Health of unpaid carers in Wales, UK: a population data linkage study

Abstract Background The population of unpaid carers in Wales increased to record. There is no systematic approach to record unpaid caring status, resulting in limited quantitative evidence on unpaid carers’ health. The aim of this study is to: (i) create an e-cohort of unpaid carers by linking routinely collected health and administrative datasets in Wales, UK. (ii) investigate whether long-term health conditions and multimorbidity are more prevalent amongst unpaid carers than non-carers. Methods Unpaid carers were identified by linking primary care dataset, National Survey for Wales data with demographic characteristics in the Secure Anonymise Information Linkage Databank. The clinical codes identified in Cambridge Multimorbidity Score were used to explore the prevalence of long-term health conditions. Results A total of 91 220 unpaid carers in Wales were identified between 1 January 2010 and 1 March 2022. Unpaid carers were found at higher risk of managing 35 of 37 long-term health conditions and multimorbidity than non-carers, exacerbated amongst younger age groups and deprived communities. Conclusions The creation of the first e-cohort of unpaid carers in Wales provides opportunities to perform rapid analysis to systematically understand health needs and evaluate initiatives in future. To better support unpaid carers, flexible approaches focusing on early identification and prevention is crucial.


Introduction
With an ageing population and increasing proportion managing complex health needs, unpaid carers play a significant role in providing care.Unpaid carer is 'a person looks after, or give any help or support to family members, friends, neighbours or others'. 1 It was estimated that there were over 400 000 unpaid carers in 2019 in Wales, UK and approximately 700 000 during the COVID-19 pandemic with the number expected to keep increasing in a foreseeable future. 2,3roviding care can be rewarding whilst having negative impact on unpaid carers' own health but often overlooked. 4,5][8][9][10] However, there is limited quantitative studies investigating the health needs of unpaid carers.Particularly, there is a lack of systematic data collection on unpaid carers in Wales, therefore conducting large scale data analysis to understand the inequalities and unpaid carers' health condition becomes extremely difficult. 11his study aims to generate a reproducible e-cohort of unpaid carers in Wales utilizing routinely collected health and administrative data, and understand the health needs of this population group.

E-cohort of unpaid carers
Working with general practitioners, a list of Read codes was validated to identify unpaid carers. 12National Survey for Wales (NSW) 2016/17, 2017/18, 2018/19 and 2019/20 data and primary care dataset (from 1 January 2010 to 1 March 2022) were used to identify unpaid carers (Fig. 1).Anonymous Linking Field code was utilized to eliminate duplication.Demographic Service Database and Census 2011 were linked to obtain socio-demographic information.

A matched comparison group
Welsh residents who had not been identified as unpaid carers were randomly selected to form a comparison group matched with unpaid carers on sex, age and deprivation. 13

Long-term health conditions and multimorbidity
Clinical codes used to identify 37 long-term health conditions in Cambridge Multimorbidity Score were used to scan primary care records [14][15][16] (full list of 37 long-term health conditions can be accessed from CPRD@Cambardge-Codes List 16 ).Presenting with two or more conditions was defined as multimorbidity.We retrospectively tracked the primary care records of unpaid carers and flagged up cases with long-term health conditions and multimorbidity.

Statistical analysis
In this study, all data were stored in the Secure Anonymised Information Linkage (SAIL) Databank where data linkage and analysis were conducted. 17Crude prevalence rate, adjusted prevalence rate, rate ratio of long-term health conditions and multimorbidity were produced.P < 0.05 was considered as statistically significant.

E-cohort characteristics
Amongst 91 220 unpaid carers, approximately two-thirds were female (65.0%versus 50.7% in the general population) and a quarter were between 55 and 64 years old (24.7% versus 12.8%).In total, 5.5% of them were under the age of 25 and 16.5% were over 75.Distribution across deprivation quintiles was similar with mid-2018 Welsh population estimate. 18

Overall prevalence of long-term health conditions
Overall, the prevalence of all 37 long-term health conditions was significantly higher amongst unpaid carers than non-carers, except for peripheral vascular disease (P = 0.14) and multiple sclerosis (P = 0.64).The five most common conditions amongst unpaid carers were anxiety and/or depression (age-sex standardized rate 18 : 243.9 per 1000 population), hypertension (89.2), hearing loss (79.4), chronic kidney disease (72.6) and asthma (60.2) (Table 1).The  1.

By age groups and deprivation
Anxiety and/or depression were the most prevalent condition in all age groups amongst unpaid carers except 75+ years group, with higher prevalence than non-carers.The difference between unpaid carers and non-carers decreased with increasing age.Rate ratio for anxiety and/or depression between unpaid carers and non-carers was 2.3 (95% CI 2.0-2.6) for under 25 years old and reduced to 1.5 (1.5-1.6) for 45-54 and over 75 year olds (Table 1).The prevalence of 8 conditions (amongst the top 10) was higher for unpaid carers living in the most deprived areas than the least except for hearing loss and cancer, e.g.anxiety and/or depression (287.6 per 1000 population versus 209.3 per 1000 population) and hypertension (88.1 versus 87.1).The difference between unpaid carers and non-carers was marginally higher in the most deprived areas except for anxiety and/or depression (rate ratios, the most: the least deprived quintile 1.6: 1.9) (Table 1).

Multimorbidity
Prevalence of multimorbidity was higher in unpaid carers compared to non-carers at all ages and deprivation quintiles.

Main finding of this study
This study created a unique and reproducible e-cohort of unpaid carers in Wales with 91 220 identified.Findings suggested that unpaid carers experience poorer health comparing to non-carers and the gap exacerbated amongst younger groups and in deprived communities.The health needs of unpaid carers are usually overlooked due to the focus on the health of the carees and juggling with other responsibilities.

What is already know to this topic
Previous studies found providing care have detrimental impact on unpaid carers' own health and wellbeing, whilst with limited quantitative evidence.

What this study adds
The e-cohort can be used for rapid analysis in research and evaluation towards tailor-made support for unpaid carers in future.From policy making perspective, health services and policy should take flexible approaches for unpaid carers to access healthcare and focus on early identification and prevention for highlighted groups and conditions.

Limitations of this study
There are certain limitations with this study.The identification of unpaid carers in routine data relies on the caring status to be recognized and recorded in surveys and clinical records.There are systematic differences in recording nonhealth patient information across healthcare settings.These barriers contribute to underestimating population size and bias in quantifying health needs, but highlighting the vital need of standardized approaches in recognizing and recording unpaid caring status.

Fig. 1
Fig. 1 Flow diagram of the creation of the e-cohort of unpaid carers in Wales.

Table 1
Top 10 prevalent (crude rate and adjusted rate of per 1000 population) of long-term health condition amongst unpaid carers (whole sample and by age groups) and rate ratio against non-carers