Variability in Long COVID Definitions and Validation of Published Prevalence Rates

Lauren E Wisk; Michelle L’Hommedieu; Kate Diaz Roldan; Imtiaz Ebna Mannan; Erica S Spatz; Robert A Weinstein; Arjun K Venkatesh; Michael Gottlieb; Ryan Huebinger; Kristin L Rising; Juan Carlos C Montoy; Kari A Stephens; Robert M Rodriguez; Mandy J Hill; Kelli N O’Laughlin; Nicole L Gentile; Ahamed H Idris; Shu-Xia Li; Michelle Santangelo; Efrat R Kean; Samuel A McDonald; Kristyn Gatling; Joann G Elmore

doi:10.1001/jamanetworkopen.2025.26506

. 2025 Aug 12;8(8):e2526506. doi: 10.1001/jamanetworkopen.2025.26506

Variability in Long COVID Definitions and Validation of Published Prevalence Rates

Lauren E Wisk ^1,^✉, Michelle L’Hommedieu ¹, Kate Diaz Roldan ¹, Imtiaz Ebna Mannan ², Erica S Spatz ^3,⁴, Robert A Weinstein ^5,⁶, Arjun K Venkatesh ^3,⁷, Michael Gottlieb ⁸, Ryan Huebinger ⁹, Kristin L Rising ^10,¹¹, Juan Carlos C Montoy ¹², Kari A Stephens ^13,¹⁴, Robert M Rodriguez ¹⁵, Mandy J Hill ⁹, Kelli N O’Laughlin ^16,¹⁷, Nicole L Gentile ^13,^18,¹⁹, Ahamed H Idris ²⁰, Shu-Xia Li ³, Michelle Santangelo ⁸, Efrat R Kean ¹⁰, Samuel A McDonald ^20,²¹, Kristyn Gatling ⁵, Joann G Elmore ¹, for the INSPIRE Group

¹Division of General Internal Medicine and Health Services Research, David Geffen School of Medicine at the University of California, Los Angeles

²Codetta Bio Inc, Morrisville, North Carolina

³Center for Outcomes Research and Evaluation (CORE), Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut

⁴Department of Epidemiology, Yale School of Public Health, New Haven, Connecticut

⁵Division of Infectious Diseases, Department of Internal Medicine, Rush University Medical Center, Chicago, Illinois

⁶Division of Infectious Diseases, Department of Medicine, Cook County Hospital, Chicago, Illinois

⁷Department of Emergency Medicine, Yale School of Medicine, New Haven, Connecticut

⁸Department of Emergency Medicine, Rush University Medical Center, Chicago, Illinois

⁹Department of Emergency Medicine, UTHealth Houston, Houston, Texas

¹⁰Department of Emergency Medicine, Sidney Kimmel Medical School, Thomas Jefferson University, Philadelphia, Pennsylvania

¹¹Center for Connected Care, Thomas Jefferson University, Philadelphia, Pennsylvania

¹²Department of Emergency Medicine, University of California, San Francisco

¹³Department of Family Medicine, University of Washington, Seattle

¹⁴Department of Biomedical Informatics and Medical Education, University of Washington, Seattle

¹⁵Department of Medicine, University of California Riverside School of Medicine

¹⁶Department of Emergency Medicine, University of Washington, Seattle

¹⁷Department of Global Health, University of Washington, Seattle

¹⁸Department of Rehabilitation Medicine, University of Washington, Seattle

¹⁹Department of Laboratory Medicine and Pathology, University of Washington, Seattle

²⁰Department of Emergency Medicine, University of Texas Southwestern Medical Center, Dallas

²¹Clinical Informatics Center, University of Texas Southwestern Medical Center, Dallas

Accepted for Publication: June 16, 2025.

Published: August 12, 2025. doi:10.1001/jamanetworkopen.2025.26506

^✉

Corresponding Author: Lauren E. Wisk, PhD, Division of General Internal Medicine and Health Services Research, David Geffen School of Medicine at UCLA, 1100 Glendon Ave, Ste 850, Los Angeles, CA 90024 (lwisk@mednet.ucla.edu).

Author Contributions: Mr Ebna Mannan had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Wisk and L’Hommedieu were co–first authors.

Concept and design: Wisk, L’Hommedieu, Ebna Mannan, Gottlieb, Huebinger, Stephens, Rodriguez, Santangelo, Elmore.

Acquisition, analysis, or interpretation of data: Wisk, L’Hommedieu, Diaz Roldan, Ebna Mannan, Spatz, Weinstein, Venkatesh, Gottlieb, Huebinger, Rising, Montoy, Stephens, Rodriguez, Hill, O’Laughlin, Gentile, Idris, Li, Kean, McDonald, Gatling, Elmore.

Drafting of the manuscript: Wisk, L’Hommedieu, Diaz Roldan, Ebna Mannan, Gottlieb, Stephens, Li, Santangelo.

Critical review of the manuscript for important intellectual content: Wisk, Diaz Roldan, Ebna Mannan, Spatz, Weinstein, Venkatesh, Gottlieb, Huebinger, Rising, Montoy, Stephens, Rodriguez, Hill, O’Laughlin, Gentile, Idris, Santangelo, Kean, McDonald, Gatling, Elmore.

Statistical analysis: Wisk, Ebna Mannan, Stephens, Li.

Obtained funding: Weinstein, Gottlieb, Stephens, Idris.

Administrative, technical, or material support: Wisk, L’Hommedieu, Diaz Roldan, Weinstein, Gottlieb, Huebinger, Stephens, Rodriguez, Hill, Santangelo, Gatling, Elmore.

Supervision: Wisk, L’Hommedieu, Weinstein, Venkatesh, Gottlieb, Stephens, Rodriguez, O’Laughlin, Li, Santangelo, Elmore.

Conflict of Interest Disclosures: Dr Wisk reported receiving grants from the National Institute of Diabetes and Digestive and Kidney Diseases outside the submitted work. Dr L’Hommedieu reported receiving grants from the National Institutes of Health (NIH) during the conduct of the study and from the Patient-Centered Outcomes Research Institute and the California Breast Cancer Research Program outside the submitted work. Dr Venkatesh reported receiving grants from the SAEM Foundation and Centers for Medicare & Medicaid Services outside the submitted work. Dr Gottlieb reported receiving grants from the Bill and Melinda Gates Foundation and the Biomedical Advanced Research and Development Authority during the conduct of the study. Dr Rising reported receiving grants from the NIH, Abbott, Healgen, Prenosis, and MeMed outside the submitted work. Dr Montoy reported receiving grants from the Centers for Disease Control and Prevention (CDC), the Substance Abuse and Mental Health Services Administration, and the US Food & Drug Administration outside the submitted work. Dr Stephens reported receiving grants from the NIH outside the submitted work. Dr Rodriguez reported receiving funding from the National Institute of Allergy and Infectious Diseases and from Pfizer Inc during the conduct of the study. Dr O’Laughlin reported receiving funding from the NIH during the conduct of the study. Dr Gentile reported receiving grants from the Agency for Healthcare Research and Quality, National Center for Complementary and Integrative Health, and National Institute on Aging outside the submitted work and being codirector of a long COVID clinic at the University of Washington. Dr Idris reported receiving grants from the National Heart, Lung, and Blood Institute outside the submitted work and serving as an unpaid volunteer on both the American Heart Association Emergency Cardiovascular Care Committee and the clinical advisory board of Stryker Belfast, Inc. Dr Elmore reported serving as editor-in-chief of adult primary care topics for UpToDate and as director of the National Clinician Scholars Program at the University of California, Los Angeles. No other disclosures were reported.

Funding/Support: INSPIRE was funded by grant 75D30120C08008 from the CDC’s National Center for Immunization and Respiratory Diseases (Dr Weinstein).

Role of the Funder/Sponsor: A partner from the CDC (Sharon Saydah, MHS, PhD) assisted with the design and conduct of the study; interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The INSPIRE Group members appear in Supplement 2.

Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC.

Data Sharing Statement: See Supplement 3.

Additional Contributions: Public Health–Seattle & King County and the California Department of Public Health assisted with participant recruitment for this study. The Clinical and Translational Science Institute (CTSI) COVID Clinical Research Steering Committee and the CTSI Office of Clinical Research Patient Navigation Team and Bioinformatics Program assisted with study recruitment. The University of Washington Institute of Translational Health Sciences provided support for REDCap and for biomedical informatics resources used by the University of Washington Clinical Core and Enrolling Site to enable study recruitment, which is funded by the National Center for Advancing Translational Sciences, NIH under award number UL1TR002319.

^✉

Corresponding author.

PMCID: PMC12344537 PMID: 40794409

This cohort study uses varying definitions of long COVID from 5 published studies to estimate its prevalence in participants in the Innovative Support for Patients With SARS-CoV-2 Infections Registry.

Key Points

Question

What is the prevalence of long COVID in the Innovative Support for Patients With SARS-CoV-2 Infections Registry (INSPIRE) cohort based on definitions in published literature?

Findings

In this prospective cohort study of 4575 INSPIRE participants, when replicating long COVID definitions used in the 5 published studies with the largest sample size of reviewed studies, the prevalence of long COVID among participants who were COVID-19 positive varied at 3 months from 32% to 42% and at 6 months from 15% to 22%.

Meaning

The findings suggest that long COVID prevalence estimates vary as a function of disparate definitions used.

Abstract

Importance

Long COVID definitions vary widely, and no consensus exists on how to accurately measure its prevalence, complicating both clinical care and research.

Objective

To assess long COVID prevalence using various definitions from published literature.

Design, Setting, and Participants

This prospective, multicenter cohort study used data from the longitudinal Innovative Support for Patients With SARS-CoV-2 Infections Registry (INSPIRE). Participants aged 18 years or older with symptoms suggestive of COVID-19 illness at the time of their index SARS-CoV-2 test enrolled at 8 sites across the US from December 11, 2020, through August 29, 2022, with follow-up surveys collected through February 28, 2023.

Exposure

Positive or negative SARS-CoV-2 test result at the time of acute symptoms.

Main Outcomes and Measures

Long COVID prevalence among INSPIRE participants with a positive vs negative index SARS-CoV-2 test, based on long COVID definitions in published literature. Secondary outcomes were sensitivity and specificity of published definitions compared with self-reported long COVID.

Results

A total of 4575 INSPIRE participants were included (mean [SD] age, 40.40 [14.58] years). Most were female (3013 of 4448 [67.7%]) and aged 18 to 49 years (3338 of 4541 [73.5%]). Applying 5 published definitions for long COVID yielded a prevalence that ranged from 30.84% (95% CI, 29.33%-32.40%) to 42.01% (95% CI, 40.37%-43.66%) at 3 months and 14.23% (95% CI, 13.01%-15.55%) to 21.94% (95% CI, 20.47%-23.47%) at 6 months postinfection; in the 5 comparator studies, reported prevalence of long COVID at 1 to 5 months postinfection ranged from 2.6% (≥84 days) to 47.4% (3-5 months) and at 6 or more months postinfection ranged from 10.0% (95% CI, 8.8%-11.0%) to 61.9% (6-11 months). Using participants’ self-reported long COVID as a criterion standard, existing published definitions had low-to-moderate sensitivity (up to 66.32% [95% CI, 62.59%-69.90%] at 3 months and 45.53% [95% CI, 41.51%-49.60%] at 6 months) and high specificity (up to 81.29% [95% CI, 79.32%-83.15%] at 3 months and 94.26% [95% CI, 92.98%-95.37%]) at 6 months.

Conclusions and Relevance

In this cohort study, variability in long COVID prevalence across published definitions highlights the need for a standardized, validated definition to improve clinical recognition and research comparability, ultimately guiding more accurate diagnosis and treatment strategies.

Introduction

There were more than 100 million cases of SARS-CoV-2 infection in the US during the COVID-19 pandemic, and many individuals report the continuation of symptoms or development of new symptoms after initial infection.¹ Long COVID, a condition characterized by prolonged symptoms, has been variably defined.^2,3,4,5,6 While major organizations, including the National Academies of Sciences, Engineering, and Medicine (NASEM)⁷ and others, have developed their own definitions of long COVID, there are discrepancies between the definitions,^8,9 and many studies use their own definitions. The lack of consensus on a long COVID definition, including variations in the duration and timing of symptoms,¹⁰ is problematic and can lead to underrecognition or overrecognition by both patients and clinicians, resulting in delayed or unnecessary treatments.¹¹ Additionally, the absence of a standardized definition (ie, true criterion standard) hinders the comparability of findings across studies.^8,12

It is crucial to establish a universally accepted definition of long COVID that is informed by comprehensive research and that is both sensitive and specific. Without a validated, objective diagnostic tool to identify long COVID, most studies have relied on patients’ self-reported symptoms or those documented in the medical record to classify this condition.¹³ Long COVID prevalence rates are often higher in studies that use patients’ self-reported symptoms compared with studies using patients’ medical records.¹⁴ Patients’ self-reported data, or patient-reported outcomes, are often essential to the verification and understanding of a disease¹⁵ and may provide a better picture of patients’ actual lived experience than using documentation in the medical records. We aimed to describe the variability in long COVID definitions from published reports and to use self-reported patient data from the Innovative Support for Patients With SARS-CoV-2 Infections Registry (INSPIRE) to estimate the prevalence of long COVID in our study’s cohort using the various definitions from the literature.

Methods

Study Design and Data Source

This cohort study used data from INSPIRE, a multisite, prospective, longitudinal registry that enrolled adults who experienced acute illness with symptoms suggestive of COVID-19 from December 11, 2020, through August 29, 2022. Eight study sites across the US recruited and enrolled participants as described previously.¹⁶ Follow-up surveys were collected through February 28, 2023. Each study site’s institutional review board approved this study. All INSPIRE participants provided consent electronically for their deidentified data to be used in future studies. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.¹⁷

Cohort Definition

INSPIRE included adult participants (aged ≥18 years) who were fluent in English or Spanish, had access to an internet-enabled device, and had self-reported symptoms suggestive of acute SARS-CoV-2 infection at the time of a SARS-CoV-2 test. Eligible individuals were tested for SARS-CoV-2 with a US Food and Drug Administration–approved or authorized molecular or antigen-based assay within 42 days prior to enrolling in the study.

Participants’ COVID-19 status (COVID-19 positive or COVID-19 negative) was based on their index SARS-CoV-2 test result at enrollment. If a participant had more than 1 test within 7 days of enrollment and had discordant test results, we considered the positive test to be the true result. If a participant’s test result changed after 7 days postenrollment in the study, we kept them in their initial COVID-19 status group.

Variables

INSPIRE participants completed surveys at enrollment and every 3 months for up to 18 months after enrollment. At enrollment (baseline), participants self-reported their sociodemographic data, including age, gender (female, male, transgender or nonbinary, or other [ie, not listed, prefer not to answer, and gender nonconfirming]), race (Asian, Black, White, and multiracial or other [included American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander, multiracial, or other responses listed in eAppendix 1 in Supplement 1]), ethnicity (Hispanic, Latino, or of Spanish origin or not Hispanic, Latino, or of Spanish origin), educational level, employment status, income, marital status, and health insurance status. Self-reported data on race and ethnicity were included in the analysis because these factors have been previously associated with long COVID prevalence.¹⁸ Surveys also included questions assessing physical and mental health, symptoms suggestive of COVID-19 (eTable 1 in Supplement 1), access to care, and work-related outcomes. Participants who completed the baseline survey and the 3-month follow-up were included in our analyses; participant inclusion and loss to follow-up are shown in eFigure 1 in Supplement 1. The 3-month and 6-month study time points were chosen because they most closely align with time of assessment of long COVID symptoms in the published literature.^5,6,19,20,21 For additional analysis, we used the self-reported long COVID status from the final survey (“Regardless of whether you tested positive or negative for SARS-CoV-2 infection when you had COVID-like symptoms, do you think you had or currently have long COVID?”) to create a subset of the full sample.

Literature Search

We conducted a literature search from March to October 2023. Published studies identified in PubMed and Google Scholar meeting the following criteria were considered: (1) researchers assessed participants’ self-reported symptoms after initial SARS-CoV-2 infection, (2) researchers provided the full list of symptoms assessed via participant self-report, (3) self-reported symptom data were used to calculate the percentage of participants with persistent symptoms after initial SARS-CoV-2 infection, and (4) studies were published in English (eAppendix 2 in Supplement 1). We chose to replicate the 5 published studies^5,6,19,20,21 identified before November 2023 (eTable 2 in Supplement 1) with the largest total sample size of all the studies identified in the literature search. Examples of studies^{22,23,24,25,26,27,28,29,30} identified in our search but not included in our analyses are summarized in eTable 3 in Supplement 1.

To assess feasibility of replication, we compared the full list of symptoms assessed in each of the published studies against symptoms we assessed in the INSPIRE study at both 3 and 6 months. The INSPIRE survey included 29 symptoms on a checklist (21 COVID-19–like symptoms and 8 fatigue symptoms). Three investigators (M.L., K.D.R., I.E.M.) all independently compared the full list of symptoms assessed in each published study against the full list of symptoms assessed in the INSPIRE study and reached 100% agreement on whether the INSPIRE study assessed the same or similar symptoms. Additionally, in the INSPIRE survey, participants who self-reported experiencing other symptoms not included in our checklist survey were prompted to write in the other symptoms they experienced. The eMethods in Supplement 1 has information on recategorization of free-text responses.

Statistical Analysis

Based on the definitions of long COVID used in the 5 comparator studies^5,6,19,20,21 (Table 1), we recoded the available symptoms from the INSPIRE survey to closely match each published study. We then calculated the prevalence of long COVID for INSPIRE cohort participants using these 5 definitions. These definitions were applied to both the COVID-19–positive and COVID-19–negative samples in the INSPIRE cohorts at 3 months and 6 months; in all cases, prevalences and associated 95% CIs were calculated. The prevalences were then tested for equality based on a Z test with the Yates continuity correction applied. Variation attributable to the definition used was calculated as the range in prevalence estimates (ie, maximum minus minimum) divided by either the minimum or maximum prevalence.

Table 1. Long COVID Definitions From the 5 Studies Used for INSPIRE Replication.

Source	Study location	Study period	Long COVID definition	Duration of symptoms (time point) used for INSPIRE replication	Symptoms required to classify participants with long COVID, No.	Total sample size, No.	Symptoms assessed per original study, No.
Jones et al,¹⁹ 2021	UK	August 7, 2020, to January 22, 2021	Self-diagnosed, clinician-diagnosed, or test-confirmed COVID-19 and had symptoms of COVID-19 that lasted >4 wk	≥4 wk	≥1	31 033	12
Pagen et al,⁶ 2023^a	The Netherlands	November 2021 to January 2022	Had ≥1 of the 44 prelisted symptoms at 3 mo, thus reflecting for each participant the same period after positive test (ie, 3 mo)	3-5 mo, 6-11 mo	≥1	9797	44
Sudre et al,²⁰ 2021	UK, US, and Sweden	March 24, 2020, to September 2, 2020	Symptoms that persisted >4 wk (28 d, LC28), >8 wk (56 d, LC56) or >12 wk (84 d, LC84) between onset and end	≥84 d	≥1	4182	14
Thaweethai et al,⁵ 2023	US and Puerto Rico	December 1, 2021, to April 10, 2023	Participants met the PASC score threshold (≥12 points); the symptoms (ordered by decreasing frequencies among participants with a qualifying PASC score) were PEM, fatigue, brain fog, dizziness, GI symptoms, palpitations, changes in sexual desire or capacity, loss of or change in smell or taste, thirst, chronic cough, chest pain, and abnormal movements	6 mo	≥1	9764	12
Yoo et al,²¹ 2022	US	April 2020 to February 2021	Persistent COVID-19 symptoms on the 90-d postdischarge survey (or the 60-d survey if the 90-d survey was incomplete)	≥60 d	≥1	1038	9

Open in a new tab

Abbreviations: GI, gastrointestinal; INSPIRE, Innovative Support for Patients With SARS-CoV-2 Infections Registry; LC, long COVID; PASC, postacute sequelae of SARS-CoV-2 infection; PEM, postexertional malaise.

^{^a}

Pagen et al presented 6 definitions of long-term symptoms, but we chose to replicate definition 6 because we also defined long COVID as participants having COVID-19 symptoms after 3 months.

We calculated the sensitivity and specificity of each definition by comparing it with participants’ self-reported long COVID status as the criterion standard, treating this patient-reported outcome as the reference category; this question did not distinguish between whether the respondent ever had or was currently experiencing the syndrome, and these responses were collected approximately 23 months after enrollment. The statistical analyses were conducted using SAS, version 9.4 (SAS Institute Inc), and R, version 4.2.2 (R Project for Statistical Computing).

Results

Of 6044 INSPIRE participants (4547 [75.2%] COVID-19 positive) who completed a baseline survey, 4575 (75.7%; 3521 [77.0%] COVID-19 positive) completed a 3-month follow-up survey and were included in this study (eFigure 1 in Supplement 1). Among the 4575 included participants, the mean (SD) age was 40.40 (14.58) years. Of 4448 participants with available gender data, 3013 (67.7%) were female; 1363 (30.6%) were male; and 72 (1.6%) were transgender, nonbinary, or other. Of 4446 with available race data, 613 (13.8%) were Asian; 365 (8.2%) were Black; 3064 (68.9%) were White; and 404 (9.1%) were multiracial or other race (inclusive of 26 [0.6%] who were American Indian or Alaska Native, 24 [0.5%] Native Hawaiian or Other Pacific Islander, 242 [5.4%] multiracial, and 112 [2.5%] other). Of 4491 with available ethnicity data, 639 (14.2%) were Hispanic, Latino, or of Spanish origin and 3852 (85.8%) were not Hispanic, Latino, or of Spanish origin. Of 4541 with age data, 3338 (73.5%) were aged 18 to 49 years (Table 2). Moreover, 3897 participants (85.2%; 2986 [76.6%] COVID-19 positive) completed both a 3-month and a 6-month follow-up survey, and 3235 (83.0%) of those (2478 [76.6%] COVID-19 positive) also completed the final survey (eFigure 1 in Supplement 1).

Table 2. Baseline Demographic Characteristics of Participants in the INSPIRE Study.

Characteristic	Participants^a
Characteristic	COVID-19 positive (n = 3521)	COVID-19 negative (n = 1054)	Total (N = 4575)
Age at enrollment, y
18-34	1433/3496 (41.0)	483/1045 (46.2)	1916/4541 (42.2)
35-49	1135/3496 (32.5)	287/1045 (27.5)	1422/4541 (31.3)
50-64	649/3496 (18.6)	187/1045 (17.9)	836/4541 (18.4)
≥65	279/3496 (8.0)	88/1045 (8.4)	367/4541 (8.1)
Missing	25	9	34
Gender
Female	2291/3428 (66.8)	722/1020 (70.8)	3013/4448 (67.7)
Male	1092/3428 (31.9)	271/1020 (26.6)	1363/4448 (30.6)
Transgender, nonbinary, or other^b	45/3428 (1.3)	27/1020 (2.6)	72/4448 (1.6)
Missing	93	34	127
Race^c
Asian	452/3431 (13.2)	161/1015 (15.9)	613/4446 (13.8)
Black	232/3431 (6.8)	133/1015 (13.1)	365/4446 (8.2)
White	2434/3431 (70.9)	630/1015 (62.1)	3064/4446 (68.9)
Multiracial or other	313/3431 (9.1)	91/1015 (9.0)	404/4446 (9.1)
Multiracial	182/3431 (5.3)	60/1015 (5.9)	242/4446 (5.4)
American Indian or Alaska Native	21/3431 (0.6)	5/1015 (0.5)	26/4446 (0.6)
Native Hawaiian or Other Pacific Islander	20/3431 (0.6)	4/1015 (0.4)	24/4446 (0.5)
Other^d	90/3431 (2.6)	22/1015 (2.2)	112/4446 (2.5)
Missing	90	39	129
Ethnicity
Hispanic, Latino, or of Spanish origin	465/3459 (13.4)	174/1032 (16.9)	639/4491 (14.2)
Not Hispanic, Latino, or of Spanish origin	2994/3459 (86.6)	858/1032 (83.1)	3852/4491 (85.8)
Missing	62	22	84
Educational level
<High school	28/3447 (0.8)	16/1022 (1.6)	44/4469 (1.0)
High school graduate	197/3447 (5.7)	111/1022 (10.9)	308/4469 (6.9)
Some college	451/3447 (13.1)	196/1022 (19.2)	647/4469 (14.5)
2-y Degree	229/3447 (6.6)	82/1022 (8.0)	311/4469 (7.0)
4-y Degree	1179/3447 (34.2)	255/1022 (25.0)	1434/4469 (32.1)
>4 y of College	1363/3447 (39.5)	362/1022 (35.4)	1725/4469 (38.6)
Missing	74	32	106
Annual family income prepandemic, $
<10 000	166 (4.7)	99 (9.4)	265 (5.8)
10 000-34 999	346 (9.8)	143 (13.6)	489 (10.7)
35 000-49 999	327 (9.3)	143 (13.6)	470 (10.3)
50 000-74 999	450 (12.8)	143 (13.6)	593 (13.0)
≥75 000	2008 (57.0)	428 (40.6)	2436 (53.2)
Unknown	224 (6.4)	98 (9.3)	322 (7.0)
Vaccination status^e
No	658/2773 (23.7)	155/842 (18.4)	813/3615 (22.5)
Yes	2115/2773 (76.3)	687/842 (81.6)	2802/3615 (77.5)
Missing	748	212	960
Self-reported long COVID status^f
No	1641/2309 (71.1)	529/691 (76.6)	2170/3000 (72.3)
Yes	668/2309 (28.9)	162/691 (23.4)	830/3000 (27.7)
Missing	1212	363	1575

Open in a new tab

Abbreviation: INSPIRE, Innovative Support for Patients With SARS-CoV-2 Infections Registry.

^{^a}

Data are presented as number (percentage) or number out of total number (percentage) of participants.

^{^b}

Other included not listed, prefer not to answer, and gender nonconfirming.

^{^c}

Participants could select multiple race options on the survey.

^{^d}

This race subcategory included responses entered manually by INSPIRE participants for “some other race.” eAppendix 1 in Supplement 1 includes a list of “some other race” responses entered.

^{^e}

COVID-19 vaccination indicates participants with at least 1 dose before the index SARS-CoV-2 test.

^{^f}

Self-reported long COVID status was an item in the final survey where participants could respond with any of the following options: (1) “no,” (2) “yes,” or (3) “I don’t know/am unsure.” We treated “I don’t know/am unsure” as a missing response in this analysis; 375 participants (10.7%) from the COVID-19 positive and 125 (11.9%) from the COVID-19 negative groups responded “I don’t know/am unsure.”

The 5 comparator studies^5,6,19,20,21 each used distinct long COVID definitions, and the studies were conducted in different periods (Table 1 and eFigure 2 in Supplement 1). The number of symptoms from comparator studies that were matched to INSPIRE symptoms varied from 8 to 34 (Thaweethai et al,⁵ 8; Yoo et al,²¹ 9; Jones et al,¹⁹ 12; Sudre et al,²⁰ 12; Pagen et al,⁶ 34), with some potential for one-to-many (vs one-to-one) matching when symptom categories were broadly specified. There were 5 symptoms out of the 29 from the INSPIRE questionnaire (21 COVID-19–like symptoms and 8 fatigue symptoms) that were present in all 5 of the comparator studies (“more tired than usual,” “pain or tightness in your chest,” “diarrhea [>3 loose or looser than normal stools in 24 hours],” “decreased smell or change in smell,” and “fatigue, tiredness, or exhaustion”) (eTable 4 in Supplement 1).

Of the 5 comparator studies, reported prevalence of long COVID at 1 to 5 months postinfection ranged from 2.6% (≥84 days [Sudre et al²⁰]) to 47.4% (3-5 months [Pagen et al⁶]) and at 6 or more months ranged from 10.0% (95% CI, 8.8%-11.0% [Thaweethai et al⁵]) to 61.9% (6-11 months [Pagen et al⁶]). Applying these studies’ long COVID definitions to our INSPIRE cohort yielded a 3-month prevalence of long COVID that ranged from 30.84% (95% CI, 29.33%-32.40%) to 42.01% (95% CI, 40.37%-43.66%) among the COVID-19–positive group and from 28.08% (95% CI, 25.41%-30.92%) to 40.32% (95% CI, 37.35%-43.36%) among the COVID-19–negative group (Figure). We attributed 24.8% to 32.9% of the variation in long COVID prevalence to differences in the definitions used. Similarly, we calculated a 6-month prevalence of long COVID that ranged from 14.23% (95% CI, 13.01%-15.55%) to 21.94% (95% CI, 20.47%-23.47%) among the COVID-19–positive group and from 14.60% (95% CI, 12.40%-17.10%) to 23.27% (95% CI, 20.59%-26.18%) among the COVID-19–negative group. Despite similar prevalence estimates for long COVID among the cohorts that were COVID-19 positive vs COVID-19 negative (eTable 5 in Supplement 1), we identified a greater absolute number of symptoms for individuals who were COVID-19 positive who met various definitions for long COVID compared with individuals who were COVID-19 negative (eg, mean [SD] long COVID symptom count of 2.96 [2.8] among those who were COVID-19 positive vs 2.36 [2.8] among those who were COVID-19 negative per the definition of long COVID by Yoo et al²¹) (eTable 6 and eFigure 3 in Supplement 1).

Figure. — The long COVID definition from each study was applied to eligible Innovative Support for Patients With SARS-CoV-2 Infections Registry (INSPIRE) participants who completed the surveys at 3 months and/or 6 months. The INSPIRE cohort eligible for the 3-month follow-up survey had 4575 participants and for both the 3- and 6-month surveys, 3897 participants. Prevalences were calculated based on persistence of 1 or more symptoms (constituting source definitions) at baseline and 3 months (for 3-month prevalence) and at baseline, 3 months, and 6 months (for 6-month prevalence). Error bars indicate 95% CIs.

^aPagen et al⁶ presented 6 definitions of long-term symptoms, but we chose to replicate definition 6 because we also defined *long COVID* as participants having COVID-19 symptoms after 3 months.

Considering the INSPIRE participants’ self-reported long COVID status as the reference category, our analyses revealed that the highest sensitivity for this patient-reported outcome was observed with the definition by Pagen et al⁶ (66.32% [95% CI, 62.59%-69.90%] at 3 months; 45.53% [95% CI, 41.51%-49.60%] at 6 months), and the highest specificity was observed with the definitions by Yoo et al²¹ (81.29% [95% CI, 79.32%-83.15%] at 3 months) and Thaweethai et al⁵ (94.26% [95% CI, 92.98%-95.37%] at 6 months) among the cohort that was COVID-19 positive (Table 3 and eTable 7 in Supplement 1). Overall, both the sensitivity and specificity of all assessed definitions were low to moderate, with sensitivity lower than specificity for all definitions.

Table 3. Sensitivity and Specificity When Applying Definitions of Long COVID Used in the 5 Studies to the INSPIRE Participants Who Were COVID-19 Positive in the Index Test^a.

Source study for long COVID definition	3-mo Follow-up		6-mo Follow-up
Source study for long COVID definition	Sensitivity, % (95% CI)	Specificity, % (95% CI)	Sensitivity, % (95% CI)	Specificity, % (95% CI)
Jones et al,¹⁹ 2021	59.28 (55.45-63.03)	78.43 (76.36-80.40)	39.07 (35.16-43.09)	93.02 (91.63-94.24)
Pagen et al,⁶ 2023	66.32 (62.59-69.90)	70.57 (68.30-72.76)	45.53 (41.51-49.60)	88.52 (86.82-90.07)
Sudre et al,²⁰ 2021	59.13 (55.30-62.89)	77.27 (75.16-79.28)	39.07 (35.16-43.09)	93.09 (91.70-94.31)
Thaweethai et al,⁵ 2023	54.04 (50.18-57.87)	81.05 (79.07-82.92)	31.79 (28.09-35.67)	94.26 (92.98-95.37)
Yoo et al,²¹ 2022	55.54 (51.68-59.35)	81.29 (79.32-83.15)	35.26 (31.45-39.22)	93.93 (92.62-95.08)

Open in a new tab

Abbreviation: INSPIRE, Innovative Support for Patients With SARS-CoV-2 Infections Registry.

^{^a}

The sensitivity and specificity of prevalence estimates were calculated by considering as the criterion standard the self-reported response of INSPIRE participants who tested positive for COVID-19 in the index test to the final survey (which asked, “Regardless of whether you tested positive or negative for SARS-CoV-2 infection when you had COVID-like symptoms, do you think you had or currently have long COVID?”). Six-month results indicate results from participants who completed both 3- and 6-month surveys.

Discussion

We found that researchers’ decision to use any of a range of published long COVID definitions could result in considerable variation in the prevalence estimate obtained from their cohort. Furthermore, we identified that most definitions had less than optimal sensitivity and specificity for identifying subjectively reported long COVID, indicating that standard definitions do not align with patient-reported experience of long COVID. In addition, we found that the most relevant distinguishing feature between potential long COVID cases (ie, among COVID-19–positive cases) vs other postillness conditions (ie, among COVID-19–negative cases) was the overall symptom count, with more symptoms present in the long COVID cases.

NASEM released a 2024 long COVID definition and encourages the federal government, clinicians, and others to adopt this new definition to aid in consistent diagnosis, documentation, and treatment.⁷ As the report rightfully acknowledges, the lack of consensus on the definition of long COVID hinders accurate comparison across long COVID studies⁹ and may lead to challenges in identification and evaluation of subsequent treatment.¹¹ The suggested NASEM definition, however, is extremely broad (presence of any of hundreds of symptoms for at least 3 months as a continuous or relapsing and remitting symptom) and does not require laboratory confirmation or other proof of the initial infection.⁷ As our study only assessed 29 checklist symptoms with an “other” (free-text) option (substantially fewer symptoms than NASEM suggests), we were not able to report results using that operationalization. Given the broad NASEM definition, its specificity will likely be low with substantial false-positive results.³¹

Differences in the list of symptoms used for defining long COVID accounted for 24.8% to 32.9% of the variation in published prevalence rates among the 5 studies we assessed. The long COVID definitions from the published literature that we evaluated included from 9²¹ up to 44⁶ total symptoms, whereas the NASEM definition is far broader³¹ and is likely to lead to increased prevalence and improved sensitivity at the expense of misclassification from increased false-positives. While this may be considered a sensible tradeoff given that we found that the evaluated definitions generally performed better in terms of specificity rather than sensitivity, the goals of such a decision need to be carefully considered to determine which test characteristic should be maximized. A large number of symptoms to assess may overwhelm clinicians; thus, the ease of use of the definition in clinical practice should also be considered. Based on experiences using lists of symptoms for diagnosis of other diseases without clear-cut biologic markers, like Behçet or Still disease,³² the lists are more easily used in clinical practice when they are parsimonious and targeted or if there are clear clinical markers (which is not the case with long COVID³³). If invasive interventions are developed for treating long COVID, greater specificity will be important to avoid overtreatment as well as inappropriately labeling individuals with a diagnosis that may affect their approach to life.

Of note, the most distinguishing feature of INSPIRE participants who were COVID-19 positive and were identified as having long COVID was not simply the presence of symptoms at 3 months postinfection but the number of symptoms present; individuals who were COVID-19 positive tended to have a greater mean number of reported symptoms than did their counterparts who were COVID-19 negative but who otherwise would have met the definition for long COVID. Though we did not compare severity or functional impairment (consistent with the 2024 NASEM recommendations), modifying definitions to include a minimum number of discrete symptoms may improve the ability of such an algorithm to uniquely identify long COVID among those with a history of COVID-19. The long COVID definition will undoubtedly evolve over time as new evidence and understanding emerge.

In addition, our finding that up to one-third of the variation in reported prevalence of long COVID was likely due to the varying definitions in use suggests that the remaining two-thirds of variation may be due to differences in sample characteristics (eg, higher prevalence rates observed in studies of hospitalized patients compared with patients recruited from the general population) or potentially due to the timing of the study (Figure) and, thus, due to different SARS-CoV-2 variants or availability of vaccines.³⁴ Future work is needed to fully evaluate the extent to which certain patient characteristics, variants, and vaccination are responsible for making some subpopulations susceptible to experiencing long COVID.³⁵ This will be critically important for addressing equity issues in the diagnosis and treatment of long COVID moving forward.

Strengths and Limitations

Our study has several strengths. First, we included a comparison group comprising participants who were symptomatic but tested negative for SARS-CoV-2.³⁶ This comparison group allowed us to calculate rates of long COVID in a cohort that was COVID-19 negative, thus providing data on potential misclassification when using these definitions in a cohort without a confirmed positive SARS-CoV-2 test—especially relevant given that the 2024 NASEM⁷ definition does not require evidence of positive testing for SARS-CoV-2. We additionally evaluated the long COVID prevalence using a shorter time frame for assessment (3 months, as suggested by NASEM⁷) vs longer (6 months), highlighting not only the extent to which prevalence rates decline over time but also how the ability of various definitions for distinguishing long COVID from other postacute syndromes declines as well. In addition, our use of a geographically diverse cohort with representation across several COVID-19 variants (ie, pre-Delta, Delta, and Omicron) supports the generalizability of these findings regarding variant.

Our study also has several limitations. First, participants may be subject to recall bias when self-reporting symptoms,³⁷ and we did not examine participants’ electronic health records to see if symptoms were also documented in their medical records. However, some symptoms, such as fatigue, impaired smell and taste, and chest pain, are more likely to be reported in self-reported surveys than in health care records.³⁸ Patients’ self-reported symptoms still provide important insights into how patients’ daily activities and functioning are affected,³⁸ and their use may mitigate potential biases introduced by incomplete or inconsistent documentation in a medical record.³⁹ Second, we did not exclude patients (either COVID-19 positive or COVID-19 negative) who had a SARS-CoV-2 infection after their baseline survey; thus, we may have miscalculated the prevalence of long COVID after a single infection. Similarly, our measure of self-reported long COVID included the experience of this condition at any point up to 23 months after the initial infection and so may lead to misestimation of self-reported long COVID after a single infection. Third, we did not examine whether participant demographics, such as race and ethnicity, were associated with prevalence estimates for each long COVID definition, though prior work by our group has identified similar symptom reporting across racial and ethnic groups.⁴⁰ Fourth, we only used the long COVID definitions from 5 published studies; results may have varied if we used different definitions from other published studies. Moreover, we did not achieve perfect alignment of symptoms assessed within our cohort and the 5 studies we evaluated, which may have led to an underestimation of prevalence in our sample.

Conclusions

In this cohort study of long COVID prevalence, variability in prevalence across 5 different published definitions highlighted the need for a standardized, validated definition to improve clinical recognition and research comparability, ultimately guiding more accurate diagnosis and treatment strategies. We described significant variation in reported long COVID definitions and prevalence rates and provided data on sensitivity and specificity of these different definitions in the study’s INSPIRE cohort. Some definitions outperformed others regarding uniquely identifying patient-reported long COVID in a sample that was COVID-19 positive compared with COVID-19 negative, but differences were generally small. Future epidemiologic investigations into long COVID should seek to apply a consistent epidemiologic definition, which likely will differ from a broader clinical definition because of the subjective nature of many long COVID symptoms and the broader uses of clinical definitions (eg, for insurance decisions and reimbursement). A more parsimonious list of symptoms for defining long COVID epidemiologically, or requiring that a minimum of 2 symptoms be reported, may add needed specificity, which is important for informing which patients to include in intervention trials and in studies of pathogenesis. Further definition refinements ideally would be based on an understanding of the pathophysiologic mechanisms underlying long COVID; without such insights, a single definition for clinical and epidemiologic use (ie, for research and public health) is unlikely.

Supplement 1.

eAppendix 1. List of “Other” Races Entered by INSPIRE Participants

eTable 1. Symptoms Assessed in the INSPIRE Survey

eFigure 1. INSPIRE Participant Flow Diagram

eAppendix 2. Search Criteria Used in PubMed and Google Scholar

eTable 2. Long COVID Prevalence Rate Studies Used for INSPIRE Replication

eTable 3. A Summary of 9 Published Studies That We Did Not Replicate

eMethods. Our Methods for Recoding INSPIRE Participants’ “Other” Symptoms

eFigure 2. Timeline of Data Collection for Studies Used for INSPIRE Replication

eTable 4. Symptoms Assessed in the INSPIRE Study Compared to the Five Studies

eTable 5. Comparison of Long COVID Prevalence From the Studies With INSPIRE COVID Cohorts

eTable 6. Symptom Count by Long COVID Definition

eFigure 3. Mean Symptom Count by Long COVID Definition at 3 Months and 6 Months

eTable 7. 2 × 2 Tables Comparing Long COVID Counts From the INSPIRE Final Survey (Participant Self-Report) vs Counts After Applying the Long COVID From 5 Studies

eReferences

jamanetwopen-e2526506-s001.pdf^{(732.4KB, pdf)}

Supplement 2.

INSPIRE Group members

jamanetwopen-e2526506-s002.pdf^{(164.3KB, pdf)}

Supplement 3.

Data Sharing Statement

jamanetwopen-e2526506-s003.pdf^{(13.1KB, pdf)}

References

1.World Health Organization . COVID-19 cases, world. Accessed March 20, 2024. https://data.who.int/dashboards/covid19/cases
2.Centers for Disease Control and Prevention . Long COVID basics. Updated February 3, 2025. Accessed July 24, 2024. https://www.cdc.gov/covid/long-term-effects/
3.National Institute for Health and Care Excellence . COVID-19 rapid guideline: managing the long-term effects of COVID-19. December 18, 2020. Updated January 25, 2024. Accessed July 24, 2024. https://www.nice.org.uk/guidance/ng188/chapter/1-Identification#case-definition [PubMed]
4.Soriano JB, Murthy S, Marshall JC, Relan P, Diaz JV; WHO Clinical Case Definition Working Group on Post-COVID-19 Condition . A clinical case definition of post-COVID-19 condition by a Delphi consensus. Lancet Infect Dis. 2022;22(4):e102-e107. doi: 10.1016/S1473-3099(21)00703-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Thaweethai T, Jolley SE, Karlson EW, et al. ; RECOVER Consortium . Development of a definition of postacute sequelae of SARS-CoV-2 infection. JAMA. 2023;329(22):1934-1946. doi: 10.1001/jama.2023.8823 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Pagen DME, van Bilsen CJA, Brinkhues S, et al. Prevalence of long-term symptoms varies when using different post-COVID-19 definitions in positively and negatively tested adults: the PRIME post-COVID study. Open Forum Infect Dis. 2023;10(10):ofad471. doi: 10.1093/ofid/ofad471 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.National Academies of Sciences, Engineering, and Medicine. A Long COVID Definition: A Chronic, Systemic Disease State with Profound Consequences. The National Academies Press; 2024. [PubMed] [Google Scholar]
8.Chaichana U, Man KKC, Chen A, et al. Definition of post-COVID-19 condition among published research studies. JAMA Netw Open. 2023;6(4):e235856. doi: 10.1001/jamanetworkopen.2023.5856 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Pan D, Pareek M. Toward a universal definition of post-COVID-19 condition—how do we proceed? JAMA Netw Open. 2023;6(4):e235779. doi: 10.1001/jamanetworkopen.2023.5779 [DOI] [PubMed] [Google Scholar]
10.Chou R, Herman E, Ahmed A, et al. Long COVID definitions and models of care: a scoping review. Ann Intern Med. 2024;177(7):929-940. doi: 10.7326/M24-0677 [DOI] [PubMed] [Google Scholar]
11.Srikanth S, Boulos JR, Dover T, Boccuto L, Dean D. Identification and diagnosis of long COVID-19: a scoping review. Prog Biophys Mol Biol. 2023;182:1-7. doi: 10.1016/j.pbiomolbio.2023.04.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Wisk LE, Nichol G, Elmore JG. Toward unbiased evaluation of postacute sequelae of SARS-CoV-2 infection: challenges and solutions for the long haul ahead. Ann Intern Med. 2022;175(5):740-743. doi: 10.7326/M21-4664 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Krysa JA, Buell M, Pohar Manhas K, et al. Understanding the experience of long COVID symptoms in hospitalized and non-hospitalized individuals: a random, cross-sectional survey study. Healthcare (Basel). 2023;11(9):1309. doi: 10.3390/healthcare11091309 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Woodrow M, Carey C, Ziauddeen N, et al. Systematic review of the prevalence of long COVID. Open Forum Infect Dis. 2023;10(7):ofad233. doi: 10.1093/ofid/ofad233 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.St Clair P, Gaudette É, Zhao H, Tysinger B, Seyedin R, Goldman DP. Using self-reports or claims to assess disease prevalence: it’s complicated. Med Care. 2017;55(8):782-788. doi: 10.1097/MLR.0000000000000753 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.O’Laughlin KN, Thompson M, Hota B, et al. ; INSPIRE Investigators . Study protocol for the Innovative Support for Patients With SARS-COV-2 Infections Registry (INSPIRE): a longitudinal study of the medium and long-term sequelae of SARS-CoV-2 infection. PLoS One. 2022;17(3):e0264260. doi: 10.1371/journal.pone.0264260 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Int J Surg. 2014;12(12):1495-1499. doi: 10.1016/j.ijsu.2014.07.013 [DOI] [PubMed] [Google Scholar]
18.Jacobs MM, Evans E, Ellis C. Racial, ethnic, and sex disparities in the incidence and cognitive symptomology of long COVID-19. J Natl Med Assoc. 2023;115(2):233-243. doi: 10.1016/j.jnma.2023.01.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Jones R, Davis A, Stanley B, et al. Risk predictors and symptom features of long COVID within a broad primary care patient population including both tested and untested patients. Pragmat Obs Res. 2021;12:93-104. doi: 10.2147/POR.S316186 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Sudre CH, Murray B, Varsavsky T, et al. Attributes and predictors of long COVID. Nat Med. 2021;27(4):626-631. doi: 10.1038/s41591-021-01292-y [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Yoo SM, Liu TC, Motwani Y, et al. Factors associated with post-acute sequelae of SARS-CoV-2 (PASC) after diagnosis of symptomatic COVID-19 in the inpatient and outpatient setting in a diverse cohort. J Gen Intern Med. 2022;37(8):1988-1995. doi: 10.1007/s11606-022-07523-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Augustin M, Schommers P, Stecher M, et al. Post-COVID syndrome in non-hospitalised patients with COVID-19: a longitudinal prospective cohort study. Lancet Reg Health Eur. 2021;6:100122. doi: 10.1016/j.lanepe.2021.100122 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Carvalho-Schneider C, Laurent E, Lemaignen A, et al. Follow-up of adults with noncritical COVID-19 two months after symptom onset. Clin Microbiol Infect. 2021;27(2):258-263. doi: 10.1016/j.cmi.2020.09.052 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Desgranges F, Tadini E, Munting A, et al. ; the RegCOVID Research Group . Post-COVID-19 syndrome in outpatients: a cohort study. J Gen Intern Med. 2022;37(8):1943-1952. doi: 10.1007/s11606-021-07242-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Klein H, Asseo K, Karni N, et al. Onset, duration and unresolved symptoms, including smell and taste changes, in mild COVID-19 infection: a cohort study in Israeli patients. Clin Microbiol Infect. 2021;27(5):769-774. doi: 10.1016/j.cmi.2021.02.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Lapa J, Rosa D, Mendes JPL, Deusdará R, Romero GAS. Prevalence and associated factors of post-COVID-19 syndrome in a Brazilian cohort after 3 and 6 months of hospital discharge. Int J Environ Res Public Health. 2023;20(1):848. doi: 10.3390/ijerph20010848 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Morioka S, Tsuzuki S, Suzuki M, et al. Post COVID-19 condition of the Omicron variant of SARS-CoV-2. J Infect Chemother. 2022;28(11):1546-1551. doi: 10.1016/j.jiac.2022.08.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Peghin M, Palese A, Venturini M, et al. Post-COVID-19 symptoms 6 months after acute infection among hospitalized and non-hospitalized patients. Clin Microbiol Infect. 2021;27(10):1507-1513. doi: 10.1016/j.cmi.2021.05.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Román-Montes CM, Flores-Soto Y, Guaracha-Basañez GA, et al. Post-COVID-19 syndrome and quality of life impairment in severe COVID-19 Mexican patients. Front Public Health. 2023;11:1155951. doi: 10.3389/fpubh.2023.1155951 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Kisiel MA, Janols H, Nordqvist T, et al. Predictors of post-COVID-19 and the impact of persistent symptoms in non-hospitalized patients 12 months after COVID-19, with a focus on work ability. Ups J Med Sci. 2022;127. doi: 10.48101/ujms.v127.8794 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Jason L. Here’s what is wrong with the National Academies’ long COVID definition—overly broad criteria won’t help patients in the long run. MedPage Today. June 14, 2024. Accessed June 25, 2024. https://www.medpagetoday.com/opinion/second-opinions/110658?xid=nl_secondopinion_2024-06-16&eun=g1268782d0r
32.Feist E, Mitrovic S, Fautrel B. Mechanisms, biomarkers and targets for adult-onset Still’s disease. Nat Rev Rheumatol. 2018;14(10):603-618. doi: 10.1038/s41584-018-0081-x [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Erlandson KM, Geng LN, Selvaggi CA, et al. ; RECOVER-Adult Cohort . Differentiation of prior SARS-CoV-2 infection and postacute sequelae by standard clinical laboratory measurements in the RECOVER cohort. Ann Intern Med. 2024;177(9):1209-1221. doi: 10.7326/M24-0737 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Gottlieb M, Wang RC, Yu H, et al. ; Innovative Support for Patients With SARS-CoV-2 Infections Registry (INSPIRE) Group . Severe fatigue and persistent symptoms at 3 months following severe acute respiratory syndrome coronavirus 2 infections during the pre-Delta, Delta, and Omicron time periods: a multicenter prospective cohort study. Clin Infect Dis. 2023;76(11):1930-1941. doi: 10.1093/cid/ciad045 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Fang Z, Ahrnsbrak R, Rekito A. Evidence mounts that about 7% of US adults have had long COVID. JAMA. 2024;332(1):5-6. doi: 10.1001/jama.2024.11370 [DOI] [PubMed] [Google Scholar]
36.Wisk LE, Gottlieb MA, Spatz ES, et al. ; INSPIRE Group . Association of initial SARS-CoV-2 test positivity with patient-reported well-being 3 months after a symptomatic illness. JAMA Netw Open. 2022;5(12):e2244486. doi: 10.1001/jamanetworkopen.2022.44486 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Gottlieb M, Spatz ES, Yu H, et al. ; INSPIRE Group . Long COVID clinical phenotypes up to 6 months after infection identified by latent class analysis of self-reported symptoms. Open Forum Infect Dis. 2023;10(7):ofad277. doi: 10.1093/ofid/ofad277 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Kisiel MA, Kock C, Sultan J, Janols H, Janson C, Pingel R. Agreement between reported questionnaire data and medical records on diagnosis and COVID-19 symptoms at onset. Eur Clin Respir J. 2023;10(1). doi: 10.1080/20018525.2023.2282251 [DOI] [Google Scholar]
39.Menachemi N, Collum TH. Benefits and drawbacks of electronic health record systems. Risk Manag Healthc Policy. 2011;4:47-55. doi: 10.2147/RMHP.S12985 [DOI] [PMC free article] [PubMed] [Google Scholar]
40.O’Laughlin KN, Klabbers RE, Ebna Mannan I, et al. ; INSPIRE Group . Ethnic and racial differences in self-reported symptoms, health status, activity level, and missed work at 3 and 6 months following SARS-CoV-2 infection. Front Public Health. 2024;11:1324636. doi: 10.3389/fpubh.2023.1324636 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

eAppendix 1. List of “Other” Races Entered by INSPIRE Participants

eTable 1. Symptoms Assessed in the INSPIRE Survey

eFigure 1. INSPIRE Participant Flow Diagram

eAppendix 2. Search Criteria Used in PubMed and Google Scholar

eTable 2. Long COVID Prevalence Rate Studies Used for INSPIRE Replication

eTable 3. A Summary of 9 Published Studies That We Did Not Replicate

eMethods. Our Methods for Recoding INSPIRE Participants’ “Other” Symptoms

eFigure 2. Timeline of Data Collection for Studies Used for INSPIRE Replication

eTable 4. Symptoms Assessed in the INSPIRE Study Compared to the Five Studies

eTable 5. Comparison of Long COVID Prevalence From the Studies With INSPIRE COVID Cohorts

eTable 6. Symptom Count by Long COVID Definition

eFigure 3. Mean Symptom Count by Long COVID Definition at 3 Months and 6 Months

eTable 7. 2 × 2 Tables Comparing Long COVID Counts From the INSPIRE Final Survey (Participant Self-Report) vs Counts After Applying the Long COVID From 5 Studies

eReferences

jamanetwopen-e2526506-s001.pdf^{(732.4KB, pdf)}

Supplement 2.

INSPIRE Group members

jamanetwopen-e2526506-s002.pdf^{(164.3KB, pdf)}

Supplement 3.

Data Sharing Statement

jamanetwopen-e2526506-s003.pdf^{(13.1KB, pdf)}

[zoi250746r1] 1.World Health Organization . COVID-19 cases, world. Accessed March 20, 2024. https://data.who.int/dashboards/covid19/cases

[zoi250746r2] 2.Centers for Disease Control and Prevention . Long COVID basics. Updated February 3, 2025. Accessed July 24, 2024. https://www.cdc.gov/covid/long-term-effects/

[zoi250746r3] 3.National Institute for Health and Care Excellence . COVID-19 rapid guideline: managing the long-term effects of COVID-19. December 18, 2020. Updated January 25, 2024. Accessed July 24, 2024. https://www.nice.org.uk/guidance/ng188/chapter/1-Identification#case-definition [PubMed]

[zoi250746r4] 4.Soriano JB, Murthy S, Marshall JC, Relan P, Diaz JV; WHO Clinical Case Definition Working Group on Post-COVID-19 Condition . A clinical case definition of post-COVID-19 condition by a Delphi consensus. Lancet Infect Dis. 2022;22(4):e102-e107. doi: 10.1016/S1473-3099(21)00703-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r5] 5.Thaweethai T, Jolley SE, Karlson EW, et al. ; RECOVER Consortium . Development of a definition of postacute sequelae of SARS-CoV-2 infection. JAMA. 2023;329(22):1934-1946. doi: 10.1001/jama.2023.8823 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r6] 6.Pagen DME, van Bilsen CJA, Brinkhues S, et al. Prevalence of long-term symptoms varies when using different post-COVID-19 definitions in positively and negatively tested adults: the PRIME post-COVID study. Open Forum Infect Dis. 2023;10(10):ofad471. doi: 10.1093/ofid/ofad471 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r7] 7.National Academies of Sciences, Engineering, and Medicine. A Long COVID Definition: A Chronic, Systemic Disease State with Profound Consequences. The National Academies Press; 2024. [PubMed] [Google Scholar]

[zoi250746r8] 8.Chaichana U, Man KKC, Chen A, et al. Definition of post-COVID-19 condition among published research studies. JAMA Netw Open. 2023;6(4):e235856. doi: 10.1001/jamanetworkopen.2023.5856 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r9] 9.Pan D, Pareek M. Toward a universal definition of post-COVID-19 condition—how do we proceed? JAMA Netw Open. 2023;6(4):e235779. doi: 10.1001/jamanetworkopen.2023.5779 [DOI] [PubMed] [Google Scholar]

[zoi250746r10] 10.Chou R, Herman E, Ahmed A, et al. Long COVID definitions and models of care: a scoping review. Ann Intern Med. 2024;177(7):929-940. doi: 10.7326/M24-0677 [DOI] [PubMed] [Google Scholar]

[zoi250746r11] 11.Srikanth S, Boulos JR, Dover T, Boccuto L, Dean D. Identification and diagnosis of long COVID-19: a scoping review. Prog Biophys Mol Biol. 2023;182:1-7. doi: 10.1016/j.pbiomolbio.2023.04.008 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r12] 12.Wisk LE, Nichol G, Elmore JG. Toward unbiased evaluation of postacute sequelae of SARS-CoV-2 infection: challenges and solutions for the long haul ahead. Ann Intern Med. 2022;175(5):740-743. doi: 10.7326/M21-4664 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r13] 13.Krysa JA, Buell M, Pohar Manhas K, et al. Understanding the experience of long COVID symptoms in hospitalized and non-hospitalized individuals: a random, cross-sectional survey study. Healthcare (Basel). 2023;11(9):1309. doi: 10.3390/healthcare11091309 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r14] 14.Woodrow M, Carey C, Ziauddeen N, et al. Systematic review of the prevalence of long COVID. Open Forum Infect Dis. 2023;10(7):ofad233. doi: 10.1093/ofid/ofad233 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r15] 15.St Clair P, Gaudette É, Zhao H, Tysinger B, Seyedin R, Goldman DP. Using self-reports or claims to assess disease prevalence: it’s complicated. Med Care. 2017;55(8):782-788. doi: 10.1097/MLR.0000000000000753 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r16] 16.O’Laughlin KN, Thompson M, Hota B, et al. ; INSPIRE Investigators . Study protocol for the Innovative Support for Patients With SARS-COV-2 Infections Registry (INSPIRE): a longitudinal study of the medium and long-term sequelae of SARS-CoV-2 infection. PLoS One. 2022;17(3):e0264260. doi: 10.1371/journal.pone.0264260 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r17] 17.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Int J Surg. 2014;12(12):1495-1499. doi: 10.1016/j.ijsu.2014.07.013 [DOI] [PubMed] [Google Scholar]

[zoi250746r18] 18.Jacobs MM, Evans E, Ellis C. Racial, ethnic, and sex disparities in the incidence and cognitive symptomology of long COVID-19. J Natl Med Assoc. 2023;115(2):233-243. doi: 10.1016/j.jnma.2023.01.016 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r19] 19.Jones R, Davis A, Stanley B, et al. Risk predictors and symptom features of long COVID within a broad primary care patient population including both tested and untested patients. Pragmat Obs Res. 2021;12:93-104. doi: 10.2147/POR.S316186 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r20] 20.Sudre CH, Murray B, Varsavsky T, et al. Attributes and predictors of long COVID. Nat Med. 2021;27(4):626-631. doi: 10.1038/s41591-021-01292-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r21] 21.Yoo SM, Liu TC, Motwani Y, et al. Factors associated with post-acute sequelae of SARS-CoV-2 (PASC) after diagnosis of symptomatic COVID-19 in the inpatient and outpatient setting in a diverse cohort. J Gen Intern Med. 2022;37(8):1988-1995. doi: 10.1007/s11606-022-07523-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r22] 22.Augustin M, Schommers P, Stecher M, et al. Post-COVID syndrome in non-hospitalised patients with COVID-19: a longitudinal prospective cohort study. Lancet Reg Health Eur. 2021;6:100122. doi: 10.1016/j.lanepe.2021.100122 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r23] 23.Carvalho-Schneider C, Laurent E, Lemaignen A, et al. Follow-up of adults with noncritical COVID-19 two months after symptom onset. Clin Microbiol Infect. 2021;27(2):258-263. doi: 10.1016/j.cmi.2020.09.052 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r24] 24.Desgranges F, Tadini E, Munting A, et al. ; the RegCOVID Research Group . Post-COVID-19 syndrome in outpatients: a cohort study. J Gen Intern Med. 2022;37(8):1943-1952. doi: 10.1007/s11606-021-07242-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r25] 25.Klein H, Asseo K, Karni N, et al. Onset, duration and unresolved symptoms, including smell and taste changes, in mild COVID-19 infection: a cohort study in Israeli patients. Clin Microbiol Infect. 2021;27(5):769-774. doi: 10.1016/j.cmi.2021.02.008 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r26] 26.Lapa J, Rosa D, Mendes JPL, Deusdará R, Romero GAS. Prevalence and associated factors of post-COVID-19 syndrome in a Brazilian cohort after 3 and 6 months of hospital discharge. Int J Environ Res Public Health. 2023;20(1):848. doi: 10.3390/ijerph20010848 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r27] 27.Morioka S, Tsuzuki S, Suzuki M, et al. Post COVID-19 condition of the Omicron variant of SARS-CoV-2. J Infect Chemother. 2022;28(11):1546-1551. doi: 10.1016/j.jiac.2022.08.007 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r28] 28.Peghin M, Palese A, Venturini M, et al. Post-COVID-19 symptoms 6 months after acute infection among hospitalized and non-hospitalized patients. Clin Microbiol Infect. 2021;27(10):1507-1513. doi: 10.1016/j.cmi.2021.05.033 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r29] 29.Román-Montes CM, Flores-Soto Y, Guaracha-Basañez GA, et al. Post-COVID-19 syndrome and quality of life impairment in severe COVID-19 Mexican patients. Front Public Health. 2023;11:1155951. doi: 10.3389/fpubh.2023.1155951 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r30] 30.Kisiel MA, Janols H, Nordqvist T, et al. Predictors of post-COVID-19 and the impact of persistent symptoms in non-hospitalized patients 12 months after COVID-19, with a focus on work ability. Ups J Med Sci. 2022;127. doi: 10.48101/ujms.v127.8794 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r31] 31.Jason L. Here’s what is wrong with the National Academies’ long COVID definition—overly broad criteria won’t help patients in the long run. MedPage Today. June 14, 2024. Accessed June 25, 2024. https://www.medpagetoday.com/opinion/second-opinions/110658?xid=nl_secondopinion_2024-06-16&eun=g1268782d0r

[zoi250746r32] 32.Feist E, Mitrovic S, Fautrel B. Mechanisms, biomarkers and targets for adult-onset Still’s disease. Nat Rev Rheumatol. 2018;14(10):603-618. doi: 10.1038/s41584-018-0081-x [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r33] 33.Erlandson KM, Geng LN, Selvaggi CA, et al. ; RECOVER-Adult Cohort . Differentiation of prior SARS-CoV-2 infection and postacute sequelae by standard clinical laboratory measurements in the RECOVER cohort. Ann Intern Med. 2024;177(9):1209-1221. doi: 10.7326/M24-0737 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r34] 34.Gottlieb M, Wang RC, Yu H, et al. ; Innovative Support for Patients With SARS-CoV-2 Infections Registry (INSPIRE) Group . Severe fatigue and persistent symptoms at 3 months following severe acute respiratory syndrome coronavirus 2 infections during the pre-Delta, Delta, and Omicron time periods: a multicenter prospective cohort study. Clin Infect Dis. 2023;76(11):1930-1941. doi: 10.1093/cid/ciad045 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r35] 35.Fang Z, Ahrnsbrak R, Rekito A. Evidence mounts that about 7% of US adults have had long COVID. JAMA. 2024;332(1):5-6. doi: 10.1001/jama.2024.11370 [DOI] [PubMed] [Google Scholar]

[zoi250746r36] 36.Wisk LE, Gottlieb MA, Spatz ES, et al. ; INSPIRE Group . Association of initial SARS-CoV-2 test positivity with patient-reported well-being 3 months after a symptomatic illness. JAMA Netw Open. 2022;5(12):e2244486. doi: 10.1001/jamanetworkopen.2022.44486 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r37] 37.Gottlieb M, Spatz ES, Yu H, et al. ; INSPIRE Group . Long COVID clinical phenotypes up to 6 months after infection identified by latent class analysis of self-reported symptoms. Open Forum Infect Dis. 2023;10(7):ofad277. doi: 10.1093/ofid/ofad277 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r38] 38.Kisiel MA, Kock C, Sultan J, Janols H, Janson C, Pingel R. Agreement between reported questionnaire data and medical records on diagnosis and COVID-19 symptoms at onset. Eur Clin Respir J. 2023;10(1). doi: 10.1080/20018525.2023.2282251 [DOI] [Google Scholar]

[zoi250746r39] 39.Menachemi N, Collum TH. Benefits and drawbacks of electronic health record systems. Risk Manag Healthc Policy. 2011;4:47-55. doi: 10.2147/RMHP.S12985 [DOI] [PMC free article] [PubMed] [Google Scholar]

[zoi250746r40] 40.O’Laughlin KN, Klabbers RE, Ebna Mannan I, et al. ; INSPIRE Group . Ethnic and racial differences in self-reported symptoms, health status, activity level, and missed work at 3 and 6 months following SARS-CoV-2 infection. Front Public Health. 2024;11:1324636. doi: 10.3389/fpubh.2023.1324636 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Variability in Long COVID Definitions and Validation of Published Prevalence Rates

Lauren E Wisk, PhD

Michelle L’Hommedieu, PhD

Kate Diaz Roldan, MPH

Imtiaz Ebna Mannan, MS

Erica S Spatz, MD, MHS

Robert A Weinstein, MD

Arjun K Venkatesh, MD, MBA, MHS

Michael Gottlieb, MD

Ryan Huebinger, MD

Kristin L Rising, MD, MSHP

Juan Carlos C Montoy, MD, PhD

Kari A Stephens, PhD

Robert M Rodriguez, MD

Mandy J Hill, DrPH, MPH

Kelli N O’Laughlin, MD, MPH

Nicole L Gentile, MD, PhD

Ahamed H Idris, MD

Shu-Xia Li, PhD

Michelle Santangelo, MS

Efrat R Kean, MD, MSc

Samuel A McDonald, MD, MS

Kristyn Gatling, MA

Joann G Elmore, MD, MPH

Key Points

Question

Findings

Meaning

Abstract

Importance

Objective

Design, Setting, and Participants

Exposure

Main Outcomes and Measures

Results

Conclusions and Relevance

Introduction

Methods

Study Design and Data Source

Cohort Definition

Variables

Literature Search

Statistical Analysis

Table 1. Long COVID Definitions From the 5 Studies Used for INSPIRE Replication.

Results

Table 2. Baseline Demographic Characteristics of Participants in the INSPIRE Study.

Figure. Comparison of Long COVID Prevalence Among the INSPIRE Participants When Using 5 Different Published Definitions.

Table 3. Sensitivity and Specificity When Applying Definitions of Long COVID Used in the 5 Studies to the INSPIRE Participants Who Were COVID-19 Positive in the Index Testa.

Discussion

Strengths and Limitations

Conclusions

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Table 3. Sensitivity and Specificity When Applying Definitions of Long COVID Used in the 5 Studies to the INSPIRE Participants Who Were COVID-19 Positive in the Index Test^a.