-up was November 2024; longer horizons would preferentially exclude more recent initiators and rely on outcome elements (e.g., death) with greater latency near the cutoff. These analyses focused on early management because guidelines during the study period recommended monthly reassessment/titration until control.^4,35

Statistical Analysis

Descriptive statistics

Characteristics of the population were summarized overall and by period using descriptive statistics (Supplemental Methods). Because treatment initiation depends on pre-treatment SBP levels, we assessed outcomes within strata of pre-treatment SBP (<140, 140–160, and ≥160 mm Hg) for each period. Patients missing baseline values for the continuous variables of total cholesterol, HDL cholesterol, or eGFR values (Table S3) were excluded from corresponding summary statistics. For dichotomous variables, those missing baseline data were assumed to not have the condition of interest, although we created a missing indicator for patients with missing ethnicity or non-Hispanic patients missing race data.

Weighted outcome models

To estimate period-specific outcomes, linear regression modeled continuous post-treatment SBP and DBP levels, quasibinomial regression modeled the continuous PDC outcome, and Poisson regression with robust error variance modeled binary outcomes. We generated two weights using the weightit package: 1) inverse probability of treatment weights (IPTW) to adjust for confounding covariates and potential risk factors of the outcomes; and 2) drop-out weights to address potential bias due to missing post-treatment BP outcomes (Table S4).³⁶ Both weights incorporated sex, race, ethnicity, diabetes, chronic kidney disease (CKD), and history of cardiovascular disease (CVD). Drop-out weights were estimated in the entire sample, then outcome models were weighted and restricted to those with non-missing post-index BP (N=188,705; 69.5%). For medication adherence, outcome models included all Veterans regardless of post-index BP availability. The product of the two weights was used in the regression models, in addition to a categorical variable indicating the calendar month of the index date to account for potential seasonal variation in SBP.³⁷ Standard methods were used to assess covariate balance before and after weighting (Supplemental Methods).³⁸

To estimate period-specific means, proportions, and 95% confidence intervals (CI), we used model-based marginal estimates derived from the weighted generalized linear models using the emmeans package. We conducted pairwise comparisons between periods using adjusted contrasts of the marginal estimates and reported two-sided p-values for the comparisons between period 1 vs. period 2 and period 2 vs. period 3, with p-values <0.05 considered statistically significant (Tukey’s method adjusted for multiple comparisons to maintain the family-wise error rate). The analyses reflect early post-initiation management (1-year BP and adherence), a direct indicator of intensification and support rather than simple initiation rates.

Subgroup and sensitivity analyses

For comparability with prior VHA BP studies,^1,9,10,39 we evaluated mean post-treatment SBP and SBP control in the overall cohort, not stratified by pre-treatment SBP groups. We also assessed subgroups by age, sex, race and ethnicity, diabetes, CVD, CKD, and index combination therapy use. We also conducted a descriptive analysis evaluating overall combined SBP/DBP control by VISN, and conducted a post-hoc sensitivity analysis incorporating VISN as a random effect in the models. Although the primary analysis accounts for inherent differences between patients with and without follow-up BP values using dropout weights, we performed a sensitivity analysis without the dropout weights, similar to previously published studies.^40,41 Finally, because patients may be seen only once annually, or may receive a diagnosis and treatment in one visit, we repeated analyses among a cohort requiring only 1 valid BP measurement pre-index instead of 2.

Analyses were conducted in R (version 4.4.1).

RESULTS

Baseline Characteristics

There were 271,496 Veterans in the current analysis, with a mean (standard deviation, SD) age of 62 (14) years, 8.4% were female, and 65.8% were Non-Hispanic White (Table 1). The median [IQR] number of years between first hypertension diagnosis and the index date across Periods 1, 2, and 3, were 0.8 [IQR 0.0–3.1], 1.0 [0.0–3.7], and 1.0 [IQR 0.0–4.9], respectively (Figure S3 for histograms). Over time, across Periods 1 through 3, Veterans’ average age increased, the proportion who were current smokers decreased, total cholesterol and estimated glomerular filtration decreased, and the proportion with histories of CKD or CVD increased. Initiation of antihypertensive combination therapy decreased, and among those starting combination therapy, the proportion using a fixed-dose combination product decreased. Weighting adequately balanced the periods in terms of baseline characteristics, with all absolute standardized mean differences <0.1 (Figure S4).

Table 1:

Characteristics of Veterans with hypertension who initiated antihypertensive medication, 2014–2023.

Characteristics	Overall	Perioda
		Period 1	Period 2	Period 3
	N = 271,496	N = 104,879	N = 86,230	N = 80,387
Age, years
Mean (SD)	62 (14)	61 (13)	62 (14)	63 (14)
Median (25th, 75th percentile)	64 (53, 71)	63 (52, 69)	64 (53, 72)	64 (54, 74)
Age category, n (%)
18 to 44, n (%)	33,353 (12.3%)	13,271 (12.7%)	10,508 (12.2%)	9,574 (11.9%)
45 to 64, n (%)	112,478 (41.4%)	45,471 (43.4%)	35,019 (40.6%)	31,988 (39.8%)
65 to 74, n (%)	84,947 (31.3%)	33,831 (32.3%)	28,033 (32.5%)	23,083 (28.7%)
≥ 75, n (%)	40,718 (15.0%)	12,306 (11.7%)	12,670 (14.7%)	15,742 (19.6%)
Female, n (%)	22,735 (8.4%)	8,216 (7.8%)	7,063 (8.2%)	7,456 (9.3%)
Race and ethnicity group, n (%)
Hispanic	17,045 (6.3%)	6,352 (6.1%)	5,467 (6.3%)	5,226 (6.5%)
Non-Hispanic Black	49,836 (18.4%)	19,329 (18.4%)	15,954 (18.5%)	14,553 (18.1%)
Non-Hispanic White	178,583 (65.8%)	69,703 (66.5%)	56,505 (65.5%)	52,375 (65.2%)
Non-Hispanic Otherb	8,626 (3.2%)	3,315 (3.2%)	2,717 (3.2%)	2,594 (3.2%)
Missing ethnicity or non-Hispanic and missing race	17,406 (6.4%)	6,180 (5.9%)	5,587 (6.5%)	5,639 (7.0%)
Current smoker, n (%)	48,374 (17.8%)	20,950 (20.0%)	14,052 (16.3%)	13,372 (16.6%)
Time interval between first hypertension diagnosis code and index date, years, median (25th, 75th percentile)	0.9 (0.0, 3.8)	0.8 (0.0, 3.1)	1.0 (0.0, 3.7)	1.0 (0.0, 4.9)
Pre-treatment systolic BP category, n (%)
<140 mm Hg	153,744 (56.6%)	59,713 (56.9%)	49,623 (57.5%)	44,408 (55.2%)
140 to <160 mm Hg	103,577 (38.2%)	39,713 (37.9%)	32,280 (37.4%)	31,584 (39.3%)
≥160 mm Hg	14,175 (5.2%)	5,453 (5.2%)	4,327 (5.0%)	4,395 (5.5%)
Total cholesterol, mg/dL, median (25th, 75th percentile)	179 (152, 208)	181 (155, 210)	178 (151, 207)	178 (150, 207)
HDL-cholesterol, mg/dL, median (25th, 75th percentile)	44 (37, 54)	44 (37, 53)	44 (37, 54)	45 (38, 54)
Estimated glomerular filtration rate, mL/min/1.73m2, mean (standard deviation)	83 (19)	84 (19)	83 (19)	82 (19)
Diabetes, n (%)	68,846 (25.4%)	25,901 (24.7%)	21,876 (25.4%)	21,069 (26.2%)
Chronic kidney disease, n (%)	43,246 (15.9%)	14,550 (13.9%)	14,005 (16.2%)	14,691 (18.3%)
History of CVD, n (%)	45,305 (16.7%)	16,536 (15.8%)	14,577 (16.9%)	14,192 (17.7%)
Combination therapyc, n (%)	47,638 (17.5%)	19,902 (19.0%)	15,511 (18.0%)	12,225 (15.2%)
Initiated ≥1 single-pill combination productd, n (%)	18,120 (38.0%)	7,956 (40.0%)	5,756 (37.2%)	4,408 (36.1%)

^aPeriod 1: Pre-ACC/AHA guideline 11/2014–11/30/2017; Period 2: Post-ACC/AHA guideline, pre-COVID 12/1/2017–12/31/2019; Post-ACC/AHA guideline, post-COVID 1/1/2020–5/31/2023.

^bIncludes Multi-racial, Asian American, American Indian or Alaskan Native, and Native Hawaiian or Other Pacific Islander.

^cDefined as initiating ≥2 antihypertensive medication classes.

^dDenominator is the combination therapy N from the preceding row.

P<0.001 for each characteristic across periods.

For variables with missing values (race, total cholesterol, and cardiovascular disease risk), denominators to calculate percentages were determined by the number of Veterans with non-missing values.

Table values are mean (standard deviation) for continuous and frequency (%) for categorical variables, unless otherwise specified.

ACC/AHA: American College of Cardiology/American Heart Association; CVD: cardiovascular disease; HDL: high-density lipoprotein

Primary outcome: pre- and post-treatment BP levels

In the <140 mm Hg group, the adjusted mean (95% CI) pre-treatment SBP levels were 128.7 (128.7, 128.8), 128.7 (128.6, 128.7), and 128.8 (128.7, 128.9) mm Hg in Periods 1, 2, and 3, respectively (p=0.01 across periods) (Figure 2, Panel A; Figure S5, Panel A for box plots). In the 140–160 mm Hg group, corresponding values were 147.0 (146.6, 147.0), 146.9 (146.9, 147.0), and 147.1 (147.1, 147.2) mm Hg (p<0.001 across periods). For those with SBP ≥160 mm Hg, mean (95% CI) pre-treatment SBP levels were nearly identical across periods at 167.1 (166.9, 167.3), 167.1 (166.9, 167.3), and 167.0 (166.8, 167.2) mm Hg (p=0.76 across periods). Adjusted mean pre-treatment DBP levels decreased across periods (both p<0.001) (Figure S6, Panel A).

Figure 2: Adjusted mean pre- and post-treatment systolic blood pressure levels by period.

The figure shows adjusted mean pre- and post-treatment systolic BP values in each period, with 95% confidence intervals contained in parentheses. Values are derived from models adjusting for calendar month, age, sex, race, ethnicity, diabetes, chronic kidney disease, and history of CVD.

Abbreviations: SBP: blood pressure

For mean post-treatment SBP, in the <140 mm Hg stratum, period-specific mean (95% CI) were 128.3 (128.2, 128.4), 128.1 (128.0, 128.3), and 128.0 (127.9, 128.2) mm Hg (p=0.05); in the 140–160 mm Hg group: 135.4 (135.3, 135.6), 135.3 (135.1, 135.5), and 135.0 (134.9, 135.2) mm Hg (p=0.002 across periods), and in the ≥160 mm Hg group: 142.9 (142.4, 143.4), 142.8 (142.1, 143.4), and 142.3 (141.7, 142.9) mm Hg (p=0.30 across periods) (Figure 2, Panel B). Within all pre-treatment SBP strata, mean post-treatment DBP decreased between Periods 1 and 2 (p<0.05) but were similar between Periods 2 and 3 (p>0.05) (Figure S5, Panel B for box plots; Figure S6, Panel B).

One-year post-treatment BP control

When examining BP control to <140/90 mm Hg, control rates (95% CI) were highest among those with pre-treatment SBP <140 mm Hg and increased from 82.3% (82.0%, 82.7%) in Period 1 to 83.3% (82.9%, 83.7%) and 84.2% (83.8%, 84.7%) in Periods 2 and 3, respectively (both p<0.001) (Figure 3, Panel A). In the 140–160 mm Hg stratum, rates rose from 64.1% (63.7%, 64.5%) to 66.1% (65.6%, 66.5%) and 67.3% (66.9%, 67.8%) (both p<0.001). Among Veterans with pre-treatment SBP ≥160 mm Hg, control across periods was 45.6% (44.7%, 46.5%), 46.6% (45.5%, 47.7%), and 47.7% (46.6%, 48.8%) respectively (p=0.60 Period 2 vs. 1, p=0.08 Period 3 vs. 2).

Figure 3: One-year blood pressure control among Veterans initiating antihypertensive medication.

The figure shows adjusted one-year post-treatment SBP control in each period, with 95% confidence intervals contained in parentheses. Values are derived from models adjusting for calendar month, age, sex, race, ethnicity, diabetes, chronic kidney disease, and history of CVD.

Abbreviations: SBP: blood pressure

For the more stringent <130/80 mm Hg threshold, among those with pre-treatment SBP <140 mm Hg, control rates (95% CI) were 38.2% (38.0%, 38.4%), 40.0% (39.7%, 40.3%), and 39.9% (39.6%, 40.1%) (p<0.001 for Period 2 vs. 1, p=0.82 for Period 3 vs. 2) (Figure 3, Panel B). In the 140–160 mm Hg group, control increased from 20.7% (20.4%, 20.9%) to 22.0% (21.7%, 22.2%) (p<0.001), then to 22.6% (20.4%, 20.9%) (p=0.002). Among those with SBP ≥160 mm Hg, control across periods was 15.1% (14.6%, 15.6%), 15.4% (14.8%, 16.0%), and 15.2% (14.6%, 15.8%) (p=0.83 Period 2 vs.1, p=0.94 Period 3 vs. 2). BP control at individual SBP and DBP thresholds, separately, are shown in Table S5. Among 115,704 Veterans ≥60 years of age (61.3% of the overall period), depending on pre-treatment strata, post-treatment SBP control to <150 mm Hg ranged from 68.2% to 95.3%, with no significant differences detected across periods (p>0.05) (Figure S7).

12-month medication adherence

In each pre-treatment SBP period, mean PDC and (proportion with PDC ≥80%) rose across periods (Table S6). For example, among Veterans with pre-treatment SBP 140–160 mm Hg, mean PDC rose from 67.7% (67.4%, 68.0%) to 69.2% (68.9%, 69.5%) (p<0.001), then to 70.8% (70.5%, 71.1%) (p<0.001), with corresponding increases in PDC ≥80% from 47.3% (46.9%, 47.7%) to 49.4% (48.9%, 49.8%) (p<0.001), then to 51.5% (51.0%, 52.0%) (p<0.001).

Subgroup and Sensitivity Analyses

When combining all SBP strata together, the mean (95% CI) post-treatment SBP was 131.7 (131.6, 131.8), 131.5 (131.4, 131.6), and 131.6 (131.4, 131.8) mm Hg in Periods 1, 2, and 3 (p=0.04 Period 2 vs. 1, p=0.83 Period 3 vs. 2) (Table S7). BP control to <140/90 mm Hg was 73.7% (73.4%, 73.9%), 75.1% (74.8%, 75.4%), and 75.7% (75.4%, 76.0%) across Periods 1, 2, and 3, respectively (p<0.001 2 vs. 1, p=0.02 3 vs. 2), and corresponding values for the <130/80 mm Hg threshold were 30.5% (30.3%, 30.6%), 32.1% (31.9%, 32.3%), and 31.8% (31.6%, 32.0%) (p<0.001 2 vs 1, p=0.07 3 vs. 2).

Subgroup analyses are shown in Tables S8–S10. In the 140–160 mm Hg group (38.2% of the overall population), despite the proportion with SBP <140 mm Hg increasing in most groups between Periods 2 and 3, Non-Hispanic Black patients observed a decrease from 69.2% (68.1%, 70.2%) to 66.8% (65.8%, 67.9%) (p=0.008) (Table S9). In Non-Hispanic Other patients (i.e., multi-racial, Asian American, American Indian or Alaskan Native, and Native Hawaiian or Other Pacific Islander), SBP control to <140 mm Hg increased from 69.5% (66.7%, 72.4%) in Period 2 to 75.0% (72.2%, 78.0%) in Period 3 (p=0.02), and <130 mm Hg increased from 28.9% (27.5%, 30.5%) in Period 1 to 33.7% (31.7%, 35.7%) in Period 2 (p<0.001).

In Veterans with pre-treatment SBP ≥160 mm Hg (5.2% of the overall population), in female Veterans, SBP control <140 mm Hg increased from 56.6% (52.7%, 60.7%) in Period 1 to 64.2% (59.7%, 69.0%) in Period 2 (p=0.04), before decreasing to 52.1% (48.4%, 56.1%) in Period 3 (p<0.001) (Table S10). A decreasing proportion of female Veterans with post-treatment SBP <130 mm Hg occurred from 23.4% (20.8%, 26.4%) in Period 2 to 18.6% (16.4%, 21.0%) in Period 3 (p=0.02). Veterans of Hispanic ethnicity observed a decrease in SBP control <140 mm Hg from 54.7% (49.2%, 60.9%) in Period 2 to 44.2% (39.7%, 49.3%) in Period 3 (p=0.02). Finally, in the Non-Hispanic Other group, the mean SBP increased from 140.6 (137.2, 144.0) mm Hg in Period 1 to 147.2 (143.5, 150.8) mm Hg in Period 2 (p=0.04) – commensurately, the proportion with SBP control <130 mm Hg decreased from 22.5% (18.9%, 26.9%) to 12.0% (9.3%, 15.5%) (p<0.001).

VISN-level BP control rates varied by <5–10% depending on the control definition and period (Figures S8–S10), and including VISN as a random effect did not change the estimates (Figures S11–S12). In sensitivity analyses, removing drop-out weights did not alter the primary findings, although 95% CI were wider (Table S11). Restricting study entry criteria to only 1 valid BP measurement yielded similar results to the primary analysis, with tighter confidence intervals.

DISCUSSION

The current analysis has several important findings. First, among Veterans initiating antihypertensive medication, the pre-treatment BP level did not meaningfully change from before to after publication of the 2017 ACC/AHA BP guideline or the COVID-19 pandemic. For patients with stage 2 hypertension in particular (SBP ≥140 mm Hg), this stability despite lower recommended treatment thresholds is a classic signal of clinical inertia, the failure to initiate or intensify therapy when indicated. Second, the proportion of patients with BP control one year after medication initiation also did not change meaningfully during the same time periods, with nominal increases in medication adherence. Third, only half to two-thirds of Veterans with stage 2 hypertension in the most contemporary period of the study had BP controlled at <140/90 mm Hg, and fewer than a quarter had controlled BP at the more stringent <130/80 mm Hg threshold. Finally, we observed notable differences in BP control among Veterans who were female or from racial and ethnic minority backgrounds. As the next iteration of the clinical guidelines will begin to inform care in 2025 and beyond,²⁶ these findings underscore the need to match lower diagnostic thresholds with equity-focused, action-oriented treatment algorithms, which emphasize earlier initiation of evidence-based treatments, rapid titration, and adherence support, to accelerate BP control and close persistent sex- and race/ethnicity-based gaps in cardiovascular outcomes among Veterans. Nonetheless, the 12–25 mm Hg mean SBP reductions observed across pretreatment strata represent clinically important improvements, as prior meta-analyses show that every 10 mm Hg reduction in SBP lowers cardiovascular event risk by 20%.⁴² Even when guideline targets are not achieved, the magnitude of these SBP reductions meaningfully reduce future cardiovascular risk and highlight the importance of sustaining and intensifying therapy to maintain long-term control.

The current findings add to a body of literature evaluating the impact of guidelines on BP control. Our findings align with previous work concluding that BP control rates did not largely improve after the guidelines were published in 2017.² Among US adults with prevalent hypertension taking antihypertensive medication in the National Health and Nutrition Examination Surveys (NHANES) 2015–2016, 66.7% had controlled BP <140/90 mm Hg,² with little change in 2017–2020 (67.8%). These findings are similar to the current analysis, where BP control to <140/90 mm Hg among Veterans with pre-treatment SBP 140–160 mm Hg was 64.1% in Period 1 and 66.1% in Period 2 (p<0.001). NHANES data also noted a decreasing trend in BP control (<140/90 mm Hg) in non-Hispanic Black individuals (58.2% in 2015–2016 to 52.6% in 2017–2020, p=0.077), similar to the small decrease noted the current study during the same period (69.2% to 66.8%, p=0.008). In the highest-risk SBP stratum (≥160 mm Hg), the current analysis observed marked increases in SBP control to <140 mm Hg after the BP guidelines were published in patients who identify as female (56.6% to 64.2%, p=0.04) and non-Hispanic Black (46.3% to 51.0%, p=0.03). In contrast, male Veterans and Veterans of non-Hispanic White race and ethnicity had relatively stable BP control rates between Periods 1 and 2 in all SBP strata. These findings collectively underscore that disparities in BP control by sex, race, and ethnicity persist after publication of the clinical guidelines.⁴³

Although racial disparities in BP control in the general population are attributed to differences in healthcare access, the integrated structure of the VHA may minimize the impact of many traditional access barriers (e.g., cost of care). Other mechanisms may therefore contribute to the observed race and ethnicity differences in BP control. One potential consideration is that Black adults more commonly exhibit low-renin, salt-sensitive hypertension and may require more intensive multidrug therapy to achieve BP control.⁴⁴ In the presence of therapeutic inertia, these patients may be less likely to achieve target BP. In addition, factors such as medication acceptance, patient mistrust, perceived discrimination,^45–47 and broader social and structural influences may affect hypertension management even within integrated healthcare systems such as the VHA.^48–50

Extending into the pandemic period, NHANES data noted an overall BP control rate of 68.3% among those taking antihypertensives in 2021–2023 (p=0.65 vs. 2017–2020), similar to the current analysis (70.1% of the middle SBP stratum with SBP <140 mm Hg in Period 3). Among 1.7 million US hypertensive patients from 24 health systems included in the PCORNet BP Control Lab, BP control to <140/90 mm Hg in the pre- and post-pandemic eras was highly variable among health systems (range 46–74%), with an average decrease in BP control by 7.2% (95% CI 5.4%−9.0%) with the onset of the pandemic (p<0.001).⁶ One analysis of 464,585 US adults enrolled in an employer-sponsored wellness program reported a mean increase of 1.1 to 2.5 mm Hg during each month of the pandemic period despite stable month-to-month SBPs during the pre-pandemic period.⁷ Another multi-system analysis of 137,593 US adults, BP control decreased by 3.4% during the pandemic.⁸ Unlike these analyses, the current analysis indicated that BP control to <140/90 mm Hg stagnated or only slightly increased (max absolute change 2.3%) during COVID in most Veteran groups. The differences between our findings and these prior studies could be due to inherent differences in study population (i.e., treated vs. untreated patients) or the current analysis including more post-pandemic years (>3) compared to the other studies (max 1) that likely represent the acute peri-pandemic period. Finally, the current analysis supports and adds to findings from a separate VHA analysis of 618,497 treated hypertensive Veterans with uncontrolled BP in the pre-pandemic period demonstrating that the pandemic had minimal impact on BP control (risk ratio 0.98, 95% CI 0.98–0.99).¹⁰ Taken together, these findings highlight the importance of ongoing longitudinal surveillance to understand evolving patterns of BP control and to prevent widening disparities as the pandemic’s longer-term effects continue to unfold.

In 2012, Fletcher et al. reported improvement in BP control rates (<140/90 mm Hg) from 45.7% in 2000 to 75.3% in 2010 at 15 VHA facilities.¹ In the sensitivity analyses collapsing pre-treatment SBP strata of the current analysis, BP control rates were similar at this threshold (73.7%−75.7%). However, evaluating only those with stage 2 hypertension (i.e., pre-treatment BP ≥140 mm Hg), the current analysis demonstrated much lower BP control rates at the <140/90 mm Hg threshold in the most recent Period (67.3% for the middle SBP stratum and 47.7% for the highest SBP stratum) and even lower rates at the stricter <130/80 mm Hg threshold (39.9% for the middle SBP stratum and 14.8% for the highest SBP stratum). These findings serve to provide more contemporary data of new-users of antihypertensive medications and call attention to the low one-year BP control rates at the stricter <130/80 mm Hg threshold, which is the recommended treatment target in the 2025 AHA/ACC guidelines.²⁶ Although the current analysis focused on new-users, the findings are similar to prior estimates of prevalent users (28% of over 2.7 million Veterans with prevalent hypertension between 2016–2017 had BP <130/80 mm Hg),⁹ lending some validity to our results. The similar rates between the prevalent user and the current new-user study also indicate an opportunity to optimize treatment after the initial treatment period. Nonetheless, the BP control rates in the current study are close to the Healthy People 2030 goal to have 18.9% of hypertensive adults with controlled BP to <130/80 mm Hg,⁵¹ with notable opportunity to improve control rates for those with pre-treatment SBP ≥160 mm Hg. Although some Veterans may achieve control after 1 year, delayed titration is clinically important: earlier control and greater time in BP target range are associated with fewer cardiovascular events, independent of baseline BP.^52–55

Although the current analysis focuses on patient-level, early BP management, center and provider context matters. Prior work shows measurable between-site variability in VA BP control and prescribing,^56,57 and that team-based/pharmacist-involved models can improve control,⁵⁸ while provider-level factors contribute to therapeutic inertia.⁵⁹ In the VA, initial treatment decisions vary by provider type,⁶⁰ and resident physicians are 22–54% more likely to intensify therapy more often than attending or mid-career providers.⁶¹ Provider-level factors associated with inertia also include clinical workload, competing demands, and uncertainty about when to intensify therapy.^59,61 The current analysis demonstrates that a majority (75–80%) of Veterans are initiated on antihypertensive monotherapy, similar to prior VA analyses.⁶⁰ Veterans initiating monotherapy experience a median 2-month interval from medication initiation to first outpatient visit,⁶² double the one-month follow-up period recommended by the 2017 ACC/AHA guidelines⁴ and 2020 VA/DoD blood pressure guidelines.⁶³ Together, these findings highlight a critical and clinicially actionable window of post-initiation therapeutic inertia. In this context, the suboptimal BP control observed at one year likely reflects delays in repeat BP evaluation and intensification rather than inadequate initial treatment response alone, highlighting early post-initiation care as a critical window for improving hypertension control.

The VHA became a national leader in hypertension care by coupling evidence-based treatments with system-wide programs to drive their implementation. These programs include VHA-specific clinical guidelines, treatment algorithms, remote patient monitoring, clinical decision support, quality monitoring dashboards, performance measures, audit-and-feedback, and patient-centered medical homes, among others.^1,64 Although promotion of multi-drug regimens helped contribute to improved BP control rates at the VHA in the early 2000s,³⁹ recent data indicate that the reliance on initial monotherapy in the VHA is high (66%) and rising.⁶⁰ In the current study, the initiation of combination therapy declined from 19.0% in Period 1 to 15.2% in Period 3, with a commensurate decline in use of single-pill combination products from 40.0% to 36.1%. Persisting reliance on monotherapy may both reflect and reinforce inertia: clinicians must first initiate, or switch to, combination therapy to meet <130/80 mm Hg targets, yet such intensifications remain uncommon in clinical practice.⁶⁵ To support robust follow-up care, since 2003, the VHA’s remote telemonitoring program has enabled Veterans to access a free home BP monitor and frequent follow-up with a care coordinator to discuss readings.⁶⁶ The VHA’s rapid telehealth expansion helped to maintain primary care engagement among >1.3 million high-risk Veterans between 2020 and 2022,⁶⁷ which could be one of many potential reasons why we did not observe adverse changes in BP levels during the pandemic period. The VHA also has integrated team-based care into normal workflows via “Patient Aligned Care Teams”, one component of which includes elevating scopes of practice for pharmacists and nurses to independently manage hypertension.⁶⁸ As hypertension guidelines continue to evolve away from monotherapy and toward intensive targets, as is recommended in the most recent 2025 guideline,²⁶ the VHA must evolve its existing infrastructure to include cross-cutting strategies to accelerate improvement in BP control,⁶⁹ such as efforts to address disparities across subpopulations,⁷⁰ increasing use of single-pill combination products, and implementing clinician-facing interventions to combat inertia.

A major strength of the current study lies in the robust clinical, administrative, and pharmaceutical information available for VHA patients, allowing us to accurately assess demographics, clinical conditions, laboratory measurements, vital signs, and medication use. The VHA provides near equal access to antihypertensive medications for all Veterans via a mail-order pharmacy and preferred 90-day refill system, bolstering sustainable medication use for all Veterans. Because VA clinicians may individualize targets per different guidelines (e.g., VA/DoD, ACC/AHA 2017, or ACP/AAFP 2017), we reported multiple thresholds (combined and component SBP/DBP) and added SBP <150 mm Hg for ≥60 years to enhance generalizability across practice standards. Nonetheless, this study should be interpreted in the context of several limitations. BP was measured during clinical care and may not have been measured in a standardized manner for Veterans included in the current study. We also did not include home BP measurements, as these data are not stored in the database and require natural language processing to query from free-text clinical notes. However, clinical decisions may often have been based on patient-reported home BP measurements. Given that we sought to isolate impacts of early treatment (a time period that exerts substantial influence on BP control) through a new-user study design, we could not estimate awareness, diagnosis, and treatment rates. However, our inclusion of a variable indicating the interval between diagnosis and treatment revealed a relatively short period between initial diagnosis and treatment (1 year on the median). As with other VHA studies, the population was mostly male and non-Hispanic White, and results may differ in periods with more women or racial and ethnic minority groups. We lacked direct measures of some social determinants of health (e.g., income, education, transportation), so residual confounding is possible, and we could not evaluate whether socioeconomic factors modify observed differences in BP control across SDoH subgroups. However, population-wide analyses indicate that pandemic-era declines in BP control were largely attributable to longer intervals between follow-up visits and fewer BP measurements,^6,10 supporting our emphasis on observed care-continuity pathways while recognizing unmeasured social determinants of health may contribute to these results. Because of the large sample size, some statistically significant differences between periods were very small in magnitude and did not represent clinically meaningful differences. Further, we did not model multilevel (facility/provider) effects in this analysis to focus on patient-level data; as a result, we did not formally quantify regional or local variation in care delivery during the pandemic, although exploratory analyses indicated modest heterogeneity in control rates across VISNs. Finally, we did not know what BP target treating clinicians were trying to achieve and therefore could not directly measure clinical inertia in this large, secondary dataset, which is an area of ongoing research.

CONCLUSION

In the current study of US Veterans newly diagnosed and treated for hypertension, neither the release of 2017 ACC/AHA BP guidelines nor the COVID-19 pandemic meaningfully changed pre- or post-treatment BP levels or BP control rates, and variation by sex, race, and ethnicity remain. Only a minority of Veterans reach a stringent guideline-recommended BP goal of <130/80 mm Hg within a year of initiating treatment. Despite the VHA’s comprehensive system for hypertension care, these findings signal stalled progress in light of the VHA’s earlier gains as highlighted by prior studies, particularly for those with stage 2 hypertension, and reveal persistent equity gaps. Reversing this stasis will require re-calibration toward evidence-based interventions explicitly designed to combat clinical inertia and support adherence, such as algorithmic titration, pharmacist-led BP protocols, single-pill combinations, and performance-feedback dashboards.

CLINICAL PERSPECTIVE.

What is known?

• Blood pressure (BP) control rates among US adults have declined since approximately 2014, and clinical inertia occurs in up to half of hypertension-related clinical encounters nationwide.

• The 2017 ACC/AHA guideline lowered the recommended threshold for antihypertensive treatment initiation and intensification from ≥140/90 mm Hg to ≥130/80 mm Hg for adults at elevated cardiovascular risk, yet national BP control rates did not meaningfully improve following the guideline publication.

• The COVID-19 pandemic disrupted hypertension care delivery, with studies in the general population reporting modest decreases in BP control during the acute pandemic period; however, comparable data specifically among Veterans initiating antihypertensive medication — using a new-user study design with separate evaluation of guideline and pandemic impacts — have been lacking.

What the study adds:

• Among 271,496 US Veterans newly initiating antihypertensive medication between 2014 and 2023, pre-treatment BP levels — a marker of when clinicians decided to act — did not meaningfully change before or after the 2017 ACC/AHA guideline or the onset of the COVID-19 pandemic, indicating persistent clinical inertia across all three time periods studied.

• Despite the guideline-recommended target of <130/80 mm Hg, only two-thirds of Veterans with stage 2 hypertension (pre-treatment SBP ≥140 mm Hg) achieved BP control to <140/90 mm Hg within one year of treatment initiation, and fewer than one in four achieved the more stringent <130/80 mm Hg threshold — with these rates remaining largely static across periods.

• BP control gaps were disproportionately borne by Veterans who were female or from racial and ethnic minority backgrounds; notably, Non-Hispanic Black Veterans in the 140–160 mm Hg stratum and Hispanic and female Veterans with severe hypertension (SBP ≥160 mm Hg) experienced declining BP control rates in the post-pandemic period, underscoring the need for equity-focused, multi-level interventions to improve early hypertension management in the VHA.