Incidence rates of twelve chronic diseases/conditions in US adults: findings from a population-based study

Summary

Background

There are no contemporary studies simultaneously examining the incidence of multiple chronic diseases/conditions in US adults. We estimated one-year incidence rates for 12 chronic diseases/conditions with high public burden.

Methods

Data were from the nationally representative National Health Interview Survey 2019–2020 Longitudinal Cohort (N = 10,415). We assessed incidence rates for anxiety, arthritis, asthma, cancer, chronic obstructive pulmonary disease (COPD), chronic pain, coronary heart disease (CHD), depression, diabetes, high cholesterol, hypertension, and obesity. We calculated overall and sex-specific age-standardized incidence rates per 1000 person-years (PY) and examined associations with baseline age, sex, race/ethnicity, education, insurance, and smoking status.

Findings

In the sample, 51.7% ([95% CI: 50.3–53.1]; N = 5624, representing 129.7 million adults) were female, 20.5% ([95% CI: 20–22.2]; N = 3405, representing 21.1 million adults) were aged 65+ years, 63.2% ([95% CI: 60.9–65.4]; N = 7495, representing 158.5 million adults) were Non-Hispanic White, 16.5% ([95% CI: 14.7–18.4]; N = 1153, representing 41.5 million adults) were Hispanic/Latino, 28.8% ([95% CI: 27.4–30.2]; N = 4228, representing 72.3 million adults) graduated college, and 72.4 ([95% CI: 70.8–73.9]; N = 7287, representing 167.0 million adults) had private health insurance. One-year incidence rates were lowest for diabetes, COPD and CHD: 13.8/1000 PY (95% CI: 10.8–16.8), 14.4/1000 PY (95% CI: 11.8–17.0), and 14.7/1000 PY (95% CI: 12.7–17.0), respectively. The highest rates were observed for high cholesterol and chronic pain, 85.7/1000 PY (95% CI: 79.4–91.9) and 85.3/1000 PY (95% CI: 78.7–92.0), respectively, while all other rates were between these extremes. Females had higher rates of anxiety (69.6 [95% CI: 60.5–75.7] vs 36.4 [95% CI: 29.8–43.1]) and depression (63.7 [95% CI: 55.1–72.3] vs 34.5 [95% CI: 28.4–40.7]), while males had higher rates of hypertension (77.3 [95% CI: 72.3–82.3] vs 56.5 [95% CI: 48.7–64.3]). Incidence-related risk factors differed across diseases/conditions, with age the most consistent predictor.

Interpretation

This comprehensive assessment documented striking variation in chronic disease/condition incidence. The findings provide essential evidence for prioritizing and coordinating prevention initiatives across the chronic disease/condition spectrum.

Funding

US National Institutes of Health.

Research in context.

Evidence before this study

There appeared to be no comprehensive, contemporary studies simultaneously examining the incidence of multiple chronic conditions in the United States. To confirm this supposition, we searched PubMed, EMBASE and Google Scholar for reports published in English between January 1, 1975, through January 30, 2025. Searches terms included “incidence” or “incidence rate” and “United States”, with the addition of either “anxiety” or “arthritis” or “adult-onset asthma” or “cancer” or “chronic obstructive pulmonary disease” or “chronic pain” or “coronary heart disease” or “depression” or “diabetes” or “high cholesterol”. We also reviewed the reference lists of all identified articles for additional studies.

We found that while numerous national studies of US adults have described the prevalence (i.e., current self-reported or diagnosed diseases) of chronic disease, much less is known about their incidence (onset). Of these US incidence studies, most are limited to specific subpopulations based on age, sex, race, and ethnicity, geographic area, or SES status (e.g., those with higher education and/or income). Others are based on commercial insurance claims data that are missing demographic and health behavior data, non-longitudinal data, or surveillance data, where the true denominator (i.e., true incidence) is unknown and census data substituted. We found that to date, the best source of nationally representative U.S. longitudinal data on chronic disease is the National Health and Nutrition Examination Survey Epidemiologic Follow-up Study (NHEFS) allowing incidence studies of rheumatoid arthritis, adult-on asthma, cancer, coronary heart disease, emphysema, depression, diabetes, hypertension, and obesity. However, NHEFS data are severely outdated, with collection occurring between 1971 and 1992 and associated articles rarely presenting common measures of incidence such as cumulative incidence or incidence rates per person-years (PY) of follow-up, leaving the magnitude of related public health burden unexplored.

Added value of this study

Our study provides the first comprehensive analysis of incidence rates across 12 leading chronic diseases in a nationally representative longitudinal sample of U.S. adults, addressing a critical gap in public health research that has persisted for decades. We used a consistent group of factors to identify risk of disease by sex, age, race, ethnicity, education attainment, type of health insurance, and cigarette smoking status. We present both standardized cumulative incidence and incidence rates stratified by sex to establish a basis for international comparison, assessing prevention strategies, and informing public health initiative and policies. We found significant heterogeneity in incidence rates and risk factors among the 12 studied diseases, with diseases associated with lower mortality rates (arthritis, CP, high cholesterol) having higher incidence rates than diseases with higher mortality rates (e.g., cancer, CHD, COPD, diabetes). We expect that many other high-income countries will have similar distributions of burden.

Implications of all the available evidence

Our findings establish incidence rate estimates for leading chronic conditions as an essential complement to prevalence data for comprehensive disease burden assessment, enabling more effective health resource allocation and targeted prevention policies across the spectrum of chronic conditions. We found that age was the only consistent predictor across diseases. As the populations of high-income and some upper-middle income countries continue to age, we can expect the incidence rates for these chronic diseases to increase substantially. Progressively comprehensive prevention strategies addressing multiple risk factors simultaneously may be needed to offset aging's impact.

Introduction

Epidemiological research on chronic diseases associated with high morbidity and mortality in the United States (US) has increased substantially since early studies examining the impact of smoking status on cancer, heart disease, and hypertension.¹ Numerous national studies of US adults have described the prevalence (i.e., current self-reported or diagnosed diseases) of these chronic diseases;¹ less is known about incidence (onset) of chronic diseases/conditions at a national level.

The existing incidence studies are limited to selected geographic areas,2, 3, 4, 5, 6, 7 subpopulations,^2,3,8, 9, 10, 11, 12 age-ranges,2, 3, 4^,9,12, 13, 14 or socioeconomic status (SES) (e.g., those with lower education and/or income).^2,10,12,15 Others are based on insurance claims,¹⁶ non-longitudinal,^6,17 or surveillance data, where the true denominator (i.e., true incidence) is unknown.¹⁸ To date, the unsurpassed source of nationally representative longitudinal data on chronic diseases/conditions in the US is the National Health and Nutrition Examination Survey Epidemiologic Follow-up Study (NHEFS), allowing incidence studies of: rheumatoid arthritis,¹⁹ adult-onset asthma,¹¹ cancer,^20,21 coronary heart disease (CHD),22, 23, 24 emphysema,²⁵ depression,^26,27 diabetes,28, 29, 30, 31, 32, 33 hypertension,^8,34, 35, 36 and obesity.³⁷ However, NHEFS data are severely outdated, with collection occurring between 1971 and 1992. Associated articles have rarely expressed measures of incidence such as cumulative incidence or incidence rates per person-years (PY) of follow-up, leaving the magnitude of related public health burden unexplored.

Given that there are no contemporary studies simultaneously examining the incidence of multiple chronic diseases/conditions in US adults, the aim of the present study is to estimate one-year incidence rates for 12 chronic diseases/conditions with high public burden. Specifically, this study examines 2020 incidence rates for anxiety, arthritis, adult-onset asthma, cancer (any type), COPD, chronic pain, CHD, depression, diabetes, elevated cholesterol, hypertension, and obesity using data from the National Health Interview Survey (NHIS) 2019–2020 Longitudinal Cohort. We also present sex-stratified rates and examine possible risk factors. The nationally representative NHIS-Longitudinal Cohort allows us to capture incidence at the population level. Applying a common methodology to the analyses of the 12 chronic diseases/conditions under study allows us to assess risks in a uniform manner and for optimal cross-disease comparisons of population burden; this goal is consistent with numerous objectives from the U.S Department of Health and Human Services (DHHS) Healthy People 2030 (HP2030). HP2030 is a federal initiative that sets forth national objectives for improving the health and well-being of the U.S. population over the current decade, including those for chronic diseases/conditions associated with high burden.³⁸

Methods

Study population

The NHIS is a nationally representative, annual survey of the US civilian, non-institutionalized population from 50 states and the District of Columbia.³⁹ The National Center for Health Statistics (NCHS) staff, in conjunction with the U.S. Census Bureau, use geographically clustered sampling techniques to select a random sample of dwelling units. Multistage methods are used to partition this sample into nested levels of strata (based on population density) and clusters (based on addresses within a defined geographic region). A sample adult is randomly chosen from each dwelling unit for the NHIS. Institutionalized individuals (homeless and/or transient persons not in shelters, active-duty military personnel and civilians on military bases, persons in long-term care institutions or persons in correctional facilities) are not included. Application of NCHS-derived sampling weights according to poststratification by age/sex/race/ethnicity allows an accurate extrapolation of findings to the US civilian, non-institutionalized population.³⁹ All data used are publicly and freely available through the NCHS website: https://www.cdc.gov/nchs/nhis/documentation/2020-nhis.html. Each quarter of the sampling year (January–March, April–June, July–September, October–December) is independently nationally representative and contains approximately the same number of participants. To identify the subsample of NHIS participants also recruited into the subsequent Medical Expenditure Panel Survey (MEPS), each quarter of the NHIS is divided into 4 panels, with 2 panels randomly selected for the MEPS participant selection from each NHIS quarter.^39,40

For the 2019 NHIS, 32,997 randomly selected adults agreed to participate with a final 2019 Sample Adult Response Rate of 59.1%. For the 2020 NHIS, recruitment was at the normal rate until late March 2020, at which time NCHS reported data collection difficulties because of the COVID-19 pandemic. In response, data collection for the 2020 NHIS transitioned from mostly in-person interviews to mostly telephone interviews, with only telephone surveys from April to June. Because of possible reductions in geographic coverage and lower response rates often seen with telephone interviewing, NCHS decided to telephone reinterview adults who completed the 2019 NHIS Sample Adult during the 3rd and 4th quarters of 2020.⁴¹ This NCHS decision was based on previous compliance with the 2019 survey process, and the presence of detailed telephone information.

Prior to selection of the 2019 NHIS participants for 2020 re-interview, NCHS excluded 2019 NHIS adult participants (10,836) already chosen for 2020 MEPS participation (Fig. 1) because of potential participant overburden. Of the remaining candidate pool (21,161), 334 were excluded due to being deceased or residing in an institutional setting in 2020 (e.g., nursing home, prison, active military). NCHS considered the remaining 20,827 as eligible for 2020 follow-up. Of these, 1746 either had a proxy respondent in 2019 or were missing current contact information, while 8666 declined further participation, leaving 10,415 to be re-interviewed for the 2020 NHIS, resulting in a 50% completion rate (10,415/20,827).^39,41

Fig. 1.

Study sample flow chart.

Aside from the opportunity to increase the 2020 NHIS sample as result of Covid-related recruitment concerns, NCHS envisioned the reinterviewing of 2019 Sample Adults as an opportunity to assess longitudinal changes in health outcomes measured on both survey years and assess measures during the pandemic in a representative sample of the population with nearly complete telephone contact information and known demographic distributions.^39,41 These 2019 NHIS participants re-interviewed in 2020 became the NHIS-Longitudinal Cohort.

Follow-up time of NHIS-Longitudinal Cohort participants

As shown in Supplementary Figure S1, the length of follow-up is based on the midpoint of the quarter in which a participant was interviewed in 2019 and the midpoint of quarter in which they were interviewed in 2020; the length of follow-up thus varies from 12 to 18 months for 12681.25 total years of follow-up, which is equivalent to 313 million years of follow-up when weighted to the US population. Follow-up time is based on 3-month quarters versus monthly to account for the quarterly panel approach used for MEPS recruitment. Because the NCHS desired at least one full year of follow-up, no data was collected in the first two quarters of 2020.

All aspects of data collection for both years were approved by the NCHS Research Ethics Review Board. Verbal consent was obtained from all respondents by survey staff. Because the public use files of NHIS are completely de-identified, on August 19, 2022, the US National Institutes of Health Institutional Review Board Committee Panel 1 authorized an exemption from review. This report follows the STROBE guideline for cohort studies (Supplementary Table S1).⁴²

Operational definitions of chronic diseases/conditions

For the NHIS years 2019–2020, Sample Adults’ health status was determined with the following questions (Supplementary Table S2): 1) “Have you EVER been told by a doctor or other health professional that you had anxiety disorder, arthritis, asthma, cancer (any), COPD [including emphysema, and/or chronic bronchitis], coronary heart disease (CHD), high cholesterol, depression, or diabetes [not gestational diabetes or prediabetes]”; 2) “Were you told on two or more DIFFERENT visits that you had hypertension or high blood pressure”; 3) “In the past 3 months, how often did you have pain? Would you say never, some days, most days, or every day?” Following convention,⁴¹ chronic pain was defined as pain on most days or every day in the past 3 months; and 4) Based on self-reported weight and height, body mass index (BMI) values were calculated by NCHS staff and assigned as: underweight (BMI <18.5), normal weight (BMI from 18.5 to <25), overweight (BMI from 25 to <30), and obese (BMI ≥ 30). Identical or similar self-reported survey items have been used in US incidence studies of asthma,¹² cancer,43, 44, 45 chronic pain,⁴⁶ CHD,43, 44, 45 COPD,¹³ depression,⁴⁷ diabetes,^{5,6,31,32,48,49} hypertension,¹⁴ and stroke,⁵⁰ as well as international studies.51, 52, 53, 54, 55, 56, 57

Incidence estimates were based on participants free of the specific disease/condition in 2019 who then reported the disease/condition in 2020. No attempt was made to impute data for the small percentage (n = 141; 1.4%) without valid information for one or more chronic diseases/conditions in 2020. Missing data are presented as “Unknown”.

Statistical analysis

We used absolute standard difference (ASD) values to compare the baseline characteristics for the complete analytic sample (i.e., all 10,415 NHIS-Longitudinal Cohort participants) to 21,582 non-Longitudinal Cohort participants in the 2019 NHIS (i.e., those not randomized for invitation, randomized but found ineligible, or randomized but declined) (Table 1). Following convention, an ASD <0.1 indicates that the NHIS-Longitudinal Cohort participants did not differ substantively from the 2019 respondents who did not participate in the NHIS-Longitudinal Cohort.

Table 1.

Baseline characteristics of 2019 NHIS adult participants by enrollment status in the 2019–2020 longitudinal cohort.

Variable	2019 enrolled longitudinal cohort (longitudinal sample weights)			2019 not enrolled (sample adult weights)			Absolute standardized difference scoresd
	Raw frequency	Weighted frequency (1000's)	Weighted prevalence (95% CI)	Raw frequency	Weighted frequency (1000's)	Weighted % (95% CI)
Total	10,415	250,917	100	21,582	175,427	100
Age group (yrs)
18–34	1935	74,366	29.6 (28.2–31.1)	5123	55,311	31.50 (30.6–32.4)	0.041
35–44	1506	40,900	16.3 (15.3–17.3)	3460	28,887	15.5 (15.9–17.1)	0.005
45–54	1513	40,476	16.1 (15.2–17.1)	3273	28,352	16.1 (15.5–16.8)	0
55–64	2056	42,179	16.8 (15.8–17.8)	3755	28,014	16.0 (15.4–16.5)	0.022
65+	3405	52,996	21.1 (20.1–22.2)	5971	34,863	19.9 (19.2–20.6)	0.03
Sex
Male	4790	121,160	48.3 (46.9–49.8)	9943	84,644	48.3 (47.4–49.1)	0
Female	5624	129,741	51.7 (50.3–53.1)	11,637	90,770	51.7 (50.9–52.6)	0
Unknown	1	15	UR	2	13	UR	UR
Race/Hispanic Heritage
Non-Hispanic White	7495	158,534	63.2 (60.9–65.4)	14,420	107,916	61.5 (59.9–63.2)	0.035
Non-Hispanic Black or African American	999	29,678	11.8 (10.5–13.2)	2484	21,794	12.4 (11.4–13.4)	0.018
Non-Hispanic Asian	512	14,310	5.7 (4.9–6.5)	1136	10,483	6.0 (5.4–6.6)	0.013
Non-Hispanic American Indian or Alaska Native	157	4043	1.6 (0.8–2.4)	303	2373	1.4 (1.0–1.7)	0.016
Othera	99	2846	1.1 (0.8–1.4)	240	2139	1.2 (1.0–1.4)	0.009
Hispanic/Latino	1153	41,506	16.5 (14.7–18.4)	2999	30,722	17.5 (16.1–18.9)	0.027
Education Attainment
College graduate	4228	72,259	28.8 (27.4–30.2)	7049	46,347	26.4 (25.4–27.4)	0.054
Some college	3038	77,594	30.9 (29.6–32.2)	6384	53,811	30.7 (29.8–31.5)	0.004
High school graduate or equivalent	2016	63,966	25.5 (24.1–26.9)	5319	46,064	26.3 (25.4–27.1)	0.018
Did not graduate high school	1004	3531	14.1 (12.9–15.2)	2726	27,797	15.8 (15.0–16.7)	0.048
Unknown	39	1787	0.7 (0.4–1.0)	140	1409	0.8 (0.6–1.0)	0.012
Health insurance
Privateb	7287	169,991	72.4 (70.8–73.9)	14,312	115,122	69.2 (68.2–70.3)	0.070
Publicc	1349	35,829	15.3 (14.0–16.5)	3412	27,928	16.8 (15.9–17.7)	0.041
Uninsured	726	2860	12.2 (10.0–13.3)	2161	22,373	13.5 (12.7–14.2)	0.036
Unknown	18	481	0.2 (0.1–0.3)	90	820	0.5 (0.4–0.6)	0.051
Saw a physician in the last year
No	1212	37,845	15.1 (14.0–16.2)	3035	27,741	15.8 (15.1–16.5)	0.019
Yes	9195	212,809	84.8 (83.7–85.9)	18,356	145,900	83.2 (82.5–83.9)	0.044
Unknown	8	262	UR	191	7787	1.0 (0.8–1.2)	UR
Smoking status
Current	1259	32,715	13.0 (12.1–14.0)	3044	25,084	14.3 (13.6–15.0)	0.038
Former	2781	56,889	22.6 (21.6–23.7)	5192	36,619	20.9 (20.2–21.6)	0.044
Never	6363	161,066	64.2 (62.8–65.6)	12,568	106,571	60.7 (59.8–61.7)	0.072
Unknown	12	248	UR	778	7154	4.1 (3.7–4.5)	UR
Anxiety
No	8898	215,468	85.9 (84.9–86.9)	18,292	150,198	85.6 (84.9–86.3)	0.009
Yes	1508	35,292	14.0 (13.1–15.0)	3237	24,835	14.2 (13.5–14.8)	0.003
Unknown	9	156	UR	53	394	0.2 (0.1–0.3)	UR
Arthritis
No	7493	196,781	78.9 (77.8–80.0)	16,234	139,043	79.3 (78.6–78.0)	0.022
Yes	2195	53,950	21.5 (19.8–22.1)	5299	35,993	20.5 (19.8–21.2)	0.025
Unknown	7	77	UR	49	390	0.2 (0.1–0.3)	UR
Asthma
No	8970	215,761	86.0 (85.0–86.9)	18,748	152,040	86.7 (86.1–87.2)	0.020
Yes	1440	35,065	14.0 (13.0–14.9)	2789	23,060	13.1 (12.6–13.7)	0.026
Unknown	5	90	UR	45	327	0.2 (0.1–0.3)	UR
Cancer
No	9018	226,505	90.5 (89.8–91.3)	19,090	159,536	90.9 (90.5–91.4)	0.021
Yes	1392	24,316	9.7 (9.0–10.4)	2452	15,554	8.9 (8.4–9.3)	0.028
Unknown	5	96	UR	40	337	0.2 (0.1–0.3)	UR
CHD
No	9799	240,382	95.8 (95.4–96.2)	20,249	166,702	95.0 (94.7–95.4)	0.038
Yes	596	10,233	4.1 (3.7–4.5)	1250	8095	4.6 (4.3–4.9)	0.025
Unknown	20	302	0.1 (0.0–0.2)	83	631	0.4 (0.3–0.5)	0.060
High cholesterol
No	7064	185,833	74.1 (73.0–75.1)	15,633	133,731	76.2 (75.5–77.0)	0.049
Yes	3335	64,726	25.8 (24.7–26.9)	5844	10,891	23.3 (22.6–24.0)	0.058
Unknown	16	358	0.1 (0.06–0.2)	105	805	0.5 (0.3–0.6)	0.073
Chronic paine
No	7961	198,650	79.2 (78.0–80.3)	16,159	135,094	77.0 (76.3–77.7)	0.053
Yes	2446	52,079	20.8 (19.6–21.9)	4738	34,103	19.4 (18.7–20.1)	0.035
Unknown	8	187	UR	685	6320	3.6 (3.2–4.0)	UR
COPD
No	9807	240,048	95.7 (95.2–96.1)	20,351	167,038	95.2 (94.9–95.6)	0.024
Yes	602	10,755	4.3 (3.8–4.8)	1185	8060	4.6 (4.2–5.0)	0.015
Unknown	6	11	UR	46	329	0.2 (0.1–0.3)	UR
Depression
No	8571	209,793	83.6 (82.6–84.6)	17,937	147,993	84.4 (83.7–85.0)	0.022
Yes	1835	40,981	16.3 (15.3–17.3)	3597	27,068	15.5 (14.8–16.1)	0.025
Unknown	9	143	UR	54	366	0.2 (0.1–0.3)	UR
Diabetes
No	9311	227,733	90.8 (90.0–91.5)	19,283	159,091	90.7 (90.2–91.2)	0.003
Yes	1098	23,081	9.2 (8.5–9.9)	2257	16,009	9.1 (8.7–9.6)	0.003
Unknown	6	101	UR	42	328	0.2 (0.1–0.3)	UR
Hypertension
No	6977	181,788	72.4 (71.2–73.7)	15,047	129,598	73.9 (73.0–74.7)	0.034
Yes	3427	68,861	27.4 (26.2–28.7)	6515	45,697	26.0 (25.3–26.8)	0.032
Unknown	11	268	UR	20	132	0.1 (0.0–0.1)	UR
Obesityf
No	6829	163,148	65.0 (63.7–66.4)	14,304	116,296	66.3 (65.5–67.1)	0.027
Yes	3392	82,405	32.8 (31.5–34.2)	6.605	53,533	30.5 (29.7–31.3)	0.04954
Unknown	194	5364	2.13 (1.7–2.6)	673	5597	3.2 (2.9–3.5)	0.069
Any unknowns
No	9859	231,260	92.2 (91.1–93.2)	19,360	153,225	87.3 (86.4–94.0)	0.162
Yes	556	19,656	7.8 (6.8–8.9)	2222	22,202	12.7 (11.9–13.6)	0.162

CI: confidence interval, CHD: coronary heart disease, COPD: Chronic Obstructive Pulmonary Disease.

UR: Unreliable; Does not meet National Center for Health Statistics standards for reliability in National Surveys.⁵⁸

^aOther race includes Native Hawaiian, Pacific Islanders, those who reported more than one race, or ‘some other race'.

^bPublic Insurance (primarily Medicare or Medicaid).

^cPrivate Insurance includes those who had both Private and Public Insurance.

^dThe absolute standard difference (ASD) is based on the Cohen's Effect Size Index (Cohen J. Statistical Power Analysis for the Behavioral Sciences (2nd edn). Lawrence Erlbaum Associates Publishers: Hillsdale, NJ, 1988) and is calculated as: p1−p2/sqrt [({p1 (1−p1)} + {p2 (1−p2)})/2] were p1 = proportion 1 and p2 = proportion 2. Cohen considered ASD <0.2 to indicate small, not meaningful differences between two proportions. In the current case, an ASD <0.2 indicates that the longitudinal subsample did not differ substantively from the 2019 NHIS respondents who were not selected for the longitudinal sample.

^ePain most or every day in the last 3 months.

^fBody mass index ≥ 30.

We calculated the proportions of participants reporting each specific chronic disease in 2020 to those without it at baseline in 2019. Age-standardization was used to reduce the effects of age for subpopulation comparisons and was calculated using a direct standardization approach. The direct method averages subpopulation rates using age-specific weights based on the age distribution of the 2010 US population.^59,60 Age-standardized rates were calculated using an EXCEL spreadsheet, as well as rates by sex, for each chronic disease with 95% confidence intervals (CI) per 1000 person-years (PY).

Frequency analyses were generated using SAS Survey Procedures (version 9.4, SAS Institute Inc., Cary, NC). All proportion estimates in the text and tables were weighted to the US population using NCHS-supplied longitudinal weights calculated using complex regression modeling (Supplementary Text S1) to produce estimates representative of the US civilian, non-institutionalized population 18 years of age and older. All rates per 1000 PY (95% CI), as well as 1-year cumulative incidence (%; 95% CI) are presented in the text and tables. The alpha was set at 0.05 and no adjustments were made for multiple comparisons. Z-tests, obtained by dividing the difference between rates by the standard error of the difference, were used to compare the rates of chronic diseases/conditions between males and females. When comparing rates across diseases/health conditions (e.g., rate of hypertension vs rate of obesity), non-overlapping 95% CI were considered statistically significant.⁵⁸

Relative Risks (RR) were estimated via multivariable Poisson regression models with robust standard errors using SAS Proc GENMOD (version 9.4 SAS Institute Inc. Cary, NC). We used the Repeated Subject option in Proc GENMOD to incorporate NHIS-Longitudinal Cohort survey sampling design characteristics, PPSU (clustering) and PSTRAT (stratification). In our case, we indicated “Repeated Subject = PPSU(PSTRAT)”, which specifies that the observations for any NHIS participants are uniquely identified by both PPSU (clustering) and PSTRAT (stratification). The fully controlled multivariable regression models included age, sex, race, Hispanic/Latino ethnicity, educational attainment, health insurance status, physician visit, and smoking status. Sex was self-reported based on the question: “Are you male or female?” In the NHIS, self-reported race was coded as American Indian/Alaska Native, Asian, Black/African American, White, and Other (including Native Hawaiian, Pacific Islander, more than one race, or ‘some other race’). Hispanic/Latino ethnicity was reported dichotomously (Yes or No). For sample size reasons, we combined self-reported race and Hispanic/Latino ethnicity into 6 discrete groups: non-Hispanic White, non-Hispanic American Indian/Alaska Native, non-Hispanic Asian, non-Hispanic Black, non-Hispanic Other and Hispanic. We used non-Hispanic White as the reference group for comparisons, as they generally experience more privilege in American society. Age was coded as 18–44, 45–64 and 65+ years. Educational attainment was categorized into 4 groups (did not graduate high school; graduated high school or equivalent; attended some college; 4-year college graduate). Health insurance status was coded into 3 groups: None, Public (primarily Medicare or Medicaid), and Private, which includes those who had both Private and Public Insurance. Cigarette smoking status was coded as Never, Former, and Current. To control for healthcare utilization, we captured whether participants saw a physician within the prior 12 months.

Role of funding source

This research was supported by a grant by a grant from the National Institute of Aging, R01AG065351 and through the Intramural Research Program of the National Institutes of Health (NIH). The contributions of the NIH author were made as part of their official duties as an NIH federal employee are in compliance with agency policy requirements, and are considered Works of the United States Government. However, the findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services (HHS).

Results

At baseline (Table 1), 51.7% ([95% CI: 50.3–53.1]; N = 5624, representing 129.7 million adults) were female, 20.5% ([95% CI: 20–22.2]; N = 3405, representing 21.1 million adults) were aged 65+ years, 63.2% ([95% CI: 60.9–65.4]; N = 7495, representing 158.5 million adults) were Non-Hispanic White, 16.5% ([95% CI: 14.7–18.4]; N = 1153, representing 41.5 million adults) were Hispanic/Latino, 28.8% ([95% CI: 27.4–30.2]; N = 4228, representing 72.3 million adults) graduated college, 72.4 ([95% CI: 70.8–73.9]; N = 7287, representing 167.0 million adults) had private health insurance and 13.0% ([95% CI:12.1–14.0]; N = 1259, representing 32.7 million adults) were current smokers. Most participants (84.8% [95% CI: 83.7–85.9]; N = 9195 representing 212.8 million adults) had seen a physician in the last year. The four most prevalent chronic diseases/conditions at baseline were hypertension (27.4% [95% CI: 26.2–28.7]; N = 3427 representing 68.9 million adults), high cholesterol (25.8% [95% CI: 24.7–26.9]; N = 3335 representing 64.7 million adults), arthritis (21.5% [95% CI:19.8–22.1] N = 2195 representing 54.0 million adults), and chronic pain (20.8% [95% CI: 19.6–21.9]; N = 2446 representing 52.1 million adults), while those with the lowest prevalence were cancer (9.7% [95% CI 9.0–10.4]; N = 1392 representing 24.3 million adults), diabetes (9.2% [95% CI: 8.5–9.9] N = 1098 representing 23.1 million adults), COPD (4.3% [95% CI: 3.8–4.8]; N = 602 representing 10.8 million adults) and CHD (4.1% [95% CI: 3.7–4.5]; N = 596 representing 10.2 million adults).

There were no substantive differences (i.e., ASD <0.1) in demographic and chronic disease/condition distributions of 2019 participants by NHIS-Longitudinal Cohort participation status (Table 1), with the exception that individuals not participating in the NHIS-Longitudinal Cohort were more likely to have item nonresponse (“unknown”) for at least one variable (12.7%; 95% CI: 11.9–13.6) compared with those participating (7.8%; 95% CI: 6.8–8.9) (ASD = 0.162).

Overall incidence rates

In Table 2 and Fig. 2, incidence is shown for all diseases/conditions. Comparatively lower rates were seen for diseases/conditions often associated with high mortality—cancer, CHD, COPD, and diabetes—with rates of 21.6/1000 PY (95% CI: 18.5–24.7), 14.7/1000 PY (95% CI: 12.4–17.0), 14.4/1000 PY (95% CI: 11.8–17.0), and 13.8/1000 PY (95% CI: 10.8–16.8), respectively. Diseases/conditions not typically associated with increased mortality—arthritis, chronic pain, high cholesterol—had the highest rates: 68.2/1000 PY (95% CI: 61.7–74.7), 85.3/1000 PY (95% CI: 78.7–92.0), and 85.7/1000 PY (95% CI: 79.4–91.9). Depression and anxiety were midrange (48.6/1000 PY; 95% CI: 44.3–53.9 and 52.3/1000 PY; 95% CI: 46.5–58.1, respectively).

Table 2.

Age-standardized 1-Year Cumulative Incidence and incidence rate per 1000 person years of follow-up for 12 chronic diseases/conditions.

	Denominator data			Numerator data		Outcomes
	Number of adults without event in 2019	Weighted number of adults (1000's)a without event in 2019	Person years of follow-up (1000's)	Frequency of event in 2020	Weighted frequency (1000's)a of event in 2020	1-Year cumulative incidence (%; 95% CIb)	Incidence rate/1000 PYc (95% CIb)
Anxiety
All adults	8898	215,468	268,772	564	13,857	6.5 (5.8–7.3)	52.3 (46.5–58.1)
Male	4305	109,457	136,634	195	4964	4.6 (3.7–5.4)	36.4 (29.8–43.1)
Female	4593	106,010	132,137	369	8892	8.7 (7.5–9.9)	69.6 (60.5–78.7)
Arthritis
All adults	7493	196,781	245,367	708	15,143	8.5 (7.7–9.3)	68.2 (61.7–74.7)
Male	3687	98,957	123,410	290	6858	7.6 (6.5–8.7)	61.2 (52.2–70.3)
Female	3805	97,808	121,942	418	8285	9.4 (8.3–10.5)	75.3 (69.6–80.9)
Unknown	1	15	15	0	0
Asthma
All adults	8970	215,761	269,337	257	6949	3.3 (2.7–3.8)	26.0 (19.2–32.9)
Male	4208	104,931	130,966	107	3164	3.1 (2.3–3.9)	25.0 (18.9–31.1)
Female	4761	110,815	138,356	150	3785	3.4 (2.6–4.1)	27.2 (19.1–35.3)
Unknown	1	15	15	0	0	0	0
Cancer
All adults	9018	226,505	287,713	297	6028	2.7 (2.3–3.1)	21.6 (18.5–24.7)
Male	4225	110,694	138,190	128	2414	2.3 (1.9–2.8)	18.7 (15.2–22.3)
Female	4792	115,795	144,508	169	3614	3.1 (2.5–3.7)	24.6 (19.6–29.7)
Unknown	1	15	15	0	0	0	0
CHDd
All adults	9779	240,382	299,880	249	4591	1.8 (1.5–2.2)	14.7 (12.4–17.0)
Male	4421	114,478	142,863	125	2.363	2.1 (1.7–2.5)	16.9 (13.5–20.3)
Female	5377	125,888	157,002	124	2228	1.2 (0.9–1.6)	13.1 (9.8–16.3)
Unknown	1	15	15	0	0	0	0
High cholesterol
All adults	7064	185,833	231,788	774	17,885	10.7 (9.8–11.6)	85.7 (79.4–91.9)
Male	3216	89,084	111,194	367	9374	11.8 (10.3–13.2)	91.9 (82.4–101.5)
Female	3847	96,734	120,579	407	8511	9.7 (8.5–10.8)	77.6 (69.1–86.1)
Unknown	1	15	15	0	0	0	0
Chronic paine
All adults	7961	198,650	248,025	938	20,902	10.7 (9.8–11.5)	85.3 (78.7–92.0)
Male	3747	97,185	121,391	407	9556	10.0 (8.8–11.3)	80.5 (70.5–90.5)
Female	4214	101,464	126,634	531	11,346	11.2 (9.9–12.5)	89.9 (79.9–99.9)
COPDf
All adults	9807	240,048	299,351	204	4390	1.8 (1.5–2.1)	14.4 (11.8–17.0)
Male	4555	116,896	145,955	90	2007	1.7 (1.3–2.1)	13.8 (10.4–17.1)
Female	5251	123,136	1153,381	114	2383	1.9 (1.4–2.4)	15.1 (11.2–19.2)
Unknown	1	15	15	0	0	0	0
Depression
All adults	8571	209,793	261,758	543	12,693	6.1 (5.4–6.7)	48.6 (43.3–53.9)
Male	4195	107,504	134,188	193	4604	4.3 (3.5–5.1)	34.5 (28.4–40.7)
Female	4375	102,273	217,555	349	8073	7.9 (6.8–9.1)	63.7 (55.1–72.3)
Unknown	1	15	15	1	15	UR	UR
Diabetes
All adults	9311	227,733	284,148	166	3892	1.7 (1.4–2.1)	13.8 (10.8–16.8)
Male	4273	109,647	136,921	75	1598	1.5 (1.1–2.0)	12.2 (9.1–15.4)
Female	5037	118,071	147,212	91	2294	1.9 (1.3–2.5)	15.2 (10.2–20.1)
Unknown	1	15	15	0	0	0	0
Hypertension
All adults	6977	171,788	226,897	606	13,157	8.3 (7.5–9.1)	63.2 (57.1–69.3)
Male	3183	87,392	109,161	295	7118	9.6 (8.3–10.9)	77.3 (72.3–82.3)
Female	3793	94,381	117,721	311	6039	7.0 (6.0–8.0)	56.5 (48.7–64.3)
Unknown	1	15	15	0	0	0	0
Obesityg
All adults	6829	163,148	203,477	368	9952	6.2 (5.4–7.0)	49.9 (43.7–56.1)
Male	3218	80,206	99,921	162	4541	5.8 (4.7–6.9)	46.2 (37.8–54.7)
Female	3611	82,942	103,556	206	5411	6.7 (5.6–7.9)	53.8 (43.8–63.9)

^aBecause of rounding error, the sum of males and females may not equal the total.

^bConfidence Intervals.

^cPerson years.

^dCoronary heart disease.

^ePain most or every day in the last 3 months.

^fChronic Obstructive Pulmonary Disease.

^gBody mass index ≥ 30.

Fig. 2.

Age-standardized incidence rates per 1000 person years of follow-up for 12 chronic diseases. Legend: Age-standardization was used to reduce age's effects for subpopulation comparisisons and was calculated using a direct standardization approach. The direct method averages subpopulation rates using age-specific weights based on the age distribution of the 2010 US population. Incidence rates per 1000 person years of follow-up are presented as are the 95% Confidence Intervals (CI) for these rates. CHD: coronary heart disease. COPD: Chronic Obstructive Pulmonary Disease.

Incidence rates by sex

Generally, males and females had similar incidence rates (Table 2; Fig. 3). Hypertension was one exception, where males (77.3/1000 PY, 95% CI: 72.3–82.3) had significantly higher rates (Z-test; p = 0.028) than females (56.5/1000 PY, 95% CI: 48.7–64.3). Conversely, females had significantly higher rates of anxiety (p = 0.004) and depression (p = 0.005) than males: 69.6/1000 PY (95% CI: 60.5–78.7) vs. 36.4/1000 PY (95% CI: 29.8–43.1), and 63.7/1000 PY (95% CI: 55.1–72.3) vs. 34.5/1000 PY (95% CI: 28.8–40.2), respectively.

Fig. 3.

Age-standardized incidence rates per 1000 person years of follow-up for 12 chronic diseases by sex. Legend: Age-standardization was used to reduce age's effects for subpopulation comparisons and was calculated using a direct standardization approach. The direct method averages subpopulation rates using age-specific weights based on the age distribution of the 2010 US population. Incidence rates per 1000 person years of follow-up are presented as are the 95% Confidence Intervals (CI) for these rates. CHD: coronary heart disease. COPD: Chronic Obstructive Pulmonary Disease.

The unadjusted (Supplementary Table S3) and adjusted (Table 3) regression models produced similar results when looking at risks of disease/conditions across racial and ethnic groups with two notable exceptions: 1) Non-Hispanic Black adults had a lower incidence of anxiety, CHD, COPD, and depression, and a higher incidence of chronic pain, compared with Non-Hispanic Whites in the adjusted model, but there was no significant race difference in unadjusted models; 2) While the adjusted models demonstrated that Non-Hispanic Whites were at greater risk of anxiety and depression than Hispanic/Latinos, the unadjusted model did not identify significant differences. These results suggest one or more control variables are potential confounders, mediators or modifiers in the relationships between race/ethnicity and chronic diseases.

Table 3.

Fully adjusted relative risk^a (95% confidence intervals) of incident disease/condition in 2020 by participant characteristics.

	Anxiety	Arthritis	Asthma	Cancer	CHDb	High Cholesterol
Age
18–34	Ref	Ref	Ref	Ref	Ref	Ref
35–44	0.94; 0.72–1.23	2.12; 1.38–3.25	0.69; 0.45–1.07	5.03; 2.04–12.38	1.55; 0.42–5.78	1.44; 1.10–1.89
45–54	0.76; 0.75–1.01	4.07; 2.76–6.00	0.79; 0.52–1.19	5.0; 2.03–12.30	4.72; 1.58–14.15	2.01; 1.55–2.62
55–64	0.72; 0.55–0.93	6.63; 4.61–9.54	0.99; 0.69–1.43	15.61; 6.81–35.78	16.2; 5.91–44.53	3.06; 2.41–3.88
65+	0.47; 0.36–0.62	8.40; 5.84–12.07	0.54; 0.36–0.81	22.09; 9.62–50.69	34.81; 12.82–94.58	3.71; 2.92–4.73
Sex
Male	Ref	Ref	Ref	Ref	Ref	Ref
Female	1.83; 1.53–2.20	1.27; 1.08–1.49	1.02; 0.79–1.32	1.01; 0.79–1.28	0.57; 0.45–0.71	1.01; 0.88–1.16
Race/Hispanic-Latino
Non-Hispanic White	Ref	Ref	Ref	Ref	Ref	Ref
Non-Hispanic Black or African American	0.64; 0.47–0.86	0.63; 0.46–0.87	0.72; 0.44–1.17	0.36; 0.20–0.65	0.59; 0.36–0.96	1.21; 0.97–1.51
Non-Hispanic Asian	0.45; 0.27–0.75	0.78; 0.51–1.17	1.02; 0.57–1.81	0.16; 0.04–0.64	0.83; 0.39–1.77	1.09; 0.62–1.3
Non-Hispanic American Indian or Alaska Native	1.16; 0.65–2.07	0.84; 0.45–1.58	0.99; 0.36–2.68	0.37; 0.09–1.49	0.90; 0.37–2.18	0.78; 0.43–1.42
Hispanic/Latinoc	0.63; 0.46–0.85	0.76; 0.57–1.02	1.20; 0.83–1.76	0.26; 0.13–0.54	1.07; 0.70–1.66	1.08; 0.85–1.36
Otherd	0.44; 0.14–1.37	0.79; 0.30–2.12	1.67; 0.62–4.53	0.88; 0.22–3.55	0.58; 0.08–4.11	0.59; 0.22–1.58
Education
College graduate	Ref	Ref	Ref	Ref	Ref	Ref
Some college	1.10; 0.89–1.38	1.22; 0.99–1.50	1.28; 0.92–1.79	0.94; 070–1.28	1.23; 0.91–1.66	0.92; 0.77–1.09
High school graduate or equivalent	1.27; 1.00–1.61	1.49; 1.20–1.85	1.63; 1.15–2.31	1.28; 0.94–1.74	1.45; 1.07–1.96	1.08; 0.90–1.30
Did not graduate high school	1.60; 1.17–2.18	1.85; 1.41–2.43	2.18; 1.41–3.36	0.72; 0.43–1.19	1.08; 0.72–1.63	1.31; 1.03–1.66
Health insurance
Privatee	Ref	Ref	Ref	Ref	Ref	Ref
Publicf	1.92; 1.53–2.41	1.50; 1.21–1.85	1.54; 1.10–2.16	1.19; 0.86–1.66	2.06; 1.58–2.68	1.13; 0.93–1.37
Uninsured	0.81; 0.55–1.18.	0.84; 0.58–1.23	0.68; 0.40–1.15	1.56; 0.94–2.61	0.87; 0.40–1.90	1.02; 0.77–1.36
Dr. visit in last year
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	1.29; 0.97–1.71	1.65; 1.21–2.25	1.03; 0.70–1.52	0.61; 0.42–0.86	1.99; 1.10–3.57	1.66; 1.30–2.13
Cigarette smoking
Never	Ref	Ref	Ref	Ref	Ref	Ref
Former	1.06; 0.85–1.31	1.15; 0.96–1.39	0.93; 0.68–1.28	1.08; 0.83–1.42	1.69; 1.32–2.18	1.30; 1.11–1.53
Current	1.29; 1.0–1.67	1.20; 0.94–1.54	0.89; 0.60–1.32	1.19; 0.83–1.71	1.56; 1.09–2.22	1.45; 1.18–1.77

	Chronic paing	COPDh	Depression	Diabetes	Hypertension	Obesityi
Age
18–34	Ref	Ref	Ref	Ref	Ref	Ref
35–44	0.92; 0.78–1.08	1.42; 0.73–2.77	1.01; 0.76–1.34	0.70; 0.30–1.64	2.18; 1.9–3.20	1.22; 0.87–1.71
45–54	1.18; 1.01–1.38	2.28; 1.25–4.17	0.94; 0.70–1.26	3.12; 1.74–5.61	3.62; 2.53–5.17	1.09; 0.77–1.54
55–64	1.20; 1.03–1.0	3.29; 1.89–5.74	1.0; 0.77–1.29	3.56; 2.03–6.24	4.40; 3.12–6.22	0.89; 0.64–1.25
65+	1.43; 1.23–1.66	3.66; 2.10–6.38	0.58; 0.44–0.78	3.51; 1.98–6.20	8.07; 5.76–11.32	0.60; 0.42–0.85
Sex
Male	Ref	Ref	Ref	Ref	Ref	Ref
Female	1.12; 1.02–1.23	0.90; 0.68–1.18	1.65; 1.38–1.98	0.70; 0.52–0.95	0.69; 0.59–0.82	1.22; 0.98–1.53
Race/Hispanic-Latino
Non-Hispanic White	Ref	Ref	Ref	Ref	Ref	Ref
Non-Hispanic Black or African American	1.65; 1.44–1.90	0.59; 0.35–0.96	0.65; 0.48–0.89	2.52; 1.68–3.80	1.95; 1.49–2.55	1.62; 1.17–2.25
Non-Hispanic Asian	0.55; 0.41–0.74	0.52; 0.19–1.41	0.35; 0.19–0.64	1.25; 0.50–3.11	0.93; 0.59; 1.47	0.46; 0.24–0.88
Non-Hispanic American Indian or Alaska Native	1.16; 0.77–1.75	1.82; 0.92–3.59	1.04; 0.58–1.85	0.87; 0.21–3.54	1.22; 0.67–2.24	0.57; 0.18–1.80
Hispanic/Latinoc	1.30; 1.11–1.51	0.52; 0.29–0.94	0.63; 0.46–0.85	2.46; 1.59–3.81	1.09; 0.81–1.47	1.06; 0.70–1.59
Otherd	0.48; 0.23–0.45	0.57; 0.08–4.04	0.80; 0.33–1.96	2.97; 0.94–9.45	2.18; 1.08–4.41	2.01; 0.94–4.28
Education
College graduate	Ref	Ref	Ref	Ref	Ref	Ref
Some college	0.97; 0.86–1.10	2.27; 1.57–3.28	1.33; 1.07–1.64	2.27; 1.50–3.44	1.34; 1.09–1.66	1.56; 1.18–2.05
High school graduate or equivalent	1.23; 1.09–1.40	1.69; 1.12–2.57	1.19; 0.93–1.52	1.72; 1.08–2.75	1.27; 1.01–1.60	1.63; 1.19–2.22
Did not graduate high school	1.03; 0.85–1.23	2.08; 1.28–3.38	1.06; 0.76–1.50	2.21; 1.31–3.73	1.68; 1.25–2.26	1.73; 1.13–2.63
Health insurance
Privatee	Ref	Ref	Ref	Ref	Ref	Ref
Publicf	1.21; 1.05–1.38	2.26; 1.64–3.10	2.16; 1.71–2.73	2.10; 1.46–3.03	1.03; 0.81–1.31	1.20; 0.87–1.67
Uninsured	0.95; 0.78–1.15	1.29; 0.73–2.28	1.97; 1.47–2.64	1.66; 0.95–2.90	1.00; 0.69–1.45	1.21; 0.81–1.80
Dr. visit in last year
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	1.31; 1.11–1.53	1.08; 0.68–1.70	1.46; 1.10–1.94	3.45; 1.59–7.49	2.08; 1.51–2.85	1.02; 0.74–1.41
Cigarette smoking
Never	Ref	Ref	Ref	Ref	Ref	Ref
Former	1.07; 0.95–1.20	1.70; 1.23–2.35	1.20; 0.97–1.48	1.09; 0.76–1.56	1.13; 0.93–1.38	1.13; 0.86–1.47
Current	1.35; 1.16–1.56	2.70; 1.89–3.87	1.43; 1.10–1.84	1.08; 0.76–1.56	1.34; 1.05–1.72	0.81; 0.57–1.14

Bold values indicate statistical significance at p < 0.05.

^aRelative risks were adjusted for sex, age, race, Hispanic or Latino ethnicity, and college attainment, health insurance status, whether saw a doctor within the last 12 months, cigarette smoking status.

^bCHD, coronary heart disease.

^cHispanic or Latino ethnicity includes self-reported Central American, Cuban, Dominican, Mexican or Mexican American, Puerto Rican, South American, and other Hispanic, Latino, or Spanish.

^dOther race includes self-reported Native Hawaiian, Pacific Islander, more than 1 race, and “some other race.”

^ePrivate Insurance includes those who had both Private and Public Insurance.

^fPublic Insurance (primarily Medicare or Medicaid).

^gChronic pain, pain most or every day in the last 3 months.

^hCOPD, chronic obstructive pulmonary disease.

ⁱObesity, body mass index ≥ 30.

In both the unadjusted (Supplementary Table S3) and adjusted models (Table 3), age was the most consistent predictor of incident chronic disease/conditions, with the risk of incident arthritis, cancer, CHD, high cholesterol, chronic pain, COPD, diabetes, and hypertension increasing with age, while anxiety, asthma, depression and obesity decreased with age.

Differing relationships were seen between the 12 chronic diseases/conditions and educational attainment in the adjusted model: 1) For anxiety, arthritis, asthma, high cholesterol, and obesity, risk increased as educational attainment decreased; 2) CHD and chronic pain had highest risk in high school graduates; and 3) COPD, depression and diabetes had highest risks in those with some college education. There was no evidence of association between educational attainment and cancer. Again, the unadjusted and adjusted models produced similar results with minor discrepancies (e.g., whether there is a linear relationship between educational attainment and risk of CP).

In both unadjusted and adjusted regressions, those with public health insurance had higher risks for incident anxiety, arthritis, CHD, COPD, chronic pain depression, and diabetes, compared with those privately insured. While the unadjusted model identified significant associations between insurance status and high cholesterol and hypertension, these associations were not significant in the adjusted model.

In both unadjusted and adjusted models, participants who reported seeing a physician in the last year demonstrated higher risks of incident arthritis, high cholesterol, chronic pain, diabetes, and hypertension. Neither model type saw statistically significant associations between a physician visit and incident anxiety, COPD, or obesity. Only the adjusted models found associations between physician visits and increased risk for CHD and depression but decreased risks for cancer.

The adjusted models found that both current and former cigarette smokers had greater risks of incident CHD, high cholesterol, and COPD, with only current smokers having increased risks of chronic pain, depression, and hypertension. Smoking status was not associated with anxiety, arthritis, asthma, cancer, diabetes, or obesity in the adjusted models, but was associated with incident anxiety, arthritis, or cancer in the unadjusted analyses, suggesting that one or more of the model covariates may have confounded, mediated or modified smoking's effect.

Discussion

Our analyses establish the first comprehensive, nationally representative estimates of incidence rates for 12 chronic diseases/conditions with substantial public health burden in the US. The simultaneous examination of these diseases/conditions within a single dataset provides directly comparable estimates of disease onset and associated risk factors, addressing a critical gap in understanding the relative public health impact of these diseases/conditions to inform prevention strategies.

In our study, incidence rates varied substantially across diseases/conditions. For every 1000 adults followed for a year, approximately 85 developed chronic pain or high cholesterol, while about 14–15 received a diagnosis of CHD, COPD, or diabetes. The pattern suggests an inverse relationship between disease incidence and mortality: diseases/conditions associated with higher mortality (e.g., cancer, CHD, COPD, diabetes) demonstrated lower incidence compared with diseases less closely linked to mortality (arthritis, chronic pain, high cholesterol).

We also estimated the 2019 prevalence of the 12 diseases/conditions (Table 1) and found that the relative levels and patterns differed from those for incidence. For instance, while the 2019 prevalence of anxiety and asthma were identical (14.0%, respectively), their incidence rates differed by twofold (52.3/1000 PY vs 26.0/1000 PY). This discrepancy between prevalence and incidence highlights the critical role of disease duration and survival in shaping population health burden.

We compare our findings to previous work using the NHEFS that explored the incidence of the same or similar diseases/conditions (Supplementary Table S4). Unfortunately, few of the NHEFS studies presented incidence rates such as cumulative incidence or rates per PY. For several publications, sufficient information was presented for us to hand-calculate the cumulative incidence (Supplementary Table S4). Comparing the NHEFS and the NHIS-Longitudinal Cohort for any given disease, we found that the cumulative incidence increased as the years of follow-up increased. Therefore, it is vital to regard timeframe when interpreting the cumulative incidence literature.

Consistent with published NHEFS data,23, 24, 25 both current and former cigarette smokers had greater risks of incident CHD and COPD, with current smokers also at increased risks for chronic pain, depression, and hypertension. Like others,³¹ we did not find an association between smoking and incident diabetes. Smoking status was only associated with incident anxiety, arthritis, or cancer in the unadjusted analyses, suggesting that one or more of the model covariates substantially mediated, modified, or confounded smoking's effect. For cancer, it is also possible that different associations might have been found if we had investigated specific cancer types more closely related to smoking status (e.g., lung, trachea, mouth, etc). While the NHIS-Longitudinal Cohort does capture these details, the numbers were too small (N < 20) for reliable analyses. Different associations may have been found if we had been able to examine the number of cigarettes smoked per day. For instance, some data suggest that only very heavy smokers have increased risk of diabetes.⁴⁹ Unfortunately, the in-depth data were only captured for about one-third of smokers in the 2019 NHIS and not analyzed.

Obesity is considered both a chronic disease61, 62, 63 and a risk factor for other chronic diseases including cancer,⁶⁴ CHD,⁶⁵ and diabetes.⁶⁶ Thus, we expected that obesity would have overlapping risk factors with the other chronic diseases studied. While this is indeed the case for age, sex, race, and Hispanic-Latino ethnicity as predictors, the size and direction of the risks associated with obesity varied minimally compared with other chronic diseases. For instance, whereas participants aged 65+ had 40% lower risk of obesity than those aged 18–34, the corresponding risks for cancer, CHD and diabetes, corresponding risks were 2200%, 3481% and 351% higher. Our finding that obesity incidence peaks between 40 and 60 is consistent with national prevalence data.67, 68, 69

Also consistent with the literature is our founding that three core demographic variables, age, sex and race/Hispanic-Latino ethnicity were risk factors for several chronic diseases/conditions. For instance, age has been identified as a risk factor for CHD,^23,24 diabetes,28, 29, 30, 31, 32 COPD,¹¹ and obesity.⁶⁹ Sex is a risk factor for arthritis,¹⁹ asthma,¹¹ CHD,^23,24 COPD,¹¹ and obesity,^67,68 while race/Hispanic-Latino ethnicity is a risk factor for CHD,²⁴ COPD,¹¹ diabetes,28, 29, 30, 31, 32 and obesity.67, 68, 69

Consistent with studies from the 1990's and 2000's that used global measures of health status as outcomes when comparing types of health insurance,^70,71 we found those publicly insured had higher risks of incident chronic diseases compared with their private counterparts. We are not aware of any studies using more recent data, or studies which assessed specific chronic diseases as done in the present study.

Despite our comparison of nationally representative chronic disease/conditions incidence by core demographic characteristics, these results cannot be considered an exhaustive examination of chronic disease risk factors. Many important health behaviors associated with multiple chronic diseases, such as physical activity and alcohol use,¹ were not assessed in the NHIS-Longitudinal Cohort. The inclusion of these factors into the regression models may have modified the observed associations.

While many US incidence studies,^5,6,12, 13, 14^,31,32,43, 44, 45, 46, 47, 48, 49, 50 and those from other countries51, 52, 53, 54, 55, 56, 57 have used self-reported questionnaire data similar or identical to our NHIS-Longitudinal Cohort items, the most reliable incidence data would be based on health provider evaluations accompanied by clinical markers. In this regard, several studies have compared self-reports of diseases/conditions to physician reports. Particularly relevant are such sensitivity studies using the MEPS datasets, as MEPS participants are drawn from the previous year's NHIS adult sample. The MEPS participant survey uses questions identical or like those in the NHIS. If a MEPS participant reports that they saw a healthcare provider for a given disease/condition, the MEPS survey asks for permission to contact the healthcare provider and obtain chart data. Thus, it is possible to compare whether participant responses are consistent with provider responses. Such comparisons of MEPS self-reported and healthcare provider reported diseases/conditions found high agreement for diseases/conditions that cause pain, and those that require ongoing treatment and/or affect daily life.⁷² Similar findings have been found in clinical populations.^73,74 Relevant to the current analyses, Machlin et al., 2009⁷² found the highest concurrence between self-reports and doctor-reported conditions (>90%) for depression, diabetes and rheumatoid arthritis, greater than 85% concurrence for anxiety and asthma, greater than 80% concurrence for CHD and hypertension, and 78% concurrence for COPD. Cancer and hyperlipidemia had surprisingly low concurrence at 68% and 62%, respectively, suggesting our estimated rates for these two diseases/conditions are most prone to bias.

The NHIS-Longitudinal Cohort completion rate of 50.0% is a potential limitation that impacts on the generalizability of our study. However, this rate is on par with NHIS response rates in adjacent years (Supplementary Table S5 shows the NHIS sample adult response rates from 2017 to 2024) and it is comparable to the completion rates of other nationally-representative surveys such as Behavioral Risk Factor Surveillance System (BRFSS) survey (2020 response rate: 47.9%)⁷⁵ or the 2017–2020 NHANES (response rate: 51.0%).⁷⁶ The downward trend in NHIS response rates is consistent with declining response rates seen in all national surveys over the prior two decades^77,78 due to multiple factors such as difficulty contacting selected respondents and increased likelihood of interview refusal.^77,78

The relationships across the chronic diseases/conditions studied were not examined, particularly with one chronic disease/condition being a risk factor for another; for example the bidirectional associations of diabetes and depression;79, 80, 81, 82 asthma as a risk for factor for COPD,⁸³ and heart disease,⁸⁴ chronic pain as risk factors for heart disease,⁸⁵ and cancer, obesity, and hypertension as risk factors for many other chronic diseases.¹

The CDC's National Center for Health Statistics created the NHIS-Longitudinal Cohort specifically to help alleviate difficulties in study recruitment for the 2020 NHIS at the beginning of the COVID-10 pandemic. While re-recruiting 2019 NHIS participants did dramatically increase the number of participants taking the 2020 survey, the NCHS identified several biases in the data, as least some of which may have resulted for the high proportion of telephone vs in-person interviews.^40,41 Specifically, the re-recruited participants tended to be older, in better overall health, more likely to be homeowners, and to use preventative care versus 2019 participants who did not participate in the NHIS-Longitudinal Cohort. Although the NHIS-Longitudinal Cohort sampling weights created by NCHS and applied by our team account for these differences to some degree, the sampling weights cannot eliminate the potential for biases that might have artificially lowered disease/condition rates in our sample. This potential bias, along with the absence of institutionalized individuals in the NHIS sampling frame, impacts on the generalizability of our findings to all US adults.

A final study limitation is that since the cohort was followed into the first year of COVID-19 (2020), our data may not necessarily be transferable to non-COVID-19 years for two reasons: 1) Since we were assessing the first report of a diagnosed condition/disease (except for chronic pain and obesity), the dramatic delay or reduction in healthcare visits during the COVID pandemic86, 87, 88 may have led us to underestimate the true incidence of disease. 2) There is accumulating evidence that COVID-19 has a bi-directional association with several chronic diseases/conditions including anxiety and depression,⁸⁹ diabetes,90, 91, 92 and hypertension.⁹³ The 2020 NHIS did ask participants if they ever had COVID-19. Yet, this information was collected at the same time as the reports of the 12 chronic diseases, preventing assessment of directionality for any observed associations. Therefore, the 2020 COVID-19 data was not incorporated into the current study.

In contrast to ubiquitous prevalence studies, incidence analyses representative of the US adult population are remarkably rare. To address a critical gap left by decades of single-disease incidence studies often based on data collected nearly half a century ago, we provide the first simultaneous analysis of incidence for twelve leading chronic diseases/conditions. The differences we observed in both rates and risk factors for disease incidence establish critical benchmarks for understanding burden through future periodic nationally representative cohort studies. These results, combined with published estimates of health-related disability, mortality, and societal costs,¹ provide an essential empirical foundation for prioritizing public health initiatives, particularly vital as healthcare systems navigate both chronic and emerging infectious disease challenges.⁹⁴

Contributors

RLN: literature search, study design, data analysis and interpretation, writing, figures, tables. RLN made the decision to submit the manuscript on behalf of all other authors.

TF: literature search, data interpretation, writing.

HGP: conceptualization, investigation, methodology, writing—review and editing.

FPK: conceptualization, data curation, methodology, writing—review and editing.

KAM: investigation, visualization, writing—review and editing.

RS: data investigation, writing—review and editing.

AZ: literature review, text review and editing.

Data sharing statement

All data are deidentified and freely and publicly available on the website of the NCHS. The NCHS website contains all documentation necessary (e.g., data dictionary, sampling procedures, response rates, etc.) to use the data including coding examples for a variety of statistical packages able to handle the complex sampling design used by NCHS. Data and documentation can be downloaded at: https://www.cdc.gov/nchs/nhis/documentation/2020-nhis.html.

Declaration of interests

RLN, TF, HGP, RS, and AZ report no conflicts of interest.

Unrelated research by FPK is supported by NIH NIAMS grant R01AR071440 and the Duke Clinical Research Institute. FPK also received travel support to scientific meetings from the United States Association for the Study of Pain and the International Association for the Study of Pain, and received compensation from the NIH HEAL initiative for serving as a workshop panelist with a subsequent follow-up webinar.

KM is a full-time travel nurse with the company, Turstaff; all compensation is exclusively nursing related.