Clinical Infectious Diseases · 2012 · 72 citations · 24 references
Veterinary MedicineInfectious DiseasesVeterinary EpidemiologyInfectious Disease EpidemiologyOutbreak InvestigationSpatial EpidemiologyBacterial EpidemiologyGoat Infectious DiseasesIncidence StudiesInfection SurveillanceLivestock Infectious DiseasesAttack RateBacterial Load QuantificationCommunity-based Outbreak InvestigationCoxiella Burnetii TransmissionEpidemiological Tracing Of PathogensFever ScreeningImmunoglobulin M SeroprevalencePoint-source Outbreak Investigation
BACKGROUND: In early 2009, a dairy-goat annex care farm in South Limburg, the Netherlands, reported 220 Coxiella burnetii-related abortions in 450 pregnant goats. These preceded human cases and occurred in a region that was Q-fever free before 2009, providing a unique quasi-experimental setting for investigating regional transmission patterns associated with a Q-fever point source. METHODS: Index-farm residents/employees, visitors, and their household contacts were traced and screened for C. burnetii. Distribution of community cases was analysed using a geographic information system. True incidence, including undetected infections, was estimated regionwide by seroprevalence in a pre- versus postoutbreak sample, and near-farm by immunoglobulin M seroprevalence in a municipal population sample. Environmental bacterial load was repeatedly measured in surface and aerosol samples. RESULTS: Serological attack rate was 92% (24/26) in index-farm residents/employees, 56% (28/50) in visitors, and 50% (7/14) in household contacts, and the clinical attack rate (ie, the proportion of persons seropositive for acute infection who also had clinical illness) was ≥ 80%. Notified symptomatic community cases (n = 253) were scattered downwind from the index farm, following a significant exposure-response gradient. Observed incidence ranged from 6.3% (0-1 km) to 0.1% (4-5 km), and remained high beyond. True incidence of infections was estimated at 2.9% regionwide, extrapolating to 8941 infections; estimated near-farm incidence was 12%. Coxiella burnetii load was high on-farm (2009), and lower off-farm (2009-2010). CONCLUSIONS: Linking a single dairy-goat farm to a human Q-fever cluster, we show widespread transmission, massive numbers of undetected infections, and high attack rates on- and off-farm, even beyond a 5-km high-risk zone. Our investigation may serve as an essential case study for risk assessment in public health and related fields such as bioterrorism response and preparedness.
24
Flexible smoothing with B-splines and penalties
Paul H.C. Eilers, Brian D. Marx · Statistical Science · 1996 · 3.7K citations · Full text
Barbara Schimmer, Ronald ter Schegget, Marjolijn Wegdam et al. · BMC Infectious Diseases · 2010 · 195 citations · Full text
Veterinary Medicine, Infectious Diseases, Veterinary Epidemiology +17