A cross-sectional study of the relationship between delayed healthcare utilisation and chronic conditions among older adults during the COVID-19 pandemic in Ireland [version 1; peer review: awaiting peer review]

Background: Many healthcare services were cancelled or postponed during the coronavirus 2019 (COVID-19) pandemic, likely impacting the management of chronic conditions prevalent among older adults in Ireland. Methods: Data from the Irish Longitudinal Study on Ageing COVID-19 study and previous waves were used. Taking healthcare demand into account, the relationship between delayed healthcare utilisation among older adults (≥50) with chronic conditions was examined. Further analyses examined the reasons for delays in healthcare utilisation, and whether they were the result of the reduced availability of healthcare services or participant decisions. Results: In total, 31.6% of participants reported experiencing healthcare delay. The first analysis found that older adults with two or more chronic conditions were more likely to have experienced healthcare delay than those with no chronic conditions (odds ratio (OR): 1.46, 95% confidence interval (CI): 1.11, 1.90). The second analysis found that older adults with two or more chronic conditions were more likely to have healthcare delayed by the provider (OR: 1.73, 95% CI: 1.16, 2.56), and were also more likely to delay their own healthcare (OR: 1.62, 95% CI: 1.14, 2.29) than older adults with no chronic conditions. Additionally, people aged 70 years and over, females, those with problematic alcohol consumption, those with third-level education, those who had visited the GP and those who reported polypharmacy were all more likely to experience healthcare delay, while older adults living with others and those living outside Dublin were less likely to experience healthcare delay. Open Peer Review Reviewer Status AWAITING PEER REVIEW Any reports and responses or comments on the article can be found at the end of the article. HRB Open Research Page 1 of 12 HRB Open Research 2021, 4:112 Last updated: 13 OCT 2021


Introduction
The first case of the SARS-CoV-2 virus was diagnosed in the Republic of Ireland on the 29 th February 2020 (RTÉ). Since then, the coronavirus 2019 (COVID-19) pandemic has had an enormous impact on the lives of older adults, both in terms of the virus itself, with older adults having a higher risk of serious illness and mortality from COVID-19 (Williamson et al., 2020), and in terms of the impact of restrictions enacted to reduce its spread. In Ireland, stringent lockdown measures saw the intermittent closure or delay of all non-essential services including healthcare services such as outpatient clinics, cancer screening services, elective inpatient procedures and day care services (Health Service Executive, National Screening Service). It is likely that the closure of these services impacted the healthcare utilisation of older adults and in particular those with chronic conditions as frequent users of healthcare services. People with multimorbidity (two or more chronic conditions) are more likely to use healthcare services such as general practitioner (GP) or inpatient and outpatient hospital services (Bähler et al., 2015;Glynn et al., 2011;McDaid, 2013). This study seeks to examine the relationship between delayed healthcare utilisation and older adults with chronic conditions during the COVID-19 pandemic in Ireland, with a particular focus on those with multimorbidity.
Internationally, there is evidence of a significant decrease in healthcare utilisation during the pandemic. For example, a systematic review of studies from 20 different countries, found, on average, a 37% reduction in healthcare utilisation across all age groups (Moynihan et al., 2020), while a German study of over two million patients, aged 65 years and over, found that hospital admissions had decreased by 39% overall (Michalowsky et al., 2021). In one American study, 30.3% of participants aged 65 and over delayed routine care and 4.4% delayed emergency care during the pandemic (Czeisler et al., 2020). The English Longitudinal Study of Ageing (ELSA) COVID-19 sub-study found that 14% of participants, aged 50 years and over, who required health services did not try to access them (Zaninotto et al., 2020). While a study in Hong Kong, of adults aged 60 years and over, who had a least two chronic conditions, found that there was an increase in missed medical appointments from 16% before the COVID-19 pandemic to 22% during the pandemic (Wong et al., 2020). The reasons for these reductions in healthcare utilisation are complex. Several factors appear to contribute to the decrease including both providers and patients cancelling or delaying appointments. Importantly, several studies have found that some groups have been impacted more than others. For example, people with more than one chronic condition (Macinko et al., 2020;Topriceanu et al., 2021;Zaninotto et al., 2020), females (Macinko et al., 2020;Topriceanu et al., 2021), people with higher levels of education (Macinko et al., 2020) and those who had visited the GP in the previous twelve months (Macinko et al., 2020), were more likely to experience some type of healthcare cancellation or delay.
Research in Ireland reflects international findings. A national online survey found that 32% of participants had postponed medical treatment during the pandemic (National University of Ireland Galway, 2020). In the first few months of the pandemic, GPs in Ireland reported a reduction in non-COVID-19 related consultations for people aged 70 years and over (Homeniuk & Collins, 2021). There is also evidence of decreased use of specific types of healthcare services, such as a 32.5% decrease in the numbers of people attending emergency departments in March 2020, compared to March 2019(Brick et al., 2020; a reduction of 17.1% in trauma referrals to the National Neurosurgical Centre (Horan et al., 2021); a reduction in presentations with self-harm in March-April 2020 followed by an increase in May 2020 (McIntyre et al., 2021); and an overall reduction in small biopsy surgeries and cancer resections in two public hospitals (O'Connor et al., 2021). Crowley & Hughes (2021) provide a more detailed description of the impact of the pandemic, on both health service need and demand in Ireland, finding that there were varying degrees of reduction in service provision across different types of healthcare services and increased use and demand after the first wave of the pandemic. Equally, reductions in demand can vary between different cohorts. Fahy et al. (2020) found an overall reduction of 21% in trauma presentations to the emergency department of one Irish hospital, but only a 2% decrease in trauma presentations from people over the age of 65.
Similarly, reduced use is not necessarily an indication of reduced need but potentially of unmet need. Prior to the pandemic, there was already evidence of unmet need for healthcare services in Ireland, with one study finding that approximately 4% of participants (aged 16 years and over) reported an unmet need for healthcare (Connolly & Wren, 2017). It is likely that this figure is higher for older adults as healthcare use increases with age (Roe et al., 2020). Indeed, among older adults who report difficulties with activities of daily living, approximately 30% reported an unmet health or social care need, some or all of the time (McGarrigle & Kenny, 2020). It is likely the pandemic has impacted on levels of unmet need, with one study reporting that over 57% of patients (over the age of 70), in a large public hospital, had some type of healthcare appointment cancelled during the pandemic, with outpatient hospital appointments the most often reported service to be cancelled (Bailey et al., 2021).
The reduction in healthcare utilisation and potential increase in unmet needs may negatively impact on a range of health outcomes. A German study has already found a significant decrease in the diagnoses of specific diseases in April/May 2020 compared to April/May 2019, such as a decrease in diagnoses of dementia (38%), stroke (38%), diabetes (38%) and Parkinson's (32%) (Michalowsky et al., 2021). The pandemic has resulted both in the cancellation of healthcare services and has likely impacted the health seeking behaviours of older adults with chronic conditions, meaning that they may not be receiving the care they need and creating a pent-up demand for healthcare services once restrictions are lifted. Given the important implications of this both for the health of individuals and the health service as a whole, this study seeks to answer two questions: (1) With a primary focus on chronic conditions, what were the individual characteristics associated with delayed healthcare utilisation among older adults in Ireland during the COVID-19 pandemic?
(2) What are the factors driving the delay in healthcare utilisation, is it due to delays or cancellations by the provider or the participant?

Data and study design
The Irish Longitudinal Study on Ageing (TILDA) is a nationally representative cohort study of community-dwelling adults aged 50 years and over in Ireland. The first survey was conducted in 2009 and the study is now in Wave 6 of data collection. The information gathered is comparable to that of other longitudinal studies globally, including the Health and Retirement Study (HRS) on which TILDA is based. There are three elements to the study: a computer assisted interview conducted by trained interviewers in the participants home, a self-completion questionnaire and a health assessment conducted by trained nurses. For a more detailed description of the design and methodology of TILDA see Donoghue et al. (2018); Kearney et al. (2011);Kenny et al. (2010); Whelan & Savva (2013). In response to the COVID-19 pandemic, TILDA administered a self-completion questionnaire (SCQ) to its participants. All current TILDA participants were invited to participate in this study. The SCQ (see extended data) examined the impact of COVID-19 on the lives of older adults in Ireland, with questions on topics such as changes to daily activities, protective behaviours, health and wellbeing, perceptions of ageism, healthcare utilisation and COVID-19 symptomology and testing. The World Health Organisation's COSMO (COVID-19 Snapshot Monitoring) toolkit guided the design of the SCQ. Furthermore, many of the questions are similar to those used in previous TILDA waves and in other longitudinal studies on ageing (Ward et al., 2021). The SCQ was posted to all current TILDA participants in July 2020 and returned surveys were accepted until November 2020. The analysis presented here also includes data from Wave 5 of TILDA which was conducted in 2018. TILDA has a two-stage sampling design: the first stage involves a random selection of geographic clusters based on the Geodirectory in Ireland, and the second stage is a probability sample of addresses from within those clusters (Kenny et al., 2010). To account for this two-stage sampling design, and address potential selection bias, we applied survey weights that adjusted for clustering and stratification. These weights also use the complex survey option and, to address potential attrition bias over time, adjust for non-response due to attrition.

Consent and ethical statement
Along with the SCQ, an information leaflet and participant consent form were posted to all participants (Ward, 2021). This ensured that informed written consent was obtained from all participants. Ethical approval for the COVID-19 study was granted by the Irish National Research Ethics Committee . For the wider TILDA study, ethical approval is provided for each wave by the Dependent variable (I) -Delayed healthcare As part of the COVID-19 study, participants responded to two questions about delayed healthcare. The first asked: "Since the outbreak of the COVID-19 pandemic in March 2020, was there any time when you needed medical (including dental) care, but delayed getting it, or did not get it at all? Yes or no.
And secondly: "What type(s) of care or health services did you delay?" Major surgery (requiring a hospital stay of one or more nights), minor surgery (as an outpatient or day case), seeing your General Practitioner, getting a prescription filled, getting medications, dental care, optician, public health of Community Nurse, occupational therapy, physiotherapy services, psychological/counselling services, hearing services, respite services and other.
Delayed healthcare was the outcome of interest and was created by combining responses to these two questions as a binary variable: (1) No healthcare delayed (2) Healthcare delayed. 'No healthcare delayed' consisted of participants who had answered no to question one and had not listed any services in question two. 'Healthcare delayed' consisted of participants who had answered yes to question one or who had named any of the services listed in question two, even if they had answered no to question one.
Dependent variable (II) -Reasons for delayed healthcare A second analysis was then conducted to explore research question two by examining the reasons behind why healthcare was delayed, whether it was cancellations or delays by the provider or the participant. The dependent variable was constructed by using dependent variable I (as outlined above) and combining it with the question: Why did you delay or not get that care?

I could not get an appointment
Response categories one to three were coded as healthcare delayed by the participant and responses four to six as healthcare delayed by the provider. The dependent variable consists of three categories: 'No healthcare delayed', 'healthcare delayed by provider' and 'healthcare delayed by participant'. Some participants (n=100) had both healthcare delayed by the provider and had delayed their own healthcare. These were included in the third category 'healthcare delayed by participant'. This is a significant number of participants, and initially we explored including this as an additional category in the dependent variable however this was not feasible given the sample size. For the purposes of transparency and clarity, we include an additional analysis in the extended data which categorises the above participants (n=100) under 'healthcare delayed by provider' (Hennelly et al., 2021). Participants who had healthcare delayed but did not give a reason why, were excluded from the second analysis (n=133).

Independent variables
In this analysis, having a chronic condition 1 (none, one or two or more) was the main independent variable of interest. Additional, independent variables were chosen through examination of the existing literature and consist of socioeconomic variables including: age (50-69 years or 70 years +), gender (male or female), education level (primary, secondary or third), living arrangements (living alone or with others), location (living in Dublin, another urban setting (town or city), or a rural location) and marital status (married, never married, separated/divorced or widowed). Additional variables 2 theorised to have a relationship with healthcare utilisation were included. These were: having a cardiovascular condition 3 (none, one or two or more); being a current smoker (yes or no); problematic alcohol consumption (yes or no) assessed using the CAGE questionnaire (Mayfield et al., 1974); health insurance coverage (no cover, medical/GP visit card only, health insurance only or both a medical/GP visit card and health insurance); one or more GP visits in the previous 12 months (yes or no); and polypharmacy, where participants report taking five or more medications, including supplements (yes or no). A time variable was included to account for differences that may have existed as the pandemic and related-restrictions changed over time. This is a binary variable which divides participants into two groups, those who returned the survey before the 1 st August 2020, and those who returned it on or after the 1 st August 2020. This date was chosen as Ireland slowly transitioned out of lockdown around this time which could potentially impact responses. The same group of independent variables were used in both analyses with two exceptions. In the second analysis marital status was recoded as a binary variable of married or other (never married, separated/divorced and widowed) and cardiovascular conditions was recoded also as a binary variable, to be none or one or more. Both of these changes were made to ensure adequate sample size in the respective categories.

Statistical analyses
The survey weights described above were applied to all analyses and 95% confidence intervals (CI) are reported throughout. The first analysis of delayed healthcare used logistic regression, as the dependent variable was a binary variable. The second analysis used a multinomial logit regression. In both analyses, estimates are expressed in odds ratio (OR) with 95% confidence intervals. All analyses were conducted in Stata 15 (StataCorp, 2017).

Participant characteristics
A total of 5,535 questionnaires were posted out to participants, with 3,922 participants responding. The final analytic sample included 3,001 participants. A description of study participants is provided in Table 1. Overall, 31.6% of older adults experienced healthcare delay, 30% of those aged 50-69 years and 33.8% of those aged 70 years and over. In the total sample, there were more females (56.9%) than males (43.1%) and equally a higher percentage of females (34.1%) than males (28.3%) experienced healthcare delay. The highest percentage of participants had second-level education, 44.4%. A large majority of the sample, 75%, lived with others and 42.9% lived in a rural area. The majority of the participants were married, 72%. Sixty-two percent had at least one chronic condition and 47.6% of participants had at least one cardiovascular condition. A minority of participants (9.6%) were current smokers and 11.8% reported problematic alcohol consumption. Nearly all of the participants had public and/ or private health insurance, with 8.6% having no healthcare coverage. Ninety-two percent of participants had visited the GP in the previous twelve months (prior to 2018) and 30.6% reported taking five or more medications. The majority of participants (66.1%) returned the survey before the 1 st August 2020.
Analysis of the relationship between delayed healthcare utilisation and older adults with chronic conditions Table 2 shows the results of the first analysis which examined the relationship between delayed healthcare utilisation and chronic conditions among older adults. The reference category was having no healthcare delayed. Older adults with two or more chronic conditions were significantly more likely to have healthcare delayed than participants with no chronic condition (OR: 1.46, 95% CI: 1.11, 1.90). Females were more likely to have healthcare delayed than males (OR: 1.25, 95% CI: 1.01, 1.56). Older adults with third-level education were more likely to have healthcare delayed than those with primary-level education (OR: 1.61, 95% CI: 1.19, 2.17). Older adults living in another urban setting (town or city) in Ireland were less likely to have healthcare delayed than those living in Dublin (OR: 0.69, 95% CI: 0.52, 0.91). Participants with problematic alcohol consumption were more likely to have healthcare delayed than participants with no problematic alcohol consumption (OR: 1.54, 95% CI: 1.13, 2.10). Participants who had visited the GP were more likely to experience healthcare delay than those who hadn't (OR: 2.10, 95% CI: 1.32, 3.34). Participants who reported polypharmacy were more likely to experience healthcare delay than those who didn't (OR: 1.37, 95% CI: 1.08, 1.74).  Table 3 presents the results of the second analysis which examined the relationship between older adults with chronic conditions and the reasons for healthcare delay, whether it was delayed or cancelled by the provider or the participant. Older adults with two or more chronic conditions were more likely to have healthcare delayed by the provider than older adults with no chronic condition (OR: 1.73, 95% CI: 1.16, 2.56). In addition, older adults with a third-level education were more likely to have healthcare delayed by the provider than older adults with primary-level education (OR: 1.69, 95% CI: 1.07, 2.65). Older adults living in another urban setting and those living in a rural area were less likely to have healthcare delayed by the provider than older adults living in Dublin (OR: 0.60, 95% CI: 0.41, 0.88 and OR: 0.69, 95% CI: 0.48, 1.00). Participants with problematic alcohol consumption were more likely to have healthcare delayed by the provider than those with no problematic alcohol consumption (OR: 1.58, 95% CI: 1.04, 2.40). Participants who had visited the GP were more likely to have healthcare delayed by the provider than those who hadn't (OR: 2.56, 95% CI: 1.12, 5.83). Participants who reported polypharmacy were more likely to have healthcare delayed by the provider than those who didn't (OR: 1.53, 95% CI: 1.09, 2.14).

Multivariable analysis of reasons for delayed healthcare utilisation
In relation to healthcare being delayed by the participant, older adults with two or more chronic conditions were more likely to cancel or delay their own healthcare than those with no chronic condition (OR: 1.62, 95% CI: 1.14, 2.29). Older adults aged 70 years and over were more likely to cancel or delay their own healthcare than older adults aged 50-69 years (OR: 1.43, 95% CI: 1.00, 2.03). Older females were more likely to cancel or delay their own healthcare than males (OR:

Discussion
This study found that older adults with two or more chronic conditions were more likely to have healthcare delayed during the early months of the COVID-19 pandemic. This is similar to findings reported in other studies, among different cohorts and in different healthcare systems (Macinko et al., 2020;Topriceanu et al., 2021). It is not surprising that people with chronic conditions were more likely to experience healthcare delay, given that they are likely to have a higher demand for healthcare services (Bähler et al., 2015;Glynn et al., 2011;McDaid, 2013). However, healthcare need is accounted for, to some extent in this analysis, by the GP visit and polypharmacy variables, yet our analysis still shows that those with two or more chronic conditions were more likely to experience healthcare delay. Studies prior to the pandemic, based on self-reported difficulties accessing healthcare, showed that individuals with poorer health were more likely to feel unable to access care (Cylus & Papanicolas, 2015;Schneider & Devitt, 2018). Additionally, these findings likely reflect risk averseness on both the providers' and participants' sides.
In this instance, depending on their chronic conditions, participants may have been at high risk or very high risk of COVID-19 (Health Service Executive). Research on adherence to physical distancing guidance in Ireland found that those in high risk and very high risk groups were more likely to avoid social gatherings than those in the lower risk group (Durand et al., 2021). It is therefore reasonable to suspect that this may have also impacted on healthcare utilisation.
In the first analysis, similar to Topriceanu et al. (2021), we did not find an association between age and delayed healthcare. However, when we looked at who, among those who did delay healthcare, made the decision to do so, we found that participants over the age of 70 were more likely to delay or cancel their own care, providing some evidence that the above government advice on very high risk groups (Health Service Executive), including advising those over 70 to cocoon, impacted on their health-seeking behaviours and resulted in cancelled or delayed healthcare. Interestingly, however, we did not see the same relationship between age and healthcare being delayed or cancelled by the provider. Again, this is similar to previous research on accessing healthcare where age is not found to be a factor in self-reported difficulties in accessing healthcare (Cylus & Papanicolas, 2015;Schneider & Devitt, 2018).
While the above two findings make sense in the context of risk factors for COVID-19 some of our other findings are not as clear cut. We found a relationship between higher-level education and delayed healthcare. This is similar to Macinko et al. (2020) which found that those with nine or more years of education were more likely to have scheduled care cancelled. On the providers' side, it may be the case that non-essential healthcare was more likely to be delayed and that people with higher education are more likely to use non-essential healthcare. However, if this was the case, one would have expected to see a similar relationship with private health insurance in this analysis. On the participants' side it may be that participants with third level education are more likely to adhere to precautionary behaviours and are more risk averse, however, to date, there is mixed evidence around the relationship between education and adherence to precautionary behaviours (Moran et al., 2021). One unusual finding was that older adults with problematic alcohol consumption were more likely to experience healthcare delay than those without. Heavy alcohol consumption has a significant negative impact on health, disability and mortality (World Health Organization, 2019), and reduced access to healthcare during the COVID-19 pandemic is likely to exacerbate existing health issues for this cohort and reduce the chance of early intervention. There are concerns from addiction professionals in Ireland that both the stress of the pandemic and the closure and changes to addiction services has exacerbated symptoms for people with addiction (Columb et al., 2020), and there is some evidence of an increase in alcohol and mental health-related admissions to healthcare facilities after the first wave of COVID-19 in 2020 (Crowley & Hughes, 2021;McIntyre et al., 2021).
In the first analysis, we found that older adults living in another urban setting (town or city) in Ireland were less likely to experience healthcare delay than older adults living in Dublin. Interestingly, in the second analysis, we found that providers were less likely to delay or cancel appointments for older adults living in another urban setting and in rural areas compared to participants living in Dublin. This could reflect the impact of increased pressure and prioritisation of patients with COVID-19 on healthcare services in Dublin, patient flows and the size of the patient population may have been a factor here.
Our study found that older females were more likely to delay or cancel their own healthcare than older males, similar to two other studies (Macinko et al., 2020;Topriceanu et al., 2021). Differences between genders is reflected in studies on precautionary behaviours where women were slightly more likely to comply with COVID-19 related measures (Durand et al., 2021;Galasso et al., 2020;Moran et al., 2021), and perhaps therefore more likely to be risk averse and cancel their own appointments out of fear of contracting COVID-19 and/or giving it to others. Equally, females are more likely to be family carers, with some family carers in Ireland expressing concern about the repercussions of contracting COVID-19 themselves on those they care for (Family Carers Ireland, 2020). However, it may also mean that the pandemic has exacerbated existing inequalities as, prior to the pandemic, females were more likely to report an unmet healthcare need (Connolly & Wren, 2017;Schneider & Devitt, 2018). We also found that those who live with others were less likely to cancel their own healthcare than those who live alone, pointing to the potential positive impact of others on the health-seeking behaviour of older adults. Similar to Macinko et al. (2020), we found that participants who had visited the GP were more likely to experience healthcare delay, this is unsurprising given that GP visits and polypharmacy are a potential reflection of healthcare need as a whole.
Interestingly, there is no evidence that a lack of access to a medical card/GP visit card or health insurance had any impact on delayed healthcare utilisation during the pandemic. This is very interesting given Ireland's dual healthcare system 4 , the fact that those without access to free primary care or private health insurance are more likely to report an unmet healthcare need (Connolly & Wren, 2017), and that eligibility for public healthcare is related to a higher number of GP visits amongst TILDA participants (Hudson & Nolan, 2015). This negative finding may have been influenced by the fact that the Irish Government purchased the services of the private hospitals, where they essentially functioned as additional public hospitals, from March to June 2020 (Kennelly et al., 2020). This was a particularly novel occurrence within the Irish healthcare system.
Overall, our study found that 31.6% of participants had some type of healthcare delayed during the first few months of the pandemic, adding to concerns that the pandemic has increased levels of unmet need for healthcare services. Fundamentally, availability of and access to healthcare is of vital importance, with international evidence showing the negative impact of delayed access to healthcare on health outcomes, diagnoses and mortality (Prentice & Pizer, 2007;Wang et al., 2012). Additionally, healthcare delay is likely to have had a significant impact on those providing care to older adults both family carers and professional carers, adding additional stress and concern due to delayed and cancelled appointments. Moynihan et al. (2020) reflect on the lessons that can be learnt from the pandemic around healthcare provision and healthcare waste. Globally, there is evidence that the overuse of medical services is a common occurrence (Brownlee et al., 2017), and there is evidence of healthcare waste in some healthcare services in Ireland (Ryan et al., 2019). The pandemic provides an opportunity to fully examine healthcare utilisation within the Irish healthcare system in order to better understand and assess both healthcare waste and unmet need and ensure that the provision of healthcare is timely, effective, and appropriate. Equally, further research is required to fully understand the impact of the COVID-19 pandemic on health outcomes for older adults in Ireland and internationally.
Limitations This is a cross-sectional study; therefore, the findings cannot be interpreted as causal. Given the phrasing of the first question on delayed healthcare it is possible that some of those who answered no and who have no chronic conditions, did not need to seek care. Therefore, some of the findings could reflect the difference in the underlying probability of needing care in the first place. However, it was not possible to separate out these groups any further. We have included an additional sensitivity analysis in the extended data which solely looks at those who had healthcare delayed. However, this does not resolve the issue, but may be of interest to the reader. As it is difficult to capture all services, there may be more health and social care delays and cancellations that were not picked up on by the survey. The answers of the survey were self-reported so there is a risk of recall bias.

Conclusion
This paper found that older adults with two or more chronic conditions were more likely to have healthcare delayed than those with no chronic conditions during the COVID-19 pandemic in Ireland. The paper also found other factors which had a positive relationship with healthcare delay such as people aged 70 years and over, females, older adults with problematic alcohol consumption and those with third-level education. Further research is required to understand in more detail the impact of these healthcare delays on older adults in Ireland.

Data availability
Underlying data TILDA data cannot be made fully publicly available due to data protection issues concerning participants' personal information. De-identified participant data can be accessed via the following methods. TILDA provides access to the researcher microdata files (RMF), including the COVID-19 study and previous waves of TILDA, through a physical hot desk system based in the TILDA offices in Dublin. Researchers can request an application form through tilda.hotdesk@ tcd.ie. Applications are reviewed by the management team and where access is granted, researchers can book time on a hot desk to work on the TILDA RMF datasets.
The publicly accessible TILDA COVID-19 SCQ dataset has been deposited to the Irish Social Science Data Archive (ISSDA, University College Dublin) and will be publicly available when they publish it in accordance with established TILDA data sharing processes that adhere to European Union and national data protection regulations (GDPR and HRR). Due to an unexpected delay, data was not made available in Q2 as predicted in the protocol (Ward et al., 2021), but is currently in the process of being made publicly accessible.
The data file, codebook, and accompanying documentation will be made available via the ISSDA alongside existing public TILDA data files, from previous waves of TILDA. To access the Covid-19 Study and previous waves of TILDA, please complete an ISSDA Data Request Form for Research Purposes, sign it, and send it to ISSDA by email (issda@ucd. ie). For teaching purposes, please complete the ISSDA Data Request Form for Teaching Purposes, and follow the procedures, as above. Teaching requests are approved on a onceoff module/workshop basis. Subsequent occurrences of the module/workshop require a new teaching request form. The fully pseudonymised COVID-19 study public dataset will also be made available via the Inter-university Consortium for Political and Social Research (ICPSR) at the University of Michigan for the purpose of long-term data storage and dissemination. There is no specific date for when this will be complete. International researchers and educators from within and outside the European Economic Area can apply to access this data for teaching and research purposes. Again, individual identifiers will not be included in these datasets and data will be shared in line with participant consent and relevant data protection legislation. Contractual arrangements will be put in place prior to any data transfers.
Extended data Figshare: A cross-sectional study of the relationship between delayed healthcare utilisation and chronic conditions among older adults during the COVID-19 pandemic in Ireland, appendix. https://doi.org/10.6084/m9.figshare.15173709.v1 (Hennelly et al., 2021) This project contains the following files: -Appendix.pdf. This file contains extended data as listed above.
Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).

Author contributions
All authors conceived of the study and guided the study methodology. MW and RK were responsible for data collection. NH and MW conducted the analysis. GL and NH were responsible for the literature review. All authors contributed to the findings, discussion, and final manuscript.