Academic Journal
View in EDS
PDF Full Text
HTML Full Text

Sensitivity of contact-tracing for COVID-19 in Thailand: a capture-recapture application

Bibliographic Details
Title: Sensitivity of contact-tracing for COVID-19 in Thailand: a capture-recapture application
Authors: R. Lerdsuwansri, P. Sangnawakij, D. Böhning, C. Sansilapin, W. Chaifoo, Jonathan A. Polonsky, Victor J. Del Rio Vilas
Source: BMC Infectious Diseases, Vol 22, Iss 1, Pp 1-10 (2022)
Publisher Information: BMC, 2022.
Publication Year: 2022
Collection: LCC:Infectious and parasitic diseases
Subject Terms: COVID-19, Contact tracing, Thailand, Capture-recapture, Sensitivity, Infectious and parasitic diseases, RC109-216
More Details: Abstract Background We investigate the completeness of contact tracing for COVID-19 during the first wave of the COVID-19 pandemic in Thailand, from early January 2020 to 30 June 2020. Methods Uni-list capture-recapture models were applied to the frequency distributions of index cases to inform two questions: (1) the unobserved number of index cases with contacts, and (2) the unobserved number of index cases with secondary cases among their contacts. Results Generalized linear models (using Poisson and logistic families) did not return any significant predictor (age, sex, nationality, number of contacts per case) on the risk of transmission and hence capture-recapture models did not adjust for observed heterogeneity. Best fitting models, a zero truncated negative binomial for question 1 and zero-truncated Poisson for question 2, returned sensitivity estimates for contact tracing performance of 77.6% (95% CI = 73.75–81.54%) and 67.6% (95% CI = 53.84–81.38%), respectively. A zero-inflated negative binomial model on the distribution of index cases with secondary cases allowed the estimation of the effective reproduction number at 0.14 (95% CI = 0.09–0.22), and the overdispersion parameter at 0.1. Conclusion Completeness of COVID-19 contact tracing in Thailand during the first wave appeared moderate, with around 67% of infectious transmission chains detected. Overdispersion was present suggesting that most of the index cases did not result in infectious transmission chains and the majority of transmission events stemmed from a small proportion of index cases.
Document Type: article
File Description: electronic resource
Language: English
ISSN: 1471-2334
Relation: https://doaj.org/toc/1471-2334
DOI: 10.1186/s12879-022-07046-6
Access URL: https://doaj.org/article/a6e621a9b1e0458988723246900a4467
Accession Number: edsdoj.6e621a9b1e0458988723246900a4467
Database: Directory of Open Access Journals
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHjPtM4BHU3ZchRwgzYmadcigk49r9CVlbU7V5F6lgH7WwGuZc9vbCTRlxkAtA6EQz5yAAAA4jCB3wYJKoZIhvcNAQcGoIHRMIHOAgEAMIHIBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDCXZaLuEy3Ps_4B2RgIBEICBmvAabHYNQrJc-4RhqWPjjrQXHxtxJ5UPYLsYgBPC09L-phh1qCmZtPzs4rqKUEfy6Fa009dMuGJS742fs2oDk6V6PXoGWMDUc0LQEWRBjAWjRtGLAlrQPDaVCFknW5dY7t7_tMcI3fisKOoU5rBQuQAY1U4qvsrLGaOrfLjojqXS2RpZrZUXgwJuTG2Lbx_k8juvtWv0yP1d8wU=
Text:
  Availability: 1
  Value: <anid>AN0154978801;[1chu]29jan.22;2022Feb02.06:23;v2.2.500</anid> <title id="AN0154978801-1">Sensitivity of contact-tracing for COVID-19 in Thailand: a capture-recapture application </title> <p>Background: We investigate the completeness of contact tracing for COVID-19 during the first wave of the COVID-19 pandemic in Thailand, from early January 2020 to 30 June 2020. Methods: Uni-list capture-recapture models were applied to the frequency distributions of index cases to inform two questions: (<reflink idref="bib1" id="ref1">1</reflink>) the unobserved number of index cases with contacts, and (<reflink idref="bib2" id="ref2">2</reflink>) the unobserved number of index cases with secondary cases among their contacts. Results: Generalized linear models (using Poisson and logistic families) did not return any significant predictor (age, sex, nationality, number of contacts per case) on the risk of transmission and hence capture-recapture models did not adjust for observed heterogeneity. Best fitting models, a zero truncated negative binomial for question 1 and zero-truncated Poisson for question 2, returned sensitivity estimates for contact tracing performance of 77.6% (95% CI = 73.75–81.54%) and 67.6% (95% CI = 53.84–81.38%), respectively. A zero-inflated negative binomial model on the distribution of index cases with secondary cases allowed the estimation of the effective reproduction number at 0.14 (95% CI = 0.09–0.22), and the overdispersion parameter at 0.1. Conclusion: Completeness of COVID-19 contact tracing in Thailand during the first wave appeared moderate, with around 67% of infectious transmission chains detected. Overdispersion was present suggesting that most of the index cases did not result in infectious transmission chains and the majority of transmission events stemmed from a small proportion of index cases.</p> <p>Keywords: COVID-19; Contact tracing; Thailand; Capture-recapture; Sensitivity</p> <hd id="AN0154978801-2">Background</hd> <p>Following the notification of the first COVID-19 cases in Thailand on 11 January 2020, the Department of Disease Control (DDC), Ministry of Public Health Thailand started recording essential information to monitor the epidemic. By early May 2020, the epidemic had receded from a daily peak of 188 cases in mid-March 2020 to single digit daily counts. The first wave of the epidemic was under control. At the time of writing Thailand was experiencing a second wave that started in early December 2020, with a cumulative number of just over 26,000 cases as of 5 March 2021 (https://ddc.moph.go.th/viralpneumonia/eng/index.php).</p> <p>Thailand's successful initial response to COVID-19 was aided by a strong national capacity to trace and quarantine contacts using Rapid Response Teams and Village Health Volunteers who were trained during earlier major infectious disease outbreaks such as H1N1, SARS, and Avian Influenza [[<reflink idref="bib1" id="ref3">1</reflink>]]. Despite the prompt reaction by local health authorities, the Intra-Action-Review (IAR) on Thailand's response to COVID-19 highlighted the need for a sensitive COVID-19 surveillance system to facilitate detection of individual cases, small clusters and monitor trends [[<reflink idref="bib3" id="ref4">3</reflink>]].</p> <p>Contact tracing (CT) aims to identify, assess and manage contacts exposed to disease to prevent onward transmission [[<reflink idref="bib4" id="ref5">4</reflink>]]. In this capacity, CT remains a critical function towards the control of infectious diseases. Similar to other surveillance efforts, sensitivity, or the ability to detect all the events of interest, is one of the most relevant technical attributes towards the assessment of CT performance [[<reflink idref="bib5" id="ref6">5</reflink>]]. For COVID-19, timeliness and sensitivity are the most cited performance attributes [[<reflink idref="bib6" id="ref7">6</reflink>]]. Whereas timeliness can be directly measured (and it is normally decomposed in multiple metrics to reflect the many steps in the flow of information and biological samples that constitute the surveillance system), that is not the case for sensitivity. Several approaches towards its estimation have been suggested [[<reflink idref="bib5" id="ref8">5</reflink>], [<reflink idref="bib7" id="ref9">7</reflink>]]. Here we focus on capture-recapture (CRC) models [[<reflink idref="bib8" id="ref10">8</reflink>]]. Broadly, this family of methodological approaches estimates the number of individuals missing from identifying mechanisms such as disease surveillance systems (SS). The estimation of the SS sensitivity and probability of event detection follows.</p> <p>CRC approaches have been extensively used to estimate disease SS sensitivity [[<reflink idref="bib10" id="ref11">10</reflink>]]. Specifically on CT, Polonsky and colleagues applied uni-list CRC models to Ebola Virus Disease (EVD) data from the 2018–2020 EVD outbreak in North Kivu Province, Democratic Republic of the Congo (DRC) [[<reflink idref="bib11" id="ref12">11</reflink>]]. The authors addressed two specific questions: (<reflink idref="bib1" id="ref13">1</reflink>) what is the true number of index cases with unobserved contacts (in effect assessing the sensitivity of contact identification efforts), and (<reflink idref="bib2" id="ref14">2</reflink>) what is the true number of index cases with secondary cases among their contacts (in effect assessing the sensitivity of case detection among contacts). CRC approaches, on country aggregated case data, were also applied to estimate the true number of COVID-19 infections, estimated to be three to eight times larger than those reported [[<reflink idref="bib12" id="ref15">12</reflink>]].</p> <p>Here we first describe Thailand's first wave of COVID-19 CT data, and then the application of uni-list CRC models to quantify the number of unobserved index cases, and CT sensitivity. Specifically, we aim to answer the following: question (<reflink idref="bib1" id="ref16">1</reflink>) how many index cases with contacts were missed by CT, and question (<reflink idref="bib2" id="ref17">2</reflink>) how many index cases with infected contacts were missed by the CT mechanism.</p> <hd id="AN0154978801-3">Methods</hd> <p></p> <hd id="AN0154978801-4">Materials</hd> <p>Our data stems from Thailand's regular COVID-19 CT operations. Figure 1 presents a flowchart of the contact tracing process undertaken by the local communicable disease control units (CDCU) and joint investigation teams (JIT) from DDC. Once the patient is diagnosed as being infected with SARS-CoV-2, so called the confirmed case, contact tracing will be conducted to obtain the list of contacts. The identified contacts are classified as either high-risk contacts or low-risk contact following investigation guidelines [[<reflink idref="bib13" id="ref18">13</reflink>]]. High-risk contact is defined as a contact who is more likely to contract the virus through exposure to respiratory secretions of the confirmed case while not wearing PPE according to standard precautions. Low-risk contact is defined as a contact who is less likely to contract the virus from the confirmed case. This includes contacts who have not met the definition for high-risk contact. Only high-risk contacts were quarantined in the designated places and basic demographic information such as age, sex, and nationality were collected and recorded in the contact form. Our data set comprises the period 11 January 2020 to 30 June 2020. A total of 352 cases were identified through contact tracing system leading to 6359 high risk contacts and 4299 low risk contacts.</p> <p>Graph: Fig. 1 A brief flowchart of the contact tracing process undertaken by the local communicable disease control units (CDCU) and the Department of Disease Control (Compiled by the authors)</p> <hd id="AN0154978801-5">Initial analysis</hd> <p>We describe the data according to the available demographic predictors associated with the index cases (age, sex, and nationality) and the number of contacts per index case. We applied logistic regression to assess whether any of the above predictors (with age as a continuous variable), had any effect on the probability of identifying secondary cases. Using the best fitting models of the count distributions (see next section), we regressed the observed covariates on the number of contacts per index case with at least one contact (<emph>n</emph> = 341) to assess whether we should adjust for covariates in our capture-recapture calculations. We also applied a zero-inflated negative binomial model. By the full distribution of all index cases we mean the following: out of the 352 index cases only 30 had infectious contacts (secondary infections), namely 16 index cases had 1 infected contact, 9 had 2, 4 had 3 and 1 had 4 infectious contacts. The large number of zeros is reflected in the zero-inflated negative-binomial modeling which adds simply an additional parameter just for those with zero infectious contacts.</p> <hd id="AN0154978801-6">Capture-recapture modelling</hd> <p>We are interested in deriving an estimate of the unknown true number of COVID-19 cases with contacts that entered the CT mechanism. This would address question 1 (Q1) as above. The tracing of contacts is likely to lead to the identification of secondary infections for a subset of index cases. This data informs question 2 (Q2). For both questions, the data can be binned into the number of index cases with one listed (Q1) or infected (Q2) contact ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>1</mn></msub></math> </ephtml> ), two listed or infected contacts ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>2</mn></msub></math> </ephtml> ), and so on up to the number of index cases with the maximum number of listed or infected contacts ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mi>m</mi></msub></math> </ephtml> ). Here, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>0</mn></msub></math> </ephtml> , the frequency of index cases with unobserved contacts (for Q1) or unobserved infected contacts (for Q2) is unknown and the target of the inference. Statistically, the identification process leads to a zero-truncated count distribution of cases with at least one listed or infected contact, i.e. with positive integers (ones, twos, threes, etc.), but no zeros. By applying CRC approaches, we can infer <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>0</mn></msub></math> </ephtml> , the number of unobserved cases with at least one listed or infected contact.</p> <p>For both questions, we fit parametric models (Poisson, Negative Binomial, and Geometric) to the observed counts using the maximum likelihood method. Then, the smallest Akaike and Bayesian Information Criterion (AIC and BIC, respectively) are used for model selection. After estimating model parameters, we can estimate <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>0</mn></msub></math> </ephtml> as</p> <p>1 <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mover accent="true"><mi>f</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub><mo>=</mo><mfrac><mrow><mi>n</mi><msub><mi>p</mi><mn>0</mn></msub></mrow><mrow><mn>1</mn><mo>-</mo><msub><mi>p</mi><mn>0</mn></msub></mrow></mfrac><mo>,</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml></p> <p>Graph</p> <p>where <emph>n</emph> is the observed sample size and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>p</mi><mn>0</mn></msub></math> </ephtml> is the estimated probability of missing an index case with non-zero contacts as computed from the models. The population size estimator <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mo>=</mo><mi>n</mi><mo>+</mo><msub><mover accent="true"><mi>f</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub></mrow></math> </ephtml> follows.</p> <p>In addition to the model-based estimators we consider two further alternatives for comparison purposes: the Turing's estimator [[<reflink idref="bib14" id="ref19">14</reflink>]] and Chao's lower bound estimator [[<reflink idref="bib15" id="ref20">15</reflink>]]. Turing's estimator is formulated under a homogeneous Poisson distribution with parameter <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mi>λ</mi></math> </ephtml> . Let <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>p</mi><mn>0</mn></msub></math> </ephtml> be the probability of zero count or missing an observation. We have</p> <p>2 <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mi>p</mi><mn>0</mn></msub><mo>=</mo><msup><mi>e</mi><mrow><mo>-</mo><mi>λ</mi></mrow></msup><mo>=</mo><mfrac><mrow><msup><mi>e</mi><mrow><mo>-</mo><mi>λ</mi></mrow></msup><mi>λ</mi></mrow><mi>λ</mi></mfrac><mo>=</mo><mfrac><msub><mi>p</mi><mn>1</mn></msub><mi>λ</mi></mfrac><mo>.</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml></p> <p>Graph</p> <p>The estimate of <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>p</mi><mn>0</mn></msub></math> </ephtml> can be calculated from observed frequencies as follows</p> <p>3 <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mover accent="true"><mi>p</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub><mo>=</mo><mfrac><mrow><msub><mi>f</mi><mn>1</mn></msub><mo stretchy="false">/</mo><mi>N</mi></mrow><mrow><mi>S</mi><mo stretchy="false">/</mo><mi>N</mi></mrow></mfrac><mo>=</mo><mfrac><msub><mi>f</mi><mn>1</mn></msub><mi>S</mi></mfrac><mo>,</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml></p> <p>Graph</p> <p>where <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>S</mi><mo>=</mo><msubsup><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>0</mn></mrow><mi>m</mi></msubsup><mi>i</mi><msub><mi>f</mi><mi>i</mi></msub></mrow></math> </ephtml> . Thus, Turing's estimator for estimating the population size is given by</p> <p>4 <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mrow><mi mathvariant="italic">Turing</mi></mrow></msub><mo>=</mo><mfrac><mi>n</mi><mrow><mn>1</mn><mo>-</mo><msub><mi>f</mi><mn>1</mn></msub><mo stretchy="false">/</mo><mi>S</mi></mrow></mfrac><mo>.</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml></p> <p>Graph</p> <p>Chao (1987) suggested a mixed Poisson model with <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>p</mi><mi>i</mi></msub><mo>=</mo><msubsup><mo>∫</mo><mn>0</mn><mo movablelimits="true">inf</mo></msubsup><mfrac><mrow><msup><mi>e</mi><mrow><mo>-</mo><mi>λ</mi></mrow></msup><msup><mi>λ</mi><mi>i</mi></msup></mrow><mrow><mi>i</mi><mo>!</mo></mrow></mfrac><mi>g</mi><mrow><mo stretchy="false">(</mo><mi>λ</mi><mo stretchy="false">)</mo></mrow><mi>d</mi><mi>λ</mi></mrow></math> </ephtml> for <emph>i</emph> = 0, 1, 2,... and arbitrary density <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>g</mi><mo stretchy="false">(</mo><mi>λ</mi><mo stretchy="false">)</mo></mrow></math> </ephtml> [[<reflink idref="bib15" id="ref21">15</reflink>]]. Chao's estimator incorporates not only the unobserved heterogeneity in the Poisson parameter but also leads to a very simple nonparametric estimator by applying the Cauchy–Schwarz inequality to the lower bound for the probability of a not observed event</p> <p>5 <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mi>p</mi><mn>0</mn></msub><mo>≥</mo><mfrac><msubsup><mi>p</mi><mn>1</mn><mn>2</mn></msubsup><mrow><mn>2</mn><msub><mi>p</mi><mn>2</mn></msub></mrow></mfrac><mo>.</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml></p> <p>Graph</p> <p>Replacing these probabilities by observed frequencies, the lower bound for the estimate of zero counts is computed as <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mover accent="true"><mi>f</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub><mo>≥</mo><msubsup><mi>f</mi><mn>1</mn><mn>2</mn></msubsup><mo stretchy="false">/</mo><mrow><mo stretchy="false">(</mo><mn>2</mn><msub><mi>f</mi><mn>2</mn></msub><mo stretchy="false">)</mo></mrow></mrow></math> </ephtml> . As a result, Chao's lower bound estimator for the population size is</p> <p>6 <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mrow><mi mathvariant="italic">Chao</mi></mrow></msub><mo>=</mo><mi>n</mi><mo>+</mo><mfrac><msubsup><mi>f</mi><mn>1</mn><mn>2</mn></msubsup><mrow><mn>2</mn><msub><mi>f</mi><mn>2</mn></msub></mrow></mfrac><mo>.</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml></p> <p>Graph</p> <p>Clearly, (<reflink idref="bib6" id="ref22">6</reflink>) uses only part of the available information, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>1</mn></msub></math> </ephtml> and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>2</mn></msub></math> </ephtml> , as opposed to Turing estimator that uses all the information in the sample by means of <emph>S</emph>. In addition, a mixing distribution <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>g</mi><mo stretchy="false">(</mo><mi>λ</mi><mo stretchy="false">)</mo></mrow></math> </ephtml> is not required to be specified and estimated showing the non-parametric nature of this estimator.</p> <hd id="AN0154978801-7">Confidence Interval for the unknown population size</hd> <p>To estimate 95% confidence intervals (95% CIs), we use resampling techniques as described in the CRC literature [[<reflink idref="bib16" id="ref23">16</reflink>]]. Suppose that <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover></math> </ephtml> is the estimated size under a fitted model. Then, we generate <emph>B</emph> samples of size <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover></math> </ephtml> using the fitted model and its estimated parameter(s). For each sample, all zeros are truncated and the size estimate <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mi>b</mi></msub></math> </ephtml> computed, for each of the samples <emph>b</emph> = 1, 2,..., <emph>B</emph> leading to a sample of estimates <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mn>1</mn></msub></math> </ephtml> , <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mn>2</mn></msub></math> </ephtml> ,..., <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mi>B</mi></msub></math> </ephtml> . We choose <emph>B</emph> = 10,000 to minimize bootstrap simulation random error, and then use two methods towards CI construction:</p> <p></p> <ulist> <item> <bold> The normal approximation method, using the median a robust estimator for the mean where _HT_ <math xmlns="<ulink href="http://www.w3.org/1998/Math/MathML">http://www.w3.org/1998/Math/MathML</ulink>"><mover accent="true"><mrow><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover></mrow><mrow><mo stretchy="false">¯</mo></mrow></mover></math> _ht_ = median( _HT_ <math xmlns="<ulink href="http://www.w3.org/1998/Math/MathML">http://www.w3.org/1998/Math/MathML</ulink>"><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mi>b</mi></msub></math> _ht_ | _I</bold>i_ = 1, 2,..., <emph>B</emph>) and calculate the bootstrap standard error as</item> <p></p> <item> 7 <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><mi>S</mi><mi>E</mi><mo>=</mo><msqrt><mrow><mtext>median</mtext><mo stretchy="false">(</mo><msup><mrow><mo stretchy="false">(</mo><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mi>b</mi></msub><mo>-</mo><mover accent="true"><mrow><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover></mrow><mrow><mo stretchy="false">¯</mo></mrow></mover><mo stretchy="false">)</mo></mrow><mn>2</mn></msup><mo stretchy="false">|</mo><mi>b</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>,</mo><mo>...</mo><mo>,</mo><mi>B</mi></mrow></msqrt><mrow><mo stretchy="false">)</mo><mo>.</mo></mrow></mrow></mtd></mtr></mtable></mrow></math> </ephtml></item> </ulist> <p>Graph</p> <p></p> <ulist> <item> The 95% confidence interval for the true population size can then be constructed by means of <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mo>±</mo><mn>1.96</mn><mo>×</mo><mi>S</mi><mi>E</mi></mrow></math> </ephtml> .</item> <p></p> <item> The percentile method where we use the 2.5th percentile of the distribution of <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mi>b</mi></msub></math> </ephtml> as the lower end and the 97.5th percentile as the upper end.</item> </ulist> <hd id="AN0154978801-8">Results</hd> <p></p> <hd id="AN0154978801-9">Descriptive analyses</hd> <p>Graph: Fig. 2 Epidemic curve of the COVID-19 outbreak in Thailand from 4 Jan 2020 to 30 June 2020 (3171 confirmed cases)</p> <p>In the period 4 Jan 2020 to 30 June 2020, 3171 cases were confirmed (Fig. 2). Of those, 352 (11.1%) index cases were followed through CT leading to the identification of subsequent contacts for 341 of them. Among these 341 index cases with non-zero contacts from which 6,359 high risk contacts were listed, there were 44 index cases with one contact ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>1</mn></msub></math> </ephtml> = 44), 22 with two contacts ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>2</mn></msub></math> </ephtml> = 22), 24 with three contacts ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>3</mn></msub></math> </ephtml> = 24), and so on. Table 1 shows the complete distribution of index cases with traced contacts for the first 50 index cases. For infected contacts, the complete distribution is as follows: index cases with one infected contact ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>1</mn></msub></math> </ephtml> = 16), two infected contacts ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>2</mn></msub></math> </ephtml> = 9), three infected contacts ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>3</mn></msub></math> </ephtml> = 4), and four infected contacts ( <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>4</mn></msub></math> </ephtml> = 1).</p> <p>Table 1 Frequency distribution of counts of index cases with contacts (only first 50 counts)</p> <p> <ephtml> <table frame="hsides" rules="groups"><tbody><tr><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>1</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq36.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>2</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq37.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>3</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq38.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>4</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq39.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>5</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq40.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>6</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq41.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>7</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq42.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>8</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq43.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>9</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq44.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mrow xmlns=""><msub><mi>f</mi><mn>1</mn></msub><mn>0</mn></mrow></math><inline-graphic href="12879_2022_7046_Article_IEq45.gif" /></p></td></tr><tr><td align="left"><p>44</p></td><td align="left"><p>22</p></td><td align="left"><p>24</p></td><td align="left"><p>16</p></td><td align="left"><p>15</p></td><td align="left"><p>10</p></td><td align="left"><p>11</p></td><td align="left"><p>10</p></td><td align="left"><p>9</p></td><td align="left"><p>9</p></td></tr><tr><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>11</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq46.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>12</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq47.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>13</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq48.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>14</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq49.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>15</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq50.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>16</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq51.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>17</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq52.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>18</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq53.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>19</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq54.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>20</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq55.gif" /></p></td></tr><tr><td align="left"><p>9</p></td><td align="left"><p>10</p></td><td align="left"><p>6</p></td><td align="left"><p>7</p></td><td align="left"><p>8</p></td><td align="left"><p>16</p></td><td align="left"><p>3</p></td><td align="left"><p>2</p></td><td align="left"><p>3</p></td><td align="left"><p>8</p></td></tr><tr><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>21</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq56.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>22</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq57.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>23</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq58.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>24</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq59.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>25</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq60.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>26</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq61.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>27</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq62.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>28</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq63.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>29</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq64.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>30</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq65.gif" /></p></td></tr><tr><td align="left"><p>5</p></td><td align="left"><p>3</p></td><td align="left"><p>5</p></td><td align="left"><p>6</p></td><td align="left"><p>1</p></td><td align="left"><p>6</p></td><td align="left"><p>1</p></td><td align="left"><p>2</p></td><td align="left"><p>4</p></td><td align="left"><p>5</p></td></tr><tr><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>31</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq66.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>32</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq67.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>33</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq68.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>34</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq69.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>35</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq70.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>36</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq71.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>37</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq72.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>38</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq73.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>39</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq74.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>40</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq75.gif" /></p></td></tr><tr><td align="left"><p>4</p></td><td align="left"><p>4</p></td><td align="left"><p>1</p></td><td align="left"><p>3</p></td><td align="left"><p>2</p></td><td align="left"><p>2</p></td><td align="left"><p>1</p></td><td align="left"><p>1</p></td><td align="left"><p>1</p></td><td align="left"><p>6</p></td></tr><tr><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>41</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq76.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>42</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq77.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>43</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq78.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>44</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq79.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>45</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq80.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>46</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq81.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>47</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq82.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>48</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq83.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>49</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq84.gif" /></p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mi>f</mi><mn>50</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq85.gif" /></p></td></tr><tr><td align="left"><p>0</p></td><td align="left"><p>3</p></td><td align="left"><p>1</p></td><td align="left"><p>1</p></td><td align="left"><p>1</p></td><td align="left"><p>2</p></td><td align="left"><p>0</p></td><td align="left"><p>1</p></td><td align="left"><p>0</p></td><td align="left"><p>2</p></td></tr></tbody></table> </ephtml> </p> <p>Of the 341 index cases with at least one contact, 196 (57.48%) were males and 145 (42.52%) were females. At a 5% level of significance, there was sufficient evidence to conclude that there was a difference between the proportions of these contacts from male and female (goodness-of-fit Chi-square test with P-value = 0.007). See more details of the goodness-of-fit test in [<reflink idref="bib18" id="ref24">18</reflink>]. The median age was 37 years (mean = 39.62, interquartile range (IQR) = 28–50, min = 0.3, max = 83). The statistics of age by gender are given in the following:</p> <p></p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>Age of male</p></th><th align="left"><p>Age of female</p></th></tr></thead><tbody><tr><td align="left"><p>Median</p></td><td align="left"><p>40</p></td><td align="left"><p>34</p></td></tr><tr><td align="left"><p>IQR</p></td><td align="left"><p>29–52</p></td><td align="left"><p>26–47</p></td></tr></tbody></table> </ephtml> </p> <p>These showed median age for males was significantly greater than that of females (Wilcoxon signed-rank test with P-value = 0.004) [[<reflink idref="bib18" id="ref25">18</reflink>]]. The vast majority of cases (<reflink idref="bib290" id="ref26">290</reflink>, 85.04%) cases were Thai. Meanwhile, 51 (14.96%) were foreign nationals: 26 cases (7.62%) from China, 5 cases (1.46%) from Japan, 4 cases (1.17%) from Denmark, and 61 cases from other locations.</p> <p>From 341 index cases with non-zero contacts noted before, 30 (8.8%) index cases had at least one infected contact. The median age of this set of index cases was 44 years (mean <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mo>=</mo><mn>42.87</mn></mrow></math> </ephtml> , IQR <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mo>=</mo><mn>29.25</mn></mrow></math> </ephtml> –56, min <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mo>=</mo><mn>6</mn></mrow></math> </ephtml> and max <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mo>=</mo><mn>80</mn></mrow></math> </ephtml> ). Summary statistics of age by gender are concluded as follows:</p> <p></p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>Age of male</p></th><th align="left"><p>Age of female</p></th></tr></thead><tbody><tr><td align="left"><p>Median</p></td><td align="left"><p>45.5</p></td><td align="left"><p>36</p></td></tr><tr><td align="left"><p>IQR</p></td><td align="left"><p>38.5–46.75</p></td><td align="left"><p>28.25–46.75</p></td></tr></tbody></table> </ephtml> </p> <p>Furthermore, almost all index cases with infected contacts were Thai (28 cases, 93.33%). We also show summary statistics for the set of 30 index cases with infected contacts in Table 2.</p> <p>Table 2 Summary statistics for the 30 index cases with infected contacts</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Number of non-zero contacts</p></th><th align="left"><p>1</p></th><th align="left"><p>2</p></th><th align="left"><p>3</p></th><th align="left"><p>4</p></th></tr></thead><tbody><tr><td align="left"><p>Frequency</p></td><td align="left"><p>16</p></td><td align="left"><p>9</p></td><td align="left"><p>4</p></td><td align="left"><p>1</p></td></tr><tr><td align="left"><p>Median age</p></td><td align="left"><p>40.5</p></td><td align="left"><p>41</p></td><td align="left"><p>59.5</p></td><td align="left"><p>66</p></td></tr><tr><td align="left"><p>Minimum age</p></td><td align="left"><p>6</p></td><td align="left"><p>20</p></td><td align="left"><p>26</p></td><td align="left"><p>66</p></td></tr><tr><td align="left"><p>Maximum age</p></td><td align="left"><p>64</p></td><td align="left"><p>80</p></td><td align="left"><p>68</p></td><td align="left"><p>66</p></td></tr><tr><td align="left"><p>95% CIs for mean of age</p></td><td align="left"><p>32.4–48.42</p></td><td align="left"><p>26.24–53.76</p></td><td align="left"><p>23.47–83.03</p></td><td align="left"><p>66</p></td></tr><tr><td align="left"><p>%Female</p></td><td align="left"><p>50%</p></td><td align="left"><p>55.56%</p></td><td align="left"><p>25%</p></td><td align="left"><p>0</p></td></tr></tbody></table> </ephtml> </p> <p>A zero-truncated Poisson regression (best fitting model for this reduced dataset (30 observations)) showed no significant covariate effects on the number of contacts per index case with at least one secondary case. For the larger dataset of the number of contacts per index case, the best fitting model (a zero-truncated negative binomial regression) also showed no significant effects of the covariates. These results support that no consideration of observed heterogeneity in our capture-recapture models was required. The logistic regression showed no significant covariate effects either.</p> <p>The application of a zero-inflated negative binomial model to the full distribution of all index cases (<emph>n</emph> = 352 that includes all cases detected through the CT mechanism and their contacts; we note that for 311 of such cases there were zero infected contacts, hence the use of a zero-inflated model) allows the estimation of the average number of secondary infections over the course of the outbreak that equates to an average effective reproductive number (RE = 0.14; 95% CI: 0.09–0.22), and dispersion parameter (<emph>k</emph> = 0.1). We note that our estimate of the reproduction number applies to the entirety of the period under study and is sensitive to the implementation of several public and health social measures in country at different times.</p> <hd id="AN0154978801-10">Applications of capture-recapture models towards the estimation of contact tracing sensitivit...</hd> <p>Graph: Fig. 3 Frequency distribution of number of index cases with contacts (n = 341)</p> <p>Graph: Fig. 4 Frequency distribution of number of index cases with contacts (n = 341)</p> <p>As seen in Fig. 3, the distribution of index cases with contacts presents a long tail. Clearly, this long-tailed distribution is fitted a lot better by the negative binomial than by the Poisson distribution and the geometric distribution (see Fig. 4). The best fit (in effect addressing Q1) is given by a zero-truncated negative binomial model (Table 3) leading to an estimate of unobserved index cases with contacts of <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mover accent="true"><mi>f</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub></math> </ephtml> = 98.18, and an estimated size of the overall count of index cases with contacts of <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover></math> </ephtml> = 439.18. In the appendix we derive population size estimators for Turing and Chao approaches for the chosen negative binomial distribution, and in Table 4 we present the results of the three estimators including 95% confidence intervals. Note that Chao's estimator is slightly higher than the other two, indicating potential residual heterogeneity. However, this might be also still within random error variation as the confidence intervals in Table 4 express. Using the estimated <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>0</mn></msub></math> </ephtml> from the zero-truncated negative binomial model, we estimate the CT sensitivity to detect index cases with contacts as 341/(341 + 98) = 0.776 or 77.6%.</p> <p>Table 3 Model results and fit criteria for the count data of index cases with all contacts</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Model</p></th><th align="left"><p>Log-likelihood</p></th><th align="left"><p>AIC</p></th><th align="left"><p>BIC</p></th></tr></thead><tbody><tr><td align="left"><p>Poisson (<math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true" xmlns=""><mi>λ</mi><mo stretchy="false">^</mo></mover></math><inline-graphic href="12879_2022_7046_Article_IEq93.gif" /> = 18.648)</p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mo xmlns="">-</mo></math><inline-graphic href="12879_2022_7046_Article_IEq94.gif" />4663.285</p></td><td char="." align="char"><p>9328.569</p></td><td char="." align="char"><p>9332.401</p></td></tr><tr><td align="left"><p>Negative Binomial (mue = 14.479, size = 0.4197)</p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mo xmlns="">-</mo></math><inline-graphic href="12879_2022_7046_Article_IEq95.gif" /><bold>1304</bold>.<bold>133</bold></p></td><td char="." align="char"><p><bold>2612</bold>.<bold>266</bold></p></td><td char="." align="char"><p><bold>2619</bold>.<bold>93</bold></p></td></tr><tr><td align="left"><p>Geometric (<math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true" xmlns=""><mi>p</mi><mo stretchy="false">^</mo></mover></math><inline-graphic href="12879_2022_7046_Article_IEq96.gif" /> = 0.0536)</p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mo xmlns="">-</mo></math><inline-graphic href="12879_2022_7046_Article_IEq97.gif" />1329.368</p></td><td char="." align="char"><p>2660.735</p></td><td char="." align="char"><p>2664.567</p></td></tr></tbody></table> </ephtml> </p> <p>Bold values indicate that Negative Binomial is the best fit for the distribution of the number of index cases with contacts (n=341) due to the smallest AIC and BIC</p> <p>Table 4 Estimates of unobserved contacts, population size and 95% CI (<emph>n</emph> = 341) for the three approaches (based upon the negative-binomial model)</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Estimators</p></th><th align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mover accent="true"><mi>f</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq98.gif" /></p></th><th align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true" xmlns=""><mi>N</mi><mo stretchy="false">^</mo></mover></math><inline-graphic href="12879_2022_7046_Article_IEq99.gif" /></p></th><th align="left"><p>Bootstrap median</p></th><th align="left"><p>95% CI normal approximation</p></th><th align="left"><p>CI from percentile BT</p></th></tr></thead><tbody><tr><td align="left"><p>MLE</p></td><td char="." align="char"><p>98.18</p></td><td char="." align="char"><p>439.18</p></td><td char="." align="char"><p>438.271</p></td><td char="–" align="char"><p>402.07–474.47</p></td><td char="–" align="char"><p>391.25–510.05</p></td></tr><tr><td align="left"><p>Turing</p></td><td char="." align="char"><p>101.78</p></td><td char="." align="char"><p>442.78</p></td><td char="." align="char"><p>438.0309</p></td><td char="–" align="char"><p>401.35–474.72</p></td><td char="–" align="char"><p>390.74–511.09</p></td></tr><tr><td align="left"><p>Chao</p></td><td char="." align="char"><p>148.84</p></td><td char="." align="char"><p>489.84</p></td><td char="." align="char"><p>439.5344</p></td><td char="–" align="char"><p>382.01–497.06</p></td><td char="–" align="char"><p>376.29–566.81</p></td></tr></tbody></table> </ephtml> </p> <p>Fifth and sixth column show bootstrap CI by the normal approximation and percentile methods, respectively</p> <p>Graph: Fig. 5 Frequency distribution of number of index cases with infected contacts (n = 30)</p> <p>Next, we address Q2. As can be seen in Fig. 5, the best fitting model is given by the zero-truncated Poisson model (Table 5) with <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true"><mi>λ</mi><mo stretchy="false">^</mo></mover></math> </ephtml> = 1.126. Table 6 provides the estimated frequency of index cases with infected but unobserved contacts for the zero truncated Poisson <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mover accent="true"><mi>f</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub><mo>=</mo><mfrac><mrow><mi>n</mi><msup><mi>e</mi><mrow><mo>-</mo><mover accent="true"><mi>λ</mi><mo stretchy="false">^</mo></mover></mrow></msup></mrow><mrow><mn>1</mn><mo>-</mo><msup><mi>e</mi><mrow><mo>-</mo><mover accent="true"><mi>λ</mi><mo stretchy="false">^</mo></mover></mrow></msup></mrow></mfrac></mrow></math> </ephtml> = 14.37, and those from the Turing and Chao approaches for reference. Using the estimated <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>0</mn></msub></math> </ephtml> from the zero truncated Poisson model we estimated the sensitivity of contact tracing to detect index cases with infected contacts as 30/(30 + 14) = 0.676 or 67.6%.</p> <p>Table 5 Model performance for the count index cases with infected contacts</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Model</p></th><th align="left"><p>Log-likelihood</p></th><th align="left"><p>AIC</p></th><th align="left"><p>BIC</p></th></tr></thead><tbody><tr><td align="left"><p>Poisson (<math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true" xmlns=""><mi>λ</mi><mo stretchy="false">^</mo></mover></math><inline-graphic href="12879_2022_7046_Article_IEq103.gif" /> = 1.126)</p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mo xmlns="">-</mo></math><inline-graphic href="12879_2022_7046_Article_IEq104.gif" /><bold>32</bold>.<bold>6684</bold></p></td><td align="left"><p><bold>67</bold>.<bold>3368</bold></p></td><td align="left"><p><bold>68</bold>.<bold>738</bold></p></td></tr><tr><td align="left"><p>Negative Binomial (mue = 1.176, size = 1175.709)</p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mo xmlns="">-</mo></math><inline-graphic href="12879_2022_7046_Article_IEq105.gif" />32.6914</p></td><td align="left"><p>69.3829</p></td><td align="left"><p>72.1853</p></td></tr><tr><td align="left"><p>Geometric (<math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true" xmlns=""><mi>p</mi><mo stretchy="false">^</mo></mover></math><inline-graphic href="12879_2022_7046_Article_IEq106.gif" /> = 0.6)</p></td><td align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mo xmlns="">-</mo></math><inline-graphic href="12879_2022_7046_Article_IEq107.gif" />33.6506</p></td><td align="left"><p>69.3012</p></td><td align="left"><p>70.7024</p></td></tr></tbody></table> </ephtml> </p> <p>Bold values indicate show that Poisson is the best fit for the distribution of the number of index cases with infected contacts (n=30) due to the smallest AIC and BIC</p> <p>Table 6 Estimates of unobserved index cases with infected contacts, population size and 95% CI (<emph>n</emph> = 30) for the three approaches (based upon the Poisson model)</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Estimators</p></th><th align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msub xmlns=""><mover accent="true"><mi>f</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub></math><inline-graphic href="12879_2022_7046_Article_IEq108.gif" /></p></th><th align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mover accent="true" xmlns=""><mi>N</mi><mo stretchy="false">^</mo></mover></math><inline-graphic href="12879_2022_7046_Article_IEq109.gif" /></p></th><th align="left"><p>BT median</p></th><th align="left"><p>95% CI</p></th><th align="left"><p>CI from percentile BT</p></th></tr></thead><tbody><tr><td align="left"><p>MLE</p></td><td align="left"><p>14.37</p></td><td align="left"><p>44.38</p></td><td align="left"><p>45.51</p></td><td align="left"><p>36.62–54.40</p></td><td align="left"><p>33.87–64.17</p></td></tr><tr><td align="left"><p>Turing</p></td><td align="left"><p>14.12</p></td><td align="left"><p>44.12</p></td><td align="left"><p>45.47</p></td><td align="left"><p>36.01–54.94</p></td><td align="left"><p>33.29–64.63</p></td></tr><tr><td align="left"><p>Chao</p></td><td align="left"><p>14.22</p></td><td align="left"><p>44.22</p></td><td align="left"><p>45.5</p></td><td align="left"><p>33.21 <math xmlns="http://www.w3.org/1998/Math/MathML"><mo xmlns="">-</mo></math><inline-graphic href="12879_2022_7046_Article_IEq110.gif" />57.79</p></td><td align="left"><p>31.64–77.40</p></td></tr></tbody></table> </ephtml> </p> <p>Fifth and sixth column show bootstrap CI by the normal approximation and percentile methods, respectively</p> <hd id="AN0154978801-11">Discussion</hd> <p>Our results show a moderately sensitive CT system in Thailand, able to detect more than two thirds of infectious transmission chains during this first wave. The capacity of the system to detect index cases with at least one contact was even higher at 77.6%. Further, it was straightforward to estimate the average intensity of the transmission; this appeared low as shown by the estimated RE (0.14; 95% CI: 0.09–0.22). As reported by an increasing number of works [[<reflink idref="bib19" id="ref27">19</reflink>]], we have also found substantial overdispersion in our data suggesting that most of the index cases did not result in infectious transmission chains and the majority of transmission events stemmed from a small proportion of index cases.</p> <p>The magnitude of the unobserved fraction of COVID-19 cases has been estimated as substantial. Here we propose a mechanism towards the estimation of such undetected population but stress that as the unit of study is the index case once they enter the CT mechanism, which allows the repeated identification of the index case through his/her contacts and the subsequent generation of the count distributions of interest, our inference is therefore limited to CT. In other words, we cannot estimate the overall size of under-reporting that may be associated with other forms of COVID-19 surveillance. Moreover, reiterating the index case as our unit of inference, our results cannot inform the number of contacts (infected or not) missed by CT, from the missed number of index cases estimated by our models. Other approaches have recently been suggested towards the estimation of the under-reported fraction of COVID-19 cases. Lawson and Kim (2021) have recently modelled the spatio-temporal distribution of COVID-19 in South Carolina (US) and considered the role of asymptomatic transmission as a latent effect, and suggested the use of scaling factors to account for the missing cases as done for seasonal influenza [[<reflink idref="bib21" id="ref28">21</reflink>]]. As our data did not specify whether the index cases were symptomatic or not, our estimates of <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>f</mi><mn>0</mn></msub></math> </ephtml> are likely to include both.</p> <hd id="AN0154978801-12">Statistical considerations</hd> <p>For each index case, the number of observed contacts allowed to derive a count distribution which has then been modelled parametrically. Using the best fitting model, the number of index cases with unobserved contacts could be determined and, thus, the completeness of CT. Clearly, the estimate of the frequency of index cases with undetected contacts depends on the model of choice. Hence, we also considered alternative estimators including those of Chao and Turing which weaken the assumption of the chosen model. Chao's estimator allows for heterogeneity in the parameter of the probability model whereas Turing's estimator avoids maximum likelihood estimation. If these alternative approaches lead to substantially different estimates of the size, the choice of the model might be questionable. In all our analyses, the approaches led to similar size estimates. We have also considered whether the distribution was affected by the observed heterogeneity as captured by the available covariates gender, age, or nationality. A generalized linear model analysis (using Poisson and logistic regression) showed no significant association to any of these covariates. Hence, we did not consider a stratified capture-recapture modeling. This is not to say that these variables have no effect on the sensitivity of CT, just that for our dataset such predictors did not show any significance in the unobserved number of index cases. We note that a recent study on EVD showed different patterns in the number of contacts and the probability of zero contacts between two well-defined waves in DRC, and suggested possible improvements in CT as teams become more accustomed over time [[<reflink idref="bib11" id="ref29">11</reflink>]]. In our case, there was no clear break in the time series of cases to support such analysis. However, comparing first and subsequent waves of cases in Thailand would be feasible.</p> <p>We assumed a closed population which is a reasonable assumption under lock-down conditions, and typically met in these kinds of applications by steering the observational window to be small enough. We also assumed independence in the observation (sampling) of index cases. This would be typically violated if these would occur in clusters. Heterogeneity and clustering work in the same way so that Chao's lower bound estimator would still be a conservative approach to the estimation of completeness. In all cases, the parametric modelling and Chao' estimator have returned similar findings which supports our assumption of independence.</p> <hd id="AN0154978801-13">Perspectives</hd> <p>Several countries have used different CT mechanisms, e.g., traditional CT, use of CCTV systems, mobile applications, for the purpose of identifying contacts. In such situations, multi-list CRC models might merit study to assess multiple identification of contacts by more than one data stream.</p> <p>Hook and Regal (1995) stated that the application of CRC methods had very little impact in the public health arena. In other words, their policy value might be small [[<reflink idref="bib22" id="ref30">22</reflink>]]. Providing more informative outputs with indication of where under-reporting is occurring, and what population groups might be more affected would increase the policy value [[<reflink idref="bib23" id="ref31">23</reflink>]]. However, our limited dataset did not present significant heterogeneity to inform such questions. Richer datasets would be required to that effect. A related challenge is the timing of these types of evaluations, with their retrospective nature also limiting their policy value. More real-time applications of CRC across the operational units engaged in the deployment of CT would merit study. These studies might support the identification and quantification of the impact of operational constraints (e.g., size of contact tracing teams, experienced processes and teams) in the sensitivity of CT. Such efforts to extract more value from CT data might provide additional stimulus to strengthen this critical and neglected public health capacity.</p> <hd id="AN0154978801-14">Conclusion</hd> <p>Capture-recapture models have been used for more than four decades for the estimation of disease surveillance sensitivity. This study provides a relatively simple approach for the estimation of the sensitivity of COVID-19 contact tracing efforts. Completeness of COVID-19 contact tracing in Thailand during the first wave appeared moderate, with around 67% of infectious transmission chains detected. Overdispersion was present suggesting that most of the index cases did not result in infectious transmission chains and the majority of transmission events stemmed from a small proportion of index cases.</p> <hd id="AN0154978801-15">Acknowledgements</hd> <p>All authors are grateful to the Ministry of Public Health Thailand for providing access to the Covid-19 contact tracing data. The authors would like to thank the Editor and referees for reviewing the manuscript and providing valuable comments.</p> <hd id="AN0154978801-16">Authors' contributions</hd> <p>VDR and DB conceived the study. RL, PS and DB were responsible for the statistical analysis. RL, DB and VDR wrote the first draft the manuscript. All authors had access to data and RL, PS and DB verified the data. CS, WC, and JP contributed to the data interpretation and revision of the manuscript. All authors read and approved the final manuscript.</p> <hd id="AN0154978801-17">Funding</hd> <p>No funding was received for conducting this study.</p> <hd id="AN0154978801-18">Availability of data and materials</hd> <p>The data that underlie the results reported in this study are available from the corresponding author on reasonable request.</p> <hd id="AN0154978801-19">Ethics approval and consent to participate</hd> <p>No ethical issues were raised as the data available to the research team had been previously collected and had no individual identiers.</p> <hd id="AN0154978801-20">Consent for publication</hd> <p>Not applicable.</p> <hd id="AN0154978801-21">Competing interests</hd> <p>All authors declare no competing interests.</p> <hd id="AN0154978801-22">Appendix: Turing and Chao estimators under Negative Binomial distribution</hd> <p>Under Negative Binomial distribution, the probability function is given by</p> <p> <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mi>p</mi><mi>x</mi></msub><mo>=</mo><mfrac><mrow><mi mathvariant="normal">Γ</mi><mo stretchy="false">(</mo><mi>x</mi><mo>+</mo><mi>κ</mi><mo stretchy="false">)</mo></mrow><mrow><mi mathvariant="normal">Γ</mi><mo stretchy="false">(</mo><mi>x</mi><mo>+</mo><mn>1</mn><mo stretchy="false">)</mo><mi mathvariant="normal">Γ</mi><mo stretchy="false">(</mo><mi>κ</mi><mo stretchy="false">)</mo></mrow></mfrac><msup><mi>π</mi><mi>κ</mi></msup><msup><mrow><mo stretchy="false">(</mo><mn>1</mn><mo>-</mo><mi>π</mi><mo stretchy="false">)</mo></mrow><mi>x</mi></msup><mo>,</mo><mi>x</mi><mo>=</mo><mn>0</mn><mo>,</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>,</mo><mo>...</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml> </p> <p>Graph</p> <p>where <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi mathvariant="normal">Γ</mi><mo stretchy="false">(</mo><mo>.</mo><mo stretchy="false">)</mo></mrow></math> </ephtml> is the Gamma function, and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mi>π</mi></math> </ephtml> and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mi>κ</mi></math> </ephtml> are the model parameters. Since <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>p</mi><mn>0</mn></msub><mo>=</mo><msup><mi>π</mi><mi>κ</mi></msup></mrow></math> </ephtml> , <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>p</mi><mn>1</mn></msub><mo>=</mo><mi>κ</mi><msup><mi>π</mi><mi>κ</mi></msup><mrow><mo stretchy="false">(</mo><mn>1</mn><mo>-</mo><mi>π</mi><mo stretchy="false">)</mo></mrow></mrow></math> </ephtml> and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>E</mi><mo stretchy="false">(</mo><mi>X</mi><mo stretchy="false">)</mo><mo>=</mo><mi>κ</mi><mo stretchy="false">(</mo><mn>1</mn><mo>-</mo><mi>π</mi><mo stretchy="false">)</mo><mo stretchy="false">/</mo><mi>π</mi></mrow></math> </ephtml> . We have <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>p</mi><mn>1</mn></msub><mo stretchy="false">/</mo><mi>E</mi><mrow><mo stretchy="false">(</mo><mi>X</mi><mo stretchy="false">)</mo></mrow><mo>=</mo><msup><mi>π</mi><mrow><mi>κ</mi><mo>+</mo><mn>1</mn></mrow></msup></mrow></math> </ephtml> and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>p</mi><mn>0</mn></msub><mo>=</mo><msup><mrow><mo stretchy="false">(</mo><msup><mi>π</mi><mrow><mi>κ</mi><mo>+</mo><mn>1</mn></mrow></msup><mo stretchy="false">)</mo></mrow><mrow><mi>κ</mi><mo stretchy="false">/</mo><mo stretchy="false">(</mo><mi>κ</mi><mo>+</mo><mn>1</mn><mo stretchy="false">)</mo></mrow></msup><mo>=</mo><msup><mrow><mo stretchy="false">(</mo><msub><mi>p</mi><mn>1</mn></msub><mo stretchy="false">/</mo><mi>E</mi><mrow><mo stretchy="false">(</mo><mi>X</mi><mo stretchy="false">)</mo></mrow><mo stretchy="false">)</mo></mrow><mrow><mi>κ</mi><mo stretchy="false">/</mo><mo stretchy="false">(</mo><mi>κ</mi><mo>+</mo><mn>1</mn><mo stretchy="false">)</mo></mrow></msup></mrow></math> </ephtml> . In spirit of Turing estimator, we get</p> <p> <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mrow><mi mathvariant="italic">Turing</mi></mrow></msub><mo>=</mo><mfrac><mi>n</mi><mrow><mn>1</mn><mo>-</mo><msub><mover accent="true"><mi>P</mi><mo stretchy="false">^</mo></mover><mn>0</mn></msub></mrow></mfrac><mo>=</mo><mfrac><mi>n</mi><mrow><mn>1</mn><mo>-</mo><msup><mrow><mo stretchy="false">(</mo><msub><mi>f</mi><mn>1</mn></msub><mo stretchy="false">/</mo><mi>S</mi><mo stretchy="false">)</mo></mrow><mrow><mi>κ</mi><mo stretchy="false">/</mo><mo stretchy="false">(</mo><mi>κ</mi><mo>+</mo><mn>1</mn><mo stretchy="false">)</mo></mrow></msup></mrow></mfrac><mo>,</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml> </p> <p>Graph</p> <p>where <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>S</mi><mo>=</mo><msubsup><mo>∑</mo><mrow><mi>x</mi><mo>=</mo><mn>0</mn></mrow><mi>m</mi></msubsup><mi>x</mi><msub><mi>f</mi><mi>x</mi></msub></mrow></math> </ephtml> . In addition, the negative binomial distribution is part of the power series family <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>p</mi><mi>x</mi></msub><mo>=</mo><msub><mi>a</mi><mi>x</mi></msub><msup><mi>t</mi><mi>x</mi></msup><mi>A</mi><mrow><mo stretchy="false">(</mo><mi>t</mi><mo stretchy="false">)</mo></mrow></mrow></math> </ephtml> with <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msub><mi>a</mi><mi>x</mi></msub><mo>=</mo><mi mathvariant="normal">Γ</mi><mrow><mo stretchy="false">(</mo><mi>x</mi><mo>+</mo><mi>κ</mi><mo stretchy="false">)</mo></mrow><mo stretchy="false">/</mo><mrow><mo stretchy="false">(</mo><mi mathvariant="normal">Γ</mi><mrow><mo stretchy="false">(</mo><mi>x</mi><mo>+</mo><mn>1</mn><mo stretchy="false">)</mo></mrow><mi mathvariant="normal">Γ</mi><mrow><mo stretchy="false">(</mo><mi>κ</mi><mo stretchy="false">)</mo></mrow><mo stretchy="false">)</mo></mrow></mrow></math> </ephtml> and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>A</mi><mrow><mo stretchy="false">(</mo><mi>t</mi><mo stretchy="false">)</mo></mrow><mo>=</mo><msup><mrow><mo stretchy="false">(</mo><mn>1</mn><mo>-</mo><mi>t</mi><mo stretchy="false">)</mo></mrow><mi>κ</mi></msup></mrow></math> </ephtml> . Considering mixing the negative binomial together with some arbitrary mixing density <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>λ</mi><mo stretchy="false">(</mo><mi>t</mi><mo stretchy="false">)</mo></mrow></math> </ephtml> , we have</p> <p> <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mi>g</mi><mi>x</mi></msub><mrow><mo stretchy="false">(</mo><mi>λ</mi><mo stretchy="false">|</mo><mi>κ</mi><mo stretchy="false">)</mo></mrow><mo>=</mo><msubsup><mo>∫</mo><mrow><mn>0</mn></mrow><mn>1</mn></msubsup><mfrac><mrow><mi mathvariant="normal">Γ</mi><mo stretchy="false">(</mo><mi>x</mi><mo>+</mo><mi>κ</mi><mo stretchy="false">)</mo></mrow><mrow><mi mathvariant="normal">Γ</mi><mo stretchy="false">(</mo><mi>x</mi><mo>+</mo><mn>1</mn><mo stretchy="false">)</mo><mi mathvariant="normal">Γ</mi><mo stretchy="false">(</mo><mi>κ</mi><mo stretchy="false">)</mo></mrow></mfrac><msup><mi>t</mi><mi>x</mi></msup><msup><mrow><mo stretchy="false">(</mo><mn>1</mn><mo>-</mo><mi>t</mi><mo stretchy="false">)</mo></mrow><mi>κ</mi></msup><mi>λ</mi><mrow><mo stretchy="false">(</mo><mi>t</mi><mo stretchy="false">)</mo></mrow><mi>d</mi><mi>t</mi><mo>=</mo><msubsup><mo>∫</mo><mrow><mn>0</mn></mrow><mn>1</mn></msubsup><msub><mi>a</mi><mi>x</mi></msub><msup><mi>t</mi><mi>x</mi></msup><mi>A</mi><mrow><mo stretchy="false">(</mo><mi>t</mi><mo stretchy="false">)</mo></mrow><mi>λ</mi><mrow><mo stretchy="false">(</mo><mi>t</mi><mo stretchy="false">)</mo></mrow><mi>d</mi><mi>t</mi><mo>.</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml> </p> <p>Graph</p> <p>Since</p> <p> <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><mfrac><mrow><msub><mi>g</mi><mn>1</mn></msub><mo stretchy="false">/</mo><msub><mi>a</mi><mn>1</mn></msub></mrow><mrow><msub><mi>g</mi><mn>0</mn></msub><mo stretchy="false">/</mo><msub><mi>a</mi><mn>0</mn></msub></mrow></mfrac><mo>≤</mo><mfrac><mrow><msub><mi>g</mi><mn>2</mn></msub><mo stretchy="false">/</mo><msub><mi>a</mi><mn>2</mn></msub></mrow><mrow><msub><mi>g</mi><mn>1</mn></msub><mo stretchy="false">/</mo><msub><mi>a</mi><mn>1</mn></msub></mrow></mfrac><mo>,</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml> </p> <p>Graph</p> <p>then <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>a</mi><mn>0</mn></msub></math> </ephtml> , <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>a</mi><mn>1</mn></msub></math> </ephtml> , and <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><msub><mi>a</mi><mn>2</mn></msub></math> </ephtml> are replaced and probabilities are substituted by observed frequencies. For a mixed Negative Binomial model with arbitrary density, the new estimator is accomplished in spirit of Chao estimator as</p> <p> <ephtml> <math display="block" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable><mtr><mtd columnalign="right"><mrow><msub><mover accent="true"><mi>N</mi><mo stretchy="false">^</mo></mover><mrow><mi mathvariant="italic">Chao</mi></mrow></msub><mo>=</mo><mi>n</mi><mo>+</mo><mfrac><msup><mrow><mo stretchy="false">(</mo><msub><mi>f</mi><mn>1</mn></msub><mo stretchy="false">/</mo><msub><mi>a</mi><mn>1</mn></msub><mo stretchy="false">)</mo></mrow><mn>2</mn></msup><mrow><msub><mi>f</mi><mn>2</mn></msub><mo stretchy="false">/</mo><msub><mi>a</mi><mn>2</mn></msub></mrow></mfrac><mo>=</mo><mi>n</mi><mo>+</mo><mfenced close=")" open="("><mfrac><mrow><mi>κ</mi><mo>+</mo><mn>1</mn></mrow><mi>κ</mi></mfrac></mfenced><mfrac><msubsup><mi>f</mi><mn>1</mn><mn>2</mn></msubsup><mrow><mn>2</mn><msub><mi>f</mi><mn>2</mn></msub></mrow></mfrac><mo>,</mo></mrow></mtd></mtr></mtable></mrow></math> </ephtml> </p> <p>Graph</p> <p>(see [[<reflink idref="bib24" id="ref32">24</reflink>]]).</p> <hd id="AN0154978801-23">Abbreviations</hd> <p></p> <p>• DDC</p> <p></p> <ulist> <item> Department of disease control</item> <p></p> </ulist> <p>• RRT</p> <p></p> <ulist> <item> Rapid response teams</item> <p></p> </ulist> <p>• IAR</p> <p></p> <ulist> <item> Intra-action-review</item> <p></p> </ulist> <p>• CT</p> <p></p> <ulist> <item> Contact tracing</item> <p></p> </ulist> <p>• CRC</p> <p></p> <ulist> <item> Capture-recapture</item> <p></p> </ulist> <p>• SS</p> <p></p> <ulist> <item> Surveillance systems</item> <p></p> </ulist> <p>• EVD</p> <p></p> <ulist> <item> Ebola virus disease</item> <p></p> </ulist> <p>• DRC</p> <p></p> <ulist> <item> Democratic Republic of the Congo</item> <p></p> </ulist> <p>• CDCU</p> <p></p> <ulist> <item> Communicable disease control units</item> <p></p> </ulist> <p>• JIT</p> <p></p> <ulist> <item> Joint investigation teams</item> <p></p> </ulist> <p>• AIC</p> <p></p> <ulist> <item> Akaike information criterion</item> <p></p> </ulist> <p>• BIC</p> <p></p> <ulist> <item> Bayesian information criterion</item> <p></p> </ulist> <p>• CI</p> <p></p> <ulist> <item> Confidence interval</item> <p></p> </ulist> <p>• IQR</p> <p></p> <ulist> <item> Interquartile range</item> <p></p> </ulist> <p>• RE</p> <p></p> <ulist> <item> Effective reproductive number</item> </ulist> <hd id="AN0154978801-24">Publisher's Note</hd> <p>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p> <ref id="AN0154978801-25"> <title> References </title> <blist> <bibl id="bib1" idref="ref1" type="bt">1</bibl> <bibtext> WHO: Joint intra-action review of the public health response to COVID-19 in Thailand. 2020. https://<ulink href="http://www.who.int/docs/default-source/searo/thailand/iar-covid19-en.pdf">www.who.int/docs/default-source/searo/thailand/iar-covid19-en.pdf</ulink> Accessed 20 Nov 2020.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref2" type="bt">2</bibl> <bibtext> Kaweenuttayanon N, Pattanarattanamolee R, Sornchaa. Community surveillance of COVID-19 by village health volunteers, Thailand. Bull World Health Organ. 2021; 99: 393-397. 10.2471/BLT.20.274308</bibtext> </blist> <blist> <bibl id="bib3" idref="ref4" type="bt">3</bibl> <bibtext> WHO: The Ministry of Public Health and the World Health Organization review Thailand's COVID-19 response. 2020. https://<ulink href="http://www.who.int/thailand/news/detail/14-10-2020-Thailand-IAR-COVID19">www.who.int/thailand/news/detail/14-10-2020-Thailand-IAR-COVID19</ulink> Accessed 17 Dec 2020.</bibtext> </blist> <blist> <bibl id="bib4" idref="ref5" type="bt">4</bibl> <bibtext> WHO: Contact tracing in the context of COVID-19. 2021. https://<ulink href="http://www.who.int/publications/i/item/contact-tracing-in-the-context-of-covid-19">www.who.int/publications/i/item/contact-tracing-in-the-context-of-covid-19</ulink> Accessed 2 Mar 2021.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref6" type="bt">5</bibl> <bibtext> Buehler JW, Hopkins RS, Overhage JM. Framework for evaluating public health surveillance systems for early detection of outbreaks: recommendations from the CDC Working Group. MMWR Recomm Rep. 2004; 53: 1-11. 15129191</bibtext> </blist> <blist> <bibl id="bib6" idref="ref7" type="bt">6</bibl> <bibtext> Vogt F, Kurup K, Mussleman P et al. Contact tracing indicators for COVID-19: rapid scoping review and conceptual framework. 2021. https://<ulink href="http://www.medrxiv.org/content/10.1101/2021.05.13.21257067v1.full.pdf">www.medrxiv.org/content/10.1101/2021.05.13.21257067v1.full.pdf</ulink> Accessed 30 May 2021.</bibtext> </blist> <blist> <bibl id="bib7" idref="ref9" type="bt">7</bibl> <bibtext> Bhatia R, Klausner J. Estimating individual risks of COVID-19-associated hospitalization and death using publicly available data. PLoS ONE. 2021; 15: 1-12</bibtext> </blist> <blist> <bibl id="bib8" idref="ref10" type="bt">8</bibl> <bibtext> Böhning D, van der Heijden PGM, Bunge J. Capture-recapture methods for the social and medical science. 2019: Boca Raton; CRC Press</bibtext> </blist> <blist> <bibl id="bib9" type="bt">9</bibl> <bibtext> McRea R, T MBJ. Analysis of capture-recapture data. 2015: Boca Raton; CRC Press</bibtext> </blist> <blist> <bibtext> Vergne T, Del Rio Vilas VJ, Cameron A. Capture-recapture approaches and the surveillance of livestock diseases: a review. Prev Vet Med. 2015; 120: 253-264. 10.1016/j.prevetmed.2015.04.003</bibtext> </blist> <blist> <bibtext> Polonsky JA, Böhning D, Keita M, et al. Novel use of capture-recapture methods to estimate completeness of contact tracing during an Ebola outbreak, Democratic Republic of the Congo, 2018–2020. Emerg Infect Dis. 2021. https://doi.org/10.3201/eid2712.204958.</bibtext> </blist> <blist> <bibtext> Böhning D, Rocchetti I, Maruotti A. Estimating the undetected infections in the Covid-19 outbreak by harnessing capture-recapture methods. Int J Infect Dis. 2020; 97: 197-201. 10.1016/j.ijid.2020.06.009</bibtext> </blist> <blist> <bibtext> MOPH: Guidelines for surveillance and investigation of coronavirus disease 2019 (COVID-19). 2020. https://ddc.moph.go.th/viralpneumonia/eng/file/guidelines/G%5fen%5f21022020.pdf Accessed 15 Feb 2021.</bibtext> </blist> <blist> <bibtext> Good I. On the population frequencies of species and the estimation of population parameters. Biometrika. 1953; 40: 237-264. 10.1093/biomet/40.3-4.237</bibtext> </blist> <blist> <bibtext> Chao A. Estimating the population size for capture-recapture data with unequal catchability. Biometrics. 1987; 43: 783-791. 1:STN:280:DyaL1c7gsl2qsw%3D%3D. 10.2307/2531532</bibtext> </blist> <blist> <bibtext> Norris JL, Pollock KH. Including model uncertainty in estimating variances in multiple capture studies. Environ Ecol Stat. 1996; 3: 235-244. 10.1007/BF00453012</bibtext> </blist> <blist> <bibtext> Orasa A, Böhning D, Maruotti A. Uncertainty estimation in heterogeneous capture-recapture count data. J Stat Comput Simul. 2017; 87: 2094-2114. 10.1080/00949655.2017.1315668</bibtext> </blist> <blist> <bibtext> Rey D, Neuhäuser M. Wilcoxon-Signed-Rank Test. 2011: Heidelberg; Springer. 10.1007/978-3-642-04898-2_616</bibtext> </blist> <blist> <bibtext> Adam DC, Wu P, Wong JY. Clustering and superspreading potential of SARS-CoV-2 infections in Hong Kong. Nat Med. 2020; 26: 1741-1749. 10.1038/s41591-020-1092-0</bibtext> </blist> <blist> <bibtext> Lemieux JE, Siddle KJ, Shaw BM. Phylogenetic analysis of SARS-CoV-2 in Boston highlights the impact of superspreading events. Science. 2021; 371: 1-9. 10.1126/science.abe3261</bibtext> </blist> <blist> <bibtext> Lawson AB, Kim J. Space-time covid-19 Bayesian SIR modeling in South Carolina. PLoS ONE. 2021; 16: 1-14. 10.1371/journal.pone.0242777</bibtext> </blist> <blist> <bibtext> Hook EB, Regal RRBrookmeyer R, Stroup DF. Completeness of reporting: capture-recapture methods in public health surveillance. Monitoring the Health of Populations. Statistical Principles and Methods for Public Health Surveillance. 2004: New York; Oxford University Press: 341-359</bibtext> </blist> <blist> <bibtext> Gignoux E, Idowu R, Bawo L. Use of capture-recapture to estimate underreporting of Ebola virus disease, Montserrado County, Liberia. Emerg Infect Dis. 2015; 21: 2265-2267. 10.3201/eid2112.150756</bibtext> </blist> <blist> <bibtext> Böhning D, Baksh MF, Lerdsuwansri R. Use of the ratio plot in capture-recapture estimation. J Comput Graph Stat. 2013; 22: 135-155. 10.1080/10618600.2011.647174</bibtext> </blist> </ref> <aug> <p>By R. Lerdsuwansri; P. Sangnawakij; D. Böhning; C. Sansilapin; W. Chaifoo; Jonathan A. Polonsky and Victor J. Del Rio Vilas</p> <p>Reported by Author; Author; Author; Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib10" firstref="ref11"></nolink> <nolink nlid="nl2" bibid="bib11" firstref="ref12"></nolink> <nolink nlid="nl3" bibid="bib12" firstref="ref15"></nolink> <nolink nlid="nl4" bibid="bib13" firstref="ref18"></nolink> <nolink nlid="nl5" bibid="bib14" firstref="ref19"></nolink> <nolink nlid="nl6" bibid="bib15" firstref="ref20"></nolink> <nolink nlid="nl7" bibid="bib16" firstref="ref23"></nolink> <nolink nlid="nl8" bibid="bib18" firstref="ref24"></nolink> <nolink nlid="nl9" bibid="bib290" firstref="ref26"></nolink> <nolink nlid="nl10" bibid="bib19" firstref="ref27"></nolink> <nolink nlid="nl11" bibid="bib21" firstref="ref28"></nolink> <nolink nlid="nl12" bibid="bib22" firstref="ref30"></nolink> <nolink nlid="nl13" bibid="bib23" firstref="ref31"></nolink> <nolink nlid="nl14" bibid="bib24" firstref="ref32"></nolink>
CustomLinks:
  – Url: https://resolver.ebsco.com/c/xy5jbn/result?sid=EBSCO:edsdoj&genre=article&issn=14712334&ISBN=&volume=22&issue=1&date=20220101&spage=1&pages=1-10&title=BMC Infectious Diseases&atitle=Sensitivity%20of%20contact-tracing%20for%20COVID-19%20in%20Thailand%3A%20a%20capture-recapture%20application&aulast=R.%20Lerdsuwansri&id=DOI:10.1186/s12879-022-07046-6
    Name: Full Text Finder (for New FTF UI) (s8985755)
    Category: fullText
    Text: Find It @ SCU Libraries
    MouseOverText: Find It @ SCU Libraries
  – Url: https://doaj.org/article/a6e621a9b1e0458988723246900a4467
    Name: EDS - DOAJ (s8985755)
    Category: fullText
    Text: View record from DOAJ
    MouseOverText: View record from DOAJ
Header DbId: edsdoj
DbLabel: Directory of Open Access Journals
An: edsdoj.6e621a9b1e0458988723246900a4467
RelevancyScore: 958
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 957.874755859375
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Sensitivity of contact-tracing for COVID-19 in Thailand: a capture-recapture application
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22R%2E+Lerdsuwansri%22">R. Lerdsuwansri</searchLink><br /><searchLink fieldCode="AR" term="%22P%2E+Sangnawakij%22">P. Sangnawakij</searchLink><br /><searchLink fieldCode="AR" term="%22D%2E+Böhning%22">D. Böhning</searchLink><br /><searchLink fieldCode="AR" term="%22C%2E+Sansilapin%22">C. Sansilapin</searchLink><br /><searchLink fieldCode="AR" term="%22W%2E+Chaifoo%22">W. Chaifoo</searchLink><br /><searchLink fieldCode="AR" term="%22Jonathan+A%2E+Polonsky%22">Jonathan A. Polonsky</searchLink><br /><searchLink fieldCode="AR" term="%22Victor+J%2E+Del+Rio+Vilas%22">Victor J. Del Rio Vilas</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: BMC Infectious Diseases, Vol 22, Iss 1, Pp 1-10 (2022)
– Name: Publisher
  Label: Publisher Information
  Group: PubInfo
  Data: BMC, 2022.
– Name: DatePubCY
  Label: Publication Year
  Group: Date
  Data: 2022
– Name: Subset
  Label: Collection
  Group: HoldingsInfo
  Data: LCC:Infectious and parasitic diseases
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22COVID-19%22">COVID-19</searchLink><br /><searchLink fieldCode="DE" term="%22Contact+tracing%22">Contact tracing</searchLink><br /><searchLink fieldCode="DE" term="%22Thailand%22">Thailand</searchLink><br /><searchLink fieldCode="DE" term="%22Capture-recapture%22">Capture-recapture</searchLink><br /><searchLink fieldCode="DE" term="%22Sensitivity%22">Sensitivity</searchLink><br /><searchLink fieldCode="DE" term="%22Infectious+and+parasitic+diseases%22">Infectious and parasitic diseases</searchLink><br /><searchLink fieldCode="DE" term="%22RC109-216%22">RC109-216</searchLink>
– Name: Abstract
  Label: Description
  Group: Ab
  Data: Abstract Background We investigate the completeness of contact tracing for COVID-19 during the first wave of the COVID-19 pandemic in Thailand, from early January 2020 to 30 June 2020. Methods Uni-list capture-recapture models were applied to the frequency distributions of index cases to inform two questions: (1) the unobserved number of index cases with contacts, and (2) the unobserved number of index cases with secondary cases among their contacts. Results Generalized linear models (using Poisson and logistic families) did not return any significant predictor (age, sex, nationality, number of contacts per case) on the risk of transmission and hence capture-recapture models did not adjust for observed heterogeneity. Best fitting models, a zero truncated negative binomial for question 1 and zero-truncated Poisson for question 2, returned sensitivity estimates for contact tracing performance of 77.6% (95% CI = 73.75–81.54%) and 67.6% (95% CI = 53.84–81.38%), respectively. A zero-inflated negative binomial model on the distribution of index cases with secondary cases allowed the estimation of the effective reproduction number at 0.14 (95% CI = 0.09–0.22), and the overdispersion parameter at 0.1. Conclusion Completeness of COVID-19 contact tracing in Thailand during the first wave appeared moderate, with around 67% of infectious transmission chains detected. Overdispersion was present suggesting that most of the index cases did not result in infectious transmission chains and the majority of transmission events stemmed from a small proportion of index cases.
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: article
– Name: Format
  Label: File Description
  Group: SrcInfo
  Data: electronic resource
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 1471-2334
– Name: NoteTitleSource
  Label: Relation
  Group: SrcInfo
  Data: https://doaj.org/toc/1471-2334
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1186/s12879-022-07046-6
– Name: URL
  Label: Access URL
  Group: URL
  Data: <link linkTarget="URL" linkTerm="https://doaj.org/article/a6e621a9b1e0458988723246900a4467" linkWindow="_blank">https://doaj.org/article/a6e621a9b1e0458988723246900a4467</link>
– Name: AN
  Label: Accession Number
  Group: ID
  Data: edsdoj.6e621a9b1e0458988723246900a4467
PLink https://login.libproxy.scu.edu/login?url=https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&scope=site&db=edsdoj&AN=edsdoj.6e621a9b1e0458988723246900a4467
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1186/s12879-022-07046-6
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 10
        StartPage: 1
    Subjects:
      – SubjectFull: COVID-19
        Type: general
      – SubjectFull: Contact tracing
        Type: general
      – SubjectFull: Thailand
        Type: general
      – SubjectFull: Capture-recapture
        Type: general
      – SubjectFull: Sensitivity
        Type: general
      – SubjectFull: Infectious and parasitic diseases
        Type: general
      – SubjectFull: RC109-216
        Type: general
    Titles:
      – TitleFull: Sensitivity of contact-tracing for COVID-19 in Thailand: a capture-recapture application
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: R. Lerdsuwansri
      – PersonEntity:
          Name:
            NameFull: P. Sangnawakij
      – PersonEntity:
          Name:
            NameFull: D. Böhning
      – PersonEntity:
          Name:
            NameFull: C. Sansilapin
      – PersonEntity:
          Name:
            NameFull: W. Chaifoo
      – PersonEntity:
          Name:
            NameFull: Jonathan A. Polonsky
      – PersonEntity:
          Name:
            NameFull: Victor J. Del Rio Vilas
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2022
          Identifiers:
            – Type: issn-print
              Value: 14712334
          Numbering:
            – Type: volume
              Value: 22
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: BMC Infectious Diseases
              Type: main
ResultId 1