Román Romero-Ortuño¹, Seán Kennelly²

1. Associate Professor, Discipline of Medical Gerontology, School of Medicine, Trinity College Dublin; Consultant Physician, St James’s Hospital, Dublin
2. Consultant Physician, Department of Age-Related Health Care, Tallaght University Hospital, Dublin; Clinical Associate Professor, Discipline of Medical Gerontology, School of Medicine, Trinity College Dublin

The COVID-19 pandemic has disproportionately affected nursing home residents worldwide, with Ireland having one of the highest reported proportions of COVID-19 deaths in this setting.

In Ireland, the publication of a ‘league table’ of crude number of deaths in affected care facilities sparked controversy on grounds of being potentially inaccurate and unhelpful.

In this analysis, we reviewed these published unofficial mortality data together with official data on quality standards published by the Irish regulator.

Findings:

There was substantial disagreement between the crude number of deaths and the mortality proportion per 100 beds. The association between crude number of deaths and maximum occupancy was significant with a moderate effect size (Spearman’s rho= 0.38, p<0.001, n=146). We found no significant association between occupancy-adjusted mortality and percentage of non-compliance with inspection standards (Spearman’s rho = -0.09, p=0.315, n=140). Specifically, we found no association between compliance with staffing, governance/management, premises and infection control. There was a mildly significant association between higher compliance with staff training and higher occupancy-adjusted mortality. A multivariate regression analysis on n=140 suggested a mild effect of higher overall non-compliance with lower adjusted mortality. This might seem counterintuitive, but a hypothesis is that perhaps higher mortality centres cater for more complex and vulnerable residents and hence may invest more in staff education and development. The data from which this report is based is not nuanced enough to test this hypothesis and a full understanding would require detailed case-mix data.

Conclusion:

League tables with crude COVID-19-related nursing home deaths are likely to be unhelpful and data should at least be adjusted for the size of the facilities. Even in the latter case, extrapolation to quality of care is likely to be inappropriate. Much research is needed to shed light into this complex topic and for this we urgently need a minimum dataset for care homes in Ireland.