Performances of statistical methods for the detection of seasonal influenza epidemics using a consensus-based gold standard.

Influenza epidemics are monitored using influenza-like illness (ILI) data reported by health-care professionals. Timely detection of the onset of epidemics is often performed by applying a statistical method on weekly ILI incidence estimates with a large range of methods used worldwide. However, performance evaluation and comparison of these algorithms is hindered by: (1) the absence of a gold standard regarding influenza epidemic periods and (2) the absence of consensual evaluation criteria. As of now, performance evaluations metrics are based only on sensitivity, specificity and timeliness of detection, since definitions are not clear for time-repeated measurements such as weekly epidemic detection. We aimed to evaluate several epidemic detection methods by comparing their alerts to a gold standard determined by international expert consensus. We introduced new performance metrics that meet important objective of influenza surveillance in temperate countries: to detect accurately the start of the single epidemic period each year. Evaluations are presented using ILI incidence in France between 1995 and 2011. We found that the two performance metrics defined allowed discrimination between epidemic detection methods. In the context of performance detection evaluation, other metrics used commonly than the standard could better achieve the needs of real-time influenza surveillance.

Large outbreak of urogenital schistosomiasis acquired in Southern Corsica, France: monitoring early signs of endemicization?

OBJECTIVES: Clustered cases of urogenital schistosomiasis were reported in April 2014 among French and German tourists linked to exposure in the Cavu River, Southern Corsica, France, between 2011 and 2013. We set up national surveillance for autochthonous urogenital schistosomiasis to document the largest possible number of cases in order to identify potential sites of transmission and to determine the extent of the outbreak in France and Corsica. METHODS: The early response consisted mostly of prohibiting swimming in the river, performing a nationwide serologic screening of all persons exposed to the river between 2011 and 2013 and treating confirmed cases. Physicians were asked to report all patients with one or more positive antischistosome serologic test. Cases were defined as occurring in a resident of France with serologic evidence of schistosomiasis or schistosome eggs in urine and no history of contact with freshwater in known endemic areas. We documented symptoms as well as place and time of exposure to freshwater for all subjects. To estimate the outbreak size, we modelled the effect of the 2014 nationwide screening on the 2011-2015 time series of serodiagnosed schistosomiasis cases using log-linear autoregression. RESULTS: In 2014, a total of 106 autochthonous cases were reported, including 35 symptomatic infections. All patients had swum in the Cavu during summer 2013. Over 30 000 persons were likely screened for autochthonous schistosomiasis. The model-estimated outbreak size was 338 cases, including 36 serodiagnosed in 2015. CONCLUSIONS: Besides the 2013 outbreak, there is evidence of small-scale transmission in 2015 in Corsica. Early detection and control of recurrences requires raising community and medical awareness.

Socioeconomic inequalities to accessing vaccination against human papillomavirus in France: Results of the Health, Health Care and Insurance Survey, 2012.

BACKGROUND: In France, human papillomavirus (HPV) vaccination coverage among adolescents and young women is low and decreasing. We analysed data from the 2012 Health, Health Care and Insurance Survey with the aim of identifying factors associated with this vaccination. We also compared the socioeconomic profile of unvaccinated young women to that of women who do not undergo cervical cancer screening (CCS). METHODS: Data were collected through interviews and self-administered questionnaires completed by a randomised sample of Health insurance beneficiaries. Two analyses were performed using Poisson regression: one to investigate the determinants of CCS uptake in women aged 25-65 years old (n=4508), the other to investigate the determinants of HPV vaccination in young women aged 16-24 years old (n=899). A sub-analysis was performed in 685 "daughter-mother" couples from the same household in order to analyse the association between participation to CCS in mothers and HPV vaccination in daughters. RESULTS: Factors significantly associated both to a lower CCS uptake and to an insufficient HPV vaccination were the lack of a complementary private health insurance (P=0.023 and P=0.037, respectively) and living in a family with a low household income (P<0.001 and P=0.005, respectively). A low education level was associated to a lower CCS uptake (P<0.001). The absence of CCS uptake in the last three years in mothers was associated to a lower level of HPV vaccination in their daughter (P=0.014). CONCLUSION: Women who do not undergo CCS and HPV unvaccinated young women tend to be of modest socioeconomic status. Unvaccinated young females tend to have mothers who do not undergo CCS and are therefore at risk of benefiting from none of the two cervical cancer preventive measures. The current implementation strategy concerning HPV vaccination in France may therefore increase inequalities regarding cervical cancer prevention.

Assessment of the Middle East respiratory syndrome coronavirus (MERS-CoV) epidemic in the Middle East and risk of international spread using a novel maximum likelihood analysis approach.

The emergence of the novel Middle East (ME) respiratory syndrome coronavirus (MERS-CoV) has raised global public health concerns regarding the current situation and its future evolution. Here we propose an integrative maximum likelihood analysis of both cluster data in the ME and importations in a set of European countries to assess the transmission scenario and incidence of sporadic infections. Our approach is based on a spatial-transmission model integrating mobility data worldwide and allows for variations in the zoonotic/environmental transmission and under-ascertainment. Maximum likelihood estimates for the ME, considering outbreak data up to 31 August 2013, indicate the occurrence of a subcritical epidemic with a reproductive number R of 0.50 (95% confidence interval (CI): 0.30-0.77) associated with a daily rate of sporadic introductions psp of 0.28 (95% CI: 0.12-0.85). Infections in the ME appear to be mainly dominated by zoonotic/environmental transmissions, with possible under-ascertainment (ratio of estimated to observed (0.116) sporadic cases equal to 2.41, 95% CI: 1.03-7.32). No time evolution of the situation emerges. Analyses of flight passenger data from ME countries indicate areas at high risk of importation. While dismissing an immediate threat for global health security, this analysis provides a baseline scenario for future reference and updates, suggests reinforced surveillance to limit under-ascertainment, and calls for alertness in high importation risk areas worldwide.

Barriers to implementation of the new targeted BCG vaccination in France: a cross sectional study.

The aim of the study was to assess factors influencing BCG vaccination among targeted children after the end of universal and mandatory BCG vaccination in France. A cross-sectional study was conducted in 2009 among general practitioners (GPs) of the French Sentinel Network. With the participation of 358 physician-investigators, 920 children were included. Of the 261 children (31%) identified to be at risk of tuberculosis, only 113 (44%) were vaccinated. The median number of French criteria for BCG vaccination correctly cited by the GPs was 3 of the existing 6. Of the 10 proposed, a median number of 5 regions in the world according to their level of tuberculosis risk were correctly classified by GPs. After adjustment using an alternating logistic model, 7 variables were found to be associated with the immunisation status of the target population. Six of these increased the probability of being vaccinated: children older than 6 months (OR=3.4 (CI 95% [1.4-8.6])), residents in central Paris or its suburbs (OR=14.7 [4.4-49.5]), children expected to travel to highly endemic regions (OR=3.5 [1.4-8.6]), those living in unfavourable conditions (OR=19.9 [6.2-63.9]), the GP's good knowledge of vaccination guidelines (OR=1.4 [1.1-1.9]) and the GP's perception of tuberculosis as a common disease (OR=2.2 [1.1-4.5]). Surprisingly, GPs with university training on infectious diseases tended to be more reluctant to follow vaccination guidelines (OR=0.14 [0.1-0.4]). Actions targeted at these factors could contribute to improving BCG immunisation coverage.

Herpes zoster: Burden of disease in France.

This work provides estimates of HZ incidence and HZ-related hospitalization and mortality rates in France, where no immunization programme has been implemented. Herpes zoster data was obtained from the Sentinelles surveillance general practitioners (GPs) network, the PMSI Data processing centre for hospital discharges and from the French National Mortality Database (INSERM CépiDC). The yearly HZ incidence rate averaged 382 cases per 100,000 inhabitants (95% CI 364-405) and exponentially increased with age. The annual rates of hospitalizations and mortality due to HZ varied from 4.14±0.32 to 14.42±0.39 and from 0.11±0.03 to 0.29±0.04 per 100,000 inhabitants, respectively, depending on whether HZ was coded in a 'primary' or 'primary or associated' diagnosis. One or more factors of immunodepression occurred in 43.4% of hospitalized cases and in 21.6% HZ-related deaths.

Early estimates of 2009 pandemic influenza A(H1N1) virus activity in general practice in France: incidence of influenza-like illness and age distribution of reported cases.

In the end of August 2009, an unusually elevated level of influenza-like illness (ILI) activity was reported to the French Sentinel Network. We quantified the observed excess in ILI cases in France during summer 2009 and characterised age patterns in reported cases. An excess of cases has been observed since 5 July, with a time increasing trend. The cumulated estimated excess number of ILI cases was 269,935 [179,585; 316,512], corresponding to 0.5% French population over the period. Compared to the same period in the past years, relative cumulated incidence was greater among young subjects and lower among subjects over 65 years old. Compared to past epidemics, the relative cumulated incidence was greater in children less than 5 years old. This excess of cases may reflect the current spread of the A(H1N1) virus in France, subject to the following limitations: estimates were based on clinical cases consulting a GP; large media coverage may have led to a non specific increase in consultation rates.

Complex segregation analysis accounting for GxE of bone mineral density in European pedigrees selected through a male proband with low BMD.

Osteoporosis is a common multifactorial disorder characterized by low bone mass (BMD) and high susceptibility to low-trauma fractures. Family and twin studies have found a strong genetic component in the determination of BMD, but the mode of inheritance of this trait is not yet fully understood. BMD is a complex trait whose expression is confounded by environmental influences and polygenic inheritance. Detection of potential gene-environment interactions is of great interest in the determination of bone health status. Here we have conducted segregation analyses, using the regressive class D models, in a sample of 100 European pedigrees (NEMO) with 713 subjects (524 measured for phenotypes) identified via a male with low BMD values at either the Lumbar Spine or the Femoral Neck. Segregation analyses were conducted on the residuals of LS-BMD and FN-BMD adjusted for gender, age and BMI. We tested for gene-covariate (GxE) interactions, and investigated the impact of significant GxE interactions on segregation results. Without GxE a major effect was found to be marginally significant in LS-BMD and highly significant in FN-BMD. For both traits the Mendelian hypothesis was rejected. Significant Age x gene and BMI x gene interactions were revealed. Accounting for GxE increased statistical evidence for a major factor in LS-BMD, and improved the fit of the data to the Mendelian transmission model for both traits. The best fitting models suggested a codominant major gene accounting for 45% (LS-BMD) and 44% (FN-BMD) of the adjusted BMDs. However, substantial residual correlations were also found, and these remained highly significant after accounting for the major gene.