A Literature Review of Productivity Loss Associated with Hypertension in the United States

Abstract

A literature review of peer-reviewed articles published 2000–2019 was conducted to determine the types and extent of hypertension-associated productivity loss among adults in the United States. All monetary outcomes were standardized to 2019 $ by using the Employment Cost Index. Twenty-seven articles met the inclusion criteria. Nearly half of the articles (12 articles) presented monetary outcomes of productivity loss. Absenteeism (14 articles) and presenteeism (8 articles) were most frequently assessed. Annual absenteeism was estimated to cost more than $11 billion, nationally controlling for sociodemographic characteristics. The annual additional costs per person were estimated at $63 for short-term disability, $72–$330 for absenteeism, and $53–$156 for presenteeism, controlling for participant characteristics; and may be as high as $2362 for absenteeism and presenteeism when considered in combination. The annual additional time loss per person was estimated as 1.3 days for absenteeism, controlling for common hypertension comorbidities, including stroke and diabetes; and 15.6 days for work and home productivity loss combined, controlling for sociodemographic characteristics. The loss from absenteeism alone might be more than 20% of the total medical expenditure of hypertension. Although the differences in estimation methods and study populations make it challenging to synthesize the costs across the studies, this review provides detailed information on the various types of productivity loss. In addition, the ways in which methods could be standardized for future research are discussed. Accounting for the costs from productivity loss can help public health officials, health insurers, employers, and researchers better understand the economic burden of hypertension.

Background

An estimated 47% of U.S. adults had a diagnosis of hypertension in 2015–2018, with 79% of those having uncontrolled hypertension.¹ Hypertension is a primary risk factor for cardiovascular disease (CVD).² Annual medical expenditures for hypertension were estimated at $79 billion in 2016,³ and individuals with hypertension have an estimated $1920 higher annual medical expenditure compared with those without hypertension.⁴ In addition, hypertension-associated comorbidities such as congestive heart failure and diabetes can significantly increase medical expenditures.^5,6

Hypertension is not only associated with high medical spending but can also lead to substantial indirect costs such as productivity losses. Productivity costs are measured as the paid and unpaid (eg, household chores) work lost due to sickness, disability, and premature mortality, or, according to some authors, the frictional costs associated with replacing that productivity.⁷

Productivity costs are a common and potentially substantial proportion of the total cost of hypertension and CVD.^8–10 In 2010, it was estimated that the hypertension-associated productivity loss for paid and unpaid work and lost wages due to premature mortality was $23.6 billion (2008 $).⁸ Lost future earnings from deaths in 2015 and 2016 accounted for 8% of the total cost of hypertensive disease and 39% of the total cost of CVD (inpatient, emergency department, outpatient, home health, and medication costs, and productivity loss due to premature death).¹⁰

A review conducted by Schultz et al. of presenteeism, or reduced function while at work, identified 3 studies and found that the loss from presenteeism may be a major contributor to the total cost (medical, pharmaceutical, absenteeism, and presenteeism) associated with hypertension.⁹

Hypertension is a serious chronic health condition and also a major risk factor for other costly chronic diseases. Depending on how it is assessed, estimates of productivity loss associated with hypertension may include hypertension among common comorbid health risks (eg, high cholesterol) and/or with sequelae conditions such as stroke.¹¹

Despite recent estimates, there have been no comprehensive literature reviews focused on hypertension-associated productivity loss to summarize the findings. A 2009 review of articles that summarized the total cost of health conditions, including presenteeism, reported on 3 articles that included hypertension.⁹ A different 2009 literature review on absenteeism focused on health risks (body mass index [BMI], cholesterol, stress, physical activity, and hypertension) and identified only 4 articles that reported estimates for hypertension.¹² Other economic reviews have focused on specific health risks and conditions, such as, obesity,^13,14 physical inactivity,^15,16 and diabetes.¹⁷ The present study will be the first review to focus on hypertension-associated productivity loss aiming at determining the types and the extent of these productivity losses for U.S. adults.

Methods

Study selection

A search was conducted in November 2019 to identify peer-reviewed articles published starting in January 2000 by using PubMed, Medline, CINAHL, and EconLit databases. Hypertension keywords included: hypertension, high blood pressure, elevated blood pressure, or systolic pressure. Productivity loss terms included: indirect cost, economic burden, productivity, absenteeism, missed work, presenteeism, cost of illness, role impairment, disability, sick days, premature mortality, and premature death.

Figure 1 shows the search strategy, exclusion criteria, and inclusion of articles. The search identified 434 potentially relevant articles published from January 2000 to October 2019. The search was supplemented by a reference check of relevant articles. This added 50 potentially relevant articles. Exclusion criteria were applied to all stages of screening, and articles were excluded: if the populations were outside the United States or included children; were not full-length, peer-reviewed journal articles (eg, dissertations, conference abstracts); if articles did not present original estimates (eg, reviews, commentaries); and if the estimates were not for hypertension or productivity loss.

FIG. 1.

Selection of articles on productivity loss associated with hypertension in the United States, January 2000 to October 2019. *Other types of outcomes include labor outcomes, percentage of loss, and ratings of impairment level.

After removing duplicates (n = 73), 411 titles and abstracts were screened. Of the 411, 52 articles were potentially relevant and the full text was screened. Of the 52, 42 full-text articles were reviewed and discussed among the authors of this review. Twenty-seven articles met the selection criteria and were included in this review.

Summarizing costs of productivity loss

Information on the study populations, data sources, measurement methods, analysis methods, and productivity loss estimates was abstracted. Because hypertension is a common comorbidity and a risk factor for several health conditions, also abstracted was whether studies accounted for participant characteristics in their analyses: sociodemographic characteristics; work characteristics; health behaviors and risks (eg, tobacco, diet, physical activity); health conditions related to hypertension (eg, obesity, diabetes, cholesterol, depression, heart disease, stroke),^5,18 referred to as “hypertension comorbidities” for the remainder of the article; and health conditions unrelated to hypertension (eg, allergies, pain).

This review focused on hypertension-related productivity losses: losses of paid work or unpaid home activities; with monetary or non-monetary outcomes. Sources of productivity losses can include absences from work, reduced function while at work, and the inability to work due to disability or premature death.⁷

A recent review of productivity loss and obesity summarized the results in the following categories: absenteeism, “temporary work loss such as sick leave”; presenteeism, “reduced productivity while being present at work”; disability, as indicated by short- and long-term disability claims; and premature death.¹³ Another review of productivity loss used similar categories: work loss, “synonymous with absenteeism, defined as time off work”; work limitation, “synonymous with presenteeism, defined as time lost because of diminished capacity while at work”; and work disability, “the permanent partial or complete disablement for work purposes.”¹⁹

As unpaid productivity loss was included, the outcomes were summarized into more categories. First, literature was classified into 2 main groups: (1) articles that presented monetary outcomes and (2) articles that did not report monetary outcomes. Within those 2 groups, articles were summarized by the 7 categories of productivity loss (Supplementary Appendix SA): (1) work disability, indicated by disability claims; (2) absenteeism or sick days; (3) presenteeism, reduced function and performance while at work; (4) work productivity, where the study presents a combination of absenteeism and presenteeism; (5) home/activity productivity; (6) work and home productivity, where the article presents a combined estimate for work and home productivity loss; and (7) employment status. None of the articles presented losses due to premature death.

For articles that presented both monetary and nonmonetary outcomes (eg, time lost and the monetary values), only monetary results were reported. Productivity losses with monetary outcomes were standardized to 2019 $ by using the Employment Cost Index for total compensation for civilian workers²⁰ and are presented in 2019 $ unless otherwise noted. The estimates from the articles are presented, as available: (1) loss per person with hypertension; (2) additional loss associated with hypertension per person; (3) loss associated with hypertension per person in the study; and (4) macro-level (eg, national, state-level). Nonmonetary outcomes were organized by outcome type, including time loss per period of time and percent time loss.

This activity was determined not to be human subjects research requiring CDC's Institution Review Board approval.

Results

The articles included in this review were published from 2000 to 2019 and analyzed data that represent a span of 20 years (1995–2015; Table 1). A majority of the articles focused on employees, either of a single company or specified industries (12 articles) or of multiple industries unspecified (12 articles). Health care and insurance (n = 4), financial services and banking (n = 2), manufacturing (n = 2), oil (n = 2), beverage (n = 1), and customer service (n = 1) were among the industries reported. Articles reporting the age ranges of the populations studied included the following (in years): 18–64 (n = 10), 18 and older (n = 3), 25–54 (n = 1), 25–64 (n = 1), 17 and older (n = 1), and 18–70 (n = 1).

Table 1.

Article and Study Characteristics of Productivity Loss Associated with Hypertension in the United States, Published 2000–2019 (n = 27)

First author, year		Data year(s)	National or employee subpopulations (no. of companies)	Population characteristics (n)	Outcome typea
					$	Time	Other
Employees from a single company or specified industries
1	Allen, 201830	2011–2015	Health care (n = 1)	(n = 22,893)	X	X
2	Burton, 200436	2002	Financial services (n = 1)	Age 18–64 (n = 16,651)			X
3	Burton, 200537	2002, 2004	Financial services (n = 1)	Age 18–64 (n = 28,375)			X
4	Druss, 200043	1995	Manufacturing (n = 1)	(n = 9398)		X
5	Henke, 201024	2004–2006	Beverage (n = 1)	Age 18–64, fee for service (n = 11,217)	X
6	Kirkham, 201526	2007–2010	Manufacturing (n = 1)	Age 18–64, not pregnant (n = 17,089)	X	X
7	Kowlessar, 201127	2005–2008	Multinational company (n = 1)	Age 18–64 (n = 155,213 absence; n = 37,654 presenteeism)	X	X
8	Merrill, 201240	2010	Insurance, health care (n = 3)	(n = 19,803)			X
9	Shi, 201341	2010–2011	Finance, insurance, manufacturing, health care (n = 5)	(n = 19,121)			X
10	Tsai, 200347	1990–1999	Oil (n = 1)	(n = 2203)		X
11	Tsai, 200548	1994–2003	Oil (n = 1)	(n = 2550)		X
12	Unmuessig, 201635	2010	Health care (n = 1)	Age 18+ (n = 2216)		X	X
Employees from multiple industries not specified
13	Anesetti-Rothermel, 201142	2007	National	Age 18–64 (n = 12,860)		X
14	Asay, 201625	2008–2011	National	Age 18–64, not pregnant (n = 229,615 claims; n = 24,006 survey)	X	X
15	Druss, 200129	1996	National	Age 18+ (n = 23,230)	X
16	Goetzel, 200323	1999	Large employers (n = 6)	(n = 275,201 claims; n = 112,493 absence)	X
17	Goetzel, 200422	1995–1999, 2003	Multiple	APA (workers, n ∼ 25,000); Bank One Worker Productivity Index (n = 1039); Employer Health Coalition (member employers, n = 3910); MarketScan (n = 374,799); MIDUS (age 25–54, n = 2074); telecommunication (n = 619)	X	X	X
18	Jetha, 201644	2008–2012	Multiple, from private disability insurance companies	Age 25–64 (n = 1069 hypertension)		X
19	Kannan, 200817	2006	National	Age 18–64, BMI 27+, full-time worker (n = 19,759)			X
20	Lamb, 200632	2001–2002	Multiple (n = 27) from 1 health fair	(n = 8267)	X
21	Lenneman, 201131	2005–2008	Multiple (n = 35 + 10 health plans)	Age 17+, 20+ h/week (n = 577,186)	X		X
22	Mitchell, 201133	2007–2009	Multiple	Age 18–70, participants of large HRA (n = 1,264,117)	X
23	Trogdon, 201528	2004–2008	National	Age 18–64 (n not reported or found)	X
24	Ward, 201546	2011	National	Age 18–64 (n = 25,458; n = 16,096 employed only)		X	X
With or without employment
25	Druss, 200939	2001–2003	National	Age 18+ (n = 5692)			X
26	Kessler, 200138	1995–1996	National	Age 25–54 (n = 2074)		X
27	Merikangas, 200745	2001–2003	National	Age 18+ (n = 5692)		X

^a Time includes time loss from work and other activities. Other includes, for example, percent loss, level of impairment, any loss versus not, and employment status.

APA, American Productivity Audit; BMI, body mass index; HRA, Health Risk Assessment; MIDUS, Midlife Development in the United States.

Table 2 reports the data sources for productivity loss and the named survey instruments if applicable. Some articles included more than 1 data source to examine multiple categories of productivity loss. Seven articles summarized administrative data, such as short-term disability claims or a company's personnel records to indicate absences from work. Nearly all articles presented survey data to assess some category of productivity loss. Several articles assessed absenteeism by using a single question asking for days of work missed due to health issues (Supplementary Appendix SA).

Table 2.

Data Used in the Studies of Productivity Loss Associated with Hypertension in the United States (n = 27)

Data source for productivity loss	No. of articlesa	Instrument or questionnaire, if applicable
Administrative
MarketScan
Absence	2
Short-term disability claims	3
Insurance
Nonwork disability claims	1
Company
Absence, payroll	1
Absence, personnel	1
Health surveillance system	2
Health Risk Assessment
HealthMedia succeed	1	WPAI (n = 1)
Healthier people	2	WLQ (n = 2)
Mayo Clinic	2	WLQ (n = 2)
OptumHealth	1	WLQ (n = 1)
Not specified	2	HPLI (n = 1), WPI (n = 1)
National surveys
American Productivity Audit	1
Medical Expenditures Panel Survey	4
Midlife Development in the United States	2	HPQ (n = 1)
National Comorbidity Survey	1	SDS (n = 1)
National Health Interview Survey	1
National Health and Wellness Survey	1	WPAI (n = 1)
Other surveys	6	HPQ (n = 3), WBA-P (n = 1), WHQ (n = 1), WPAI (n = 1), WPSI (n = 2)

^a Some articles include multiple sources within the same category and/or across different categories of data sources.

HPLI, Health-related Productivity Loss Inventory; HPQ, Health and Work Performance Questionnaire; SDS, Sheehan Disability Scale; WBA-P, Well-Being Assessment-Productivity; WHQ, Work Health Questionnaire; WLQ, Work Limitations Questionnaire; WPAI, Work Productivity and Activity Impairment Questionnaire; WPI, Work Productivity Index; WPSI, The Work Productivity Short Inventory.

The most commonly used instruments using multiple questions were the Work Limitations Questionnaire (WLQ) (n = 5), the Health and Work Performance Questionnaire (n = 4), the Work Productivity and Activity Impairment (WPAI) Questionnaire (n = 3), and the Work Productivity Short Inventory (WPSI; n = 2). Of these, the WPSI attempts to attribute loss to 15 specific health conditions (study participants are asked the number of days each condition is present and eg, the number of unproductive hours in a typical 8-hour workday due to the condition and related symptoms).^9,21 Of the instruments listed here, only the WPAI measures unpaid activities.²¹

Table 3 summarizes the 12 articles reporting monetary outcomes, and Table 4 summarizes the 15 articles reporting nonmonetary outcomes. Among the 27 articles and 7 categories of productivity loss, absenteeism (14 articles) and presenteeism (8 articles) were the categories most frequently assessed. Half of the monetary articles and 80% of the non-monetary articles presented analyses controlling for participant characteristics. Across the 27 articles, the following categories of participant characteristics were included in the analyses: demographic or sociodemographic characteristics (n = 17), work characteristics (eg, type of work, exempt status; n = 6), hypertension comorbidities (eg, obesity, diabetes, cholesterol, depression, and sequelae conditions) (n = 8), health or chronic conditions unrelated to hypertension (n = 7), and conditions that were not named (n = 1).

Table 3.

Productivity Loss Associated with Hypertension from Articles with Monetary Outcomes (n = 12)

First author, year	N	Statistical methods	Covariates	Monetization (source of $ values assigned)	Cost measure, annual	Estimate ($ yeara)	2019 $ Estimateb
Work disability
1. Henke, 201024	11,217	2-part: logit, ECM	Demog, job, (1)	Short-term disability insurance payment	Additional per person	50ns (2008)	63.33
2. Goetzel, 200323	275,201	Descriptive		Short-term disability insurance payment or hourly wage + benefits, 60% (source: literature)	Per employee	8.27 (1999)	14.33
3. Goetzel, 200422	380,367	Descriptive		Short-term disability insurance payment	Per employee	7.45 (2001)	11.88
Absenteeism
3. Goetzel, 200422	380,367	Descriptive		Hourly wage + benefits (source: BLS)	Per person with hypertension	170 (2001)	271.02
4. Kirkham, 201526	17,089	GEE (poisson)—time	Sociodemog, work	Daily salary − average (source: hypothetical)	Additional per person	96 (2010c)	117.63
5. Asay, 201625	229,615; 24,006	Zero-inflated poisson—time	Sociodemog, work	Hourly compensation (source: BLS)	Additional per person	298 (2015)	330.40
6. Kowlessar, 201127	155,213	GLM (negative binomial)—time	Demog, job, (2)	Annual compensation + benefits − average (source: company)	Additional per person	57 (2008c)	72.20
2. Goetzel, 200323	112,493	Descriptive		Hourly wage + benefits (source: BLS)	Per employee	60.52 (1999c)	104.66
3. Goetzel, 200422	380,367	Descriptive		Hourly wage + benefits (source: BLS)	Per employee	46.70 (2001)	74.45
7. Trogdon, 201528		Negative binomial—time extrapolated	Non-sequelae diseases	Daily earnings including fringe − average (source: CPS)	Median State (additional)	78 M (2010)	95.57 M
5. Asay, 201625	229,615; 24,006	Est extrapolated—US employed	Sociodemog, work	Hourly compensation (source: BLS)	US employed (additional)	10.3 B (2015)	11.42 B
8. Druss, 200129	23,230	Weighted sum		Daily income (source: MEPS)	US	11.5 B (1996)	21.97 B
Presenteeism
9. Allen, 201830	22,893	Descriptive		Hourly wage − site average (source: company)	Per person with hypertension	126 (2015c)	139.70
4. Kirkham, 201526	17,089	GEE (logit) —time	Sociodemog, work	Daily salary − average (source: hypothetical)	Additional per person	43 (2010c)	52.69
6. Kowlessar, 201127	37,654	GLM (gamma)—time	Demog, job, (2)	Annual compensation + benefits − average (source: company)	Additional per person	123 (2008c)	155.80
3. Goetzel, 200422	5568	Descriptive		Hourly wage + benefits (source: BLS)	Per employee	246.73 (2001)	393.35
9. Allen, 201830	22,893	Descriptive		Hourly wage − site average (source: company)	Company	400,000 (2015c)	443485.21
Work productivity
10. Mitchell, 201133	1,264,117	Propensity score matched, GLM—time	Demographic, (3)	Daily compensation − average (source: BLS), Multiplier (1.61) for absence	Additional per person	230 (2008)	291.33
11. Lenneman, 201131	577,186	Descriptive, t-test—time		Annual earnings − average (source: BLS)	Additional per person	1890*** (2009)	2361.65
12. Lamb, 200632	8267	Descriptive		Hourly compensation (source: literature)	Per employee	105 (2002c)	161.82

^a Dollar year unless otherwise noted.

^b Standardized to 2019.

^c Last data year.

(1) Overweight/obesity, high blood glucose, high total cholesterol, physical inactivity, poor diet, depression, tobacco use, alcohol consumption, and stress.

(2) Overweight/obesity, high blood glucose, high cholesterol, inadequate exercise, poor nutrition, poor emotional health, tobacco use, high alcohol consumption, poor safety practices, and high triglycerides.

(3) Overweight/obesity, high cholesterol, arthritis, heart disease, diabetes, asthma, allergies, back pain, bronchitis, cancer, depression, heartburn, migraine, osteoporosis, pain, and smoker.

B, billion; BLS, Bureau of Labor Statistics; CPS, Current Population Survey; demog, demographic; ECM, exponential conditional mean; GEE, generalized estimating equation; GLM, general linear model; M, million; ns, not significant; MEPS, Medical Expenditures Panel Survey; sociodemog, sociodemographic.

^*** P < 0.001 for time comparisons of hypertension versus normal.

Table 4.

Productivity Loss Associated with Hypertension from Articles Without Monetary Outcomes (n = 15)

First author, year	N	Statistical methods	Covariates	Outcome	Cost measure, if applicable	Estimate
Work disability
1. Jetha, 201644	1069 HTN	OLS	Sociodemog, work	Length of disability in days annually	Per person with hypertension	41.5
Absenteeism
2. Shi, 201341	19,121	Logistic	Sociodemog (baseline), person changes	Any absence, past 4 weeks—HPQ		OR = 0.99ns
3. Unmuessig, 201635	2216	GLM, mean	Sociodemog	Hours, past 2 weeks—WHQ	Per person with hypertension	1.04**
4. Tsai, 200347	2203	Descriptive		Days per year	Per person with hypertension (per person without)	6.2 (4.0)
5. Druss, 200043	9398	OLS	Sociodemog	Days per year	Per person with hypertension	5.39
6. Tsai, 200548	2550	Descriptive		Days per year	Per person with hypertension (per person without)	11.7 M (7.1) 16.7 F (9.4)
7. Anesetti-Rothermel, 201142	12,860	Linear regression	Sociodemog, (1)	Days per year	Additional per person	1.26
8. Ward, 201546	16,096	OLS	Sociodemog, work, no. of conditions (2)	Days per year	Additional per person	−0.261ns
Presenteeism
9. Burton, 200436	16,651	Logistic	Demog, (3)	Any presenteeism, past 2 weeks—WLQ Output		OR = 1.067ns
10. Merrill, 201240	19,803	Prevalence Ratio	Sociodemog, job type	Score: 20% versus 80%, past 4 weeks—HPQ		PR = 1.23*
2. Shi, 201341	19,121	Linear regression	Sociodemog (baseline), person changes	Score, past 4 weeks—WBA-PW	Additional per person	B = 1.13**
11. Burton, 200537	28,375	GLM (neg binomial)	Demog, (4)	Score, past 2 weeks—WLQ output	Additional per person	0.0661*
3. Unmuessig, 201635	2216	GLM, Mean	Sociodemog	Hours per 2 weeks—WHQ	Per person with hypertension	1.33ns
Work productivity
12. Kannan, 200834	19,759	Linear regression	Demog, smoking, cholesterol, diabetes	% impair, past 7 days—WPAI	Additional per person	B = 5.187***
13. Druss, 200939	5692	Descriptive, survey weighted		% severe impairment, past month—SDS Score, past month—SDS 0–10	Per person with hypertension (per person without)	5.4 (18.9) 1.0 (2.7)
Home/activity productivity
12. Kannan, 200834	19,759	Linear regression	Demog, smoking, cholesterol, diabetes	% activity impair, past 7 days—WPAI	Additional per person	B = 5.057***
13. Druss, 200939	5692	Descriptive, survey weighted		% severe impairment, past month—SDS Score (home) past month—SDS 0–10	Per person with hypertension (per person without)	4.0 (17.5) 1.0 (2.8)
Work and home productivity
14. Kessler, 200138	2074	Logistic	Sociodemog	Any loss, past 30 days		OR = 1.4ns
14. Kessler, 200138	2074	Unstandardized linear	Sociodemog	Days per month	Additional per person	1.3*
15. Merikangas, 200745	5692	Linear regression	Sociodemog	Days per year	Additional per person	15.6*
Employment Status
8. Ward, 201546	25,458	Probit	Sociodemog, no. chronic conditions (2)	Employed versus unemployed	Additional per person	B = −0.07*

(1) BMI, smoking, cancer, diabetes, asthma, allergies, COPD, lower respiratory, upper respiratory, heart disease, stroke, arthritis/joint disorders, chronic neck/back pain, anxiety, and mood disorder.

(2) Arthritis, asthma, cancer, CHD, COPD, diabetes, hepatitis, weak/failing kidneys, and stroke.

(3) Allergy, arthritis, asthma, back pain, cancer, depression, diabetes, heart disease, heartburn, irritable bowel, kidney disease, menopause, and osteoporosis.

(4) Smoking, physical activity, seat belt, alcohol, relaxation medication, life dissatisfaction, physical health, job dissatisfaction, stress, cholesterol, obesity, and all other reported diseases.

^* P < 0.05; ^**P < 0.01; ^***P < 0.001 for statistical comparisons of hypertension versus normal.

B, beta; Demog, demographic characteristics; F, female; GLM, general linear model; HTN, hypertension; M, male; ns, not significant; OLS, ordinary least squares; sociodemog, sociodemographic characteristics.

Among articles that provided estimates without controlling for participant characteristics, some used approaches to attribute loss more specifically to hypertension. For example, the use of the WPSI²² attempts to attribute loss to specific conditions. Other articles presented estimates of loss imputed from temporally linked inpatient, outpatient, and pharmacy records.^22,23

Productivity loss from studies with monetary outcomes

The monetary estimates in this section are described in 2019 dollars and shown in Table 3 in both the original study-year dollars and in 2019 dollars. Three articles provided estimates for work disability. From 2 articles, annual per person or employee work disability costs ranged from $12 to $14 based on short-term disability insurance payments that were temporally linked to hypertension by using inpatient, outpatient, and pharmacy data.^22,23 One article presented the estimated annual additional loss per person for work disability at $63 for short-term disability, controlling for demographic, work, and health behavior and risks.²⁴

Seven articles presented monetary estimates for hypertension-associated absenteeism in different cost measures. One article presented the estimated annual cost of absenteeism at $271 per person with hypertension.²² From 2 articles, the annual additional cost of absenteeism per person controlling for sociodemographic and work characteristics ranged from $118 to $330.^25,26 One article presented the annual additional cost per person, also controlling for health behaviors and risks, at $72.²⁷

One article presented the annual median additional cost estimate at nearly $96 million at the state-level, controlling for health conditions unrelated to hypertension.²⁸ Two articles presented national annual totals for absenteeism: an estimated additional $11.4 billion controlling for sociodemographic and work characteristics based on data from 2008 to 2011²⁵ and nearly $22 billion without controlling for participant characteristics and based on data from 1996.²⁹

Three articles presented per person cost estimates for presenteeism. The annual cost of presenteeism was reported at $140 per person, with hypertension in 1 article.³⁰ From another article, the annual additional cost of presenteeism per person was estimated at $53, controlling for sociodemographic and work characteristics among manufacturing employees.²⁶ From the third article, the annual additional cost of presenteeism per person was estimated at $156, also controlling for health behavior and risks among employees of a multinational company.²⁷

Three articles presented work productivity loss estimates (overall or different types combined). The first article presented the annual additional cost per person at an estimated $2362 without controlling for participant characteristics and using the WPAI to determine a percentage applied to an annual salary of $44,901.³¹ The second article presented the annual per employee cost at an estimated $162 without controlling for participant characteristics and using the WPSI to attribute loss to hypertension.³²

The third article estimated that the annual additional cost per person with hypertension was $291, controlling for patient characteristics, hypertension comorbidities, health behaviors and risk, and other chronic conditions. This study also used a multiplier to add a cost to an employer for the absence. Work productivity loss was based on a question about absenteeism and presenteeism by using the WLQ.³³

Productivity loss from studies without monetary outcomes

Among the 15 articles that provided summaries of nonmonetary outcomes only, the outcomes included time (n = 9), percentage of time or impairment (n = 2), and severity or level of loss (eg, 0–10 score or rating; n = 4). The assessment recall or reporting periods included past 7 days,³⁴ past 2 weeks,^35–37 past 4 weeks or a month,^38–41 past year,^42–47 and a variable amount of time.⁴⁸

Among the 7 articles reporting days lost, 1 reported on work disability, 4 reported on absenteeism, 1 reported on monthly work and home productivity, and 1 reported on work and home productivity. One article presented the average annual estimated loss at 41.5 days in work disability due to hypertension, controlling for sociodemographic and work characteristics.⁴⁴

From 2 articles, annual absenteeism ranged from 6.2 to 16.7 days per person with hypertension without controlling for participant characteristics.^47,48 From 1 article, annual absenteeism was estimated at 5.4 days per person with hypertension controlling for sociodemographic characteristics,⁴³ and from 1 article as 1.3 additional days annually per person controlling for sociodemographic, hypertension comorbidities, health behavior and risks, and other chronic conditions.⁴² Work and home productivity loss combined amounted to 1.3 additional days per person monthly³⁸ and 15.6 additional days per person annually,⁴⁵ controlling for sociodemographic characteristics.

Among the 15 articles, statistical comparisons between people with hypertension and people without hypertension were presented in 9 articles, with 4 articles reporting more than 1 type of productivity loss. Two articles presented greater presenteeism loss among those with hypertension compared with those without hypertension, controlling for various characteristics (sociodemographic⁴⁰ and health promoting behaviors, hypertension comorbidities, and health conditions unrelated to hypertension³⁷) and 2 articles presented more work and home productivity loss among those with hypertension compared with those without hypertension, controlling for sociodemographic characteristics.^38,45 The remaining article did not report significant differences in presenteeism by hypertension status.³⁶

In addition, 4 articles presented more than 1 productivity loss type: 1 article presented significant time loss differences for absenteeism but not presenteeism among health care employees, controlling for sociodemographic characteristics³⁵; 1 article presented within-person differences over time in presenteeism but not absenteeism, controlling for baseline sociodemographic characteristics⁴¹; 1 article presented significant differences in both work productivity and home/activity productivity, controlling for demographic characteristics, smoking, cholesterol, and diabetes³⁴; and 1 article presented results that found that hypertension was negatively associated with past 12-month employment but no significant differences in absences per year for the employed, controlling for chronic conditions and hypertension comorbidities.⁴⁶

Productivity loss with and without controlling for participant characteristics

Across the 27 articles, 3 articles presented hypertension-related absenteeism estimates after controlling for other variables, contributing to the level of absenteeism (Supplementary Appendix SB). Using Medical Expenditures Panel Survey (MEPS) data, raw estimates of hypertension-associated productivity loss from absenteeism could be lowered by 15%–36% after controlling for sociodemographic characteristics, work characteristics, and comorbid risks: obesity, smoking, physical inactivity, and diabetes.²⁵

Also using MEPS, estimated hypertension-associated productivity loss from absenteeism was lowered by 51% by controlling for hypertension comorbidities (specifically, heart disease, stroke, diabetes, mood disorders) and health conditions unrelated to hypertension (eg, cancer, arthritis) compared with estimates that only controlled for sociodemographic characteristics, BMI, and smoking.⁴²

Using MarketScan data, the estimated hypertension-associated productivity loss from absenteeism was lowered by 13%–69% by controlling for sociodemographic characteristics, work characteristics, and comorbid risk factors (obesity, smoking, physical inactivity, and diabetes).²⁵ Finally, using the National Health Interview Survey, estimated absenteeism did not significantly differ by hypertension status after controlling for heart disease, stroke, diabetes, and health conditions unrelated to hypertension (eg, cancer, arthritis).⁴⁶

Discussion

There are different forms of productivity loss that can include short-term absences from work, reduced function while at work, the inability to work due to disability and premature mortality, and impairments to activities of daily living. A literature review of hypertension-associated productivity loss among U.S. adults was conducted to summarize the types and the extent of these losses. This review included 27 articles, with 24 articles focused on employed populations. Given the high prevalence of hypertension, research included in this review suggests that national estimates of associated absenteeism could amount to more than $11 billion annually.²⁵ Other categories of loss add to this cost.

The present review of hypertension-associated productivity loss found that the annual additional cost per person, controlling for various characteristics: was $63 for short-term disability²⁴; ranged from $72 to $330 for absenteeism^25–27; ranged from $53 to $156 for presenteeism^26,27; and could be as high as $2362 from combined absenteeism and presenteeism.³¹ Although this is the first review to focus on hypertension and productivity loss, a recent review of the international literature for overweight and obesity and productivity loss provides some comparison to other health conditions.¹³

That review found that the additional cost per person of disability ranged from $30 to $41 for overweight status and $21 to $439 for obesity; that the additional cost per person of absenteeism ranged from $54 to $161 for overweight status and from $89 to $1586 for obesity; that the additional cost per person of presenteeism ranged from $429 to $4175 for obesity; and that the combined cost of absenteeism and presenteeism was $5515 and from $6402 to $9104 for overweight status and obesity, respectively (2016 US $).¹³

From the present review, work characteristics were considered in 9 articles and several articles focused on specific companies and industries. Occupation and the work conditions may impact productivity loss outcomes. First, presenteeism and absenteeism may be affected by the types of occupations that allow for absences. Second, occupation and work conditions may affect the extent to which health risks and health outcomes can lead to absenteeism and disability. For example, longer working hours were associated with cardiovascular events in several prospective cohorts.^49–51 Finally, workplaces may differ in their offerings of benefits, policies, programs, low-sodium food options, and opportunities for physical activity. It was outside the scope of this review to assess work characteristics that may affect health and productivity loss.

Medication and lifestyle changes can reduce the impact of hypertension on productivity loss. Hypertension medication adherence and hypertension control has been associated with less productivity impairment.^35,52,53 Among the 27 articles reviewed, a majority did not classify participants with hypertension based on medication use (Supplementary Appendix SC) and the hypertension control status was largely unclear. In addition, estimates from articles included in this review can reflect the effects of hypertension alone or hypertension with comorbidities.

For hypertension, comorbid depression, obesity, diabetes, and high cholesterol can increase productivity costs.^5,18 In this review, 30% of the articles presented estimates controlling for hypertension comorbidities that may also impact productivity loss and a few articles summarized attributed loss based on survey questions asking specifically for losses due to hypertension.

Research needs and recommendations

This review found that a wide variety of methods were used to measure, monetize, and summarize losses, making it difficult to compare across the studies. Literature reviews of productivity loss from other health risks have also noted the heterogeneity in methods.¹³ Efforts to standardize measurement can aid in comparing indirect cost across studies and health conditions. As has been suggested, the measurement of absenteeism, presenteeism, and unpaid work within the same validated questionnaire can help address this issue.⁷

Monetization may also help standardize estimates across studies and disciplines. Fifteen of the 27 articles included in this review presented nonmonetary losses. Nine articles presented time outcomes, and 6 articles presented outcomes as scales or percentages. Having estimates of nonmonetary outcomes may be useful for the occupational health literature. Estimates, such as cost or time, make it easier to interpret productivity loss across both the occupational health and health economics literatures. This could be addressed by using standardized questionnaires or by standardizing outcomes.

It has been suggested, for example, a rating of 8 on a scale of 1–10 where 10 indicates that no distraction could be converted to 80% of a typical workday or 20% time lost perwork day.⁷ Percentages can be applied to a workday, week, month, or year to derive time and then monetized with wage.

Identifying and articulating the study perspective with comprehensive measurement for that perspective will assist with standardization of losses across studies and disciplines. The literature on hypertension-associated productivity losses was largely focused on employed populations and absenteeism and presenteeism. Virtually none of the articles included in this review explicitly stated the study perspective. Therefore, impacts to the employer that could include a combination of output delays, overtime pay to other employees, and/or payments to a hiring agency were largely not considered.^54,55

However, Asay et al. used wage multipliers on national data to provide lower estimates to account for employees being able to compensate for their absences and upper estimates to account for impacts to employers.²⁵ For economic evaluations using the societal perspective, Krol et al. recommend comprehensively measuring productivity, including unpaid work, and applying friction and human capital approaches.⁷ Unpaid work and other activities have value to society, and comprehensive measurement will aid in standardizing methods.

There are also a number of opportunities to address the research limitations and research needs of hypertension-associated productivity loss. In 2017–2018, more than 20% of U.S. adults with hypertension were unaware of this and would not report having it.⁵⁶ Therefore, studies relying on self-reported diagnosis of hypertension could lead to underestimates of hypertension-associated productivity losses. In addition, nearly half of the adults in the United States have hypertension with 79% of those adults with uncontrolled hypertension.¹

Because work characteristics may affect the extent to which health risks and health outcomes can lead to absenteeism and disability and affect employers' health insurance costs, employers have an interest in the health and productivity of their employees. Future evaluations of workplace hypertension interventions could engage employers and include measures of productivity loss as outcomes. In addition, standardization of the classification of hypertension would facilitate an understanding of hypertension-associated productivity loss.

Hypertension can be defined in different ways (Supplementary Appendix SC). Some health care professionals diagnose patients with hypertension if their blood pressure is consistently 140/90 mmHg or higher. This limit is based on guidelines and recommendations released in 2003 and 2014.^57,58 Other health care professionals diagnose patients with hypertension if their blood pressure is consistently 130/80 mmHg or higher. This limit is based on the most recent American College of Cardiology and American Heart Association's (ACC/AHA) recommendation released in 2017.

Studies on hypertension should incorporate the 2017 ACC/AHA guidelines, which include stages and different recommended treatment and follow-up by the stages of elevated blood pressure.⁵⁹ Self blood pressure monitoring may be helpful for identifying hypertension and patterns of elevated blood pressure⁶⁰ and these could be utilized in future evaluations.

A majority of articles presented hypertension-associated productivity loss without controlling for health risks and chronic conditions. Results from a small number of analyses using national data indicate that hypertension-associated productivity loss from absenteeism was lowered by 13%–69% by controlling for sociodemographic characteristics, health risk factors, and chronic conditions.^25,42 Specific to hypertension, it may be useful to provide estimates by hypertension alone, hypertension with comorbid modifiable risk factors (eg, physical activity, BMI, smoking), and hypertension with sequelae conditions such as stroke and coronary heart disease.

Social determinants of health are a major contributor of hypertension.⁶⁰ Identifying different subgroups may help to inform the contribution of socioeconomic factors and hypertension and subpopulations with multiple conditions and modifiable risk factors.

Review limitations

The present review focused on describing the types and the extent of hypertension-associated productivity loss among adults in the United States reported in recent articles providing original estimates. None of the articles included in this review reported on premature death. It may be that the data are limited for such investigations or that studies tend to focus on sequelae conditions (eg, stroke) for premature mortality rather than hypertension. In addition, since it was outside the scope of this review to include interventions, studies simulating or estimating the mortality impacts of an intervention were excluded.

It is also possible that premature mortality is a type of productivity loss that is not commonly studied. A recent review of 50 international studies for overweight and obesity and productivity loss included only 2 studies of premature mortality.¹³

Conclusion

The evidence from the literature suggests substantial productivity loss associated with hypertension. The loss due to absenteeism alone might be more than 20% of the total medical expenditure on hypertension.¹⁰ Although differences in methods and populations make it difficult to average estimates across studies and combine different forms of productivity loss, this review provides detailed information on the various types of productivity loss associated with hypertension. In addition, methods that could be standardized for future research are presented.

Accounting for the costs from productivity loss can help public health officials and employers better understand the total economic burden of hypertension. This information may be used for setting intervention priorities, resource allocation decisions, and expanding the literature in the productivity loss area.

Authors' Contributions

Dr. MacLeod made substantial contributions to the conception and design of the work; data acquisition, analysis, and interpretation; drafted the article; revised the article; and approved the version to be published. Dr. Ye made substantial contributions to the analysis and interpretation of the data; provided critical revisions to the article; and approved the version to be published. Mr. Donald made substantial contributions to the analysis and interpretation of the data; provided critical revisions to the article; and approved the version to be published. Dr. Wang made substantial contributions to the conception and design of the work; data analysis and interpretation; provided critical revisions to the article; and approved the version to be published.

Disclaimer

The findings and conclusions presented in this review do not necessarily represent the official position of the Centers for Disease Control and Prevention. The use of trade names and commercial sources is for identification only and does not imply endorsement by the United States Department of Health and Human Services.

Author Disclosure Statement

All authors conducted this work as employees or contractors of the Centers for Disease Control and Prevention. The authors declare that there are no conflicts of interest.

Funding Information

No funding was received for this article.

Supplementary Material

Supplementary Appendix SA

Supplementary Appendix SB

Supplementary Appendix SC