PREDICTORS OF SEVERITY AND DURATION OF INFECTIOUS DISEASE OUTBREAKS IN LONG-TERM CARE FACILITIES IN THE SASKATOON HEALTH REGION

Introduction Infectious disease outbreaks in long-term care facilities (LTCFs) pose a significant threat to the health and well-being of residents. While prevention of all outbreaks is not realistic, minimising the severity and duration of outbreaks can be achieved through the prompt implementation of infection control policies, outbreak reporting to Public Health, and outbreak management. In this study, we investigated variables associated with outbreak severity and duration. Objective The objective was to determine if reporting time, the days between illness onset date and date of outbreak reporting to Public Health Services, is a significant predictor of infectious disease severity and duration, after controlling for LTCF location, outbreak type, year and staff to resident ratio. Methods The Public Health Observatory of the Saskatoon Health Region provided LTCF infectious disease outbreak data from 2005 to 2011. A total of 130 outbreaks were eligible for analysis, but only 60 outbreaks were assessed because they had complete data. Generalised linear models were used to test associations of reporting time with outbreak duration and attack rate. These data were modelled using normal and negative binomial distributions, respectively. Outbreak duration was the time, in days, between the index case and the outbreak conclusion. Attack rate was modelled as the number of residents ill, with the number of susceptible residents as the model offset. Model covariates included LTCF location (urban, rural), outbreak type (respiratory, gastrointestinal), year and the LTCF staff to resident ratio. Results Sixty outbreaks occurring in 17 facilities were analysed. Outbreak duration ranged from 4 to 32 days, with a mean of 16.2 days and a median of 16.5 days. The attack rate ranged from 0.0% to 68.7%, with a mean of 21.0% and a median of 16.3%. Reporting time ranged from 0 to 21 days, with mean and median values of 3.6 and 3.0 days, respectively. Reporting time was significantly associated with outbreak duration (0.8±0.2, p=0.0002), while LTCF location, outbreak type, year and staff to resident ratio were not significant (p>0.05). Reporting time was not significantly associated with attack rate (p=0.38). However, urban facilities had lower attack rates (relative rate (RR)=0.6, 95% CI 0.5 to 0.8), and gastrointestinal outbreaks had higher attack rates than respiratory outbreaks (RR=2.2, 95% CI 1.7 to 2.8). Year and staff to resident ratio were not significantly associated with attack rate (p>0.05). Conclusions Our results indicate that the time to report an outbreak is associated with the outbreak duration, and for each day the outbreak is not reported the duration increases by approximately one day. We did not find evidence that reporting time is associated with the attack rate. However, LTCF location and the type of outbreak were associated with attack rate. Attack rates were lower for urban than rural LTCFs, and higher for gastrointestinal than respiratory outbreaks. The amount of missing data was substantial; future analyses will consider imputation methods that will result in unbiased estimates of model associations.