Self-Reported ADHD Diagnosis Status among Working-Age Adults in the United States: Evidence from the 2023 National Wellbeing Survey

Abstract

Objective:

To estimate the percentage of U.S. working-age (18–64-year-old) adults in 2023 who self-reported ever being diagnosed with ADHD by a health care professional.

Method:

We analyze data from the 2023 National Wellbeing Survey (N=7,053) to estimate self-reported lifetime ADHD diagnosis status among working-age adults, overall and by sex, age, race/ethnicity, nativity, education, and rural-urban residence.

Results:

Among U.S. working-age adults in 2023, we estimate that 13.9% (95% confidence interval=13.0%–15.0%) self-reported ever being diagnosed with ADHD by a health care professional. We find statistically significant variation by each of the demographic variables analyzed, with higher rates among working-age adults who are female, younger, non-Hispanic White, U.S.-born, less well-educated, and residing in metro areas with 250,000–1 million people (relative to those living in metro areas with 1+ million population).

Conclusion:

The percentage of U.S. working-age adults who self-report in 2023 that they have ever been diagnosed with ADHD by a health care professional (13.9%) is substantially higher than estimates from 2012 (4.25%) and a 2023 estimate of 7.8% among adults of all ages (18+ years). The increase over time may reflect changes in diagnostic criteria for children and adults, greater acceptance of adult diagnosis, over- and mis-diagnosis, and methodological issues. The difference between the 2023 estimates likely reflects study-specific differences in the constructs measured, the age range of the samples, and methodological differences in the online panels used for sampling, in quality control approaches, and in post-survey weight construction. Additional data collection and empirical research is needed to validate or refine provisional estimates based on samples drawn from online panels, and to determine explanations for the observed increase over time.

Introduction

Despite increased recognition that attention deficit/hyperactivity disorder (ADHD) diagnosis and its treatment among U.S. adults have increased over the past decade, with important implications for the well-being of working-age (18–64-year-old) adults (National Academy of Science, Engineering, and Medicine, 2024), contemporary population-level estimates of the prevalence of ADHD among working-age adults in the United States are conspicuously absent in the literature. A focus on working-age adults is critical given the adverse effects of ADHD on employment, financial stress, relationship quality, health functioning, substance misuse, and premature mortality (due in large part to high rates of injury and drug overdose) (Catalá-López et al., 2022; Das et al., 2012; Kosheleff et al., 2023). Indeed, except for one estimate from a recent study that asked U.S. adults of all ages whether they had ever been diagnosed with ADHD as a screening question for the assessment of current ADHD diagnosis (Staley et al., 2024), the contemporary percentage of U.S. working-age adults who self-report that they have ever been diagnosed with ADHD by a health care professional (i.e., self-reported ADHD) is largely unknown.

The few available estimates of the prevalence of ADHD among adults largely come from meta-analytic studies that pool older data to approximate prevalence. These studies, using multi-national data from individual studies published between 2005 and 2011, report that adult ADHD prevalence is between 2.5% and 5.0% (Simon et al., 2009; Willcutt, 2012). More recently, Song et al. (2021) conducted a meta-analysis of data from 40 unique studies published between 2005 and 2020 and representing 30 countries, and reported a pooled prevalence estimate of 2.58% for persistent adult ADHD (childhood onset) and 6.76% for symptomatic adult ADHD. The rate of lifetime adult ADHD diagnosis that can be inferred from Song et al.’s (2009) study is similar to the lifetime U.S. prevalence of self-reported ADHD among adults of all ages that can be inferred from Staley et al.’s (2024) recent study (7.8%). Individually and together, these studies suggest an increase over time in the percentage of adults ever diagnosed with ADHD. Notably, however, they do not allow for the direct estimation of the prevalence of self-reported ADHD among working-age adults. Additionally, they reveal considerable cross-national variation, which authors and others attribute mostly to differences in study methodology, such as sample definition and selection procedures, concept definition, and measurement approach (Polanczyk et al., 2014; Simon et al. 2009; Song et al., 2021; Willcutt, 2012).

It is also notable that only six of the 40 studies included in this most-recent meta-analysis included data from the United States. Moreover, only six studies included data that were collected after 2012, and none of those included data from the United States (Song et al., 2021; Table S2). Using data from the 2007 and 2012 National Health Interview Survey (NHIS), London & Landes (2021) estimated that the percentage of non-institutionalized working-age U.S. adults who self-reported ever being diagnosed with ADHD by a health care professional increased from 3.41% in 2007 to 4.25% in 2012. Until recently, no national survey has collected data on self-reported lifetime ADHD diagnosis status among U.S. adults since the 2012 NHIS. The one recent U.S. study used data collected in 2023 from commercial online survey panels as part of the U.S. National Center for Health Statistics (NCHS) Rapid Surveys System (RSS) to focus on the self-reported prevalence of current ADHD diagnosis among adults of all ages (6.0%). That study suggests a lifetime prevalence of 7.8% among adults of all ages (inferred from the footnote to Table 1 indicating that 1.8% self-reported ever being diagnosed with ADHD and not currently having it) (Staley et al., 2024).

Table 1:

Population Description, U.S. Population ages 18–64 in 2023, National Wellbeing Survey.

	N1	Original NWS Weight %2	95% CI3	Adjusted NWS Weight %4	95% CI3
Total	7,053	100.0	-----	100.0	-----
Sex
Male	3,314	50.1	[48.8–51.4]	54.1	[52.5–55.6]
Female	3,739	49.9	[48.6–51.2]	46.0	[44.4–47.5]
Age (in Years)
18–29	1,844	25.3	[24.2–26.4]	27.3	[25.9–28.6]
30–39	1,706	22.8	[21.8–23.9]	23.8	[22.5–25.1]
40–49	1,500	21.1	[20.0–22.2]	21.3	[20.0–22.6]
50–64	2,003	30.8	[29.6–32.0]	27.7	[26.4–29.1]
Race/Ethnicity
Non-Hispanic White	4,247	58.8	[57.5–60.0]	57.1	[55.5–58.6]
Non-Hispanic Black/African American	911	11.9	[11.1–12.7]	12.9	[12.0–13.9]
Hispanic (All Races)	1,321	18.7	[17.6–19.6]	19.0	[17.9–20.2]
Non-Hispanic Asian/Pacific Islander	367	6.1	[5.5–6.8]	6.6	[5.8–7.4]
Non-Hispanic Other/Multiple Race/Ethnicity	207	4.6	[4.0–5.3]	4.4	[3.7–5.2]
Nativity
Foreign-Born	503	8.4	[7.6–9.2]	8.8	[8.0–9.8]
U.S.-Born	6,550	91.6	[90.9–92.4]	91.2	[90.2–92.0]
Education
Less than High School	716	10.5	[9.8–11.3]	8.3	[7.6–9.0]
High School	2,171	26.4	[25.4–27.5]	25.0	[23.8–26.2]
Associate Degree/Some College	2,464	31.3	[30.2–32.5]	29.7	[28.4–31.0]
Bachelor’s Degree or More	1,702	31.8	[30.5–33.1]	37.1	[35.5–38.7]
Rural-Urban Continuum
Metro, 1 Million or More Population	3,196	56.9	[55.6–58.1]	59.5	[58.0–60.9]
Metro, 250,000–1 Million Population	1,149	21.0	[19.9–22.1]	19.8	[18.6–21.1]
Metro, <250,000 Population	859	8.9	[8.3–9.5]	8.4	[7.7–9.1]
Large Nonmetro, Adjacent to Metro Area	565	4.0	[3.7–4.4]	3.6	[3.3–4.0]
Large Nonmetro, Not Adjacent to Metro Area	210	1.5	[1.3–1.7]	1.3	[1.1–1.5]
Medium Nonmetro, Adjacent to Metro Area	522	4.2	[3.9–4.6]	4.0	[3.6–4.5]
Medium Nonmetro, Not Adjacent to Metro Area	287	2.3	[2.0–2.5]	2.1	[1.8–2.4]
Small Nonmetro, Adjacent to Metro Area	116	0.6	[0.5–0.7]	0.6	[0.5–0.8]
Small Nonmetro, Not Adjacent to Metro Area	149	0.7	[0.6–0.8]	0.7	[0.6–0.9]

Notes:

^1.Unweighted number of cases in the sample.

^2.Weighted percentage. Original NWS Weight refers to the post-stratification weight included in the data. Using this weight yields estimates that are demographically representative of the U.S. population ages 18–64 by sex, age, race/ethnicity, education, and rural-urban continuum (Monnat et al., 2024).

^3.95% confidence interval.

^4.Weighted percentage. Adjusted NWS Weight refers to the adjustment we made to the original weight to make the NWS self-rated health distribution similar to the self-rated health distribution in the 2023 NHIS. See the Data and Methods section for details on how adjusting the weight was accomplished.

There are several reasons to hypothesize that lifetime self-reported ADHD diagnosis among U.S. working-age adults may have increased since the 2012 NHIS data were collected. First, over time, there has been an expansion and refinement of the criteria for diagnosing ADHD. The 5^th edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM) was revised to include more signs and symptoms of adult ADHD (American Psychiatric Association, 2013; Epstein & Loren, 2013). While some would refer to this critically as medicalization of deviance that increases the potential for over- and mis-diagnosis (Conrad & Potter, 2000), others would see it as necessary medical-scientific refinement of diagnostic criteria deriving from advances in knowledge about the etiology of ADHD (e.g., genetics) and its presentation in different subpopulations (e.g., girls and women). Over time, sub-types of ADHD more common among girls and women have been defined (Schiros, London, & Antshel, 2023). While growth over time in the prevalence of self-reported ADHD diagnosis might reflect community-based over-diagnosis and mis-diagnosis of related conditions among children and adults (Chamberlain, Cortese, & Grant, 2021; Cotugno 1993; Desgranges et al., 1995), those processes and outcomes likely reflect the influence of the expansion and refinement of the official diagnostic criteria for ADHD on lay people and clinicians across a range of settings.

Second and related to the foregoing, the percentage of children and adolescents who are reported to have ever diagnosed with ADHD has increased over time. For example, from 1997 to 2016, the percentage of U.S. children and adolescents who are reported to have been diagnosed with ADHD is estimated to have increased from 6.1% to 10.2% (Xu et al., 2018). Using data from the 2022 National Survey of Children’s Health (NSCH), Danielson et al. (2024) estimated that 11.4% of 3–17 year old children in the United States are reported to have ever been diagnosed with ADHD. Even accounting for over- and mis-diagnosis in childhood, adolescence, and adulthood, which may be substantial (Chamberlain, Cortese, & Grant, 2021; Polanczyk et al., 2007, 2014; Sciutto & Eisenberg, 2007), mis-reporting by adults, and differential adult mortality by ADHD diagnosis status (London & Landes, 2016, 2022), the documented increase over time (across cohorts) in the percentage of children and adolescents who are reported to have ADHD should somewhat mechanically translate into an increase in the percentage of U.S. working-age adults who self-report in 2023 that they have ever been diagnosed with ADHD.

A third factor is increasing acceptance that not all persons diagnosed by a health care professional with ADHD are first diagnosed in childhood and that ADHD symptoms sometimes persist into adulthood. Although there is debate about how much we should expect ADHD to be initially diagnosed in adulthood and the consequences of age at diagnosis for functioning and well-being across the life course (Hutt Vater et al., 2024), initial diagnosis of ADHD during adulthood has been occurring officially since 1994 when the American Psychiatric Association first included adult ADHD as a diagnostic category in the DSM-IV (American Psychiatric Association, 2008). Driven by many factors, over time, the diagnosis and consequences of ADHD among adults have become more accepted and commonly recognized (Conrad & Potter, 2000), and possibly less stigmatized due to becoming more normative. Normalization and destigmatization could increase the willingness of adults to self-report ADHD on surveys. Notably and perhaps representing a “correction” for the under-diagnosis of girls in earlier cohorts, there is evidence of larger intra-cohort increases over time in self-reported ADHD diagnosis among 18–64-year-old women relative to men (London & Landes, 2021). In their recent study, Staley et al. (2024) reported that 55% of adults of all ages with current ADHD received their diagnosis during adulthood. Relatedly, the demand for prescription stimulants used to treat ADHD has been rising for about two decades, especially among women aged 20 to 49 (Castle et al., 2007; Sibley, 2023).

Finally, there are empirical reasons to believe that there has been an increase in the prevalence of self-reported lifetime ADHD diagnosis status among working-age adults. The recent estimates of 6.0% with current ADHD diagnosis and 7.8% ever being diagnosed with ADHD among adults of all ages (18+ years old) (Staley et al., 2024) are substantially higher than estimates of lifetime ADHD from 2012 among working-age (18–64-year-old) adults (4.25%) (London & Landes, 2021). Given that, for historical reasons related to diagnostic criteria and practice, the prevalence of self-reported ADHD diagnosis drops substantially at older ages, it is likely that the overall prevalence of self-reported lifetime ADHD among working-age adults in the Staley et al. (2024) study is even higher than the estimated 7.8% among adults of all ages.

Estimating the prevalence of both self-reported lifetime and current ADHD is important for different reasons. Point-in-time estimates of the percentage of U.S. adults who self-report ever being diagnosed with ADHD by a health care professional are not equivalent to estimates of the current prevalence of clinician-verified ADHD among adults, may not align with diagnoses based on state-of-the-science clinical assessments (Chamberlain, Cortese, & Grant, 2021), and may not be directly applicable to clinical planning processes (Staley et al., 2024). In the absence of information on age at diagnosis and current symptomology, it is not possible to estimate persistent adult ADHD, symptomatic adult ADHD, or differentiate subtypes of ADHD. While self-reported ADHD lifetime and current diagnosis status is subject to a range of potential biases and errors that may lead to the over-estimation of the prevalence of ADHD (Chamberlain, Cortese, & Grant, 2021), estimating the prevalence of self-reported lifetime ADHD diagnosis is important because self-reported belief that one has ever been diagnosed with ADHD is a catalyst for self-identification, health care service use, treatment decision-making, accommodation seeking, social connection or isolation, potentially harmful self-management activities such as substance use, and stigma management. Estimating lifetime ADHD diagnosis in conjunction with current ADHD diagnosis is also necessary to identify, at a point in time, those who no longer identify with the diagnosis given to them by a health care professional.

Based on the foregoing considerations, we expect that the percentage of U.S. working-age adults who self-report in 2023 that they have ever been diagnosed with ADHD by a health care professional will be considerably higher than extant estimates based on data from 2012 and earlier. In this paper, we use data from a national sample of non-institutionalized U.S. working-age (18–64-year-old) adults collected in 2023 to estimate self-reported ADHD diagnosis status, overall and by selected demographic characteristics. Additionally, in a supplemental analysis, we use public-use data from the 2023 U.S. National Center for Health Statistics (NCHS) Rapid Surveys System (RSS) (version 2)—the data used by Staley et al. (2023)—to directly estimate the prevalence of lifetime self-reported ADHD among working-age adults overall and by selected demographic subgroups. These estimates from the NCHS RSS serve as a point of comparison to estimates derived from the NWS and have not previously been published.

DATA and METHODS

Data

We use data from the 2023 National Wellbeing Survey (NWS) (Monnat et al., 2023) – one of the few contemporary national surveys of U.S. adults that measures self-reported ADHD diagnosis status. The NWS is an annual, cross-sectional, web-based survey of non-institutionalized 18–64-year-old U.S. adults who are able to read English. NWS participants are recruited online by, and the survey is administered through, Qualtrics Panels (see Monnat et al., 2024 for details about the NWS design, sample recruitment, and demographic representativeness). The 2023 survey was the first to include a question about ADHD diagnosis status. For the 2023 NWS, 167,451 Qualtrics panel members were initially e-mailed. Of those, 38,519 clicked the link to the survey landing page and 18,499 were deemed eligible to participate. Overall, 14,891 panel members completed the survey. Collectively, Qualtrics and the NWS research team dropped 7,786 surveys due to concerns about data quality (e.g., speeding, straight-lining), resulting in 7,105 completed quality surveys. Accordingly, the overall response rate as traditionally defined is 8.9%, while the quality survey completion rate among those who accessed the survey landing page and were eligible to complete it is 38.4%. Online surveys historically have low overall response rates. The NWS response rate is in line with other national online surveys, including the U.S. Census Bureau’s Household Pulse Survey, which ranges from 6.5% to 7.0%, and is higher than the 3.8% and 4.0% response rates for the NCHS panels from which the recent Staley et al. (2024) estimates are derived. The 2023 NWS quality survey completion rate is near the estimated average (44.1%) for web-based surveys (Wu, Zhao, & Fils-Aime, 2022).

Participants completed the survey between June and September of 2023. The NWS oversamples nonmetropolitan residents, but the dataset includes a post-stratification weight to make analyses demographically representative of the U.S. non-institutionalized working-age population with respect to sex, age group, race, Hispanic ethnicity, education, and rural-urban continuum.

Measures

The dependent variable is self-reported ADHD diagnosis status (yes=1, no=0). Participants were asked: “Have you ever been told by a doctor, nurse, or other health care professional that you have any of the following medical conditions?” This question was followed by a list of fourteen medical conditions that included “Attention Deficit Hyperactivity Disorder (ADHD) or Attention Deficit Disorder (ADD).”

We examine the percentage who self-reported ever being diagnosed with ADHD by a health care professional, overall and by sex, age, race, Hispanic ethnicity, nativity, educational attainment, and rural-urban residence. We measure sex as a binary variable (female, male); we excluded from the analytic sample a small number of participants who identified their sex as non-binary. We measure age categorically (18–29, 30–39, 40–49, 50–64 years) and self-reported race/ethnicity as: non-Hispanic White; non-Hispanic Black/African American; non-Hispanic Asian/Pacific Islander; Hispanic (all races); and non-Hispanic multiple or other race. We include non-Hispanic Alaskan Natives/American Indians in the multiple or other race category because there were too few to maintain a separate subgroup. We measure nativity dichotomously (foreign-born, U.S.-born). Education is a four-category variable: less than high school; high school graduate; Associate degree/some college; Bachelor’s degree or higher). Urban-rural residence is measured using the U.S. Department of Agriculture’s Economic Research Service 2023 Rural-Urban Continuum Codes (RUCC) based on county classifications: (1) metro, population of 1 million or more; (2) metro 250,000–1 million; (3) metro, <250,000; (4) large nonmetro (urban population ≥20,000), adjacent to metro area; (5) large nonmetro, not adjacent to metro area; (6) medium nonmetro (urban population 2500–19,999), adjacent to metro area; (7) medium nonmetro, not adjacent to metro area; (8) small nonmetro (urban population <2,500), adjacent to metro area; and (9) small nonmetro, not adjacent to metro area.

Analytic Approach

We estimate descriptive statistics for the prevalence of self-reported ADHD diagnosis status, overall and by demographic subgroups. We report point estimates and 95% confidence intervals (95% CIs). After removing 52 participants who had missing information on one or more of the analytic variables, the final analytic sample includes 7,053 participants. We assess statistical significance using a design-based F statistic and by examining 95% CIs. All analyses were conducted using Stata 18 and all estimates are weighted.

In preliminary analyses, we compared the distribution of self-rated physical health in the 2023 NWS to the distribution of self-rated health among working-age adults in the 2023 NHIS, the 2022 Behavioral Risk Factor Surveillance System (BRFSS), and the 2022 National Survey on Drug Use and Health (NSDUH). BRFSS and NSDUH data for 2023 have not yet been released. Although the NWS self-rated physical health measure is not identical to the self-rated health measure in the other three surveys (i.e., the NWS asks specifically about physical health, while the other surveys ask about health without specifying physical health), our aim was to check that the 2023 NWS yielded a reasonably comparable estimate, which would help validate it as a data source for estimating the percentage of adults who self-reported ever being diagnosed with ADHD by a health care professional. This exercise proved instructive.

As seen in the left-most column in Figure 1, the population represented by the originally weighted 2023 NWS is less healthy than the populations represented by the three other weighted survey samples (the middle three columns in Figure 1). Specifically, 2023 NWS participants were less likely to report their physical health to be excellent or very good and were more likely to report their physical health to be fair or poor. This is consistent with the 2021 NWS Methodology Report, which compared the same three data sources collected in 2021 to the 2021 NWS (Monnat et al., 2024), but it is counter to expectations given the difference in measurement. Because ADHD is known to be correlated significantly with negative health outcomes and fair/poor self-rated health (Das et al., 2012; Landes & London, 2021), there is a risk of overestimating the percentage self-reporting ever being diagnosed with ADHD by a health care professional in the NWS using the post-stratification weight that adjusts only for demographic representativeness.

Figure 1: Percent Distribution of Self-Reported Health in Five National Surveys.¹.

Notes:

1. NWS (original) refers to the weighted sample that is representative of the U.S. working-aged (18–64 year old) adult population by sex, age, race/ethnicity, education, and rural-urban continuum. NWS (adjusted) refers to the re-weighted sample with a self-rated health distribution similar to the self-rated health distribution in the 2023 NHIS. See the Data and Methods section for details on how re-weighting was accomplished.

To address this concern, we standardized the original NWS post-stratification weight to the 2023 NHIS self-rated health distribution. For each self-rated health category, we estimated the percentage in that category in the 2023 NHIS and divided it by the percentage in the 2023 NWS measure. We then multiplied the NWS post-stratification weight for each respondent by the relevant ratio. For example, for excellent health, the ratio was 2.5 because the 2023 NWS included a smaller percentage of participants reporting excellent health than the 2023 NHIS. Thus, for each participant who reported excellent physical health in the 2023 NWS, we multiplied the original NWS weight by 2.5. The adjustment factors for the other self-rated health categories are 1.3 (very good), 0.8 (good), 0.5 (fair), and 0.4 (poor). As seen in the fifth column from the left in Figure 1, this adjustment yielded a self-rated health distribution from the 2023 NWS that is very similar to that from the 2023 NHIS.

We use both the original and the adjusted NWS weights when estimating self-reported ADHD diagnosis. We present both sets of results to demonstrate the impact of the adjustment and because we cannot directly validate the accuracy of any NWS-based estimate of self-rated ADHD diagnosis status since there is no other contemporary national survey that measures and systematically reports self-reported ADHD diagnosis status among working-age adults. Our preferred estimates are those obtained using the adjusted NWS weight; adjustment makes the population more representative of the national population with respect to health (i.e., “healthier” than the originally weighted sample) and should yield conservative estimates of self-reported ADHD diagnosis among working-age adults and change in its prevalence since 2012.

In addition to our main analysis, we present the results of a supplemental analysis using public-use data from the 2023 U.S. National Center for Health Statistics (NCHS) Rapid Surveys System (RSS) (version 2), which is available here: Rapid Surveys System - Data Files and Documentation. The NCHS RSS uses similar but not identical methods to the NWS and thus provides a useful comparison to the estimates obtained from the 2023 NWS.

RESULTS

Table 1 presents a description of the population. The first set of results is based on the original NWS weight and the second set is based on the adjusted weight. Overall, the differences in the demographic characteristics represented by the differently weighted samples are relatively small (i.e., 4 percentage points or less). In most cases, overlapping 95% CIs suggest that the estimates from the two samples are within the range of sampling error. Relative to the originally weighted sample, the population represented by the sample using the adjusted NWS weight is significantly (based on non-overlapping 95% CIs) more likely to be male, younger (fewer 50–64 year olds), more highly educated (fewer less than high school, more Bachelor’s degree or more), and more urban (more metro area, 1 million or more). Although there are significant differences between the two populations represented by the differently weighted samples, gaps between the 95% CIs for the same characteristic are small even when they are non-overlapping.

Since these demographic factors are known to be associated with the prevalence of ADHD (Song et al., 2021; Landes & London, 2021), albeit to some extent in countervailing directions, it is important to consider the potential influence of the demographic differences in the populations represented by the differently weighted samples on the obtained estimates of self-reported ADHD diagnosis status. While changes to the educational distribution in the adjusted population relative to the original population will tend to lower the self-reported ADHD diagnosis prevalence estimate, the higher percentage of males and the lower percentage of adults aged 50–64 years old will tend to increase it.

Here, we focus on describing the age 18–64 population represented using the adjusted NWS weight since we focus on the self-reported ADHD diagnosis status estimates obtained using the adjusted weight. The population is 54.1% male, and approximately half are between the ages of 18 and 39 years. Over half (57.1%) are classified as non-Hispanic White, 12.9% are non-Hispanic Black/African American, 19.0% are Hispanic (all races), 6.6% are non-Hispanic Asian/Pacific Islander, and 4.4% are non-Hispanic other or multiple race/ethnicity. Approximately 9% are foreign born. One-third have a high school diploma or less, slightly less than one-third have an Associate degree/some college, and slightly more than one-third have a Bachelor’s degree or more. More than half live in metro counties with 1 million or more inhabitants, 19.8% live in metro counties with 250,000 to 1 million inhabitants, 8.4% live in metro counties with under 250,000 inhabitants, and 12.3% live in nonmetro counties.

Table 2 presents estimates of the percentage of non-institutionalized U.S. working-age adults who self-reported in 2023 that they had ever been diagnosed with ADHD by a health care professional, overall and by demographic subpopulation. The first set of columns present estimates using the original NWS weight; the second set presents estimates using the adjusted weight. Unless otherwise indicated, we discuss the results obtained using the adjusted weight. As expected based on the self-rated health adjustment we made, the estimated percentage is lower when using the adjusted NWS weight (13.9% [95% CI=13.0%–15.0%]) than when using the original weight (15.2% [95% CI=14.3%–16.1%]). Notably, our preferred estimate of the prevalence of self-reported ADHD diagnosis in 2023—13.9%—is substantially higher than previous estimates for this population (4.25% in 2012; London & Landes, 2021). We consider our preferred estimate to be plausible given the levels of diagnosis reported for children and youth in recent years (e.g., 11.4% of 3–17 year old U.S. children in 2022; Danielson et al., 2024) and the likely increase in original diagnosis in adulthood over the past decade (Staley et al., 2024).

Table 2:

Estimated Prevalence of Self-Reported ADHD Diagnosis Status in the U.S. Population ages 18–64 in 2023, Overall and by Selected Demographic Subpopulations, National Wellbeing Survey.

	Original NWS Weight1			Adjusted NWS Weight2
	%3 ADHD	[95% CI]4	p5	%3 ADHD	[95% CI]4	p5
Total	15.2	[14.3–16.1]	---	13.9	[13.0–15.0]	---
Sex
Male	14.2	[13.0–15.5]	<.05	12.9	[11.6–14.4]	<.05
Female	16.1	[14.9–17.5]		15.1	[13.7–16.2]
Age (in Years)
18–29	22.0	[20.0–24.1]	<.001	20.9	[18.7–23.3]	<.001
30–39	19.6	[17.6–21.8]		16.2	[14.3–18.4]
40–49	13.2	[11.5–15.3]		11.9	[10.0–14.2]
50–64	7.6	[6.4–9.0]		6.6	[5.4–18.1]
Race/Ethnicity
Non-Hispanic White	16.1	[14.9–17.3]	<.001	14.2	[13.0–15.6]	<.001
Non-Hispanic Black/African American	10.7	[8.7–13.0]		10.8	[8.6–13.4]
Hispanic (All Races)	15.1	[13.2–17.3]		14.5	[12.3–17.0]
Non-Hispanic Asian/Pacific Islander	9.4	[6.7–13.0]		9.0	[6.0–13.8]
Non-Hispanic Other/Multiple Race/Ethnicity	23.2	[17.3–30.4]		24.3	[17.6–32.5]
Nativity
Foreign-Born	8.2	[5.9–11.2]	<.001	6.2	[4.4–8.8]	<.001
U.S.-Born	15.8	[14.9–16.8]		14.7	[13.6–15.8]
Education
Less than High School	23.9	[20.7–27.5]	<.001	22.7	[19.2–26.7]	<.001
High School	18.2	[16.5–20.0]		17.5	[15.7–19.6]
Associate Degree/Some College	13.9	[12.5–15.5]		13.9	[12.3–15.7]
Bachelor’s Degree or More	11.0	[9.5–12.8]		9.9	[8.0–11.4]
Rural-Urban Continuum
Metro, 1 Million or More Population	13.4	[12.2–14.7]	<.001	12.6	[11.3–14.0]	<.01
Metro, 250,000–1 Million	18.2	[15.9–20.6]		16.4	[14.0–19.0]
Metro, <250,000	17.5	[15.0–20.3]		15.4	[12.9–18.3]
Large Nonmetro, Adjacent to Metro Area	15.7	[12.9–19.0]		16.3	[13.0–20.2]
Large Nonmetro, Not Adjacent to Metro Area	17.2	[12.7–22.9]		15.6	[11.0–21.8]
Medium Nonmetro, Adjacent to Metro Area	18.0	[14.9–21.6]		15.8	[12.6–19.7]
Medium Nonmetro, Not Adjacent to Metro Area	18.5	[14.4–23.5]		17.7	[13.2–23.3]
Small Nonmetro, Adjacent to Metro Area	10.9	[6.0–19.0]		8.5	[4.2–16.2]
Small Nonmetro, Not Adjacent to Metro Area	10.6	[6.6–16.4]		11.0	[6.3–18.7]

Notes:

^1.Weighted percentage. Original NWS Weight refers to the post-stratification weight included in the data. Using this weight yields estimates that are demographically representative of the U.S. population ages 18–64 by sex, age, race/ethnicity, education, and rural-urban continuum (Monnat et al., 2024).

^2.Weighted percentage. Adjusted NWS Weight refers to the adjustment we made to the original weight to make the NWS self-rated health distribution similar to the self-rated health distribution in the 2023 NHIS. See the Data and Methods section for details on how adjusting the weight was accomplished.

^3.Weighted percent.

^4.95% confidence interval.

^5.Probability value based on a design-based F-statistic.

Table 2 also shows that the estimated prevalence of self-reported ADHD diagnosis status among working-age adults varies significantly by sex, age, race/ethnicity, nativity, educational attainment, and rural-urban residence. The estimated prevalence is significantly higher among females, younger adults, adults identifying as other and multiple race/ethnicity, the U.S.-born, adults with less education, and adults residing in metro areas with 250,000–1 million people (relative to adults living in metro areas with 1 million or more people). Notably, the estimated sex difference in self-reported ADHD diagnosis among working-age adults is relatively small (2.2 percentage points) and, somewhat unexpectedly, is higher among women than men. This difference is statistically significant (p=0.03), although the 95% CIs for men and women overlap (suggesting equivalence). Overall, these patterns are generally consistent with the extant literature and hold regardless of whether the original or adjusted NWS weight is used for the analysis.

Here, we present the results of a supplemental analysis of the 2023 NCHS RSS (version 2) data used by Staley et al. (2023) to estimate current self-reported ADHD diagnosis status among adults of all ages. For this supplemental analysis, we focus on lifetime self-reported ADHD diagnosis status and limit the sample to working-age adults with no missing data on the variables we include. Focusing on self-reported lifetime diagnosis status and working-age adults will lead to higher estimates than those reported by Staley et al. (2023). Those estimates will be directly comparable to the 2023 NWS estimates we report in this paper. The results we report are weighted using the weight provided in the NCHS RSS public-use file. Since this weight takes the 2023 NHIS self-rated health distribution into account, we included the NCHS RSS self-rated health distribution in Figure 1. As seen in the right-most column of Figure 1, the self-rated health distribution in the 2023 NCHS RSS includes a higher percentage of adults reporting good health and lower percentage reporting excellent health than either the 2023 NHIS or the 2023 NWS.

Table 3 presents the estimates of self-reported lifetime ADHD diagnosis status among U.S. working-age adults, overall and by demographic subpopulation, based on the 2023 NCHS RSS data. Overall, using this data source, we estimate that 9.6% [95% CI=8.6%–10.6%] of working-age adults self-report that they have ever been diagnosed with ADHD by a health care professional. Statistically significant variation is evident by sex, age, race/ethnicity, and education, with higher rates observed among male, younger, non-Hispanic White, and lesser-educated adults.

Table 3:

Estimated Prevalence of Self-Reported ADHD Diagnosis Status in the U.S. Population ages 18–64 in 2023, Overall and by Selected Demographic Subpopulations, NCHS Rapid Surveys System

	RSS1
	%2 ADHD	[95% CI]3	P4
Total	9.6	[8.6–10.6]	---
Sex
Male	10.6	[9.2–12.3]	<.05
Female	8.5	[7.3–9.9]
Age (in Years)
18–29	15.2	[12.8–18.1]	<.001
30–39	11.3	[9.4–13.6]
40–49	8.5	[6.6–10.7]
50–64	4.0	[3.1–5.2]
Race/Ethnicity
Non-Hispanic White	11.0	[9.7–12.5]	<.01
Non-Hispanic Black/African American	6.9	[4.8–9.7]
Hispanic (All Races)	8.3	[6.2–10.8]
Non-Hispanic Other Race/Ethnicity	6.8	[4.6–10.0]
Education
High School or Less	11.0	[9.2–13.1]	<.01
Some College	10.4	[8.6–12.4]
Bachelor’s Degree or More	7.3	[6.1–8.7]
Metropolitan Status
Metropolitan Area	9.3	[8.3–10.4]
Non-Metropolitan Area	11.3	[8.6–14.7]

Notes:

^1.N=4,983 working-age adults with no missing data.

^2.Weighted percent.

^3.95% confidence interval.

^4.Probability value based on a design-based F-statistic.

DISCUSSION

Drawing on newly available 2023 NWS data, we directly estimated the percentage of non-institutionalized U.S. working-age (18–64-year-old) adults who self-report ever being diagnosed with ADHD by a health care professional, overall and by demographic subpopulation. Our preferred estimate of the prevalence of self-reported lifetime ADHD diagnosis among working-age adults is 13.9% (95% CI=13.0%–15.0%), which, consistent with expectations, is substantially higher than extant estimates of the prevalence of self-reported lifetime ADHD diagnosis among working-age adults dating from 2012 or earlier. The finding that the prevalence of self-reported lifetime ADHD diagnosis among working-age adults has increased over time is consistent with: increasing prevalence among children (Xu et al., 2018; Danielson et al. 2024); increasing diagnosis among adults (London & Landes, 2021; Staley et al., 2024); and increasing demand for ADHD medication among adults (Castle et al., 2007; Sibley, 2023). While differences in study methodology between the 2023 NWS and the 2012 NHIS (i.e., sampling strategy, question structure) likely accounts for some of this estimated increase over time, it likely does not account for all of it. Both surveys measured self-reported lifetime ADHD diagnosis status with a single question.

Notably, the prevalence estimate we report for working-age adults is higher than the recently reported estimates of self-reported lifetime and current ADHD diagnosis prevalence for U.S. adults of all ages—7.8% and 6.0%, respectively (Staley et al., 2024). While the estimates from these two studies appear to be substantially different, a close analysis suggests that the results are more similar than they appear to be at first glance. Two factors account for much, but not all, of the difference in the estimates: (1) we measure self-reported ever diagnosed with ADHD and Staley et al. (2024) measure self-reported current ADHD diagnosis; and (2) we include working-age (18–64-year old) adults only and Staley et al. (2024) include adults of all ages (18+ years old). The estimates in our study are higher than those in Staley et al. (2024) in part because ever being diagnosed with ADHD is a broader construct than current ADHD diagnosis, and in part because ADHD diagnosis is more prevalent among individuals born in more-recent cohorts (working-age adults) than among individuals born earlier (adults age 65+ years old). To make this more explicit, we conducted a supplemental analysis using the same data source as Staley et al. (2024) and estimated that the self-reported lifetime ADHD diagnosis prevalence among U.S. working-age adults in the United States is 9.6% [95% CI=8.6%–10.6%], Thus, once we take those design differences into account, the prevalence estimates from the two studies differ by about 4 percentage points.

What accounts for the remaining difference in the estimates from the two studies is not exactly clear. While the methods used in the two studies are similar in many respects, there are methodological differences as well (National Center for Health Statistics, 2024). Similarities between the two studies include: (1) use of online panels; (2) use of self-report measures; (3) online data collection; (4) similar response rates; (5) the development of post-survey weighting schemes to enhance generalizability; and (6) standardization to the self-rated health distribution in the 2023 NHIS. However, there are nuanced differences embedded within these methodological similarities: (1) different online panels were used; (2) one study allowed for telephone interviews as well as online survey data collection; and (3) the quality control measures and post-survey weighting approaches taken in the two studies differed. It may also be the case that participants are less willing to report current than lifetime ADHD diagnosis status because acknowledging current diagnosis might carry more stigma. The methodological approaches used in the two studies allow for generalization to distinct target populations that are demographically representative, but it is likely that there was sampling bias in each study since neither used gold-standard random sampling methods. These differences may have contributed to observed differences in the estimates of self-reported ADHD diagnosis status.

In our main NWS analyses, we also find significant variation in the prevalence of self-reported ADHD diagnosis among working-age adults by sex, age, race/ethnicity, nativity, education, and rural-urban residence, with observed patterns similar to those commonly observed in the literature (and to those observed in our supplemental analyses based on the NCHS RSS data). Somewhat surprisingly, we find a significantly higher prevalence of self-reported ADHD diagnosis among women than among men. The sex difference is quite narrow (2.2 percentage points) and the 95% CIs overlap. On balance, the evidence suggests that the sex difference has narrowed and a similar percentage of women and men now self-report ever being diagnosed with ADHD by a health care professional. This narrowing of the sex gap is consistent with evidence that the prevalence of self-reported ADHD diagnosis among women of all ages has increased rapidly (London & Landes, 2021).

Overall, the estimates reported in this paper represent a set of current estimates of self-reported lifetime prevalence of ADHD among working-age U.S. adults. Having these estimates is valuable for the research community and for planning across a range of contexts. More broadly, future research that aims to validate or challenge these estimates and determine explanations for the observed increase is warranted (National Academy of Science, Engineering, and Medicine, 2024). Such research might use similar and different methodologies to determine the extent to which sample selection, weighting, measurement approach (i.e., question wording), and survey modality affect estimates of prevalence. Future research aimed at understanding what accounts for the discrepancy between lifetime and current self-reports of ADHD diagnosis is also needed.

This study has some limitations that should be addressed in future research. The main limitation of the NWS is that the sample is less healthy (at least in terms of self-rated health) than the randomly collected samples included in gold-standard, nationally representative surveys. While we adjusted the NWS weight to standardize the self-reported health distribution in the 2023 NWS to that of the 2023 NHIS distribution, we cannot say with certainty to what extent this alone biased our estimates. Future data collection based on gold-standard, population-representative surveys, such as the planned inclusion of ADHD measures in the 2025 NHIS (National Academy of Science, Engineering, and Medicine (2024), would address this concern. The fact that we have only a self-reported measure of ever being diagnosed with ADHD by a health care professional is also a limitation since such measures are not equivalent to clinically assessed ADHD (Chamberlain, Cortese, & Grant, 2021). Ideally, future research should include more-detailed information about the timing of diagnosis, current symptoms, and other aspects of the experience of ADHD. Studies that do not rely solely on participant’s self-report of ADHD diagnosis would also be useful for validation purposes.

Despite these limitations, this study provides contemporary estimates of the prevalence of self-reported lifetime ADHD diagnosis status in the non-institutionalized U.S. working-age adult population, which can serve as a starting point for additional research. For the purposes of health care-related policy and planning, there is a need for updated information on the prevalence of self-reported as well as clinically assessed ADHD among U.S. adults because both types of estimates are meaningful. Documentation of both self-reported and clinically assessed ADHD diagnosis status and current symptomology in the U.S. adult population is important for policy and planning related to the institutions in which working-age adults with ADHD spend time, such as universities/colleges, workplaces, the military, jails/prisons, and community-based voluntary organizations. Documenting an increase in the prevalence of self-reported ADHD diagnosis among adults has the potential to stimulate additional population-representative data collection, which can be used to validate the estimates we provide and enhance our understanding of the consequences of ADHD among adults across a broad range of contexts. Given the plausibly high prevalence of self-reported ADHD diagnosis among U.S. working-age adults that we document in this study and the significant social, economic, and health consequences of ADHD during working adulthood (Catalá-López et al., 2022; Das et al., 2012; Kosheleff et al., 2023), we recommend that national health surveys routinely collect data on self-reported lifetime and current ADHD diagnosis status among adults. The rationale for not collecting such data—that the prevalence of self-reported ADHD diagnosis status among adults is too low and cannot be reliably estimated in national samples—no longer seems valid.