Contributions to the force of infection of SARS-CoV-2 in Dutch long-term care facilities

BackgroundResidents of long-term care facilities (LTCFs) have been disproportionately affected during the COVID-19 pandemic. To inform decision-making around interventions, we quantified the SARS-CoV-2 infection risk for residents and the relative contribution of different infection sources. We estimated the force of infection (FOI) experienced by Dutch LTCF residents over time and quantified the contribution of residents, LTCF healthcare workers (HCWs), and the general population. Methods & findingsCase data were obtained by Municipal Health Services as part of the Dutch national surveillance program. During the study period (1 October 2020 to 10 November 2021), testing policies included symptom-based testing, exposure-based testing, and facility-wide serial testing. We used a data augmentation approach to include uncertainty in the timing of infection, while taking account of different testing policies. We constructed a Bayesian generalized linear model to estimate group-specific transmission rate parameters and contributions to the FOI experienced by residents. During the study period 36,877 cases were registered among residents and 19,676 among HCWs. The total daily FOI towards residents was highest in December 2020 (1.7*10-3, 95% CI: 1.5*10-3 - 1.9*10-3) and lowest in June 2021 (1.1*10-5 95%CI: 7.6*10-6 - 1.7*10-5). Resident-directed type-reproduction numbers and FOI declined as COVID-19 vaccination rollout started in residents, HCWs, and the older general population (February-May 2021). Most resident infections in spring and summer 2021 were attributable to infections in the general population. The relative contribution of the general population to the FOI decreased in summer 2021 when vaccination was available population-wide. In October-November 2021, type-reproduction numbers and FOI increased again. We observed an increase in residents susceptibility to infection in this period, which was only partially explained by the emergence of the Delta variant. Sensitivity analyses showed that the temporal trends in relative contributions to the FOI were not impacted by assumptions about immunity build-up among residents, nor by underreporting of infections. ConclusionsCOVID-19 vaccination appears to have been effective in reducing SARS-CoV-2 transmission towards residents, although other factors such as seasonality or non-pharmaceutical interventions may also have contributed to this. This effect seemed to have decreased by autumn 2021, which could be due to waning of immunity or changes in control practices. Our estimates of temporal trends in relative contributions to the FOI in LTCF residents can help target intervention efforts.