Capture-Recapture Estimators in Epidemiology with Applications to Pertussis and Pneumococcal Invasive Disease Surveillance

Abstract

IntroductionSurveillance networks are often not exhaustive nor completely complementary. In such situations, capture-recapture methods can be used for incidence estimation. The choice of estimator and their robustness with respect to the homogeneity and independence assumptions are however not well documented.MethodsWe investigated the performance of five different capture-recapture estimators in a simulation study. Eight different scenarios were used to detect and combine case-information. The scenarios increasingly violated assumptions of independence of samples and homogeneity of detection probabilities. Belgian datasets on invasive pneumococcal disease (IPD) and pertussis provided motivating examples.ResultsNo estimator was unbiased in all scenarios. Performance of the parametric estimators depended on how much of the dependency and heterogeneity were correctly modelled. Model building was limited by parameter estimability, availability of additional information (e.g. covariates) and the possibilities inherent to the method. In the most complex scenario, methods that allowed for detection probabilities conditional on previous detections estimated the total population size within a 20–30% error-range. Parametric estimators remained stable if individual data sources lost up to 50% of their data. The investigated non-parametric methods were more susceptible to data loss and their performance was linked to the dependence between samples; overestimating in scenarios with little dependence, underestimating in others. Issues with parameter estimability made it impossible to model all suggested relations between samples for the IPD and pertussis datasets. For IPD, the estimates for the Belgian incidence for cases aged 50 years and older ranged from 44 to58/100,000 in 2010. The estimates for pertussis (all ages, Belgium, 2014) ranged from 24.2 to30.8/100,000.ConclusionWe encourage the use of capture-recapture methods, but epidemiologists should preferably include datasets for which the underlying dependency structure is not too complex, a priori investigate this structure, compensate for it within the model and interpret the results with the remaining unmodelled heterogeneity in mind.

Invasive Pneumococcal Disease in Tuscany Region, Italy, 2016–2017: Integrating Multiple Data Sources to Investigate Underreporting

“…When considering the capture–recapture analysis, the estimated incidence raised to 45.6/100,000 IPD cases per year, with a certain degree of geographical variations within the region, highlighting a higher level of under reporting for older age groups. Our estimate was comparable to that of a Belgian study applying different models of capture–recapture (44 to 58/100,000 IPD cases in 2010 in >65 years old) [ 20 ]. The observed higher level of underreporting among older subjects has already been described for IPD [ 20 , 21 ] and may be linked to a lower level of hospitalization in this population group or, in the case of severe illness in the elderly, to the tendency not to give a specific diagnosis.…”

“…Our estimate was comparable to that of a Belgian study applying different models of capture–recapture (44 to 58/100,000 IPD cases in 2010 in >65 years old) [ 20 ]. The observed higher level of underreporting among older subjects has already been described for IPD [ 20 , 21 ] and may be linked to a lower level of hospitalization in this population group or, in the case of severe illness in the elderly, to the tendency not to give a specific diagnosis. Furthermore, a geographical variability in IPD underreporting has been described in the literature [ 20 , 22 ], correlated to the distance between laboratories and hospitals [ 20 ].…”

“…The integration of epidemiological, microbiological, and clinical sources via data linkage and capture–recapture analysis is a well-established method to assess the completeness of surveillance of IPD [ 19 , 20 , 21 ] and other infection diseases at a regional or national level in Italy and elsewhere [ 18 , 23 , 33 , 34 , 35 ]. Using data linkage, we estimated a mean annual incidence rate for IPD in Tuscany during the review period substantially higher (12.1/100,000 vs. 1.7/100,000) than that resulting from the sole IPD-NSS.…”

“…Capture–recapture analysis is a well-established method to estimate the quality of reporting systems, and it was previously applied to IPD and other invasive bacterial diseases [ 18 , 19 , 20 , 21 , 22 , 23 , 24 ].…”

Pertussis diagnosis in Belgium: results of the National Reference Centre for Bordetella anno 2015

“…In the peak incidence year 2014, 1056 cases were reported in Flanders, 830 cases in Wallonia and 117 cases in Brussels Capital region [18]. Using a log-linear capture-recapture method, the total number of cases of pertussis in Belgium in 2014 (all ages) was estimated to range from 24•2 to 30•8/100 000 [19]. As reported in detail in this paper, a total 1246 cases of pertussis were diagnosed by the Belgian NRC consortium in 2015: 844 cases by serology, 394 cases by PCR and eight cases by both methods.…”