Measuring Mortality Using Public Data Sources 2019 – 2020

by Gazette Reporter

Experimental analysis shows promising results using the website RIP.ie to provide close to ‘real time’ mortality trends in Ireland

  • There was a strong correlation, of more than 99%, between CSO data and notices placed on RIP.ie for the years 2016 and 2017 indicating that RIP.ie can be used to accurately measure trends in mortality
  • There was a pronounced increase in death notices placed in April 2020, with the numbers of death notices rising from 2,861 in March to 3,502 in April before decreasing to 2,635 in May
  • Based on the analysis of thousands of death notices, excess mortality is estimated at 1,072 (i.e. deaths above those likely to be experienced under normal circumstances)
  • There were significant increases in death notices related to Health Information and Quality Authority (HIQA) registered Older Persons’ Facilities, which increased to 1,237 in April from 722 in March
  • Dublin and surrounding counties were heavily impacted, with notices for Dublin increasing from 763 in March to 1,103 in April, while there were also rises in Cavan, Donegal and Monaghan
  • The use of terms such as ‘private’, ‘broadcast’ and ‘webcam’ in death notices increased significantly from March 2020, with ‘webcam’ mentioned almost 500 times between March and May 2020 compared with only 30 mentions in the five months to the end of February

The Central Statistics Office (CSO) has published an experimental analysis on Measuring Mortality Using Public Data Sources. The release examines the use of the website www.RIP.ie to monitor trends in mortality in Ireland.

Commenting on the results, Statistician, John Flanagan, said: ‘The CSO has a policy of careful and considered investigation of a data source before employing it to inform official statistics. However, due to the COVID-19 pandemic, we began to explore experimental ways of obtaining up-to-date mortality data. Since the end of March 2020, the CSO has been using the website, RIP.ie to keep track of death notices. Analysis of death notices was conducted as far back as 1 October 2019, to include the last month before the first global cases of COVID-19 were notified. The Death Events Publishing Service (DEPS) of the General Register Office (GRO) has been monitored in tandem, to validate the volumes of death notices published.

Due to the Irish custom of holding funerals within two to three days following death, these notices are usually placed in a fast and efficient manner, providing a valuable crowd-sourced means of tracking deaths. The notices are placed close to ‘real time’, we found that the average length of time between date of death and publication is about 1.1 days. In comparison, the statutory time limit is three months for the registrations of deaths in the State.

The analysis conducted for October 2019 to June 2020 shows some important trends. Most notable is the increase in death notices in April which stands in contrast with previous years. Numbers of deaths notices increased to 3,502 in April from 2,861 in March. In comparison the average number of deaths for April for the years 2013-2017 was approximately 2,500.

Based on the analysis of death notices, the estimate for excess mortality to 30 June 2020 is 1,072. This assumes that, in the absence of COVID-19 deaths, mortality would have followed a trajectory similar to previous years. At this stage, this is a speculative estimate based on experimental data. It is important to put this figure of 1,072 deaths within the context of around 30,000 deaths per year. It is possible that there will be lower mortality levels later in the year given the concentration of these deaths in nursing homes and the older population (over 90% of COVID-19 deaths based on Health Protection Surveillance Centre (HPSC) data have been in the 65+ age group). In that sense this analysis can only give an indication of excess mortality at a point in time and data for the full year of 2020 would be required to provide a more definitive picture of excess mortality.

The death notices also provide useful information on gender and place of death, especially nursing homes, at such a critical time. The majority of approximately 26,000 death notices for the period under investigation were read. In addition to county or country, it was possible to determine the place of death in most cases. Although reading each death notice is initially onerous, it allows for the building of a model for automatic classification. Information on place of death shows increases in death notices were clearly concentrated in Dublin and surrounding counties. There were also additional death notices placed in Cavan, Monaghan and Donegal. Increases in death notices related to HIQA registered Older Persons’ Facilities were also observed. In April, 35.3% of all death notices were related to older persons’ facilities, compared with 25.2% in March.

Death notices also showed not only the impact COVID-19 has had on mortality trends in Ireland but also the way in which it has shaped behaviour in the changing language observed in death notices. References to ‘private’ gatherings increased dramatically from March onwards. The term ‘broadcast’ was used once between October 2019 and February 2020 but was mentioned more than 100 times between March and May 2020. ‘Webcam’ was used less than 30 times in the five months up to the end of February 2020, but was mentioned nearly 500 times between March and May.

However, while death notices placed on RIP.ie appear to accurately follow the increased mortality associated with COVID-19, it is not used by everyone in Ireland. Contact with organisations representing minorities in Ireland indicated that there is less usage of RIP.ie amongst minority communities. In this sense, this note represents an initial experimental analysis designed to indicate trends in mortality. When published, official statistics on mortality will allow for more thorough and definitive analysis and it is only in the context of full year statistics for 2020 that the full impact of the COVID-19 pandemic on mortality might be understood.

Related Articles