Projected trends in risk of COVID-19–related death among LTC residents were generated using model estimations that incorporated time × LTC resident status as an interaction term (Figure 1). The median IRR for COVID-19 death in LTC residents compared with community-dwelling individuals rose from 8.03 (95% CrI, 1.96-23.32) on March 29 to 87.28 (95% CrI, 6.44-769.76) by April 11, 2020.
In analyses focused on risk of death within LTC facilities, we found that lagged infections among staff were associated with death among residents (Figure 2) and were significant at all lags (0 to 7 days) after adjustment for date and numbers of residents with infection. The strongest associations were seen with staff with infection at a 2-day lag (relative increase in risk of death per staff member with infection, 1.20; 95% CI, 1.14-1.26) and a 6-day lag (relative increase in risk of death per staff member with infection, 1.17; 95% CI, 1.11-1.26). In contrast, the association between infection in residents and subsequent death was variable and far weaker than the associations seen for staff. It was statistically significant only at a 0-day lag (increased risk of death per infected resident, 1.08; 95% CI, 1.01-1.15).
Discussion
In this analysis, we documented the rapid spread of COVID-19 through Ontario’s LTC system, with a marked increase in risk of death among older residents with frailty during a brief period from late March to early April 2020. Issues such as crowding, use of communal space, low staffing ratios, and high care needs (with resultant high density of physical contact between residents and staff) have long been recognized as key drivers of susceptibility to outbreaks in the LTC facility setting.4,7 In the context of COVID-19, this susceptibility has proven particularly deadly, with (as we demonstrate here) an incidence of mortality more than 13 times greater than that seen in community-living adults older than 69 years during a similar period, with relative risk of death rising sharply over time.
We also found that documented infection in facility staff, as opposed to residents, is a strong identifiable risk factor for mortality in residents, with temporality suggesting that residents are infected by staff and not vice versa. Although it might be argued that the limited nature of testing and the tendency to test staff and residents for COVID-19 after a resident dies might lead to spurious associations between identified infections and deaths, such spurious associations might result in equivalent effect sizes between residents and staff, rather than the divergent effect sizes observed here. The greater mobility and connectedness of staff, compared with residents, lends biological plausibility to this association.11
Transmission of infection is not the only mechanism by which infection in staff could result in increased mortality in an older population. Fear of COVID-19 could result in absenteeism by staff, which could itself lead to death through dehydration and other mechanisms in a high-needs population, which would be consistent with the lagged associations we observed here. Such tragedies have been documented in Canada recently.12 While this analysis was completed in April 2020, the subsequent course of events in Ontario has validated the trends we identified; as of June 10, 2020, 1766 COVID-19–related deaths have occurred among LTC facility residents in Ontario; deaths among residents of LTC facilities constituted 71% of all COVID-19 deaths in Ontario. Furthermore, 7 LTC facility staff and volunteers have died of the disease.13 Across Canada, it is estimated that more than 80% of COVID-19 deaths have occurred in the LTC facility setting; similar epidemiology is documented in the United States14 and several European countries,15,16 suggesting that our findings are likely to be generalizable.
The prevention of such deaths requires strategic guidance from health regions and the provision of sufficient testing and personal protective equipment. Provision of personal protective equipment has benefits both in bidirectional prevention of SARS-CoV-2 transmission and in providing workers peace of mind to stay on the job. Expanded testing, including testing of minimally symptomatic infection, will facilitate the early identification of infection and the implementation of effective infection control strategies. Integrated regional approaches to LTC facility human resource management, such as limiting workers to a single facility and ensuring that these workers earn a living wage to prevent the need for multiple jobs while at the same time maintaining adequate staffing levels, are also needed.17
Limitations
Like any observational study, this study has limitations, including possible incompleteness of data collected rapidly during an outbreak, inconsistency in testing across Ontario, and absence of individual-level data on LTC facility infections and deaths. We have been unable to explicitly structure autocorrelation in our time series owing to effects on standard errors resulting from the limited size of our data set. We regard our outcome of interest, ie, death from COVID-19 among residents of LTC facilities, to be less likely misclassified than nonfatal infection among staff and residents. If misclassification of infection status in these individuals occurs at random, that would likely mean the associations reported here are lower-bound effect sizes. If underidentification of both fatal and nonfatal infections among residents and staff are clustered by home, that would result in association estimates that are biased upward. The temporality in the associations we observed provides a degree of reassurance in this regard.
Conclusions
This study documented that the rapid movement of COVID-19 through Ontario’s LTC facility system has resulted in a marked surge in mortality in that population relative to community-living older adults. We found evidence that associates mortality with infection among LTC staff, highlighting the urgent need for improved infection control, more widespread testing, access to personal protective equipment, and economic protections and support for those who do this important work.
Back to top
Article Information
Accepted for Publication: June 14, 2020.
Published: July 22, 2020. doi:10.1001/jamanetworkopen.2020.15957
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Fisman DN et al. JAMA Network Open.
Concept and design: Fisman, Tuite.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Fisman, Bogoch.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Fisman.
Obtained funding: Fisman.
Administrative, technical, or material support: Bogoch, McCready.
Supervision: McCready.
Conflict of Interest Disclosures: Dr Bogoch reported serving as a consultant for BlueDot. No other disclosures were reported.
Funding/Support: The study was supported by grant OV4-170360 to Dr Fisman from the Canadian Institutes for Health Research.
Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
References
No comments:
Post a Comment