Data relating to COVID-19 in RoI is publicly available through several bodies. Data in this study was sourced from the DATA.GOV.IE website, which collates data from several national organisations, namely the Central Statistics Office (CSO), the Health Protection Surveillance Centre (HPSC), the Health Service Executive (HSE), the Department of Health and the Open Data Unit of the Department of Public Expenditure and Reform (DPER). The data are publicly available.

YLL is the product of the number of deaths (M) and the average remaining life expectancy (RLE) at the time of death: YLL = M x RLE [6].

COVID-19 case definitions in RoI have been routinely updated during the pandemic in accordance with the European Centre for Disease Prevention and Control (ECDC) guidance and updates [8]. Mortality data independently stratified by sex and age-group for the required period was published by the CSO on 10 March 2021 for events created on the Computerised Infectious Disease Reporting (CIDR) system up to midnight Friday 05 March 2021 [9]. To exclude data from 01 March 2021 to 05 March 2021 and to allow for mortality reclassification, the DoH daily briefings were reviewed up to and including 31 March 2021. To complete the sex data, the sex ratio from the defined data was applied to the “unknown” cases. To complete the age-group data the age-group ratio from the defined data was applied to the “unknown” cases. Finally, the formulated sex ratio was applied to each age-group (See Supplementary Appendices S1, S3).

GBD standard life tables for 2019 were used to calculate remaining life expectancy (RLE) for each sex separately [10]. GBD life tables are widely used internationally in BoD studies. As this study uses population level data as opposed to individual level data the average RLE was calculated for each age-group by assuming that the average age of death was the midpoint in each age-group, except for the 80+ category where the average age of death was assumed as the average age of persons in the 80+ category recorded in the Irish Life Tables [11]. Population data was taken from the CSO’s most recent Population and Migration Estimates which were calculated for April 2020 [12].

YLD was defined as the product of the number of incident cases (N), the average duration until recovery or death (D), and the disability weight (DW), which is reflective of the reduction in one’s health [measured on a scale from 0 (no impact on full health) to 1 (death)]: YLDinc = N × D × DW [6].

YLD in this study included 5 health states: asymptomatic, moderate, severe, critical, and post-acute consequences (PAC). Categorisation of cases by severity is necessary as both duration and disability weight are included in the calculation of YLDs (it would not be expected that a person with moderate COVID-19 would contribute similar YLDs as a person with severe COVID-19). Several assumptions were made in relation to informing a person’s categorisation into one of these health states, as seen in Table 1. Polymerase Chain Reaction (PCR) testing criteria for the general public in RoI continues to be based on persons presenting with at least one symptom (being symptomatic) [13], therefore, asymptomatic in this study was defined as a person infected with COVID-19 who did not present for PCR testing; however, this did not account for infections discovered during screening programmes. Asymptomatic cases carried no value for YLD estimation. As described by the European Burden of Disease Network/ECDC consensus disease model, these health states, their description, and their disability weights are based on those from the GBD 2019 study for infectious diseases of the lower respiratory tract, except for those requiring intensive care (Critical), which was defined by the European Disability Weight study [6, 14] (See Table 1).

Morbidity data for all health states, independently stratified by sex and age-group for the required period remains openly available on data.gov.ie [15], except for the health state “asymptomatic,” which was assumed to have a 3:1 ratio based on seroprevalence studies [16, 17]. Additionally, the transition probability for the health state “PAC” was assumed to be 13.3% of the overall symptomatic incidence with a duration of 28 days [18]. For estimated mean duration relating to health state “moderate,” the mean duration for lower respiratory tract infections from the GBD 2019 study was used [14]. Estimated mean duration for health states “severe” and “critical” were provided by a recent study based on Irish hospital data [19]. Similar to YLL estimates above, a formulated sex ratio was applied to each age-group (See Supplementary Appendices S2, S4).

DALYs are calculated simply by summing the YLLs and YLDs (DALY = YLL + YLD). (See Supplementary Appendix S5).

Ethical approval for this study was obtained from the Social Research Ethics Committee of University College Cork (UCC).

As DALYs are a unit of time based on person-years, our incident data are scaled by a factor of 1 365.25 , this reflects the duration of the year that the health loss is suffered. For ease of comparison with previous studies the YLLs, YLDs and DALYs were also estimated per 100,000 population. The uncertainty intervals (UI) from the GBD 2019 life tables were used to calculate the formulate the upper and lower intervals of our YLL. The respective disability weight UIs from the GBD 2019 study for infectious diseases of the lower respiratory tract and the European Disability Weight study were used to estimate the upper and lower intervals of the YLD, these UI were subsequently used to estimate the upper and lower estimates of the DALY. Uncertainties were present in our analyses, particularly in relation to the estimation of the proportion of cases who transitioned to health state “PAC” and also the duration of health state “PAC.” A sensitivity analysis assessed the effect of a 5% and 25% transition probability of symptomatic cases to “PAC.” We also assessed “PAC” durations of 14 and 56 days. Additionally, a combination of assumptions relating to “PAC” were applied to minimise and maximise the YLD. For minimisation we halved the transition probability and the duration of “PAC.” For maximisation we doubled the transition probability to “PAC” and doubled the duration of “PAC” (See Table 6) (See Supplementary Appendix S6).

There were 220,273 PCR confirmed cases of COVID-19 with a total of 4,500 deaths (confirmed or probable) as a direct result of COVID-19 within the studies stated parameters. The incidence numbers for each health state are seen in Table 2.

A fatality rate of 0.49% was estimated when considering all infected persons (asymptomatic + symptomatic). Of symptomatic cases only, we estimate a fatality rate of 2.0% and a hospitalisation rate of 6.5%. Of those hospitalised, 10.8% required treatment in an intensive care unit (ICU).

The sex of each case was defined for 4,154 cases and was “unknown” for 346 cases. Age-group was defined for 4,149 cases and was “unknown” for 351 cases. We estimate YLL of 50,823.1 (50,186, 51,309.7) [males: 24,982.2 (24,635.6, 25,164); females: 25,840.9 (25,550.4, 26,145.7)]. Based on age, persons in the 65–79 age-groups had the highest YLL at 21,153.4 (20,868.2, 21,443.4) [males: 10,459.8 (10,319.4, 10,600.2); females: 10,663.5 (10,528.7, 10,800.9)]. The highest number of deaths by sex and age-group was seen in males 80+ (1,507, 2143.7 deaths per 100,000 persons). The average age of death was 79 (males: 78.8; females: 79.2) (See Table 3).

Sex was defined in 219,507 cases and was “unknown” in 766 cases. Age-group was defined for 219,466 cases and was “unknown” in 807 cases. We estimate YLD of 799.8 (95% UI 535.7, 1,126.2). Males constituted YLD of 375.9 (251.8, 529.3) and females 423.9 (283.9, 596.9). The largest contributing age-group was “25–44” with YLD of 271.9 (182.1, 382.9) and the largest contributing health state is “PAC” with YLD of 491.8 (332.4, 691.7). The largest contributing combined sub-population (sex, age-group, and health state) were “females” “25–44” “PAC” with YLD of 88.6 (59.9, 124.6), this population consisted of 5,279 persons (See Table 4).

We estimated that COVID-19 caused 51,622.8 DALYs (50,721.7, 52,435.8) in the full year period. Overall, YLL contributed 98.5% towards the DALYs with the remaining 1.5% attributed to YLD. The sub-population with highest DALYs were “Females” “65–79” with DALYs of 10,663.5 (10,528.7, 10,800.9), which translates to 3,797.5 (3,749.5, 3,846.5) DALYs per 100,000 persons. However, the largest DALYs per 100,000 persons was seen in the Male 80+ population with a total of 12,888.8 (12,665.6, 12,899.7). We estimated 11.5 (11.3, 11.7) DALYs per death (See Tables 5, 6).

For contextualisation, we compared this study’s findings with RoI GBD data for 2019 [26]. COVID-19 is likely to be the second highest cause of death in the RoI over our study’s duration, with only ischemic heart disease (IHD) causing more deaths. Due to mortality being such a high contributing factor to the DALYs, COVID-19 is also likely to have the second highest YLL. Additionally, our COVID-19 DALY estimates are comparable in magnitude to estimates due to “Unintentional Injuries” (54,835.6) and “Communicable, Maternal, Neonatal, and Nutritional Diseases” (48,853.6), while our YLD estimates are comparable in magnitude to Non-Hodgkin Lymphoma (682.0) and Idiopathic Developmental Intellectual Disability (642.5) estimates. Whilst not a direct aim of this study, this contextualisation offers a glimpse of the severity of COVID-19 within the first year of the pandemic.

There were several methodological challenges observed in this study. Publicly available data was identified through the CSO and the HPSC; however, no data was publicly available collated by sex and age, nor was it available compiled specifically for this study’s date range. This paucity of data inputs on specific variables is a constraint, and therefore, calibrations and modelling assumptions were made. Furthermore, 10% of missingness is acceptable in scientific literature whereby the missingness is presumed to be random. Aggregate and population-level data often have this inherent limitation of granularity. It is envisioned that future research will include individual-level data when accessible so that this methodological gap can be addressed. However, in this study in order to overcome such issues, necessary assumptions were made. Where data were recorded as “Unknown” (i.e., sex), the ratio from the defined data was applied to the “unknown” data. Age of death was taken as the midpoint in each age category except for the 80+ category where the average age of death was assumed as the average age of persons in that category as per Irish life tables. Privacy restrictions mean that publicly available mortality data with sub-population counts of <5 (i.e., males 0–14) remain unavailable, therefore this study’s YLL estimations are applicable only for those aged 24 and above; this in turn will result in DALY underestimation. We estimated a maximum underestimation of 1,117.2 (95% UI 1,112.4, 1,122.4) DALYs. To allow for mortality reclassification, the Department of Health daily briefings were reviewed up to and including 31 March 2021; however, there may have been further reclassification post this date. Despite the above uncertainties, our estimates of COVID-19 mortality are unlikely to significantly differ from the final mortality estimates. Due to the recency and ongoing nature of the COVID-19 pandemic, no attempt was made to account for multimorbidity, additionally, accounting for co-morbidities within DALYs is hotly debated at present and introduces a number of ethical issues whose discussion is beyond the scope of our paper.

BoD studies use DALYs as a tool to assist international comparability; however, in respect of COVID-19 direct comparability between countries remains difficult as studies may use dissimilar life valuation tables. Additionally, due to existing pre-pandemic vulnerabilities specific to each country’s population, the exact same response applied across countries would likely yield differences in DALY estimates. Therefore, comparisons will not be straightforward, as to achieve the same results proportionately, any differences in investment in interventions would be required.

Comparison of COVID-19 cases with 2019 GBD Study is for contextualisation only. Given the age profile, it is implausible that all COVID-19 deaths are additional (i.e., deaths in the higher age-groups may have occurred irrespectively due to other causes). A lack of clarity remains in relation to “dying from COVID” and “dying with COVID.” However, case definitions in RoI have been routinely updated during the pandemic in accordance with ECDC guidance and updates [8].

This study estimates the direct burden relating to COVID-19, and no attempt was made to measure the indirect burden resulting from COVID-19, or the behavioural changes and implemented control measures (i.e., mental health, reduced screening capacity, missed developmental checks etc.).

COVID-19 has had a detrimental effect on the health of the Irish population [5, 27]. Despite extensive control measures, COVID-19 is likely to be the second highest cause of death from a single disease in the RoI within this study’s duration. It would seem highly likely that the synergistic effects of public health measures implemented during this time period could have had a significant positive population health impact; older adults bore an unequal health burden which ultimately resulted in greater DALYs for this population which were overwhelming informed by YLL. The most obvious strategy for DALY reduction relating to COVID-19 would be to focus on mortality reduction, with particular focus on high-risk groups [28].

Future research directions must include, in the first place, longitudinal studies from other phases of the pandemic which will likely generate very different values than this one, and hence paint a more accurate picture of the COVID-19 burden overall. Additionally, future research must also focus on estimating COVID-19 specific DWs, an extensive BoD study relating to the indirect effects of the COVID-19 pandemic, and an extensive study relating to the profile and timeline of “Long-COVID.”

Similar to comparable nations, the health of the Irish population has been impacted by the COVID-19 pandemic, the older populations bearing an unequal burden, particularly in relation to mortality. This work sets out to develop and adapt a standardised COVID-19 BoD methodology framework for Ireland and other comparable EU nations and beyond.