Abstract
Background:
Data from the US National Health and Nutrition Examination Survey (NHANES) are freely available and can be analyzed to produce hypertension statistics for the non-institutionalized US population. The analysis of these data requires statistical programming expertise and knowledge of NHANES methodology.
Methods:
We developed a web-based application that provides hypertension statistics for US adults using 10 cycles of NHANES data, 1999–2000 through 2017–2020. We validated the application by reproducing results from prior publications. The application’s interface allows users to estimate crude and age-adjusted means, quantiles and proportions. Population counts can also be estimated. To demonstrate the application’s capabilities, we estimated hypertension statistics for non-institutionalized US adults.
Results:
The estimated mean systolic blood pressure (BP) declined from 123 mmHg in 1999–2000 to 120 mmHg in 2009–2010 and increased to 123 mmHg in 2017–2020. The age-adjusted prevalence of hypertension (i.e., systolic BP≥130 mmHg, diastolic BP≥80 mmHg or self-reported antihypertensive medication use) was 47.9% in 1999–2000, 43.0% in 2009–2010, and 44.7% in 2017–2020. In 2017–2020, an estimated 115.3 million US adults had hypertension. The age-adjusted prevalence of controlled BP, defined by the 2017 American College of Cardiology/American Heart Association BP guideline, among non-pregnant US adults with hypertension was 9.7% in 1999–2000, 25.0% in 2013–2014, and 21.9% in 2017–2020. After age-adjustment and among non-pregnant US adults who self-reported taking antihypertensive medication, 27.5%, 48.5%, and 43.0% had controlled BP in 1999–2000, 2013–2014, and 2017–2020, respectively.
Conclusions:
The application developed in the current study is publicly available and produced valid, transparent, and reproducible results.
Keywords: Hypertension, statistics, blood pressure, blood pressure control
Graphical Abstract
The National Health and Nutrition Examination Survey (NHANES) is a program conducted by the US National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC) and is designed to assess the health and nutritional status of the non-institutionalized US population.1 NHANES data have been analyzed to provide hypertension statistics for non-institutionalized US adults with public health and policy implications. For example, NHANES data have been used to estimate the impact of the lower blood pressure (BP) levels that define hypertension and controlled BP in the 2017 American College of Cardiology/American Heart Association (ACC/AHA) BP guideline compared with the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High BP (JNC7) guideline.2 Additionally, NHANES data have been used to track the proportion of US adults with hypertension that have controlled BP, which were included in a 2020 Call-to-Action to Control BP from the US Surgeon General.3
NHANES data are publicly available and accessible through the CDC website.4 However, analyzing NHANES data requires specific statistical techniques to account for the multi-stage sampling design used to select participants, and analyses need to be weighted to produce nationally representative estimates. Also, NHANES data are collected in two-year periods referred to as cycles and the data collection protocols have changed over time for some variables. These changes need to be incorporated when estimating statistics from multiple NHANES cycles. These challenges may present barriers to analyzing NHANES data.
We developed a web-based application that provides nationally representative BP and hypertension statistics for non-institutionalized US adults using NHANES data without requiring users to conduct statistical programming. The goal of the application is to make the ability to produce statistics using NHANES data more accessible with a focus on ensuring results are valid and reproducible. The application is open-source, which means its code is publicly available and may be re-distributed or modified by anyone.5,6 In this manuscript, we review the design, development and validation of the application and present BP and hypertension statistics for US adults that were generated using the application.
METHODS
All data and materials have been made publicly available and can be accessed at https://jhs-hwg.github.io/cardioStatsUSA/.
The NHANES program was initiated in the early 1960s and beginning in 1999 has been conducted continuously, in two-year cycles.1 In each cycle, participants were identified using a multi-stage sampling process. The protocols for each cycle were approved by the NCHS Institutional Review Board. Written informed consent was obtained from each participant.
NHANES data were collected through an in-home interview and a study examination conducted at a mobile examination center. The interview included questions about demographics, health behaviors, medical history, and medication use. During the interview, the labels of medications that participants reported taking in the preceding 30 days were recorded. Antihypertensive medication classes were defined based on being commercially available according to the US Food and Drug Administration’s Orange Book, Drugs@FDA, and Lexicon Plus® databases, and either: 1) Being listed as an antihypertensive medication in the 2017 ACC/AHA BP guideline or 2) Having a labeled indication for the treatment of essential (primary) hypertension.7 Only orally-administered antihypertensive medications were included. During the study examination, height, weight and BP were measured and blood and spot urine samples were collected. Of relevance to the current application, blood samples were used to measure total and high-density lipoprotein cholesterol, glycated hemoglobin and serum creatinine, and the urine sample was used to measure albumin and creatinine and to conduct a pregnancy test. The protocol for measuring BP is available online.8 In brief, systolic and diastolic BP (SBP and DBP, respectively) were measured up to three times by trained and certified physicians. BP was measured using a mercury sphygmomanometer from 1999–2000 through 2015–2016 and using a validated oscillometric device in 2017–2020.9–11 The mean SBP and DBP levels were computed over all available measurements for each participant. The oscillometric SBP and DBP values were calibrated to the mercury device.11,12 For the current analysis, we defined hypertension, BP control, and apparent treatment resistant hypertension according to the 2017 ACC/AHA BP guideline.7 The application also has these variables defined according to the JNC7 guideline definitions and provides variables for hypertension, being recommended for antihypertensive medication and BP control according to the 2018 European Society of Cardiology/ European Society of Hypertension guideline.13,14 A list of BP, hypertension, and antihypertensive medication variables available in the application is provided in Table 1 with full definitions for all variables included in the application provided in Table S1.
Table 1:
Blood pressure, hypertension and antihypertensive medication variables that are available in the web application.
| Variable |
|---|
| Blood pressure domain |
| Systolic blood pressure (SBP), mm Hg |
| Diastolic blood pressure (DBP), mm Hg |
| Blood pressure category |
| Blood pressure category including antihypertensive medication use as a group |
| Blood pressure control defined by the JNC7 guideline |
| Blood pressure control defined by the 2017 ACC/AHA blood pressure guideline |
| Blood pressure control defined by the 2018 ESC/ESH arterial hypertension guideline’s first goal |
| Blood pressure control defined by the 2018 ESC/ESH arterial hypertension guideline’s second goal |
| Blood pressure control (SBP < 140 mm Hg and DBP < 90 mm Hg) |
| Blood pressure control (SBP < 130 mm Hg and DBP < 80 mm Hg) |
| Uncontrolled blood pressure defined by the JNC7 guideline |
| Uncontrolled blood pressure defined by the 2017 ACC/AHA blood pressure guideline |
| Uncontrolled blood pressure control defined by the 2018 ESC/ ESH arterial hypertension guideline’s first goal |
| Uncontrolled blood pressure control defined by the 2018 ESC/ ESH arterial hypertension guideline’s second goal |
| Uncontrolled blood pressure (SBP ≥ 140 mm Hg or DBP ≥ 90 mm Hg) |
| Uncontrolled blood pressure (SBP ≥ 130 mm Hg or DBP ≥ 80 mm Hg) |
| Hypertension domain |
| Hypertension defined by the JNC7 guideline |
| Hypertension defined by the 2017 ACC/AHA blood pressure guideline |
| Hypertension defined by the 2018 ESC/ESH arterial hypertension guideline |
| Awareness of hypertension |
| Apparent treatment resistant hypertension defined by the JNC7 guideline |
| Apparent treatment resistant hypertension defined by the 2017 ACC/AHA blood pressure guideline |
| Apparent treatment resistant hypertension defined by the JNC7 guideline, requires thiazide diuretic |
| Apparent treatment resistant hypertension defined by the 2017 ACC/AHA blood pressure guideline, requires thiazide diuretic |
| Antihypertensive medication domain |
| Self-reported antihypertensive medication use |
| Antihypertensive medication use recommended by the JNC7 guideline |
| Antihypertensive medication use recommended by the 2017 ACC/AHA blood pressure guideline |
| Antihypertensive medication use recommended by the 2018 ESC/ESH arterial hypertension guideline |
| Number of antihypertensive medication classes being taken |
| Fixed-dose combination use |
| Number of antihypertensive medication pills |
| Taking two or more antihypertensive medication pills |
| Antihypertensive medication classes |
| Angiotensin converting enzyme inhibitors |
| Aldosterone antagonists |
| Alpha-1 blockers |
| Angiotensin-II receptor blockers |
| Beta blockers |
| Calcium channel blockers |
| Dihydropyridine calcium channel blockers |
| Non-dihydropyridine calcium channel blockers |
| Central alpha1 agonist and other centrally acting agents |
| Direct renin inhibitors |
| Direct vasodilators |
| Loop diuretics |
| Potassium sparing diuretics |
| Thiazide or thiazide-type diuretics |
Abbreviations: ACC = American College of Cardiology; AHA = American Heart Association; DBP = diastolic blood pressure; ESC = European Society of Cardiology; ESH = European Society of Hypertension; JNC7 = Seventh Joint National Committee; and SBP = systolic blood pressure
There were 107,622 NHANES participants in the 10 cycles from 1999–2000 to 2017–2020. We restricted the dataset to adults ≥ 18 years of age. This exclusion was applied because hypertension is defined differently for children compared to adults.15 We further restricted the population to participants who completed the in-home interview and study examination, with one or more SBP and DBP measurements, and who had data on self-reported antihypertensive medication use. After these exclusions were applied, the population in the current application included 56,017 participants (Figure S1).
Features of the web application
A full summary of the features, instructions and associated tutorials are available within the application by clicking “instructions” or the “?” symbols (Figure S2).6 Briefly, users can select NHANES cycles from 1999–2000 to 2017–2020 to be analyzed. Estimates are weighted to represent the non-institutionalized US population. Users can incorporate age-adjustment through direct standardization by selecting the age-distribution for the standard population and applying this to obtain a weighted average of age-specific estimates.16,17 Users can restrict analyses to subsets of participants (e.g., those who self-reported taking antihypertensive medication). When population count estimates are requested, survey weights are calibrated within race, gender, and age groups to account for missing information on SBP, DBP or self-reported antihypertensive medication use.18 The results can be presented in tables or figures and for the overall population or in subgroups. All tables and figures created with the application can be downloaded and saved. Following CDC recommendations, unreliable statistical estimates are automatically suppressed.19 To increase precision and reliability of estimates, contiguous NHANES cycles can be combined.20
Design, Development and validation of the web application
The web application was designed to increase access to BP-related statistics produced using NHANES data (see Supplemental Methods) and was developed using Shiny, an open-source software package that translates code from the R programming language into HTML, CSS, or JavaScript commands and creates a website interface.21–24 We validated the web application by using it to reproduce statistics reported in two prior studies and one CDC report.12,25,26 We created the “cardioStatsUSA” R package to provide additional details on the web application’s design, documentation of its components, and validation.6 The online documentation for cardioStatsUSA includes detailed summaries of the web application’s validation process.27–29
Statistical analysis
We performed statistical analyses to demonstrate core features of the application. We estimated the mean SBP for US adults by NHANES cycle, 1999–2000 through 2017–2020, with points and error bars representing the estimated means and 95% confidence intervals (CI), respectively. We made bar charts presenting the age-adjusted prevalence of hypertension and the estimated number of US adults with hypertension. For age adjustment, we used the estimated age distribution of US adults from 1999 to 2020 as the standard (49.3%, 33.6%, 10.1% and 7.0% being 18 to 44, 45 to 64, 65 to 74 and ≥ 75 years of age, respectively). We created a table of the estimated race/ethnicity distribution of US adults with and without hypertension, separately. We demonstrated stratification by estimating the prevalence of hypertension by NHANES cycle for US adults with and without chronic kidney disease. We showed the application’s ability to suppress output when statistical estimates are unstable by attempting to estimate the distribution of BP categories (SBP/DBP < 120/80 mm Hg, 120–129/<80 mm Hg, 130–139/80–89 mm Hg, 140–159/90–99 mm Hg and ≥ 160/100 mm Hg) among pregnant women in 2017–2020. We then showed that reliable estimates can be obtained for the distribution of BP categories among pregnant women by pooling NHANES cycles from 2009–2010 through 2017–2020.
We illustrated how core features of the application can be combined to perform customized analyses. Specifically, we estimated the age-adjusted proportion of US adults with controlled BP by NHANES cycle among non-pregnant US adults with hypertension, overall and who self-reported taking antihypertensive medication. We also estimated the age-adjusted prevalence of apparent treatment resistant hypertension by NHANES cycle for non-pregnant US adults with hypertension who self-reported taking antihypertensive medication and had ≥ 1 classes of antihypertensive medication identified during the medication inventory. This was repeated among those with ≥ 3 antihypertensive medication classes identified during the medication inventory. For age adjustment in the analyses of BP control and apparent treatment resistant hypertension, we set the age distribution of the standard population to represent US adults with hypertension from 1999 to 2020: 26.4%, 43.4%, 17.0% and 13.2% being 18 to 44, 45 to 64, 65 to 74, and ≥ 75 years of age, respectively. Participants with missing pregnancy status were assumed to be non-pregnant for these analyses.
RESULTS
Among non-institutionalized US adults ≥ 18 years of age, the estimated mean SBP was 123 (95% CI 121, 124) mm Hg in 1999–2000, 120 (95% CI 120, 121) mm Hg in 2009–2010 and 123 (95% CI 122, 124) mm Hg in 2017–2020 (Figure 1). The age-adjusted prevalence of hypertension defined according to the 2017 ACC/AHA BP guideline was highest in 1999–2000 (47.9%), lowest in 2009–2010 and 2013–2014 (43.0%), and 44.7% in 2017–2020 (Figure 2). In 1999–2000, there were an estimated 89.8 (95% CI 77.9, 101.7) million US adults with hypertension (Figure 3). The number of US adults with hypertension increased to 115.3 (95% CI 107.4, 123.2) million in 2017–2020. In each NHANES cycle, the estimated prevalence of hypertension was higher among US adults with versus without chronic kidney disease (Figure S3). In 2017–2020, a higher percentage of US adults with versus without hypertension were non-Hispanic Black (13.5% versus 9.3%) while a lower percentage of US adults with versus without hypertension were Hispanic (12.3% versus 18.1%) (Table 2). Among pregnant women, the distribution of BP categories could not be estimated reliably in 2017–2020 (Figure S4; Panel A), but it could be estimated reliably after pooling NHANES cycles from 2009–2010 through 2017–2020 (Figure S4; Panel B)
Figure 1. Mean systolic blood pressure for US adults by calendar year.
Dots represent mean systolic blood pressure. Vertical lines represent the 95% confidence interval.
The graph is identical to the web application’s output. The estimated prevalence and the upper and lower limits of the 95% confidence interval can be displayed in the application by hovering the cursor over the dot.
The inputs used to generate this graph are available online
Figure 2. Age-adjusted prevalence of hypertension for US adults by calendar year.
Age adjustment was performed through direct standardization, using the estimated age distribution of US adults from 1999 to 2020 as the standard (49.3%, 33.6%, 10.1% and 7.0% being 18 to 44, 45 to 64, 65 to 74 and ≥75 years of age, respectively).
The graph is identical to the web application’s output. The estimated prevalence and the upper and lower limits of the 95% confidence interval can be displayed in the application by hovering the cursor over the bar.
The inputs used to generate this graph are available online
Figure 3. Number of US adults with hypertension by calendar year.
The graph is identical to the web application’s output. The estimated prevalence and the upper and lower limits of the 95% confidence interval can be displayed in the application by hovering the cursor over the bar.
The inputs used to generate this graph are available online
Table 2:
Race/ethnicity distribution of US adults with and without hypertension in 2017–2020
| svy_year | htn_accaha | demo_race | statistic | estimate | std_error | ci_lower | ci_upper | n_obs | unreliable_status | unreliable_reason | review_needed | Review_reason |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2017–2020 | No | Hispanic | percentage | 18.1 | 1.8 | 14.5 | 21.7 | 953 | FALSE | FALSE | ||
| 2017–2020 | No | Non-Hispanic Asian | percentage | 5.9 | 0.9 | 4.2 | 7.6 | 502 | FALSE | FALSE | ||
| 2017–2020 | No | Non-Hispanic Black | percentage | 9.3 | 1.2 | 7.0 | 11.5 | 802 | FALSE | FALSE | ||
| 2017–2020 | No | Non-Hispanic White | percentage | 62.8 | 2.3 | 58.2 | 67.4 | 1,325 | FALSE | FALSE | ||
| 2017–2020 | No | Other | percentage | 4.0 | 0.4 | 3.2 | 4.7 | 200 | FALSE | FALSE | ||
| 2017–2020 | Yes | Hispanic | percentage | 12.3 | 1.2 | 10.0 | 14.7 | 752 | FALSE | FALSE | ||
| 2017–2020 | Yes | Non-Hispanic Asian | percentage | 5.0 | 0.8 | 3.5 | 6.6 | 425 | FALSE | FALSE | ||
| 2017–2020 | Yes | Non-Hispanic Black | percentage | 13.5 | 1.8 | 9.9 | 17.1 | 1,343 | FALSE | FALSE | ||
| 2017–2020 | Yes | Non-Hispanic White | percentage | 64.9 | 2.8 | 59.4 | 70.3 | 1,510 | FALSE | FALSE | ||
| 2017–2020 | Yes | Other | percentage | 4.3 | 0.6 | 3.2 | 5.4 | 198 | FALSE | FALSE |
The table format is identical to the web application’s output.
svy_year indicates the NHANES cycle
htn_accaha indicates the statistic is being estimated stratified by having hypertension according the 2017 American College of Cardiology/American Heart Association blood pressure guideline.
demo_race indicates the statistic is being estimated stratified by race/ethnicity subgroup.
statistic indicates the type of statistic computed.
estimate is the point estimate of the given statistic
std_error is the standard error for estimate
ci_lower is the lower bound of a 95% confidence interval for estimate
ci_upper is the upper bound of a 95% confidence interval for estimate
n_obs is the number of observations (unweighted) used for the computation
unreliable_status is TRUE if the estimate is unreliable, FALSE if the estimate is reliable
unreliable_reason is blank if the estimate is reliable, and lists reason(s) for not being reliable otherwise
review_needed is TRUE if the estimate should be reviewed by a statistical expert, FALSE if the estimate does not need to be reviewed. The estimate may or may not need to be suppressed depending on the outcome of the review. This is determined by the National Center for Health Statistics estimates of reliable estimates.
review_reason is blank if the result does not require review, and lists reason(s) for requiring review otherwise
The age-adjusted prevalence of BP control defined according to the 2017 ACC/AHA BP guideline among non-pregnant US adults with hypertension was lowest in 1999–2000 (9.7%), highest in 2013–2014 (25.0%), and 21.9% in 2017–2020 (Figure 4; Panel A). Among US adults with hypertension who self-reported taking antihypertensive medication, 27.5%, 48.5%, and 43.0% had controlled BP in 1999–2000, 2013–2014, and 2017–2020, respectively (Figure 4; Panel B). The age-adjusted prevalence of apparent treatment resistant hypertension defined according to the 2017 ACC/AHA BP guideline among non-pregnant US adults with hypertension who self-reported taking antihypertensive medication and had one or more classes of antihypertensive medication identified during the medication inventory was lowest in 1999–2000 (13.6%), highest in 2005–2006 (20.1%), and 14.6% in 2017–2020 (Figure S5; Panel A). Further restricting this analysis to those who were taking three or more classes of antihypertensive medication, the prevalence of apparent treatment resistant hypertension was 78.1%, 76.2%, and 70.1% in 1999–2000, 2005–2006, and 2017–2020, respectively (Figure S5; Panel B).
Figure 4. Age-adjusted prevalence of blood pressure control by calendar year.
A. Among non-pregnant US adults with hypertension
Age adjustment was performed through direct standardization, using the estimated age distribution of US adults with hypertension from 1999 to 2020 as the standard (26.4%, 43.4%, 17.0% and 13.2% being 18 to 44, 45 to 64, 65 to 74, and ≥75 years of age, respectively).
The graph is identical to the web application’s output. The estimated prevalence and the upper and lower limits of the 95% confidence interval can be displayed in the application by hovering the cursor over the bar.
Participants with missing values for pregnancy status were assumed to be non-pregnant in this analysis.
The inputs used to generate this graph are available online
B. Among non-pregnant US adults who self-reported taking antihypertensive medication.
Age adjustment was performed through direct standardization, using the estimated age distribution of US adults with hypertension from 1999 to 2020 as the standard (26.4%, 43.4%, 17.0% and 13.2% being 18 to 44, 45 to 64, 65 to 74, and ≥ 75 years of age, respectively).
The graph is identical to the web application’s output. The estimated prevalence and the upper and lower limits of the 95% confidence interval can be displayed in the application by hovering the cursor over the bar.
Participants with missing values for pregnancy status were assumed to be non-pregnant in this analysis.
The inputs used to generate this graph are available online
DISCUSSION
In the current manuscript, we present a web application with a point and click interface that allows the calculation of US nationally representative estimates for BP, hypertension and antihypertensive medication-related statistics using NHANES data. Using this application, we generated crude and age-adjusted BP and hypertension statistics. Also, we generated statistics stratified by characteristics of US adults and demonstrated how multiple NHANES cycles can be pooled to obtain more precise estimates when working with small sub-groups (e.g., pregnant women). Following its validation, version 0.0.1 of the application was released and deployed on a publicly available server on January 6, 2023.30 Anyone can use the application to generate customized BP and hypertension statistics for US adults to inform research, public health programs and policy decisions.
NHANES is an ideal data source to obtain statistics related to hypertension, as it was designed to obtain nationally representative estimates of the health and nutrition status of non-institutionalized US adults. SBP and DBP were measured following a standardized protocol by trained and certified physicians. Although NHANES data are publicly available to download, working with these data requires understanding variable definitions and advanced programming and statistical knowledge to analyze complex survey data. Additional challenges of analyzing NHANES data may include downloading and merging multiple data files, even for a single NHANES cycle; combining multiple variables to create outcome definitions, which may require dealing with missing data and questionnaire skip patterns; and harmonizing variables across multiple NHANES cycles. Although multiple manuscripts based on the NHANES data have been published, they may not provide the degree of granularity needed by some readers, policy makers, and the public. The application that we present in the current manuscript addresses these challenges, increasing the ability of clinicians, public health practitioners, researchers, and the public to generate statistics using the NHANES data.
Several design decisions have been incorporated into the application. First, we required participants to have at least one SBP and DBP to be included in the dataset while some prior analyses of NHANES data required three SBP and DBP measurements.2 We chose the current approach as it is consistent with several CDC analyses.26,31 However, mean BP and the prevalence of high BP may be lower if we required multiple BP measurements.32 Any bias resulting from this decision is likely to be small as over 95% of adult NHANES participants with at least one SBP and DBP measurement had three SBP and DBP measurements. Second, the application re-calibrates the NHANES weights for the estimation of population counts. This was done because participants missing data on SBP, DBP or antihypertensive medication use cannot have BP or hypertension-related outcomes. Weights were not re-calibrated when estimating proportions as participants missing data are removed from the numerator and denominator. Medication classes were coded using generic names and the drug classes after review by two pharmacists. We recognize that the NCHS recommends using Lexicon Plus®, a proprietary database, to categorize medication classes.33 While the categorization of most medications is identical using generic drug names and Lexicon Plus®, some differences existed. Many additional decisions were made regarding the definitions of variables, inclusion of study participants, and analytic approaches. We sought to make decisions that would be widely acceptable and transparent to ensure the results could be described accurately.
Using the application, we replicated results from several prior manuscripts.27,28,29 However, some prior manuscripts may not be replicable using the application due to differences in design and variable definitions.34 For example, a prior manuscript reported the prevalence of apparent treatment resistant hypertension defined by the 2017 ACC/AHA BP guideline to be 19.7% in 2009–2014.34 When estimated by the application, the prevalence of apparent treatment resistant hypertension over this time period was 17.6%. The difference in the prevalence estimates can be attributed to the approach used to categorize medications into antihypertensive classes. The prior publication used Lexicon Plus®, which categorizes spironolactone as two drug classes, a potassium-sparing diuretic and an aldosterone receptor antagonist. We included spironolactone as a single drug class, an aldosterone receptor antagonist. Additionally, Lexicon Plus® includes sotalol (a non-selective beta blocker used exclusively for treatment of atrial or ventricular tachycardias) and nitroglycerin (a direct vasodilator used to relieve angina) as antihypertensive medications, and although these medications lower BP, we did not include these drugs as antihypertensive medications as neither is listed in the 2017 ACC/AHA BP guideline or approved by the US Food and Drug Administration with an indication for treating hypertension.7 The differences in results between the application and the previously published manuscript emphasize that users should be aware of the choices made in defining variables as this may affect the statistical estimates generated.
The application presented in the current manuscript complements an application on the CDC website with additional variables and features that can be used in combination to create highly customized statistics (Table S2).35 Variables can be analyzed as outcomes, used to generate stratified results, or conduct sub-group analyses. Combining the variables, over 300 million unique statistics can be estimated using the application. In addition, users can generate statistics for US adults from 1999–2000 to 2017–2020, pooling or stratifying the periods. All of the results from the application can be saved as images and included in scientific proposals or presentations. In addition, the ability to download results as a data set from the application allows users to further customize their results in tabular or graphical formats.
Strengths/limitations
This study has a number of strengths. We used NHANES data, which are publicly available, rigorously collected, and allow for the estimation of nationally representative statistics. Also, we leveraged open-source software to ensure that our application is transparent and freely available. This study also has several limitations. NHANES participants had their mean BP measured during a single visit, and BP guidelines recommend averaging BP from values obtained during at least two visits conducted on separate days.7 Many estimates, including antihypertensive medication use, were derived from participant self-report, and we cannot exclude the possibilities of reporting or recall bias. NHANES did not collect data on medication adherence and did not conduct ambulatory BP monitoring.36 As such, the application’s hypertension indicators do not account for white-coat and masked hypertension. The response rate for NHANES has declined and the percentage of participants with missing BP measurements or information on antihypertensive medication use increased from 1999–2000 through 2017–2020. The primary reason for missing information on BP was time constraints during the examination.
Perspectives
We developed a web-based application for the analysis of hypertension statistics among non-institutionalized adults living in the US from 1999–2000 through 2017–2020. The application is publicly available and produces valid, transparent, and reproducible results.
Supplementary Material
Novelty and Relevance.
What Is New?
We developed and validated a web application that allows users to generate customized statistics using the publically available National Health and Nutrition Examination Survey data.
In addition to describing the application, we provide hypertension statistics for US adults.
What Is Relevant?
Many clinicians, policy makers and public health professionals use statistics from the National Health and Nutrition Examination Survey.
The web application presented in the current manuscript provides an approach for people without statistical expertise to generate customized statistics using the National Health and Nutrition Examination Survey data.
Making data accessible and transparent has the potential to build trust.
Clinical/Pathophysiological Implications?
The web application can be used to generate statistics that can guide policy and clinical decision-making.
Sources of Funding
Drs. Muntner and Jaeger receive support through grant R01HL144773 from the National Heart, Lung, and Blood Institute. Drs. Muntner and Foti receive support through grant R01HL117323 from the National Heart, Lung, and Blood Institute. Dr. Muntner receives support through grant R01HL139716 from the National Heart, Lung, and Blood Institute. Dr. Hardy is supported by K01HL164763 from the National Heart, Lung, and Blood Institute. Dr. Bress is supported by R01HL139837 (National Heart, Lung, and Blood Institute, Bethesda, MD), R01AG0658505 (National Institute on Aging, Bethesda, MD), and R01AG074989 (National Institute on Aging, Bethesda, MD).
Abbreviations and acronyms
- ACC
American College of Cardiology
- AHA
American Heart Association
- BP
Blood Pressure
- CDC
Centers for Disease Control and Prevention
- JNC7
Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure
- NHANES
National Health and Nutrition Examination Survey
- NCHS
National Center for Health Statistics
Footnotes
Disclosures
Dr. Muntner reported receiving grant funding and consulting fees from Amgen Inc. unrelated to the current study. Dr. Colantonio reported receiving grant funding from Amgen Inc. unrelated to the current study. No other disclosures were reported.
REFERENCES
- 1.Centers for Disease Control and Prevention (CDC). National Center for Health Statistics (NCHS). About the national health and nutrition examination survey. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. https://www.cdc.gov/nchs/nhanes/about_nhanes.htm. Accessed October 26, 2022. [Google Scholar]
- 2.Muntner P, Carey RM, Gidding S, Jones DW, Taler SJ, Wright JT, Jr., Whelton PK. Potential U.S. Population Impact of the 2017 ACC/AHA High Blood Pressure Guideline. J Am Coll Cardiol. 2018;71:109–118. doi: 10.1016/j.jacc.2017.10.073 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.U.S. Department of Health and Human Services. The Surgeon General’s call to action to control hypertension. In: Washington, D.C.: Department of Health and Human Services, Office of the Surgeon General; 2020. [Google Scholar]
- 4.NHANES. National health and nutrition examination survey homepage, available at https://www.cdc.gov/nchs/nhanes/index.htm. Accessed October 3, 2022.
- 5.Perens B The open source definition. In: Open sources: voices from the open source revolution. O’Reilly Media; 1999:171–188. [Google Scholar]
- 6.Jaeger B, Chen L, Muntner P. cardioStatsUSA: Analysis and Visualization of Cardiometabolic Outcomes Using NHANES. https://jhs-hwg.github.io/cardioStatsUSA/. Accessed December 22, 2022. [Google Scholar]
- 7.Whelton PK, Carey RM, Aronow WS, Casey DE Jr., Collins KJ, Dennison Himmelfarb C, DePalma SM, Gidding S, Jamerson KA, Jones DW, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol. 2018;71:e127–e248. doi: 10.1016/j.jacc.2017.11.006 [DOI] [PubMed] [Google Scholar]
- 8.Ostchega Y, Prineas RJ, Paulose-Ram R, Grim CM, Willard G, Collins D. National Health and Nutrition Examination Survey 1999–2000: effect of observer training and protocol standardization on reducing blood pressure measurement error. J Clin Epidemiol. 2003;56:768–774. doi: 10.1016/s0895-4356(03)00085-4 [DOI] [PubMed] [Google Scholar]
- 9.Omboni S, Riva I, Giglio A, Caldara G, Groppelli A, Parati G. Validation of the Omron M5-I, R5-I and HEM-907 automated blood pressure monitors in elderly individuals according to the International Protocol of the European Society of Hypertension. Blood Press Monit. 2007;12:233–242. doi: 10.1097/MBP.0b013e32813fa386 [DOI] [PubMed] [Google Scholar]
- 10.Ostchega Y, Nwankwo T, Sorlie PD, Wolz M, Zipf G. Assessing the validity of the Omron HEM-907XL oscillometric blood pressure measurement device in a National Survey environment. J Clin Hypertens (Greenwich). 2010;12:22–28. doi: 10.1111/j.1751-7176.2009.00199.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ostchega Y, Nwankwo T, Chiappa M, Wolz M, Graber J, Nguyen DT. Comparing Blood Pressure Values Obtained by Two Different Protocols: National Health and Nutrition Examination Survey, 2017–2018. Vital Health Stat 1. 2021:1–26. [PubMed] [Google Scholar]
- 12.Muntner P, Miles MA, Jaeger BC, Hannon Iii L, Hardy ST, Ostchega Y, Wozniak G, Schwartz JE. Blood Pressure Control Among US Adults, 2009 to 2012 Through 2017 to 2020. Hypertension. 2022;79:1971–1980. doi: 10.1161/HYPERTENSIONAHA.122.19222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr., Jones DW, Materson BJ, Oparil S, Wright JT Jr., et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289:2560–2572. doi: 10.1001/jama.289.19.2560 [DOI] [PubMed] [Google Scholar]
- 14.Williams B, Mancia G, Spiering W, Agabiti Rosei E, Azizi M, Burnier M, Clement DL, Coca A, de Simone G, Dominiczak A, et al. 2018 ESC/ESH Guidelines for the management of arterial hypertension. Eur Heart J. 2018;39:3021–3104. doi: 10.1093/eurheartj/ehy339 [DOI] [PubMed] [Google Scholar]
- 15.Flynn JT, Kaelber DC, Baker-Smith CM, Blowey D, Carroll AE, Daniels SR, de Ferranti SD, Dionne JM, Falkner B, Flinn SK, et al. Clinical Practice Guideline for Screening and Management of High Blood Pressure in Children and Adolescents. Pediatrics. 2017;140:e20171904. doi: 10.1542/peds.2017-1904 [DOI] [PubMed] [Google Scholar]
- 16.Anderson RN, Rosenberg HM. Age adjustment. National Center for Health Statistics,. https://www.cdc.gov/nchs/hus/sources-definitions/age-adjustment.htm. 1998. Accessed February 18, 2023. [Google Scholar]
- 17.Anderson RN, Rosenberg HM. Age standardization of death rates: implementation of the year 2000 standard. Natl Vital Stat Rep. 1998;47:1–16, 20. [PubMed] [Google Scholar]
- 18.Coresh J, Astor BC, Greene T, Eknoyan G, Levey AS. Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey. Am J Kidney Dis. 2003;41:1–12. doi: 10.1053/ajkd.2003.50007 [DOI] [PubMed] [Google Scholar]
- 19.Parker JD, Talih M, Malec DJ, Beresovsky V, Carroll M, Gonzalez JF, Hamilton BE, Ingram DD, Kochanek K, McCarty F, et al. National Center for Health Statistics Data Presentation Standards for Proportions. Vital Health Stat 2. 2017:1–22. [PubMed] [Google Scholar]
- 20.National Center for Health Statistics. NHANES Tutorials - module 3 - weighting. https://wwwn.cdc.gov/nchs/nhanes/tutorials/module3.aspx. Accessed October 24, 2022.
- 21.Chang W, Cheng J, Allaire J. Shiny: Web Application Framework for R. https://CRAN.R-project.org/package=shiny. 2021. Accessed October 24, 2022.
- 22.Perrier V, Meyer F, Granjon D. shinyWidgets: Custom Inputs Widgets for Shiny. https://CRAN.R-project.org/package=shinyWidgets. 2022. Accessed October 24, 2022.
- 23.Attali D, Edwards T. Shinyalert: Easily Create Pretty Popup Messages (Modals) in Shiny. https://CRAN.R-project.org/package=shinyWidgets. 2021. Accessed October 24, 2022. [Google Scholar]
- 24.R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. https://www.R-project.org/. 2022. Accessed October 24, 2022. [Google Scholar]
- 25.Muntner P, Hardy ST, Fine LJ, Jaeger BC, Wozniak G, Levitan EB, Colantonio LD. Trends in Blood Pressure Control Among US Adults With Hypertension, 1999–2000 to 2017–2018. JAMA. 2020;324:1190–1200. doi: 10.1001/jama.2020.14545 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Fryar CD, Ostchega Y, Hales CM, Zhang G, Kruszon-Moran D. Hypertension Prevalence and Control Among Adults: United States, 2015–2016. NCHS Data Brief. 2017:1–8. [PubMed] [Google Scholar]
- 27.Jaeger BC. cardioStatsUSA: Replicating analysis from Muntner et al, 2020. https://jhs-hwg.github.io/cardioStatsUSA/articles/replicate_jama_2020.html. 2022. Accessed February 18, 2023.
- 28.Jaeger BC. cardioStatsUSA: Replicating analysis from Muntner et al, 2022. https://jhs-hwg.github.io/cardioStatsUSA/articles/replicate_hypertension_2022.html. 2022. Accessed February 18, 2023.
- 29.Jaeger BC. cardioStatsUSA: Replicating analysis from the NCHS data brief, 2017. https://jhs-hwg.github.io/cardioStatsUSA/articles/replicate_nchs_2017.html. 2022. Accessed February 18, 2023.
- 30.Jaeger B, Chen L, Muntner P. Cardiometabolic statistics for US adults. https://bcjaeger.shinyapps.io/nhanesShinyBP/. Accessed December 22, 2022.
- 31.Yoon SS, Carroll MD, Fryar CD. Hypertension Prevalence and Control Among Adults: United States, 2011–2014. NCHS Data Brief. 2015:1–8. [PubMed] [Google Scholar]
- 32.Handler J, Zhao Y, Egan BM. Impact of the number of blood pressure measurements on blood pressure classification in US adults: NHANES 1999–2008. J Clin Hypertens (Greenwich). 2012;14:751–759. doi: 10.1111/jch.12009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.National Center for Health Statistics. NHANES 1988–2016 data documentation, codebook, and frequencies: Prescription medications - drug information (RXQ_DRUG). https://wwwn.cdc.gov/Nchs/Nhanes/1999-2000/RXQ_DRUG.htm. Accessed October 27, 2022.
- 34.Carey RM, Sakhuja S, Calhoun DA, Whelton PK, Muntner P. Prevalence of Apparent Treatment-Resistant Hypertension in the United States. Hypertension. 2019;73:424–431. doi: 10.1161/HYPERTENSIONAHA.118.12191 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.National Center for Health Statistics. NHANES Interactive Data Visualizations. Designed by N. Ansai and A. Lipphardt, National Center for Health Statistics. https://www.cdc.gov/nchs/nhanes/visualization/index.htm. 2023. Accessed March 7, 2023. [Google Scholar]
- 36.Giles TD, Black HR, Messerli F, White WB. Ambulatory blood pressure monitoring should be included in the National Health and Nutritional Examination Survey (NHANES). J Am Soc Hypertens. 2012;6:364–366. doi: 10.1016/j.jash.2012.07.002 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.