Long COVID and Recovery Among US Adults
Rishi M. Shah, Kavya M. Shah, Ji Chen, Mitsuaki Sawano, Bornali Bhattacharjee, Harlan M. Krumholz
TL;DR
This study looks at how common long COVID is and how US adults recover from it using national health survey data.
Contribution
The study provides new insights into long COVID prevalence and recovery patterns in the US adult population.
Findings
Long COVID is prevalent among US adults according to the National Health Interview Survey data.
Recovery patterns vary, with some individuals experiencing prolonged symptoms.
Abstract
This cross-sectional study examines the prevalence and recovery patterns of long COVID among US adults using data from the National Health Interview Survey.
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
| Characteristic | Year | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2022 | 2023 | 2024 | ||||||||||
| No LC, No. (%) (n = 7149) | LC, No. (%) (n = 1797) | AOR (95% CI) | No LC, No. (%) (n = 12 893) | LC, No. (%) (n = 2398) | AOR (95% CI) | No LC, No. (%) (n = 16 008) | LC, No. (%) (n = 2627) | AOR (95% CI) | ||||
| Age group, y | ||||||||||||
| 18-34 | 2053 (36) | 416 (29) | <.001 | 1 [Reference] | 3253 (33) | 526 (28) | <.001 | 1 [Reference] | 3804 (31) | 578 (29) | .006 | 1 [Reference] |
| 35-49 | 1985 (28) | 525 (31) | 1.60 (1.35-1.89) | 3296 (27) | 642 (29) | 1.36 (1.16-1.59) | 4044 (27) | 699 (29) | 1.19 (1.04-1.37) | |||
| 50-64 | 1683 (22) | 483 (27) | 1.64 (1.38-1.95) | 3110 (23) | 675 (27) | 1.48 (1.27-1.73) | 3804 (23) | 693 (25) | 1.18 (1.02-1.37) | |||
| ≥65 | 1417 (14) | 370 (13) | 1.15 (0.92-1.44) | 3216 (17) | 554 (16) | 1.00 (0.84-1.19) | 4336 (19) | 651 (17) | 0.91 (0.76-1.09) | |||
| Sex | ||||||||||||
| Male | 3270 (49) | 638 (37) | <.001 | 1 [Reference] | 5902 (49) | 839 (37) | <.001 | 1 [Reference] | 7258 (48) | 926 (38) | <.001 | 1 [Reference] |
| Female | 3879 (51) | 1159 (63) | 1.53 (1.34-1.74) | 6991 (51) | 1559 (63) | 1.49 (1.34-1.67) | 8750 (52) | 1701 (62) | 1.46 (1.31-1.63) | |||
| Race and ethnicity | ||||||||||||
| Hispanic | 1187 (19) | 321 (21) | <.001 | 0.96 (0.79-1.17) | 2060 (18) | 419 (20) | <.001 | 0.98 (0.84-1.15) | 2427 (18) | 402 (18) | <.001 | 0.87 (0.74-1.01) |
| Non-Hispanic Asian | 407 (5) | 44 (2) | 0.40 (0.28-0.58) | 766 (6) | 77 (3) | 0.57 (0.43-0.76) | 1004 (6) | 86 (4) | 0.59 (0.44-0.77) | |||
| Non-Hispanic Black | 663 (10) | 162 (9) | 0.80 (0.64-1.00) | 1130 (10) | 208 (9) | 0.86 (0.70-1.05) | 1252 (9) | 220 (10) | 0.95 (0.78-1.16) | |||
| Non-Hispanic White | 4703 (63) | 1209 (64) | 1 [Reference] | 8625 (64) | 1622 (64) | 1 [Reference] | 10 874 (64) | 1817 (65) | 1 [Reference] | |||
| Other | 189 (3) | 61 (4) | 1.27 (0.89-1.81) | 312 (2) | 72 (3) | 1.07 (0.78-1.46) | 451 (3) | 102 (4) | 1.09 (0.83-1.43) | |||
| Education | ||||||||||||
| College degree | 3992 (50) | 913 (47) | .20 | 1 [Reference] | 7271 (51) | 1227 (47) | .007 | 1 [Reference] | 9175 (53) | 1382 (49) | .03 | 1 [Reference] |
| HS diploma | 2671 (42) | 748 (43) | 0.99 (0.87-1.14) | 4702 (40) | 986 (44) | 1.04 (0.91-1.19) | 5835 (40) | 1054 (43) | 1.02 (0.90-1.15) | |||
| <HS | 461 (8) | 128 (10) | 0.98 (0.73-1.31) | 869 (9) | 175 (9) | 0.81 (0.63-1.03) | 936 (8) | 181 (9) | 1.02 (0.81-1.29) | |||
| Urbanicity | ||||||||||||
| Metropolitan | 6112 (87) | 1481 (85) | .02 | 1 [Reference] | 11 089 (87) | 1969 (84) | <.001 | 1 [Reference] | 12 807 (87) | 1933 (85) | <.001 | 1 [Reference] |
| Nonmetropolitan | 1037 (13) | 316 (15) | 1.12 (0.94-1.33) | 1804 (13) | 429 (16) | 1.18 (1.01-1.38) | 3201 (13) | 694 (17) | 1.23 (1.06-1.44) | |||
| Employment status | ||||||||||||
| Employed | 4899 (75) | 1152 (71) | .002 | 1 [Reference] | 8339 (71) | 1443 (66) | <.001 | 1 [Reference] | 10 118 (69) | 1629 (68) | .39 | 1 [Reference] |
| Unemployed | 1992 (25) | 579 (29) | 1.07 (0.91-1.25) | 4073 (29) | 876 (34) | 1.13 (0.98-1.30) | 5374 (31) | 930 (32) | 0.99 (0.86-1.14) | |||
| Household income | ||||||||||||
| ≥400% FPL | 3513 (48) | 712 (39) | <.001 | 1 [Reference] | 6233 (48) | 908 (38) | <.001 | 1 [Reference] | 7847 (49) | 1005 (39) | <.001 | 1 [Reference] |
| 200%-399% FPL | 2034 (29) | 565 (33) | 1.41 (1.21-1.64) | 3711 (29) | 736 (31) | 1.30 (1.14-1.48) | 4626 (29) | 849 (33) | 1.45 (1.27-1.66) | |||
| <200% FPL | 1602 (23) | 520 (28) | 1.44 (1.19-1.74) | 2949 (23) | 754 (31) | 1.49 (1.27-1.74) | 3535 (22) | 773 (28) | 1.56 (1.32-1.83) | |||
| Insurance status | ||||||||||||
| Private | 4941 (70) | 1145 (65) | <.001 | 1 [Reference] | 8826 (70) | 1477 (62) | <.001 | 1 [Reference] | 10 927 (70) | 1686 (67) | .10 | 1 [Reference] |
| Public | 1677 (21) | 513 (26) | 1.19 (1.00-1.41) | 3247 (23) | 742 (29) | 1.24 (1.08-1.42) | 4070 (23) | 754 (25) | 1.00 (0.87-1.16) | |||
| Uninsured | 515 (9) | 137 (9) | 1.07 (0.83-1.38) | 781 (7) | 172 (9) | 1.23 (0.98-1.54) | 978 (8) | 182 (8) | 1.01 (0.82-1.25) | |||
| Usual source of care | ||||||||||||
| Yes | 6412 (89) | 1636 (91) | .06 | 1 [Reference] | 11 595 (89) | 2149 (89) | .62 | 1 [Reference] | 14 630 (90) | 2413 (91) | .39 | 1 [Reference] |
| No | 726 (11) | 155 (9.0) | 0.91 (0.72-1.15) | 1263 (11) | 243 (11) | 1.06 (0.87-1.29) | 1345 (10) | 209 (9.0) | 0.93 (0.76-1.14) | |||
| US region | ||||||||||||
| Northeast | 1263 (19) | 277 (17) | .09 | 1 [Reference] | 2193 (19) | 317 (15) | .002 | 1 [Reference] | 2743 (19) | 366 (15) | .005 | 1 [Reference] |
| Midwest | 1599 (21) | 433 (22) | 1.02 (0.82-1.26) | 2741 (20) | 527 (21) | 1.24 (1.03-1.50) | 3545 (20) | 633 (22) | 1.27 (1.05-1.53) | |||
| South | 2642 (38) | 643 (36) | 0.91 (0.74-1.13) | 4570 (36) | 934 (39) | 1.31 (1.11-1.55) | 5566 (37) | 930 (38) | 1.20 (1.01-1.42) | |||
| West | 1645 (22) | 444 (25) | 1.17 (0.93-1.46) | 3389 (25) | 620 (25) | 1.31 (1.10-1.57) | 4154 (25) | 698 (26) | 1.33 (1.11-1.59) | |||
| Characteristic | Year | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2022 | 2023 | 2024 | ||||||||||
| Not recovered, No. (%) (n = 919) | Recovered, No. (%) (n = 872) | AOR (95% CI) | Not recovered, No. (%) (n = 1063) | Recovered, No. (%) (n = 1319) | AOR (95% CI) | Not recovered, No. (%) (n = 1106) | Recovered, No. (%) (n = 1501) | AOR (95% CI) | ||||
| Age group, y | ||||||||||||
| 18-34 | 167 (23) | 247 (33) | <.001 | 1 [Reference] | 184 (23) | 339 (32) | <.001 | 1 [Reference] | 175 (22) | 399 (34) | <.001 | 1 [Reference] |
| 35-49 | 281 (34) | 243 (29) | 0.57 (0.42-0.78) | 256 (28) | 385 (30) | 0.71 (0.52-0.97) | 258 (28) | 436 (30) | 0.71 (0.54-0.94) | |||
| 50-64 | 257 (28) | 223 (26) | 0.65 (0.48-0.90) | 318 (29) | 353 (25) | 0.60 (0.46-0.79) | 323 (29) | 365 (22) | 0.50 (0.38-0.67) | |||
| ≥65 | 211 (15) | 158 (12) | 0.68 (0.46-1.00) | 305 (20) | 241 (13) | 0.65 (0.48-0.89) | 348 (21) | 298 (14) | 0.45 (0.32-0.63) | |||
| Sex | ||||||||||||
| Male | 295 (33) | 341 (40) | .001 | 1 [Reference] | 347 (34) | 486 (40) | .02 | 1 [Reference] | 380 (35) | 538 (40) | .07 | 1 [Reference] |
| Female | 624 (67) | 531 (60) | 0.77 (0.60-0.97) | 716 (66) | 833 (60) | 0.81 (0.65-1.00) | 726 (65) | 963 (60) | 0.86 (0.70-1.05) | |||
| Race and ethnicity | ||||||||||||
| Hispanic | 133 (17) | 187 (24) | .02 | 1.47 (1.08-1.99) | 168 (19) | 250 (21) | .61 | 1.29 (0.95-1.73) | 137 (16) | 264 (20) | .19 | 1.23 (0.90-1.67) |
| Non-Hispanic Asian | 21 (2) | 23 (3) | 1.47 (0.73-2.93) | 28 (3) | 47 (3) | 0.99 (0.54-1.83) | 26 (3) | 60 (4) | 1.59 (0.88-2.87) | |||
| Non-Hispanic Black | 75 (8) | 86 (10) | 1.66 (1.11-2.50) | 90 (9) | 118 (10) | 1.31 (0.90-1.89) | 97 (10) | 122 (9) | 0.92 (0.62-1.38) | |||
| Non-Hispanic White | 661 (69) | 544 (59) | 1 [Reference] | 745 (67) | 865 (63) | 1 [Reference] | 803 (68) | 996 (63) | 1 [Reference] | |||
| Other | 29 (4) | 32 (5) | 1.45 (0.66-3.17) | 32 (3) | 39 (3) | 1.08 (0.63-1.85) | 43 (3) | 59 (4) | 1.11 (0.63-1.93) | |||
| Education | ||||||||||||
| College degree | 457 (47) | 456 (48) | .01 | 1 [Reference] | 511 (43) | 708 (49) | .01 | 1 [Reference] | 565 (49) | 805 (49) | .60 | 1 [Reference] |
| HS diploma | 406 (46) | 337 (41) | 0.93 (0.72-1.20) | 460 (46) | 518 (44) | 0.87 (0.70-1.08) | 461 (43) | 585 (42) | 0.93 (0.75-1.16) | |||
| <HS | 54 (7) | 74 (12) | 1.68 (0.99-2.85) | 88 (11) | 87 (7) | 0.67 (0.43-1.04) | 74 (8) | 107 (9) | 1.21 (0.76-1.91) | |||
| Urbanicity | ||||||||||||
| Metropolitan | 732 (83) | 743 (86) | .18 | 1 [Reference] | 855 (82) | 1101 (85) | .08 | 1 [Reference] | 775 (81) | 1142 (85) | .03 | 1 [Reference] |
| Nonmetropolitan | 187 (17) | 129 (14) | 0.91 (0.65-1.29) | 208 (18) | 218 (15) | 0.89 (0.68-1.16) | 331 (19) | 359 (15) | 0.85 (0.67-1.07) | |||
| Employment status | ||||||||||||
| Employed | 566 (68) | 582 (72) | .05 | 1 [Reference] | 556 (57) | 879 (72) | <.001 | 1 [Reference] | 637 (65) | 981 (71) | .006 | 1 [Reference] |
| Unemployed | 317 (32) | 260 (28) | 0.89 (0.67-1.18) | 483 (43) | 388 (28) | 0.63 (0.49-0.80) | 446 (35) | 475 (29) | 0.99 (0.77-1.27) | |||
| Household income | ||||||||||||
| ≥400% FPL | 354 (39) | 355 (39) | .83 | 1 [Reference] | 373 (36) | 531 (40) | .008 | 1 [Reference] | 407 (39) | 592 (39) | .99 | 1 [Reference] |
| 200%-399% FPL | 291 (32) | 274 (34) | 0.96 (0.73-1.27) | 315 (29) | 414 (32) | 0.99 (0.78-1.26) | 351 (33) | 489 (34) | 0.98 (0.77-1.25) | |||
| <200% FPL | 274 (29) | 243 (27) | 1.00 (0.71-1.41) | 375 (35) | 374 (28) | 0.86 (0.65-1.14) | 348 (28) | 420 (28) | 0.95 (0.71-1.26) | |||
| Insurance status | ||||||||||||
| Private | 582 (65) | 558 (65) | .05 | 1 [Reference] | 609 (59) | 859 (64) | <.001 | 1 [Reference] | 692 (66) | 982 (68) | .08 | 1 [Reference] |
| Public | 281 (28) | 231 (24) | 0.80 (0.61-1.06) | 388 (34) | 348 (26) | 0.90 (0.71-1.16) | 349 (27) | 397 (23) | 0.95 (0.73-1.25) | |||
| Uninsured | 55 (7.4) | 82 (11) | 1.13 (0.71-1.81) | 63 (6.8) | 108 (10) | 1.55 (1.00-2.41) | 63 (7.2) | 119 (9.3) | 1.21 (0.78-1.88) | |||
| Usual source of care | ||||||||||||
| Yes | 846 (92) | 784 (89) | .19 | 1 [Reference] | 963 (90) | 1172 (88) | .30 | 1 [Reference] | 1024 (92) | 1370 (91) | .55 | 1 [Reference] |
| No | 69 (8.3) | 86 (11) | 1.12 (0.73-1.72) | 97 (10) | 144 (12) | 0.99 (0.68-1.42) | 81 (8.5) | 127 (9.3) | 0.81 (0.56-1.18) | |||
| US region | ||||||||||||
| Northeast | 130 (15) | 143 (18) | .14 | 1 [Reference] | 144 (15) | 171 (15) | .90 | 1 [Reference] | 144 (15) | 221 (15) | .81 | 1 [Reference] |
| Midwest | 226 (22) | 206 (23) | 0.90 (0.62-1.31) | 233 (20) | 288 (21) | 1.05 (0.73-1.50) | 281 (21) | 348 (22) | 1.10 (0.76-1.59) | |||
| South | 351 (39) | 291 (33) | 0.66 (0.45-0.96) | 421 (39) | 509 (39) | 0.99 (0.72-1.36) | 393 (37) | 532 (38) | 1.09 (0.77-1.53) | |||
| West | 212 (24) | 232 (26) | 0.82 (0.55-1.23) | 265 (26) | 351 (25) | 0.89 (0.62-1.27) | 288 (27) | 400 (25) | 0.94 (0.66-1.34) | |||
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsLong-Term Effects of COVID-19 · COVID-19 Clinical Research Studies · Intensive Care Unit Cognitive Disorders
Introduction
Long COVID (LC) remains a public health challenge, with millions of US adults continuing to experience its debilitating effects. Although studies have identified characteristics linked to LC and recovery from LC,^1,2^ questions remain about its scope and trajectory. Using nationally representative data, we estimate LC prevalence and recovery and assess how associated factors have evolved.
Methods
We used data from the 2022 to 2024 National Health Interview Survey (NHIS), an annual cross-sectional survey representative of the US noninstitutionalized civilian population.^3^ We used NHIS-provided sampling weights, strata, and primary sampling units to yield nationally representative estimates.
Our sample consisted of adults 18 years or older who self-reported either a positive COVID-19 test or a physician’s diagnosis of COVID-19 and had complete data on LC and recovery status. LC was defined as an affirmative response to the question, “Did you have any symptoms lasting 3 months or longer that you did not have prior to having coronavirus or COVID-19?” Recovery from LC was defined as an affirmative response to the previous question and a negative response to the question, “Do you have symptoms now?” (asked as a follow-up to those ever-reporting LC). We compared sociodemographic characteristics between adults with and without LC, and between those who had and had not recovered, using Rao-Scott χ^2^ tests (2-sided P < .05 indicating significance). Multivariable logistic regression was used to identify factors associated with LC and recovery, with potential covariates chosen based on prior knowledge of sociodemographic factors associated with COVID-19 infection.^1,2^ Multiplicity adjustment was not performed, given the study’s descriptive nature. Analyses were conducted using R, version 4.3.1 (R Foundation) with the survey package. This study was exempt from review by the Yale University Institutional Review Board because data are publicly available and deidentified. Analyses followed STROBE guidelines.
Results
Our sample included 9022 adults (weighted mean [SE] age, 44.4 [0.21] years; weighted 52.7% female) in 2022, 15 354 (45.7 [0.19] years; 53.0% female) in 2023, and 18 691 (46.3 [0.18] years; 53.6% female) in 2024. The weighted proportion of US adults reporting prior COVID-19 infection rose from 39.6% (95% CI, 38.8%-40.3%) in 2022 to 60.4% (95% CI, 59.7%-61.1%) in 2024. Among those infected, the weighted prevalence of LC since the onset of the pandemic declined from 19.7% (95% CI, 18.7%-20.7%) to 13.7% (95% CI, 13.1%-14.4%), while weighted LC recovery rates rose from 51.2% (95% CI, 48.4%-53.9%) to 59.7% (95% CI, 57.1%-62.2%).
Female sex, age 35 to 64 years, and low household income were consistently associated with higher odds of LC, while non-Hispanic Asian adults had lower odds (Table 1). Recovery from LC remained less likely among adults 35 years or older. No consistent differences by sex, race and ethnicity, education, urbanicity, employment status, household income, insurance status, usual source of care, or US region were observed (Table 2).
Discussion
In 2024, 8.3% of US adults—an estimated 21.3 million—reported ever having LC, among whom nearly 6 in 10 reported recovery, consistent with RECOVER initiative^4^ findings showing similar LC prevalence in 2023 and 2024 and longitudinal Veterans Affairs data demonstrating declining LC prevalence.^5^ Yet many adults, particularly those 35 years or older, continue to experience lasting symptoms. With no LC treatment demonstrating clear efficacy, greater investment in understanding biological mechanisms, including immunotypic differences between those who recover and those who do not, may provide insights into pathways of persistence and potential targets for intervention.^6^
This study has limitations. Self-reported LC and recovery may be misclassified, potentially biasing associations with sociodemographic factors; individuals with intermittent symptoms may be misclassified as recovered, inflating recovery estimates; the repeated cross-sectional design across survey years may reflect population-level prevalence rather than individual recovery trajectories; timing of acute COVID-19 infection and recovery was unavailable, precluding inclusion of time-dependent covariates such as COVID-19 vaccination status; and the narrow definition of LC, assessed at any time since the pandemic’s start, may affect prevalence estimates. Nevertheless, self-report remains the only way to assess LC in population data.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Wu Y, Sawano M, Wu Y, . Factors associated with long COVID: insights from two nationwide surveys. Am J Med. 2024;137(6):515-519. doi:10.1016/j.amjmed.2024.02.03238490304 · doi ↗ · pubmed ↗
- 2Shah KM, Shah RM, Sawano M, . Factors associated with long COVID recovery among US adults. Am J Med. 2024;137(9):896-899. doi:10.1016/j.amjmed.2024.04.01738649004 · doi ↗ · pubmed ↗
- 3Centers for Disease Control and Prevention. National Center for Health Statistics: 2022-2024 National Health Interview Survey. Accessed August 1, 2025. https://www.cdc.gov/nchs/nhis/documentation/index.html
- 4Geng LN, Erlandson KM, Hornig M, ; RECOVER Consortium. 2024 update of the RECOVER-Adult Long COVID Research Index. JAMA. 2025;333(8):694-700. doi:10.1001/jama.2024.2418439693079 PMC 11862971 · doi ↗ · pubmed ↗
- 5Xie Y, Choi T, Al-Aly Z. Postacute sequelae of SARS-COV-2 infection in the Pre-Delta, Delta, and Omicron eras. N Engl J Med. 2024;391(6):515-525. doi:10.1056/NEJ Moa 240321139018527 PMC 11687648 · doi ↗ · pubmed ↗
- 6Klein J, Wood J, Jaycox JR, . Distinguishing features of long COVID identified through immune profiling. Nature. 2023;623(7985):139-148. doi:10.1038/s 41586-023-06651-y 37748514 PMC 10620090 · doi ↗ · pubmed ↗