Cohort-based smoothing methods for age-specific contact rates

The use of social contact rates is widespread in infectious disease modeling since it has been shown that they are key driving forces of important epidemiological parameters. Quantification of contact patterns is crucial to parameterize dynamic transmission models and to provide insights on the (basic) reproduction number. Information on social interactions can be obtained from population-based contact surveys, such as the European Commission project POLYMOD. Estimation of age-specific contact rates from these studies is often done using a piecewise constant approach or bivariate smoothing techniques. For the latter, typically, smoothness is introduced in the dimensions of the respondent's and contact's age (i.e., the rows and columns of the social contact matrix). We propose a smoothing constrained approach-taking into account the reciprocal nature of contacts-introducing smoothness over the diagonal (including all subdiagonals) of the social contact matrix. This modeling approach is justified assuming that when people age their contact behavior changes smoothly. We call this smoothing from a cohort perspective. Two approaches that allow for smoothing over social contact matrix diagonals are proposed, namely (i) reordering of the diagonal components of the contact matrix and (ii) reordering of the penalty matrix ensuring smoothness over the contact matrix diagonals. Parameter estimation is done in the likelihood framework by using constrained penalized iterative reweighted least squares. A simulation study underlines the benefits of cohort-based smoothing. Finally, the proposed methods are illustrated on the Belgian POLYMOD data of 2006. Code to reproduce the results of the article can be downloaded on this GitHub repository https://github.com/oswaldogressani/Cohort_smoothing.

Evaluation of symptoms in respiratory syncytial virus infection in adults: psychometric evaluation of the Respiratory Infection Intensity and Impact Questionnaire™ symptom scores

BACKGROUND

The Respiratory Infection Intensity and Impact Questionnaire (RiiQ™) is a patient-reported outcome measure designed to assess symptoms and impacts of respiratory syncytial virus (RSV) infection. This study evaluated the construct validity, reliability, and responsiveness of the RiiQ™ Respiratory and Systemic Symptoms Scale scores.

METHODS

Prospective data were analyzed from a total of 1795 participants, including from non-hospitalized patients with acute respiratory infection (ARI) and no coinfections enrolled in a Phase 2b RSV vaccine study (RSV-positive: n = 60; RSV-negative: n = 1615), and two observational studies of patients hospitalized with RSV (n = 20; n = 100). Descriptive statistics, confirmatory factor analysis (CFA), test-retest intraclass correlation coefficients (ICCs), construct validity correlations (between a clinician-assessed clinical questionnaire and the RiiQ™ symptoms scale), known-groups validity, and responsiveness (correlations of change scores) were evaluated.

RESULTS

Mean patient age ranged from 66.5 to 71.5 years and the majority of patients were female. Initial assessments in the vaccine trial (ARI Day 1) were suggestive of less severe illness than in the observational studies with hospitalized patients. CFA loadings (> 0.40) supported summary scores. ICCs exceeding the recommended threshold of 0.70 supported test-retest reliability for Respiratory and Systemic Symptoms, except in the small observational study. At the scale level, correlations were moderate to strong (|r| ≥ 0.3) and positive between the Respiratory Symptoms Scale and the related clinical questionnaire scores, reflecting measurement of similar symptoms in support of convergent validity. Correlations with change in Patient Global Impression of Severity > 0.30 supported responsiveness.

CONCLUSIONS

Psychometric tests applied to the RiiQ™ Symptoms scales provide evidence of its reliability, construct validity, discriminating ability, and responsiveness for use in clinical studies to assess the onset and severity of RSV symptoms.

Measuring respiratory syncytial virus infection severity in hospitalized children using the Pediatric Respiratory Syncytial Virus Electronic Severity and Outcome Rating System (PRESORS)

BACKGROUND

The Pediatric Respiratory Syncytial Virus Electronic Severity and Outcome Rating System (PRESORS) was developed to assess the severity of respiratory syncytial virus (RSV) infection in children. Because young children cannot report how they feel or function, ratings are based on observations by the child's caregiver (Observer-Reported Outcome questionnaire [ObsRO]) and clinician (Clinician-Reported Outcome questionnaire [ClinRO]). This prospective study aimed to evaluate the psychometric properties of the PRESORS.

METHODS

The PRESORS version 6 ObsRO and ClinRO were evaluated in children with RSV infection requiring hospitalization in centers in the United States, Argentina, and Chile. Assessments were performed from days 1 to 7 by the child's caregiver and clinician. To assess inter-rater reliability, two clinicians independently performed the ClinRO near in time.

RESULTS

A total of 124 children aged ≤36 months were enrolled (mean age, 8 months). Factor analysis demonstrated that RSV severity consists of two dimensions, respiratory signs and illness behavior, and that these dimensions were consistent over time. The inter-rater reliability for the ClinRO was 0.66 (95% confidence interval [CI], 0.55-0.75) but improved to 0.79 (95% CI, 0.71-0.86) after removing one outlying site, suggesting that quantifying RSV severity is not trivial, even using qualified raters, but that an adequate inter-rater reliability is achievable with the PRESORS through adequate training. ClinRO and ObsRO displayed acceptable internal consistency and acceptable convergent validity with the Respiratory Syncytial Virus Network Scale, global impression scores, and key hospital characteristics.

CONCLUSIONS

The PRESORS is relevant and appropriate for assessing the severity of RSV infection in infants requiring hospitalization.

Risk factors for disease severity and increased medical resource utilization in respiratory syncytial virus (+) hospitalized children: A descriptive study conducted in four Belgian hospitals

Background

We aimed to provide regional data on clinical symptoms, medical resource utilization (MRU), and risk factors for increased MRU in hospitalized respiratory syncytial virus (RSV)-infected Belgian pediatric population.

Methods

This prospective, multicenter study enrolled RSV (+) hospitalized children (aged ≤5y) during the 2013–2015 RSV seasons. RSV was diagnosed within 24h of hospitalization. Disease severity of RSV (+) patients was assessed until discharge or up to maximum six days using a Physical Examination Score (PES) and a derived score based on ability to feed, dyspnea and respiratory effort (PES3). MRU (concomitant medications, length of hospitalization [LOH], and oxygen supplementation) was evaluated. Kaplan-Meier survival analysis was performed to compare MRU by age and presence of risk factors for severe disease. Association between baseline covariates and MRU was analyzed using Cox regression models.

Results

In total, 75 children were included, Median (range) age was 4 (0–41) months, risk factors were present in 18.7%, and early hospitalization (≤3 days of symptom onset) was observed in 57.3% of patients. Cough (100%), feeding problems (82.2%), nasal discharge (87.8%), and rales and rhonchi (82.2%) were frequently observed. Median (range) LOH and oxygen supplementation was 5 (2–7) and 3 (1–7) days. Oxygen supplementation, bronchodilators, and antibiotics were administered to 58.7%, 64.0%, and 41.3% of the patients, respectively. Age <3 months and baseline total PES3 score were associated with probability and the duration of receiving oxygen supplementation. LOH was not associated with any covariate.

Conclusion

RSV is associated with high disease burden and MRU in hospitalized children. Oxygen supplementation but not length of hospitalization was associated with very young age and the PES3 score. These results warrant further assessment of the PES3 score as a predictor for the probability of receiving and length of oxygen supplementation in RSV hospitalized children.

Registration

NCT02133092

Risk factors and medical resource utilization in US adults hospitalized with influenza or respiratory syncytial virus in the Hospitalized Acute Respiratory Tract Infection study

Background

Influenza and respiratory syncytial virus (RSV) are associated with substantial morbidity and mortality in the United States. We assessed risk factors for severe disease and medical resource utilization (MRU) among US adults hospitalized with influenza or RSV in the Hospitalized Acute Respiratory Tract Infection (HARTI) study.

Methods

HARTI was a prospective global (40 centers, 12 countries) epidemiological study of adults hospitalized with acute respiratory tract infections conducted across the 2017–2019 epidemic seasons. Patients with confirmed influenza or RSV were followed up to 3 months post‐discharge. Baseline characteristics, prevalence of core risk factors (CRFs) for severe disease (age ≥65 years, chronic heart or renal disease, chronic obstructive pulmonary disease, or asthma), and MRU were summarized descriptively.

Results

The US cohort included 280 influenza‐positive and 120 RSV‐positive patients. RSV patients were older (mean: 63.1 vs. 59.7 years) and a higher proportion had CRFs (87.5% vs. 81.4%). Among those with CRFs (influenza, n = 153; RSV, n = 99), RSV patients required longer hospitalizations (median length of stay: 4.5 days) and a greater proportion (79.8%) required oxygen supplementation during hospitalization compared with influenza patients (4.0 days and 59.5%, respectively). At 3 months post‐discharge, a greater proportion of RSV patients with CRFs reported use of antibiotics, antitussives, bronchodilators, and inhaled and systemic steroids versus those with influenza and CRFs. Many patients with CRFs reported hospital readmission at 3 months post‐discharge (RSV: 13.4%; influenza: 11.9%).

Conclusions

MRU during and post‐hospitalization due to RSV in adults is similar to or greater than that of influenza. Enhanced RSV surveillance and preventive and therapeutic interventions are needed.

Comparative assessment of reported symptoms of influenza, respiratory syncytial virus, and human metapneumovirus infection during hospitalization and post-discharge assessed by Respiratory Intensity and Impact Questionnaire

Background

The hospitalized acute respiratory tract infection (HARTI) study used the Respiratory Intensity and Impact Questionnaire (RiiQ™) Symptom Scale, derived from FluiiQ™, to assess and compare the burden of respiratory infection symptoms for patients with influenza, respiratory syncytial virus (RSV), and human metapneumovirus (hMPV) infection, with or without core risk factors (CRF) (age ≥65; chronic heart, renal, obstructive pulmonary disease; asthma).

Methods

This was a prospective cohort study in adult patients hospitalized with acute respiratory tract infection (40 centers, 12 countries) during two consecutive influenza/RSV/hMPV seasons (2017–2019). The RiiQ™ Symptom Scale and EuroQol 5‐Dimensions 5‐Levels (EQ‐5D‐5L) were assessed by interview at two timepoints during hospitalization and at 1, 2, and 3 months post‐discharge.

Results

Mean lower respiratory tract (LRT) symptom scores were higher for RSV and hMPV participants compared to influenza at 48 h after enrollment/early discharge (p = 0.001) and 3 months post‐discharge (p = 0.007). This was driven by LRT symptoms, including shortness of breath (SOB) (p < 0.01) and wheezing (p < 0.01) during hospitalization, and SOB (p < 0.05) and cough (p < 0.05) post‐discharge. Participants with CRF reported more moderate‐to‐severe SOB (p < 0.05) and wheezing (p < 0.05) compared to CRF(−) participants post‐discharge. EQ‐5D‐5L scores were moderately associated with RiiQ™ LRT and systemic symptoms domains.

Conclusions

Results from the HARTI study suggest that in the study population, LRT symptoms were more severe for RSV and hMPV groups and for patients with CRF. RiiQ™ Symptom Scale scores shows a moderate association with EQ‐5D‐5L indicating that the RiiQ™ may provide useful insights and offer advantages over other measures for use in interventional RSV adult clinical studies.

Spatial Modelling to Inform Public Health Based on Health Surveys: Impact of Unsampled Areas at Lower Geographical Scale

Small area estimation is an important tool to provide area-specific estimates of population characteristics for governmental organizations in the context of education, public health and care. However, many demographic and health surveys are unrepresentative at a small geographical level, as often areas at a lower level are not included in the sample due to financial or logistical reasons. In this paper, we investigated (1) the effect of these unsampled areas on a variety of design-based and hierarchical model-based estimates and (2) the benefits of using auxiliary information in the estimation process by means of an extensive simulation study. The results showed the benefits of hierarchical spatial smoothing models towards obtaining more reliable estimates for areas at the lowest geographical level in case a spatial trend is present in the data. Furthermore, the importance of auxiliary information was highlighted, especially for geographical areas that were not included in the sample. Methods are illustrated on the 2008 Mozambique Poverty and Social Impact Analysis survey, with interest in the district-specific prevalence of school attendance.

Spatial small area smoothing models for handling survey data with nonresponse

Spatial smoothing models play an important role in the field of small area estimation. In the context of complex survey designs, the use of design weights is indispensable in the estimation process. Recently, efforts have been made in these spatial smoothing models, in order to obtain reliable estimates of the spatial trend. However, the concept of missing data remains a prevalent problem in the context of spatial trend estimation as estimates are potentially subject to bias. In this paper, we focus on spatial health surveys where the available information consists of a binary response and its associated design weight. Furthermore, we investigate the impact of nonresponse as missing data on a range of spatial models for different missingness mechanisms and different degrees of missingness by means of an extensive simulation study. The computations were performed in R, using INLA and other existing packages. The results show that weight adjustment to correct for missingness has a beneficial effect on the bias in the missing at random setting for all models. Furthermore, we estimate the geographical distribution of perceived health at the district level based on the Belgian Health Interview Survey (2001). Copyright © 2017 John Wiley & Sons, Ltd.

Model-based inference for small area estimation with sampling weights

Obtaining reliable estimates about health outcomes for areas or domains where only few to no samples are available is the goal of small area estimation (SAE). Often, we rely on health surveys to obtain information about health outcomes. Such surveys are often characterised by a complex design, stratification, and unequal sampling weights as common features. Hierarchical Bayesian models are well recognised in SAE as a spatial smoothing method, but often ignore the sampling weights that reflect the complex sampling design. In this paper, we focus on data obtained from a health survey where the sampling weights of the sampled individuals are the only information available about the design. We develop a predictive model-based approach to estimate the prevalence of a binary outcome for both the sampled and non-sampled individuals, using hierarchical Bayesian models that take into account the sampling weights. A simulation study is carried out to compare the performance of our proposed method with other established methods. The results indicate that our proposed method achieves great reductions in mean squared error when compared with standard approaches. It performs equally well or better when compared with more elaborate methods when there is a relationship between the responses and the sampling weights. The proposed method is applied to estimate asthma prevalence across districts.

Estimating dynamic transmission model parameters for seasonal influenza by fitting to age and season-specific influenza-like illness incidence

Dynamic transmission models are essential to design and evaluate control strategies for airborne infections. Our objective was to develop a dynamic transmission model for seasonal influenza allowing to evaluate the impact of vaccinating specific age groups on the incidence of infection, disease and mortality. Projections based on such models heavily rely on assumed 'input' parameter values. In previous seasonal influenza models, these parameter values were commonly chosen ad hoc, ignoring between-season variability and without formal model validation or sensitivity analyses. We propose to directly estimate the parameters by fitting the model to age-specific influenza-like illness (ILI) incidence data over multiple influenza seasons. We used a weighted least squares (WLS) criterion to assess model fit and applied our method to Belgian ILI data over six influenza seasons. After exploring parameter importance using symbolic regression, we evaluated a set of candidate models of differing complexity according to the number of season-specific parameters. The transmission parameters (average R0, seasonal amplitude and timing of the seasonal peak), waning rates and the scale factor used for WLS optimization, influenced the fit to the observed ILI incidence the most. Our results demonstrate the importance of between-season variability in influenza transmission and our estimates are in line with the classification of influenza seasons according to intensity and vaccine matching.

Cost-effectiveness of seasonal influenza vaccination in pregnant women, health care workers and persons with underlying illnesses in Belgium

Risk groups with increased vulnerability for influenza complications such as pregnant women, persons with underlying illnesses as well as persons who come into contact with them, such as health care workers, are currently given priority (along with other classic target groups) to receive seasonal influenza vaccination in Belgium. We aimed to evaluate this policy from a health care payer perspective by cost-effectiveness analysis in the three specific target groups above, while accounting for effects beyond the target group. Increasing the coverage of influenza vaccination is likely to be cost-effective for pregnant women (median €6589 per quality-adjusted life-year (QALY) gained [€4073-€10,249]) and health care workers (median €24,096/QALY gained [€16,442-€36,342]), if this can be achieved without incurring additional administration costs. Assuming an additional physician's consult is charged to administer each additional vaccine dose, the cost-effectiveness of vaccinating pregnant women depends strongly on the extent of its impact on the neonate's health. For health care workers, the assumed number of preventable secondary infections has a strong influence on the cost-effectiveness. Vaccinating people with underlying illnesses is likely highly cost-effective above 50 years of age and borderline cost-effective for younger persons, depending on relative life expectancy and vaccine efficacy in this risk group compared to the general population. The case-fatality ratios of the target group, of the secondary affected groups and vaccine efficacy are key sources of uncertainty.

Eight years of the Great Influenza Survey to monitor influenza-like illness in Flanders

In 2003, an internet-based monitoring system of influenza-like illness (ILI), the Great Influenza Survey (GIS), was initiated in Belgium. For the Flemish part of Belgium, we investigate the representativeness of the GIS population and assess the validity of the survey in terms of ILI incidence during eight influenza seasons (from 2003 through 2011). The validity is investigated by comparing estimated ILI incidences from the GIS with recorded incidences from two other monitoring systems, (i) the Belgian Sentinel Network and (ii) the Google Flu Trends, and by performing a risk factor analysis to investigate whether the risks on acquiring ILI in the GIS population are comparable with results in the literature. A random walk model of first order is used to estimate ILI incidence trends based on the GIS. Good to excellent correspondence is observed between the estimated ILI trends in the GIS and the recorded trends in the Sentinel Network and the Google Flu Trends. The results of the risk factor analysis are in line with the literature. In conclusion, the GIS is a useful additional surveillance network for ILI monitoring in Flanders. The advantages are the speed at which information is available and the fact that data is gathered directly in the community at an individual level.