Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 May 18;24(5):e37931.
doi: 10.2196/37931.

Distinguishing Admissions Specifically for COVID-19 From Incidental SARS-CoV-2 Admissions: National Retrospective Electronic Health Record Study

Jeffrey G Klann  1 , Zachary H Strasser  1 , Meghan R Hutch  2 , Chris J Kennedy  3 , Jayson S Marwaha  4 , Michele Morris  5 , Malarkodi Jebathilagam Samayamuthu  5 , Ashley C Pfaff  4 , Hossein Estiri  1 , Andrew M South  6 , Griffin M Weber  7 , William Yuan  7 , Paul Avillach  7 , Kavishwar B Wagholikar  1 , Yuan Luo  2 ; Consortium for Clinical Characterization of COVID-19 by EHR (4CE)  7 ; Gilbert S Omenn  8 , Shyam Visweswaran  5 , John H Holmes  9 , Zongqi Xia  10 , Gabriel A Brat  7 , Shawn N Murphy  11
Collaborators, Affiliations

Distinguishing Admissions Specifically for COVID-19 From Incidental SARS-CoV-2 Admissions: National Retrospective Electronic Health Record Study

Jeffrey G Klann et al. J Med Internet Res. .

Abstract

Background: Admissions are generally classified as COVID-19 hospitalizations if the patient has a positive SARS-CoV-2 polymerase chain reaction (PCR) test. However, because 35% of SARS-CoV-2 infections are asymptomatic, patients admitted for unrelated indications with an incidentally positive test could be misclassified as a COVID-19 hospitalization. Electronic health record (EHR)-based studies have been unable to distinguish between a hospitalization specifically for COVID-19 versus an incidental SARS-CoV-2 hospitalization. Although the need to improve classification of COVID-19 versus incidental SARS-CoV-2 is well understood, the magnitude of the problems has only been characterized in small, single-center studies. Furthermore, there have been no peer-reviewed studies evaluating methods for improving classification.

Objective: The aims of this study are to, first, quantify the frequency of incidental hospitalizations over the first 15 months of the pandemic in multiple hospital systems in the United States and, second, to apply electronic phenotyping techniques to automatically improve COVID-19 hospitalization classification.

Methods: From a retrospective EHR-based cohort in 4 US health care systems in Massachusetts, Pennsylvania, and Illinois, a random sample of 1123 SARS-CoV-2 PCR-positive patients hospitalized from March 2020 to August 2021 was manually chart-reviewed and classified as "admitted with COVID-19" (incidental) versus specifically admitted for COVID-19 ("for COVID-19"). EHR-based phenotyping was used to find feature sets to filter out incidental admissions.

Results: EHR-based phenotyped feature sets filtered out incidental admissions, which occurred in an average of 26% of hospitalizations (although this varied widely over time, from 0% to 75%). The top site-specific feature sets had 79%-99% specificity with 62%-75% sensitivity, while the best-performing across-site feature sets had 71%-94% specificity with 69%-81% sensitivity.

Conclusions: A large proportion of SARS-CoV-2 PCR-positive admissions were incidental. Straightforward EHR-based phenotypes differentiated admissions, which is important to assure accurate public health reporting and research.

Keywords: COVID-19; SARS-CoV-2; clinical research informatics; electronic health records; health care; health data; medical informatics; patient data; phenotype; public health.

PubMed Disclaimer

Conflict of interest statement

Conflicts of Interest: The authors declare that they have no conflicts of interest. JGK reports a consulting relationship with the i2b2-tranSMART Foundation through Invocate, Inc. CJK reports consulting for the University of California, Berkeley; the University of Southern California (USC), and the University of California, San Francisco (UCSF). AMS reports funding from the National Institutes of Health (NIH)/National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) R01DK127208, NIH/National Heart, Lung, and Blood Institute (NHLBI) R01HL146818, and institutional pilot awards from the Wake Forest School of Medicine. GMW reports consulting for the i2b2-tranSMART Foundation. PA reports consulting for the Cincinnati Children’s Hospital Medical Center (CCHMC) and Boston Children’s Hospital (BCH). ZX has received research support from the NIH, the Department of Defense, and Octave Biosciences and has served on the scientific advisory board for Genentech/Roche. SNM reports professional relationships with the Scientific Advisory Board for Boston University, the Universidad de Puerto Rico, the University of California, Los Angeles (UCLA), the University of Massachusetts Medical School (UMMS), and the Kenner Family Research Fund.

Figures

Figure 1
Figure 1
The chart review process. (1-2) At each site, an equal number of patients admitted with a positive SARS-CoV-2 PCR test were sampled by quarter or by month. (3-4) A chart reviewer at the site examined primarily the admission note, discharge summary (or death note), and laboratory values for the hospitalization to classify as admitted for COVID-19, incidental SARS-CoV2, or uncertain. (5-6) These classifications were then merged with 4CE EHR data for use with shared analytic scripts in R. (7-8) The top phenotypes at each site output by the data mining algorithm were summarized, and this was used to manually construct feature sets to be used across sites by selecting components that appeared in step 7 at multiple sites. (9) The performance over time of the top multisite phenotypes was visualized. 4CE: Consortium for Clinical Characterization of COVID-19 by EHR; EHR: electronic health record; ICD-10: International Classification of Diseases, Tenth Revision; PCR: polymerase chain reaction.
Figure 2
Figure 2
Design of the phenotyping algorithm. Predictive feature sets of iteratively larger size were selected based on their sensitivity and specificity in correctly identifying COVID-19-specific admissions using 4CE EHR data and chart reviews. We chose the following parameters after testing various thresholds at all 4 sites: AND feature sets, x=0.40, y=0.20, p=0.30; OR feature sets x=0.10, y=0.50, p=0.20; and single features: x=y=p=0. 4CE: Consortium for Clinical Characterization of COVID-19 by EHR; EHR: electronic health record.
Figure 3
Figure 3
Chart-reviewed proportion of admissions specifically for COVID-19 among all chart reviews by month at each site. The bubble size shows the relative number of patient chart reviews performed that month. The trendline was weighted by bubble size and was performed using locally weighted least squares (loess) regression. Note that the y axis and 95% CI limits extend above 100%.
Figure 4
Figure 4
Performance of the top phenotyping feature sets (Table 7) over time at each site. The y axis is the number of admissions per week, the x axis is the week, and overall sensitivity and specificity are shown on each figure panel. Solid lines show the total number of weekly admissions for patients with a positive SARS-CoV-2 PCR test. Dashed lines show the number of weekly admissions after filtering to select patients admitted specifically for COVID-19 (ie, removing all patients who do not meet the phenotyping feature set criteria). The dotted line shows the difference between the solid line and the dashed line (ie, patients removed from the cohort in the dashed line). Green dots indicate correct classification by the phenotype according to chart review. Orange dots indicate incorrect classification. The dot size is proportional to the number of chart reviews. PCR: polymerase chain reaction.
See this image and copyright information in PMC

Update of

References

    1. Haendel MA, Chute CG, Bennett TD, Eichmann DA, Guinney J, Kibbe WA, Payne PRO, Pfaff ER, Robinson PN, Saltz JH, Spratt H, Suver C, Wilbanks J, Wilcox AB, Williams AE, Wu C, Blacketer C, Bradford RL, Cimino JJ, Clark M, Colmenares EW, Francis PA, Gabriel D, Graves A, Hemadri R, Hong SS, Hripscak G, Jiao D, Klann JG, Kostka K, Lee AM, Lehmann HP, Lingrey L, Miller RT, Morris M, Murphy SN, Natarajan K, Palchuk MB, Sheikh U, Solbrig H, Visweswaran S, Walden A, Walters KM, Weber GM, Zhang XT, Zhu RL, Amor B, Girvin AT, Manna A, Qureshi N, Kurilla MG, Michael SG, Portilla LM, Rutter JL, Austin CP, Gersing KR, N3C Consortium The National COVID Cohort Collaborative (N3C): rationale, design, infrastructure, and deployment. J Am Med Inform Assoc. 2021 Mar 01;28(3):427–443. doi: 10.1093/jamia/ocaa196. http://europepmc.org/abstract/MED/32805036 5893482 - DOI - PMC - PubMed
    1. Brat GA, Weber GM, Gehlenborg N, Avillach P, Palmer NP, Chiovato L, Cimino J, Waitman LR, Omenn GS, Malovini A, Moore JH, Beaulieu-Jones BK, Tibollo V, Murphy SN, Yi SL, Keller MS, Bellazzi R, Hanauer DA, Serret-Larmande A, Gutierrez-Sacristan A, Holmes JJ, Bell DS, Mandl KD, Follett RW, Klann JG, Murad DA, Scudeller L, Bucalo M, Kirchoff K, Craig J, Obeid J, Jouhet V, Griffier R, Cossin S, Moal B, Patel LP, Bellasi A, Prokosch HU, Kraska D, Sliz P, Tan ALM, Ngiam KY, Zambelli A, Mowery DL, Schiver E, Devkota B, Bradford RL, Daniar M, Daniel C, Benoit V, Bey R, Paris N, Serre P, Orlova N, Dubiel J, Hilka M, Jannot AS, Breant S, Leblanc J, Griffon N, Burgun A, Bernaux M, Sandrin A, Salamanca E, Cormont S, Ganslandt T, Gradinger T, Champ J, Boeker M, Martel P, Esteve L, Gramfort A, Grisel O, Leprovost D, Moreau T, Varoquaux G, Vie J, Wassermann D, Mensch A, Caucheteux C, Haverkamp C, Lemaitre G, Bosari S, Krantz ID, South A, Cai T, Kohane IS. International electronic health record-derived COVID-19 clinical course profiles: the 4CE consortium. NPJ Digit Med. 2020;3:109. doi: 10.1038/s41746-020-00308-0. doi: 10.1038/s41746-020-00308-0.308 - DOI - DOI - PMC - PubMed
    1. Murphy SN, Weber G, Mendis M, Gainer V, Chueh HC, Churchill S, Kohane I. Serving the enterprise and beyond with informatics for integrating biology and the bedside (i2b2) J Am Med Inform Assoc. 2010;17(2):124–30. doi: 10.1136/jamia.2009.000893. http://europepmc.org/abstract/MED/20190053 17/2/124 - DOI - PMC - PubMed
    1. Visweswaran S, Samayamuthu MJ, Morris M, Weber GM, MacFadden D, Trevvett P, Klann JG, Gainer VS, Benoit B, Murphy SN, Patel L, Mirkovic N, Borovskiy Y, Johnson RD, Wyatt MC, Wang AY, Follett RW, Chau N, Zhu W, Abajian M, Chuang A, Bahroos N, Reeder P, Xie D, Cai J, Sendro ER, Toto RD, Firestein GS, Nadler LM, Reis SE. Development of a coronavirus disease 2019 (COVID-19) application ontology for the Accrual to Clinical Trials (ACT) network. JAMIA Open. 2021 Apr;4(2):ooab036. doi: 10.1093/jamiaopen/ooab036. http://europepmc.org/abstract/MED/34113801 ooab036 - DOI - PMC - PubMed
    1. Khullar D. Do the Omicron Numbers Mean What We Think They Mean? The New Yorker. 2022. Jan 16, [2022-01-25]. https://www.newyorker.com/magazine/2022/01/24/do-the-omicron-numbers-mea... .

Publication types