Skip to main content
Download PDF

Utility of illness symptoms for predicting COVID-19 infections in children

BMC Pediatrics volume 22, Article number: 655 (2022) Cite this article

Abstract

Background

The Centers for Disease Control and Prevention and the American Academy of Pediatrics recommend that symptomatic children remain home and get tested to identify potential coronavirus disease 2019 (COVID-19) cases. As the pandemic moves into a new phase, approaches to differentiate symptoms of COVID-19 versus other childhood infections can inform exclusion policies and potentially prevent future unnecessary missed school days.

Methods

Retrospective analysis of standardized symptom and exposure screens in symptomatic children 0–18 years tested for SARS-CoV-2 at three outpatient sites April to November 2020. Likelihood ratios (LR), number needed to screen to identify one COVID-19 case, and estimated missed school days were calculated.

Results

Of children studied (N = 2,167), 88.9% tested negative. Self-reported exposure to COVID-19 was the only factor that statistically significantly increased the likelihood of a positive test for all ages (Positive LR, 5–18 year olds: 5.26, 95% confidence interval (CI): 4.37–6.33; 0–4 year olds: 5.87, 95% CI: 4.67–7.38). Across ages 0–18, nasal congestion/rhinorrhea, sore throat, abdominal pain, and nausea/vomiting/diarrhea were commonly reported, and were either not associated or had decreased association with testing positive for COVID-19. The number of school days missed to identify one case of COVID-19 ranged from 19 to 48 across those common symptoms.

Conclusions

We present an approach for identifying symptoms that are non-specific to COVID-19, for which exclusion would likely lead to limited impact on school safety but contribute to school-days missed. As variants and symptoms evolve, students and schools could benefit from reconsideration of exclusion and testing policies for non-specific symptoms, while maintaining testing for those who were exposed.

Peer Review reports

Introduction

School closures related to the coronavirus disease 2019 (COVID-19) pandemic impacted more than 57 million school-aged and 21 million in preschool or childcare in the United States in March 2020 [1]. In-person school instruction is again the expectation, but protocols for COVID-19 screening varied widely throughout the pandemic and the CDC still recommends testing as soon as possible if symptomatic and staying home [2, 3].

For schools and daycares, the Centers for Disease Control (CDC) and the American Academy of Pediatrics (AAP) recommend that children with COVID-19 symptoms stay home and be tested as soon as possible [3, 4]. However, symptoms of COVID-19 overlap extensively with other common childhood viral syndromes. The CDC and AAP’s lists of COVID-19 symptoms in children included non-specific symptoms that were not previously strict school exclusion criteria, like fatigue, headache, sore throat, and nasal congestion/rhinorrhea [4, 5]. One study previously reported that common symptoms like cough and rhinorrhea did not predict a positive severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) test in children [6]. In addition, despite widespread symptom screening during the COVID-19 pandemic, there is limited evidence on its efficacy for detecting COVID-19 cases [7]. It is possible that symptom screening led to missed school days but detected few COVID-19 cases, exacerbating learning losses for children. As the pandemic continues, with ongoing appearance of new variants and the continued stress of pandemic protocols on schools, approaches are needed to narrow the list of potential symptoms that require exclusion and testing, to optimize school attendance. Ideally such approaches could prevent overly general exclusions that exacerbate learning loss—and allow for updated symptom guidance based on new evidence as it becomes available, to account for changing symptom profiles with new variants – or future pandemics.

To assess the potential impact of exclusion and testing of symptomatic children on detection of COVID-19 and on school or child care days missed, we used cross-sectional multi-center outpatient data from symptomatic children and youth, similar to the general population of K-12 schools or daycares. For symptoms recommended by the CDC for exclusion and testing, we calculated positive likelihood ratios (+ LRs) to quantify the number of people needed to screen to identify potential infections and estimated the number of school days missed because of exclusion for symptoms. This quantifies the potential educational detriment to students [8, 9].

Methods

Study design and population

This was a cross-sectional, retrospective analysis of symptomatic children 0–18 years old who were tested for SARS-CoV-2 at three outpatient sites in Northern California, April 1 through November 30, 2020. This study was approved by the University of California San Francisco’s Committee on Human Research (CHR #20–32,287) and by the Sutter Health San Francisco Institutional Review Board (Local Board Reference #2020.168EXP). All methods were carried out in accordance with relevant guidelines and regulations. Analysis preceded outbreak of the Alpha, Delta and Omicron variants, but these variants’ pediatric symptom profile is similar to the original strain and earlier variants [10, 11].

Children tested were referred by an advice nurse or other outpatient pediatric provider for evaluation and COVID-19 testing. Given persistent PCR positivity in some individuals, we included the first positive test for each child and excluded all subsequent tests [12, 13]. Other excluded tests were those 1) with an incomplete symptom screen (Site 1: if only 1 of a 2-page screener was scanned into the patient’s record; Site 3: if < 5 symptoms were marked as present/absent); 2) with no documented symptom screen (n = 564); 3) done on asymptomatic children for procedures or school clearance (n = 219); 4) collected ≥ 7 days after symptom screen (n = 106) to only capture symptoms close to the time of test positivity or negativity, assuming that symptoms may have changed ≥ 7 days after the screen.

All tests utilized reverse transcription-polymerase chain reaction or transcription mediated amplification assay.

Symptom screening

Symptom data were gathered from site-specific, standardized symptom screeners completed by a healthcare provider during an in-person or telehealth visit prior to the COVID-19 test. As slight differences in screeners existed between sites, we grouped symptoms into two categories: those on the CDC list of COVID-19 symptoms in children [4] and other symptoms potentially used for school exclusion.

Site 1 (Benioff Children’s Hospital Oakland)

Symptoms screened included: Fever or chills, cough, nausea/vomiting/diarrhea, dyspnea, nasal congestion/rhinorrhea, muscle aches, loss of taste or smell, conjunctivitis, and rash. COVID-19 contact exposure and rash were added to this screener on July 23, 2020; we screened charts for contacts reported prior to this date. Data was extracted by manual chart review.

Site 2 (Benioff Children’s Hospital San Francisco)

Symptom screen data was extracted electronically or manually from the screener in the medical record. Compared to Site 1, additional symptoms screened include abdominal pain, headache, sore throat, and fatigue.

Site 3 (Palo Alto Medical Foundation, San Carlos)

For feasibility of data extraction from this larger cohort, all COVID-19 positive tests were identified and then matched, by age in years and month of testing, to four COVID-19 negative controls who had no previous COVID-19 positive tests; 12 cases had only three controls available. Three were excluded for < 3 matched controls. Symptoms screened included: Fever or chills, cough, nausea/vomiting/diarrhea, dyspnea, nasal congestion/rhinorrhea, muscle aches, loss of taste or smell (added April 12, 2020), and rash. Fatigue and conjunctivitis were not screened.

Site and county COVID-19 prevalence

To estimate prevalence in counties of residence for our study population, we used population-level county 7-day positivity for the 1st and 3rd weeks of each month from publicly available county Department of Public Health data [14,15,16,17,18]. For Site 3, county data was not available from the public health department or other sources; we thus used microbiology lab data from that center on 14-day positivity for children ages 0–18 tested each month (Supplemental Fig. 1).

Fig. 1
figure 1

Frequency of symptoms and COVID-19 positive versus negative test result in children with no reported COVID-19 exposure, by age. *Children at Sites 1 and 2; Site 3 symptom screener did not include. †Children tested at Sites 1 and 3; Site 2 symptom screener did not include. ‡Children tested at Site 1; Sites 2 and 3 symptom screeners did not include

Full size image

Demographics

Demographics were extracted from electronic medical records. Age was categorized as 0–4, 5–18 years to reflect childcare and K-12 school-aged groups.

Exposure to COVID-19

Recent close contact with someone who had COVID-19 was self-reported by the patient or caregiver. Data from contact tracing was not available.

Symptoms

We combined the following symptoms for analysis: Fever, chills as “fever/chills”; sinus congestion, runny nose, rhinorrhea, nasal congestion, and chest congestion as “nasal congestion/rhinorrhea”; eye redness, eye discharge, and eye pain as “conjunctivitis”; and “nausea, vomiting, or diarrhea.”

Health status

Health status was determined using the Pediatric Medical Complexity Algorithm Version 3.1 (SAS Institute Inc., Cary, NC) [19]. International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes were extracted from the medical record (last three years, inpatient and outpatient encounters) or the problem list if no recent encounters (n = 6). Children with no chronic conditions or non-chronic complex disease were categorized as “without complex chronic disease.”

Statistical analysis

For each symptom or symptom combination, we assessed (1) percent of children reporting the symptom, (2) positive LRs, which relate the change in probability of a positive test if an individual has the symptom [20], (3) number of individuals needed to screen for the symptom to identify one case [21]; and (4) number of missed school days to identify one case, due to isolation while waiting for a test, estimating 3 days missed for each positive symptom screen (day of symptom identification, day of testing, day waiting for results) (See Supplemental Table 1 for formulae).

Table 1 Children with illness symptoms tested for COVID-19, by test result, April – November 2020
Full size table

We examined each symptom individually. In addition, we examined combinations of symptoms (1) that have been more specifically associated with COVID-19 infection (e.g. fever, cough, loss of taste/smell) or (2) that are less specific to COVID-19 but are common childhood symptoms and are included in the CDC list of COVID-19 symptoms [2, 3].

To assess the clinical relevance of the calculated LRs, we used an LR nomogram; this illustrates the probability of testing COVID-19 positive in a child with that symptom (post-test probability) for select symptoms and symptom combinations [22]. We used a pre-test probability of 5%, based on community prevalence during the study period and the CDC indicator for low to moderate transmission risk for school decision-making [3].

We conducted 2 sensitivity analyses. Because the sampling scheme was different between sites, we calculated LRs for Sites 1 and 2 separately from Site 3. We also used logistic regression to calculate adjusted odds ratios for each symptom [23], adjusting for site and COVID-19 exposure. For the adjusted logistic regression models, we were unable to combine person-level data from Site 3 with that from Sites 1 and 2 due to data-sharing limitations between institutions.

We used Stata 16 (StataCorp, College Station, TX) for statistical analyses [24].

Results

The study population included 2,167 children; 83.7% were without complex chronic medical conditions, and 88.9% tested negative. Almost 20% reported a recent exposure to COVID-19 (Table 1). Across counties of residence for included children, percent positivity of COVID-19 tests ranged from 0.8 to 9.8% during the study period and was higher in some counties than others (Supplemental Fig. 1).

The most frequently reported symptoms were “fever/chills” (n = 1080, 49.8%) and “nasal congestion/rhinorrhea” (n = 1023, 47.2%). Over two-thirds of all children reported > 1 illness symptom, including 69.3% of COVID-19 positive and 68.4% of COVID-19 negative children. Amongst children with no reported COVID-19 exposure, the vast majority – including those with fever/chills, cough, or loss of taste or smell – tested negative (Fig. 1). For those children with reported exposure, the percent positive was higher across all symptoms compared to those without exposure (Fig. 1, Supplemental Fig. 2).

Fig. 2
figure 2

Impact of common illness symptoms on the probability of testing COVID-19 positive. The Fagan nomogram is a graphical tool that demonstrates the probability that a child with each symptom has a positive COVID-19 test, in a community with 5% prevalence (pre-test probability). Identifying contact exposure or loss of taste or smell in a symptom screener identifies children with a higher probability of actually having a COVID-19 infection (post-test probability) than a randomly selected child from the community. In contrast, identifying symptoms with LR 1–2 do not meaningfully increase the probability that a tested child will have COVID-19 above the background community prevalence

Full size image

Factors associated with a COVID-19 positive test

Reported exposure to someone with COVID-19 was the only factor with a statistically significantly positive likelihood ratio for all ages (Table 2).

Table 2 Utility of symptoms identified on symptom screener for identifying COVID cases in children with illness symptoms (n = 2167)
Full size table

In 5–18 year olds, of the CDC symptoms recommended for school exclusion [2, 3, 5], headache, cough, fatigue, muscle aches, and loss of taste or smell had statistically significant positive likelihood ratios (Table 2). However, the positive likelihood ratios for headache, cough, and muscle aches were less than 2, thus resulting in minimal changes to post-test probability (Fig. 2); a substantial number of school days would be missed to identify one case (14, 18, 16 respectively). Loss of taste or smell, though rare (6.0%), had a high and statistically significant positive likelihood ratio (LR: 3.57 95% CI 2.27–5.61) and substantially changed post-test probability (+ 11 percentage points) (Fig. 2).

The most common symptoms, fever/chills (46.1%), nasal congestion (43.7%), and sore throat (42.7%), had positive likelihood ratios close to 1.00, none of which were statistically significant (Table 2).

All other symptoms in both age groups (0–4 and 5–18) were either not meaningfully associated with a positive likelihood of COVID-19 (LR of < 2), were not statistically significantly associated, or had a significantly decreased association (e.g., nausea, vomiting or diarrhea; LR: 0.68, 95% CI 0.50–0.92, in 30% of 5–18 year olds). This included nasal congestion, which was very common in both groups (43.7% of 5–18 and 51.6% in 0–4) (Table 2).

In our assessment of co-occurring symptoms, in 5–18 year old children without fever, only cough and loss of taste or smell remained significantly associated with COVID-19 positivity (Table 3). In the absence of fever or cough, none of the following symptoms increased the likelihood of testing COVID-19 positive: sore throat, nausea/vomiting/diarrhea, fatigue, abdominal pain, headaches, muscle aches, shortness of breath (Table 3). Isolating and testing every 5–18 year old with no fever and only one of the following symptoms would lead to > 18 days of school missed per COVID-19 infection identified: nasal congestion/rhinorrhea, sore throat, nausea/vomiting/diarrhea, fatigue, abdominal pain, and muscle aches (Table 3).

Table 3 Diagnostic utility of common symptom combinations in children 5–18 years old with illness symptoms who were tested for COVID-19a
Full size table

Associations with COVID-19 positivity, adjusted for site and reported exposure

In sensitivity analyses of results by site, the only factors consistently associated with COVID-19 positivity across sites were COVID-19 exposure (all ages) and loss of taste/smell for 5–18 year olds (Supplemental Tables 2, 3). Fever/chills for 0–4 year olds increased the odds of being COVID-19 positive in analyses adjusted for contact exposure (Supplemental Table 3).

Discussion

This multi-site study of symptomatic children evaluated in outpatient settings during the COVID-19 pandemic demonstrates an approach that could be repeated systematically to inform school exclusion and testing policies and practices. Our study population was likely similar to those with symptoms at school or childcare: most with no reported COVID-19 exposure; few with complex medical conditions; and all tested as outpatients, so likely mildly ill. The majority of these symptomatic children tested negative for COVID-19. We found that, of the eleven COVID-19 symptoms that the CDC suggests should trigger COVID-19 testing [2, 3, 5], nasal congestion, sore throat, and “nausea, vomiting, and diarrhea” were very common (in 30–50% of children) and not significantly associated with a positive COVID-19 test in any age group; they were, however, associated with substantial missed school-days. Loss of taste or smell was strongly associated with increased likelihood of a positive test, as was self-reported COVID-19 exposure. These data suggest that, in an endemic phase of the COVID-19 pandemic with ongoing new variants [25, 26], more parsimonious lists of symptoms for exclusion might be appropriate.

Our findings that (1) COVID-19 exposure was the most significant predictor of having COVID-19, (2) other non-specific symptoms of illness either were not associated with or even reduced the likelihood of testing positive for COVID-19, and (3) loss of taste or smell was a useful predictor, are similar to findings from other studies of pediatric cohorts during the same time period, including in Italy and Canada [6, 27, 28]. As circulating COVID-19 variants change, symptom frequency may also change – and re-examining pediatric symptoms over time will be required. The use of likelihood ratios, as in this study, to help demonstrate the utility of focusing on certain symptoms could be re-utilized as the disease itself evolves.

Our analysis also aimed to quantify the impact of symptoms on school days missed, a key policy consideration in light of extensive learning losses sustained, exacerbations of educational inequities due to missed school, and the detrimental effects of hybrid and distance learning [29, 30]. We found, for example, that exclusion and non-rapid, molecular testing of those with sore throat could lead to more than 35 days of school missed for one case of COVID-19 detected, as could exclusion and testing of children with most symptoms, in the absence of fever and cough. To place that number in context, California defines truancy as > 30 min of unexcused school three times per year [31]. While illness is an excused absence, this definition highlights the importance of 35 missed school days, even spread across a group of students. Again as COVID-19 variants evolve, these calculations may need to be re-visited; but our analysis offers a simple technique for quantifying one aspect of the risks associated with school exclusions or closures: missed school days.

It is important to note that our data included infections associated with the original strain and early variants, preceding the Alpha variant B.1.1.7, Delta B.1.617.2, and Omicron B.1.1.529 variants. A prospective UK study comparing symptoms in children with Alpha and Delta variant infections found that the seven most common symptoms were the same between the two variants [10]. However, studies of the Omicron variant revealed a lower incidence of loss of smell and a higher incidence of sore throat [32, 33]. Future research with new variants could improve precision of symptom-specific point estimates, though findings from older strains could inform policymakers’ approach. For instance, these data suggests that policymakers could have recommended against school exclusion for isolated rhinorrhea, since those symptoms overlap with symptoms of common coronaviruses and rhinoviruses and have not been shown predictive of COVID-19 disease [34, 35]. Westbrook et al. found a positive predictive value of only 9% for isolated congestion/rhinorrhea, even despite their cohort’s high community COVID-19 positivity (21%) during the height of the delta variant [36]. Though recent variants (e.g., Omicron) have demonstrated increasing transmissibility [37, 38], it is critical to remember that higher community prevalence (pre-test probability) does not meaningfully change the probability that children with a symptom whose LR is not significantly different than 1 will test positive for COVID-19 (post-test probability) (Fig. 2). Policymakers uncomfortable changing symptom recommendations could consider recommending rapid antigen tests for symptomatic children in a test-to-stay strategy. Antigen tests are less sensitive than molecular tests (e.g., PCR) but would minimize days missed in the setting of low pre-test probability for non-specific symptoms [39].

Our data may also inform decisions of what criteria to maintain for potential exclusion and testing. Our data supports inquiries regarding COVID-19 contact exposure in all ages and loss of taste or smell for 5–18 year olds.

This study has limitations. Although a standardized symptom screener was used at each site, the checklist, administering healthcare provider, and ambulatory setting differed slightly between sites. However, parental or school nurse assessments of symptoms will also likely vary; our findings may thus reasonably reflect real-world settings. The accuracy of symptom screening in children 0–4 was limited to caregiver and health provider observation; however, this again reflects real-world screening. In addition, we assumed that symptoms not marked on the screener as “present” were “absent.” Since we were more confident about absence of “major” symptoms like fever and cough than less specific symptoms, we limited our analysis to reflect this. For Site 3, we also excluded screeners with fewer than 5 symptoms marked present or absent to ensure more accuracy of our assumptions. Further, COVID-19 exposure was not defined across sites; it was not possible to discriminate exposure by close versus causal contacts. In subsequent studies, it will be helpful to gather information regarding whether exposure was at school or home. Finally, we may have been underpowered to detect statistically significant likelihood ratios for some symptoms that were less common. However, point estimates were close to one for most LRs that were not statistically significant, suggesting that even with a narrow confidence interval the change to post-test probability would be minimal. In addition, COVID-19 positivity was higher at our testing sites than in surrounding communities, likely due to a relatively high threshold for testing in earlier months. As testing availability has increased and threshold for testing decreased, we would expect that children with even milder or less specific symptoms might be added to the testing pool; this would likely further decrease symptom utility for identifying COVID-19 cases. To further assess the feasibility and utility of limiting school exclusion based on mild non-specific symptoms, future studies could repeat the analysis using data from more recent variants of COVID-19. Additional data on timing of symptom onset and days of school missed due to testing could inform symptomatic testing recommendations.

Conclusion

In a large population of symptomatic children April-November 2020, the presence of most symptoms did not meaningfully increase the likelihood of testing COVID-19 positive, especially in children with no known COVID-19 exposure. This suggests that a more limited symptom list could be used as the pandemic continues to evolve – or future pandemics emerge – to create more parsimonious symptom criteria for exclusion and testing. Excluding students or staff with non-specific symptoms was unlikely to effectively or efficiently identify children with COVID-19 and likely contributed to unnecessary learning loss. The use of LRs and numbers needed to screen demonstrated that excluding and testing 5–18 year olds with loss of taste or smell, and those of all ages with exposure to COVID-19, would have been reasonable approaches to identify COVID-19 cases during the study period.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Abbreviations

COVID-19:

Coronavirus disease 2019

CDC:

Centers for Disease Control and Prevention

AAP:

American Academy of Pediatrics

SARS-CoV-2:

Severe acute respiratory syndrome coronavirus 2

LR:

Likelihood ratio ()

References

  1. Donohue JM, Miller E. COVID-19 and School Closures. JAMA. 2020;324(9):845–7. https://doi.org/10.1001/jama.2020.13092.

    Article  PubMed  CAS  Google Scholar 

  2. CDC. “COVID-19 and Your Health.” Centers for Disease Control and Prevention, February 11, 2020. https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/symptoms.html.

  3. CDC. Guidance for COVID-19 Prevention in K-12 Schools and ECE Programs. Centers for Disease Control and Prevention. Published January 13, 2022. https://www.cdc.gov/coronavirus/2019-ncov/community/schools-childcare/k-12-guidance.html. Accessed 4 Apr 2022.

  4. COVID-19 Testing Guidance. http://www.aap.org/en/pages/2019-novel-coronavirus-covid-19-infections/clinical-guidance/covid-19-testing-guidance/. Accessed 4 Apr 2022

  5. CDC. Healthcare Workers. Centers for Disease Control and Prevention. Published February 11, 2020. https://www.cdc.gov/coronavirus/2019-ncov/hcp/pediatric-hcp.html. Accessed 4 Apr 2022.

  6. King JA, Bakal JA, McAlister FA. Symptoms associated with a positive result for a swab for SARS-CoV-2 infection among children in Alberta. CMAJ. 2021;193(1):E1–9. https://doi.org/10.1503/cmaj.202065.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  7. Gostic K, Gomez AC, Mummah RO, Kucharski AJ, Lloyd-Smith JO. Estimated effectiveness of symptom and risk screening to prevent the spread of COVID-19. eLife. 2020;9. https://doi.org/10.7554/eLife.55570

  8. Meara ER, Richards S, Cutler DM. The gap gets bigger: changes in mortality and life expectancy, by education, 1981–2000. Health Aff Proj Hope. 2008;27(2):350–60. https://doi.org/10.1377/hlthaff.27.2.350.

    Article  Google Scholar 

  9. Hill RM, Rufino K, Kurian S, Saxena J, Saxena K, Williams L. Suicide Ideation and Attempts in a Pediatric Emergency Department Before and During COVID-19. Pediatrics. 2021;147(3). https://doi.org/10.1542/peds.2020-029280.

  10. Illness characteristics of COVID-19 in children infected with the SARS-CoV-2 Delta variant | medRxiv. https://www.medrxiv.org/content/10.1101/2021.10.06.21264467v1. Accessed 4 Apr 2022.

  11. CDC. Omicron Variant: What You Need to Know. Centers for Disease Control and Prevention. Published March 29, 2022. https://www.cdc.gov/coronavirus/2019-ncov/variants/omicron-variant.html. Accessed 4 Apr 2022.

  12. Wajnberg A, Mansour M, Leven E, et al. Humoral response and PCR positivity in patients with COVID-19 in the New York City region, USA: an observational study. Lancet Microbe. 2020;1(7):e283–9. https://doi.org/10.1016/S2666-5247(20)30120-8.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  13. Gombar S, Chang M, Hogan CA, et al. Persistent detection of SARS-CoV-2 RNA in patients and healthcare workers with COVID-19. J Clin Virol. 2020;129:104477. https://doi.org/10.1016/j.jcv.2020.104477.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  14. COVID-19 Testing Dashboards - Emergency Operations Center - County of Santa Clara. https://covid19.sccgov.org/dashboard-testing. Accessed 28 Jan 2021.

  15. Alameda County Health Care Services Agency Public Health Department. Data | COVID-19 | Alameda County Public Health. https://covid-19.acgov.org/data. Accessed 28 Jan 2021.

  16. COVID-19 Laboratory Testing | DataSF | City and County of San Francisco. https://data.sfgov.org/stories/s/d96w-cdge. Accessed 28 Jan 2021.

  17. San Mateo County Health. Lab Testing Data. San Mateo County Health. https://www.smchealth.org/data-dashboard/lab-testing-data. Accessed 28 Jan 2021.

  18. Contra Costa Health Services. Overview Dashboard. Coronavirus. https://www.coronavirus.cchealth.org/overview. Accessed 28 Jan 2021.

  19. Simon TD, Haaland W, Hawley K, Lambka K, Mangione-Smith R. Development and Validation of the Pediatric Medical Complexity Algorithm (PMCA) Version 3.0. Acad Pediatr. 2018;18(5):577–80. https://doi.org/10.1016/j.acap.2018.02.010.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Attia, John. “Moving beyond Sensitivity and Specificity: Using Likelihood Ratios to Help Interpret Diagnostic Tests.” https://doi.org/10.18773/austprescr.2003.082. Accessed 22 Sept 2022.

  21. Rembold CM. Number Needed to Screen: Development of a Statistic for Disease Screening. BMJ. 1998;317(7154):307–12.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  22. Caraguel CG, Vanderstichel R. The Two-Step Fagan’s Nomogram: Ad Hoc Interpretation of a Diagnostic Test Result without Calculation. Evid Based Med. 2013;18(4):125–8. https://doi.org/10.1136/eb-2013-101243.

    Article  PubMed  Google Scholar 

  23. Lee, Seung Won. “Regression Analysis for Continuous Independent Variables in Medical Research: Statistical Standard and Guideline of Life Cycle Committee.” Life Cycle 2 (2022). https://doi.org/10.54724/lc.2022.e3.

  24. StataCorp. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC; 2019.

    Google Scholar 

  25. Torjesen I. Covid-19 will become endemic but with decreased potency over time, scientists believe. BMJ. 2021;372:n494. https://doi.org/10.1136/bmj.n494.

    Article  PubMed  Google Scholar 

  26. Shaman J, Galanti M. Will SARS-CoV-2 become endemic? Science. 2020;370(6516):527–9. https://doi.org/10.1126/science.abe5960.

    Article  PubMed  CAS  Google Scholar 

  27. Armocida B, Zamagni G, Magni E, et al. Clinical, anamnestic, and sociodemographic predictors of positive SARS-CoV-2 testing in children: A cross sectional study in a tertiary hospital in Italy. PLoS ONE. 2022;17(1):e0262923. https://doi.org/10.1371/journal.pone.0262923.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  28. Weng CH, Butt WWW, Brooks MB, et al. Diagnostic value of symptoms for pediatric SARS-CoV-2 infection in a primary care setting. PLoS ONE. 2021;16(12):e0249980. https://doi.org/10.1371/journal.pone.0249980.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  29. Kaufman JH, Diliberti MK. Divergent and Inequitable Teaching and Learning Pathways During (and Perhaps Beyond) the Pandemic: Key Findings from the American Educator Panels Spring 2021 COVID-19 Surveys. Published online May 27, 2021. https://www.rand.org/pubs/research_reports/RRA168-6.html. Accessed 6 July 2021.

  30. Verlenden JV. Association of Children’s Mode of School Instruction with Child and Parent Experiences and Well-Being During the COVID-19 Pandemic — COVID Experiences Survey, United States, October 8–November 13, 2020. MMWR Morb Mortal Wkly Rep. 2021;70. https://doi.org/10.15585/mmwr.mm7011a1.

  31. Truancy - Attendance Improvement (CA Dept of Education). https://www.cde.ca.gov/ls/ai/tr/#:~:text=Definition%20of%20a%20Truant,to%20the%20proper%20school%20authority. Accessed 6 July 2021.

  32. Yang, Wei, Songliu Yang, Lei Wang, Yuxin Zhou, Yu Xin, Hongxu Li, Wenjing Mu, et al. “Clinical Characteristics of 310 SARS-CoV-2 Omicron Variant Patients and Comparison with Delta and Beta Variant Patients in China.” Virologica Sinica. 2022:S1995820X22001389. https://doi.org/10.1016/j.virs.2022.07.014.

  33. Menni C, Valdes AM, Polidori L, Antonelli M, Penamakuri S, Nogal A, Louca P, et al. Symptom Prevalence, Duration, and Risk of Hospital Admission in Individuals Infected with SARS-CoV-2 during Periods of Omicron and Delta Variant Dominance: A Prospective Observational Study from the ZOE COVID Study. Lancet. 2022;399(10335):1618–24. https://doi.org/10.1016/S0140-6736(22)00327-0.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  34. Siegel DA, Reses HE, Cool AJ, et al. Trends in COVID-19 Cases, Emergency Department Visits, and Hospital Admissions Among Children and Adolescents Aged 0–17 Years - United States, August 2020-August 2021. MMWR Morb Mortal Wkly Rep. 2021;70(36):1249–54. https://doi.org/10.15585/mmwr.mm7036e1.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  35. , Hospitalizations Associated with COVID-19 Among Children and Adolescents — COVID-NET, 14 States, March 1, 2020–August 14, 2021 | MMWR. https://www.cdc.gov/mmwr/volumes/70/wr/mm7036e2.htm. Accessed 4 Apr 2022.

  36. Westbrook AL, Benedit LC, Frediani JK, et al. Predictive value of isolated symptoms for diagnosis of SARS-CoV-2 infection in children tested during peak circulation of the delta variant. Clin Infect Dis. Published online March 10, 2022:ciac112. https://doi.org/10.1093/cid/ciac112.

  37. Madewell ZJ, Yang Y, Longini IM, Halloran ME, Dean NE. Household secondary attack rates of sars-cov-2 by variant and vaccination status: an updated systematic review and meta-analysis. Public and Global Health, 2022. https://doi.org/10.1101/2022.01.09.22268984.

  38. Baker J, Nakayama J, O’Hegarty M, McGowan A, Teran R, Bart S. SARS-CoV-2 B.1.1.529 (Omicron) Variant Transmission Within Households — Four U.S. Jurisdictions, November 2021–February 2022 | MMWR. https://www.cdc.gov/mmwr/volumes/71/wr/mm7109e1.htm?s_cid=mm7109e1_w. Accessed 4 Apr 2022.

  39. CDC. Labs. Centers for Disease Control and Prevention. Published February 11, 2020. https://www.cdc.gov/coronavirus/2019-ncov/lab/resources/antigen-tests-guidelines.html.

Download references

Acknowledgements

We are grateful to Robert Thombley (UCSF) for helping with PMCA code, Su-Ying Liang (Palo Alto Medical Foundation) for data analysis assistance, Drs. Jeanne Lee and Sohil Sud (San Francisco Department of Public Health) for review, and Dr. Thomas Newman (UCSF) for advising on LR analysis.

Funding

This work was supported in part by The Flu Lab for Dr. Emily Perito’s work on this project. The Flu Lab had no role in the study design, access to the data and no role in its interpretation, in the writing of the manuscript, or in the submission decisions.

Author information

Author notes

  1. Naomi S. Bardach and Emily R. Perito contributed equally as co-senior authors.

Authors and Affiliations

  1. Children’s Hospital Los Angeles, 4650 Sunset Blvd, Los Angeles, CA, 90027, USA

    Geena Y. Zhou

  2. Children’s Hospital Colorado, 13123 E 16Th Ave, Aurora, CO, 80045, USA

    Nicole Y. Penwill

  3. Palo Alto Medical Foundation, 370 Distel Circle, Los Altos, CA, 94022, USA

    Grace Cheng

  4. Department of Pediatrics, University of California San Francisco, 550 16th Street 4th Floor Box 0136, San Francisco, CA, 94143, USA

    Prachi Singh, Minkyung Shin, Margaret Nguyen, Shalini Mittal, William Burrough, Mia-Ashley Spad, Sarah Bourne, Naomi S. Bardach & Emily R. Perito

  5. Tufts Medical Center, 800 Washington St, Boston, MA, 02111, USA

    Ann Cheung

Authors
  1. Geena Y. Zhou
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Nicole Y. Penwill
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Grace Cheng
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Prachi Singh
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Ann Cheung
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Minkyung Shin
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Margaret Nguyen
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Shalini Mittal
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. William Burrough
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Mia-Ashley Spad
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Sarah Bourne
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Naomi S. Bardach
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Emily R. Perito
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Drs. Zhou, Penwill, Cheng, Singh, and Perito conceptualized and designed the study, collected data, analyzed and interpreted data, drafted the initial manuscript, and reviewed and revised the manuscript. No authors were paid an honorarium or other direct grant to write the manuscript. Dr. Bardach conceptualized and designed the study, analyzed and interpreted data, and reviewed and revised the manuscript. Drs. Cheung, Shin, Nguyen, Mittal, Burrough, Spad, and Bourne collected data and reviewed and revised the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Corresponding author

Correspondence to Emily R. Perito.

Ethics declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

This study was approved by the University of California San Francisco’s Committee on Human Research (CHR #20–32287) and by the Sutter Health San Francisco Institutional Review Board (Local Board Reference #2020.168EXP).

Participant data was collected via retrospective review of hospital chart data; participant consent was not obtained. The University of California San Francisco’s Committee on Human Research (CHR #20–32287) and the Sutter Health San Francisco Institutional Review Board (Local Board Reference #2020.168EXP) approved the waiver of informed consent described in the application, as the research involved no more than minimal risk to subjects; could not practicably be done without the waiver; could not practicably be done without identifiable information; will not adversely affect rights and welfare of subjects with the waiver; and will provide subjects with additional pertinent information after participation, whenever it is appropriate.

Consent for publication

Published data is entirely unidentifiable and there are no details on individuals reported within the manuscript.

Competing interest

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1.

Supplemental figures.

Additional file 2. 

Supplemental tables.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhou, G.Y., Penwill, N.Y., Cheng, G. et al. Utility of illness symptoms for predicting COVID-19 infections in children. BMC Pediatr 22, 655 (2022). https://doi.org/10.1186/s12887-022-03729-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12887-022-03729-w

Keywords

BMC Pediatrics

ISSN: 1471-2431

Contact us