Skip to main content
  • Research article
  • Open access
  • Published:

Evaluation of the impact of shigellosis exclusion policies in childcare settings upon detection of a shigellosis outbreak

Abstract

Background

In the event of a shigellosis outbreak in a childcare setting, exclusion policies are typically applied to afflicted children to limit shigellosis transmission. However, there is scarce evidence of their impact.

Methods

We evaluated five exclusion policies: Children return to childcare after: i) two consecutive laboratory tests (either PCR or culture) do not detect Shigella, ii) a single negative laboratory test (PCR or culture) does not detect Shigella, iii) seven days after beginning antimicrobial treatment, iv) after being symptom-free for 24 h, or v) 14 days after symptom onset. We also included four treatments to assess the policy options: i) immediate, effective treatment; ii) effective treatment after laboratory diagnosis; iii) no treatment; iv) ineffective treatment. Relying on published data, we calculated the likelihood that a child reentering childcare would be infectious, and the number of childcare-days lost per policy.

Results

Requiring two consecutive negative PCR tests yielded a probability of onward transmission of < 1%, with up to 17 childcare-days lost for children receiving effective treatment, and 53 days lost for those receiving ineffective treatment.

Conclusions

Of the policies analyzed, requiring negative PCR testing before returning to childcare was the most effective to reduce the risk of shigellosis transmission, with one PCR test being the most effective for the least childcare-days lost.

Peer Review reports

Background

Shigellosis is an infectious disease characterized by diarrhea, stomach cramps, and sometimes fever, starting 1–3 days after exposure to Shigella bacteria (shigellae), typically lasting 5–7 days if untreated. Antimicrobial medications can be used to reduce the duration of severe cases [1]. Shigellosis is diagnosed by laboratory testing of the stools of an infected person. Shigellae are transmitted via the fecal-oral route, and ingestion of as few as 10 bacteria can cause infection. Every year ~ 500,000 cases of shigellosis occur in the United States; outbreaks are common in childcare settings and schools [1, 2].

To try to limit transmission, state policies commonly prevent children with shigellosis from attending childcare for specified periods of time after symptoms resolve and/or following one or more negative laboratory tests. These policies affect childcare attendance and income (e.g., productivity losses of caregivers), and present a burden to schools, healthcare providers, and local public health departments. However, there is limited evidence on the impact of exclusion policies on shigellosis transmission.

We evaluated the impact of five different child exclusion policies on the likelihood that a child returning to childcare would still be infectious, and on childcare-days lost for the afflicted children, upon the detection of a shigellosis outbreak. For each exclusion policy, we quantified the probability that children with shigellosis remained infectious upon school readmission (thus posing a risk of onward transmission) and the number of childcare-days lost. Our analysis provides evidence to inform policy decisions. We also provide a user-friendly spreadsheet tool with adjustable parameters as supplementary material for public use.

Methods

We defined childcare as a facility that provides care and educational activities for around 45 children aged approximately 5 years or younger for several hours per day but not 24 h per day. We evaluated five childcare exclusion policies, reflecting policies currently used in various states (see Additional file 1: Appendix A) and policies from expert opinion (Table 1, Panel I). Because options for shigellosis management can have different levels of effectiveness, we evaluated each policy using four illustrative treatment scenarios for children with shigellosis (Table 1, Panel II). For each policy-treatment pair, we calculated the likelihood of an infectious child reentering childcare, and the expected number of childcare-days lost per child. To deal with uncertainty we included lower and upper bounds for all parameter values. Data were drawn from the literature when available, and from expert opinion when not (Table 2); the final results were calculated in 2017. Additional file 1: Appendix A and Additional file 2: Appendix B show the calculations for the probability of being infectious upon readmission to childcare.

Table 1 Main scenarios evaluated: shigellosis exclusion policies assessed and patient treatment scenarios
Table 2 Parameters used to estimate the effects of different exclusion policies

We estimated the likelihood that shigellosis patients returned to childcare while still infectious as well as the number of childcare-days lost per child (where all days lost are assumed to be childcare-days lost, not accounting for holidays or weekends), for policies based on convalescent stool tests (policies that excluded children until they had one or two consecutive negative stool tests with PCR or culture-based tests on specimens that were collected at least 24 h after completing antimicrobial therapy and/or diarrhea resolution). We evaluated test-based policies taking into account the sensitivity of PCR- and culture-based tests. Similarly, we evaluated policies that excluded children for fixed periods of time; these included exclusions for 14 days after symptom onset, 7 days after starting treatment, and 24 h after becoming symptom-free [3,4,5,6,7,8].

We evaluated each policy for patients undergoing different treatment scenarios, namely: A) immediate, effective treatment; B) effective treatment after diagnosis; C) ineffective treatment; D) no treatment. We considered as effective treatment the receipt of a course of antibiotics that the particular strain of Shigella bacteria was susceptible to, as recommended in the latest guidelines [9]. Ineffective treatment was defined as receiving an antibiotic that the Shigella strain was not susceptible to, or that did not have an effect in vivo based on pharmacokinetics [9].

For policies based on convalescent stool tests, the estimated likelihood that the returning patient reentered school while still infectious was calculated using the probability that the test would provide a false negative result, which was related to the test’s sensitivity (Additional file 1: Appendix A). We opted for this conservative estimate given variability in shedding shigellae in stool, symptom duration, and time interval before receiving the test. For policies based on a fixed time interval, the estimated likelihood that the patient returned to school while still infectious was based on the duration of shedding; in particular, we assumed that the proportion of children who remained infectious decreased linearly with each day of the shedding period, from 100% in the first day to 0% in the last day. The duration of the shedding period was estimated from the literature (Table 2).

For policies relying on negative stool sample results, the number of childcare-days lost was estimated by the number of days required to receive testing results, taking into account treatment duration, and shedding duration for policies based on convalescent stool tests. In particular, for these policies the number of childcare-days lost was calculated as a weighted average of the number of days required to obtain the required (one or two consecutive) negative results for Shigella for infectious children, and the number of days required to obtain said results for non-infectious children. The weights consisted of the likelihood of the child being infectious and non-infectious upon return to childcare. For policies based on fixed time intervals, we estimated the number of childcare-days lost by using said intervals.

We further explored the impact of exclusion policies in aggregated childcare-days lost (childcare-days lost for a group of children, assuming different children receive different treatments) for a given childcare in a separate sensitivity analysis (Additional file 3: Appendix C). We estimated the number of aggregated childcare-days lost as the attack rate multiplied by the setting size and by the number of childcare-days lost for a given combination of treatments the children receive. We considered a population of 45 children (equivalent to a small childcare facility), an attack rate of 25% [10, 11] for our reference analysis, and three different combinations of treatments. We show the results for various combinations of treatment types for our reference population of children (Additional file 3: Appendix C).

Results

Figure 1 shows the probability that an infectious child returns to childcare and the number of days that the child would be excluded, by treatment type and exclusion policy. The effectiveness of policies based on negative convalescent stool tests hinged on the test’s sensitivity, with PCR-based tests leading to lower probability of the child being infectious when readmitted. If the policy required 2 consecutive PCR-negative stool samples, the probability that the child returned to school infectious was < 1%, with the number of days the child spent at home ranging from 7 to 17 days (midpoint: 9 days) if the child received immediate, effective treatment. The maximum number of childcare-days lost per child increased to 19 days if the child received effective treatment after diagnosis; it was between 19 and 53 days if the child received ineffective treatment; and between 6 and 43 days if the child received no treatment. If only 1 Shigella-negative PCR test stool sample was required, the likelihood the child returned to school infectious was ≤6% and the number of days the child spent at home varied between 6 and 11 days (midpoint: 7 days) if the child received immediate, effective treatment; it was up to 13 days if the child received effective treatment after diagnosis; between 18 and 45 days for children receiving ineffective treatment; and between 5 and 35 days for children receiving no treatment.

Fig. 1
figure 1

The impact of 7 different exclusion policies on childcare-days lost per child and probability of infectiousness upon readmission to childcare

While the specificity of PCR and stool culture tests is the same, PCR tests are almost twice as sensitive as stool cultures (Table 2). Thus, the type of diagnostic had a larger impact on readmission of infectious children than the number of tests performed. We estimated that the likelihood of reentering school while infectious after one negative PCR test was 2 to 6%, compared with 8 to 31% for two consecutive negative stool cultures. If only one negative culture was required, the likelihood that the child returned to school infectious ranged from 28 to 56%. The number of childcare-days lost per child for the exclusion policy involving one negative culture varied from 7 to 12 days for children receiving immediate, effective treatment, and from 16 to 28 days if the child received ineffective treatment. If two negative cultures were required, this interval ranged from 9 to 19 days if the child received immediate, effective treatment and from 20 to 44 days if the child received ineffective treatment.

The policy permitting readmission 7 days after beginning antimicrobial treatment showed minimal childcare-days lost for minimum risk (0%) of infectious child readmission if the antimicrobial treatment was effective. However, all (100%) children would be readmitted while infectious if they received inappropriate treatment, because the shedding duration would be longer than 7 days after beginning antimicrobial treatment. The risk of readmitting infectious students was very variable for policies in which children returned to school 14 days after symptom onset or 24 h after being symptom-free for patients receiving ineffective (Range: 0–88%) or no treatment (Range:0–50%).

In Additional file 3: Appendix C, we further explored variation in the aggregated childcare-days lost in a shigellosis outbreak for each exclusion policy, considering a setting of 45 children and an assumed treatment mix of affected children. Results show that the cost comparison (in aggregated childcare-days lost) hinged on treatment effectiveness. When the percent of patients receiving effective treatment increases, the estimated number of aggregated childcare-days lost decreases. Conversely, when the share of patients receiving ineffective or no treatment increases, the estimated number of aggregated childcare-days lost increases.

Discussion

Exclusion policies for shigellosis patients based on convalescent testing most consistently minimized the probability of readmitting an infectious child to childcare, but varied in the number of childcare-days lost per child. PCR tests minimized the likelihood of an infectious child returning to childcare and the number of days the child was excluded. Given the different sensitivities of PCR and culture tests, the use of one PCR test more effectively minimized the probability of readmitting an infectious child than did two stool cultures. Policies based on a fixed number of exclusion days after an event (i.e., symptom onset, start of antimicrobial treatment, resolution of symptoms) exhibited greater variation for patients exposed to different treatments. Such policies resulted in fewer childcare-days lost per child only if a small percent of patients received ineffective treatment (e.g., treatment with an antimicrobial medication to which the Shigella strain was resistant).

Treatment using appropriate antimicrobial medications generally decreased the probability of readmitting an infectious child and the number of childcare-days per child lost compared with no treatment or ineffective antimicrobial treatment, and use of ineffective antimicrobials consistently maximized the probability of readmitting an infectious child and/or childcare-days lost per child. Overuse of antimicrobial treatment may also induce antimicrobial resistance, unnecessarily disrupt children’s microflora, and incur costs to the healthcare system and families. The prevalence of antimicrobial resistance is increasing among shigellae [12]; outbreaks of antimicrobial-resistant shigellosis may result in a high proportion of infectious children returning to childcare, extended days of exclusion, or both [13].

We evaluated child exclusion policies’ impact on the likelihood of readmitting infectious children to childcare and the number of childcare-days lost per child, assuming that the outbreak’s attack rate was unrelated to exclusion policies. In reality, exclusion policies that result in a higher likelihood of children returning to childcare while infectious may contribute to a higher attack rate, and thus higher number of childcare-days lost. That is, for policies resulting in a higher likelihood of returning to childcare while infectious, the current model may underestimate the number of childcare-days lost per child. While a dynamic infectious disease model would be required to quantify this bias, the current model serves as a lower bound of childcare-days lost (equivalent to assuming that after the initial outbreak is detected, caregivers’ alertness to the disease limits disease transmission, for instance, by being especially aware of symptoms in previously known patients).

Our results are also limited by the lack of data regarding the prevalence of different types of treatment and uncertainty about testing parameters. However, our supplementary material allows users to evaluate policies using new data or different assumptions. Another limitation is the absence of information regarding the timing of diagnosis and prevalence of different treatment methods. In our treatment scenarios, we opted not to mention the timing of diagnosis explicitly since the determining constraints for duration of infectiousness are the start and effectiveness of treatment, and health care providers may treat empirically in the absence of a laboratory diagnosis. To note, while culture-based diagnosis is slower to obtain than PCR-based diagnosis, culture-based diagnosis allows assessment of the resistance profile of the bacteria. Therefore, the type of diagnosis could be related to the likelihood of receiving ineffective treatment, which we have not considered. If this is the case, we may overestimate the advantages of testing via PCR.

Based on expert opinion, we assumed that children sought medical care on the second day of illness. Since the interval between symptom onset and care-seeking was assumed to be the same for children undergoing Treatments A-C, a delay in seeking medical care would shift the date of return to school equally forward for children undergoing treatment A-C relative to D. This would not affect the relative differences among most exclusion policies as applied to children undergoing most treatments. The only change in our estimates would be a relative increase in childcare-days lost per child to Treatment D, for policies requiring 24 h symptom-free or waiting 14 days after symptom onset. On another note, for the exclusion policy involving two consecutive tests, we assumed that the second test would be conducted upon receipt of the results of the first test. If the second test was conducted before the results of the first test were available, the number of childcare-days lost per child could be marginally less than we estimated.

Notably, these findings reflect scenarios with known shigellosis, such as during a shigellosis outbreak with laboratory-confirmed and epidemiologically linked cases. We assumed that the time-to-negative is equal among PCR and culture tests, which may not be the case. PCR tests may detect Shigella DNA after the bacteria are no longer viable, which would prolong childcare exclusion time, adding indirect costs to what is already a more expensive laboratory test (PCR tests cost about $27–$47; cultures cost about $9–$12 in 2015 USD [14, 15]).

Conclusions

Of the policies analyzed, exclusion policies that most effectively reduced risk of shigellosis transmission in childcare settings included the use of PCR-based tests. Our estimation suggested that the type of test (PCR or culture) was more relevant than the number of tests performed, with one PCR test being more effective than 2 cultures. The performance of policies based on fixed time intervals (i.e., waiting 14 days after onset, 7 days after beginning treatment, or 24 h without symptoms) was a function of the effectiveness and timing of treatments. Given substantial uncertainty in treatment effectiveness, comparison with other policies should be made cautiously. We hope public health officials can use these findings to establish childcare exclusion policies that effectively interrupt disease transmission while minimizing economic consequences.

Abbreviations

DNA:

Deoxyribonucleic acid

PCR:

Polymerase chain reaction

References

  1. Centers for Disease Control and Prevention. Shigella - Shigellosis. Available from: http://www.cdc.gov/shigella/general-information.html. Accessed 12 Feb 2019.

  2. Centers for Disease Control and Prevention. Outbreaks of multidrug-resistant Shigella sonnei gastroenteritis associated with day care centers--Kansas, Kentucky, and Missouri, 2005. MMWR. Morb Mortal Wkly Rep. 2006;55(39):1068.

    Google Scholar 

  3. Garrett V, Bornschlegel K, Lange D, Reddy V, Kornstein L, Kornblum J, et al. A recurring outbreak of Shigella sonnei among traditionally observant Jewish children in new York City: the risks of daycare and household transmission. Epidemiol Infect. 2006;134(06):1231–6.

    Article  CAS  Google Scholar 

  4. Jonsson J, Alvarez-Castillo MC, Sanz J, Ramiro R, Ballester E, Fernánez M, et al. Late detection of a shigellosis outbreak in a school in Madrid. Euro surveillance: bulletin europeen Sur les maladies transmissibles. Eur Commun Dis Bull. 2005;10(10):268–70.

    CAS  Google Scholar 

  5. Dabke G, Le Menach A, Black A, Gamblin J, Palmer M, Boxall N, et al. Duration of shedding of Verocytotoxin-producing Escherichia coli in children and risk of transmission in childcare facilities in England. Epidemiol Infect. 2014;142(02):327–34.

    Article  CAS  Google Scholar 

  6. Belongia EA, Osterholm MT, Soler JT, Ammend DA, Braun JE, MacDonald KL. Transmission of Escherichia coli 0157: H7 infection in Minnesota child day-care facilities. Jama. 1993;269(7):883–8.

    Article  CAS  Google Scholar 

  7. Centers for Disease Control and Prevention. Ciprofloxacin- and Azithromycin-Nonsusceptible Shigellosis in the United States. Health Alert Network, Official CDC Health Advisory; 2015. Available from: http://emergency.cdc.gov/han/han00379.asp. Accessed 12 Feb 2019.

  8. Matussek A, Jernberg C, Einemo I-M, Monecke S, Ehricht R, Engelmann I, et al. Genetic makeup of Shiga toxin-producing Escherichia coli in relation to clinical symptoms and duration of shedding: a microarray analysis of isolates from Swedish children. Eur J Clin Microbiol Infect Dis. 2017;36(8):1433–41.

    Article  CAS  Google Scholar 

  9. Kimberlin DW, Brady MT, Jackson MA, Long SS. Red book: 2018 report of the committee on infectious diseases. In: 31st ed Itasca, IL: American Academy of Pediatrics; 2018. p. 723–7.

    Google Scholar 

  10. Hoffman RE, Shillam PJ. The use of hygiene, cohorting, and antimicrobial therapy to control an outbreak of shigellosis. Am J Dis Child. 1990;144(2):219–21.

    CAS  PubMed  Google Scholar 

  11. Lerman Y, Yavzori M, Ambar R, Sechter I, Wiener M, Cohen D. Epidemic spread of Shigella sonnei shigellosis and evidence for development of immunity among children attending day-care centers in a communal settlement (kibbutz). J Clin Microbiol. 1994;32(4):1092–4.

    CAS  PubMed  PubMed Central  Google Scholar 

  12. Antibiotic Resistance Threats in the United States, 2013. Available from: http://www.cdc.gov/drugresistance/threat-report-2013/index.html. Accessed 12 Feb 2019.

  13. Karlsson MS, Bowen A, Reporter R, Folster JP, Grass JE, Howie RL, et al. Outbreak of infections caused by Shigella sonnei with reduced susceptibility to azithromycin in the United States. Antimicrob Agents Chemother. 2013;57(3):1559–60.

    Article  CAS  Google Scholar 

  14. Centers for Medicare & Medicaid services. 2015 clinical laboratory fee schedule (CLAB) public use file (PUF). Available from: https://www.cms.gov/apps/ama/license.asp?file=/Medicare/Medicare-Fee-for-Service-Payment/ClinicalLabFeeSched/Downloads/15CLAB.zip. Accessed 12 Feb 2019.

  15. Centers for Medicare & Medicaid Services. Clinical Laboratory Fee Schedule Files 2015. Available from: https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/ClinicalLabFeeSched/Clinical-Laboratory-Fee-Schedule-Files.html. Accessed 12 Feb 2019.

  16. Shane AL, Tucker NA, Crump JA, Mintz ED, Painter JA. Sharing Shigella: risk factors for a multicommunity outbreak of shigellosis. Arch Pediatr Adolesc Med. 2003;157(6):601–3.

    Article  Google Scholar 

  17. Weissman JB, Schmerler A, Gangarosa EJ, Marier RL, Lewis JN. Shigellosis in day-care centres. Lancet. 1975;305(7898):88–90.

    Article  Google Scholar 

  18. Dutta S, Chatterjee A, Dutta P, Rajendran K, Roy S, Pramanik K, et al. Sensitivity and performance characteristics of a direct PCR with stool samples in comparison to conventional techniques for diagnosis of Shigella and enteroinvasive Escherichia coli infection in children with acute diarrhoea in Calcutta, India. J Med Microbiol. 2001;50(8):667–74.

    Article  CAS  Google Scholar 

  19. Gaudio PA, Sethabutr O, Echeverria P, Hoge CW. Utility of a polymerase chain reaction diagnostic system in a study of the epidemiology of shigellosis among dysentery patients, family contacts, and well controls living in a shigellosis-endemic area. J Infect Dis. 1997;176(4):1013–8.

    Article  CAS  Google Scholar 

  20. Islam M, Hossain M, Hasan M, Rahman M, Fuchs G, Mahalanabis D, et al. Detection of Shigellae from stools of dysentery patients by culture and polymerase chain reaction techniques. J Diarrhoeal Dis Res. 1998;16(4):248–51.

Download references

Acknowledgements

Funding

This study was developed as part of the job duties of the authors, and no special funding was required for this study.

Availability of data and materials

All the data necessary to reproduce the results can be found in the manuscript’s tables; the calculations can be reproduced using the manuscript’s appendixes.

Author information

Authors and Affiliations

Authors

Contributions

CC and EAU performed the calculations. CC wrote the first draft of the manuscript. EAU revised the first version extensively. JH and AB provided the data inputs. EBK, MIM, and AB provided senior guidance. All authors read, made significant edits to the first version, and approved the final manuscript.

Corresponding author

Correspondence to Cristina Carias.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests..

Publisher’s Note

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

Additional information

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Additional files

Additional file 1:

Appendix A. "Methods and assumptions of the tool to evaluate the impact of exclusion policies associated with shigellosis". In this supplement, we provide further detail on the methods and assumptions used in the manuscript. (DOCX 81 kb)

Additional file 2:

Appendix B. "A tool for the evaluation of shigellosis exclusion policies". In this supplement, the user may change the assumptions and see how that affects the results. (XLSX 122 kb)

Additional file 3:

Appendix C. "Additional aggregate results for aggregated childcare days lost for various treatment mixes among the sample population potentially affected by a shigellosis outbreak". In this supplement, we provide additional results that can be calculated using the methods described in the paper, for a population of children. (DOCX 231 kb)

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. 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.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Carias, C., Undurraga, E.A., Hurd, J. et al. Evaluation of the impact of shigellosis exclusion policies in childcare settings upon detection of a shigellosis outbreak. BMC Infect Dis 19, 172 (2019). https://doi.org/10.1186/s12879-019-3796-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12879-019-3796-7

Keywords