Dolphin
Senior Member (Voting Rights)
Fong, W. L. E. (2026). Social inequalities in COVID-19 and post-acute infection syndromes: The role of deprivation, ethnicity, and migration status. [Doctoral Thesis, Maastricht University]. https://doi.org/10.26481/dis.20260702wf
Social inequalities in COVID-19 and post-acute infection syndromes: The role of deprivation, ethnicity, and migration status
Wing Lam Erica Fong
KEMTA
Research output: Thesis › Doctoral Thesis › Internal
21Downloads (Pure)
Overview
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
Maastricht University
Supervisors/Advisors
van Kuijk, Sander, Supervisor
Aldridge, R. W., Supervisor, External person
Beale, S, Co-Supervisor, External person
Award date2 Jul 2026
Place of PublicationMaastricht
Print ISBNs9789465376271
Publication status
Published - 2 Jul 2026
Access to Document
10.26481/dis.20260702wf
Full TextFinal published version, 18.7 MB
SummaryFinal published version, 116 KB
PropositionsFinal published version, 2.59 MB
CoverFinal published version, 764 KB
ImpactFinal published version, 979 KB
Embargoed Document
Full Text
The thesis examined the disparities in COVID-19-related hospitalisation, the risk of post- COVID-19 condition (PCC), and the consequences of PCC and other respiratory post-acute infection syndromes (PAIS). It focuses on inequalities by socioeconomic deprivation, migration status, and ethnicity.
Large-scale surveillance studies are crucial for detecting disease outbreaks, understanding pathogen epidemiology, and monitoring public health measures. However, they are often resource-intensive and difficult to maintain, especially in resource-limited settings. Chapter 2 assessed the Virus Watch study as a low-cost surveillance system by comparing its COVID-19 estimates with two national surveillance studies: the Office for National Statistics (ONS) COVID-19 Infection Survey (CIS) for positivity and incidence rates, and the Severe Acute Respiratory Infection Watch (SARI) for hospitalisation rates. Positivity and incidence rates from Virus Watch were calculated both with and without linkage to national testing data. Synchrony was measured globally (overall Spearman ) and locally (9-week rolling Spearman ). Using data from 52,526 English participants between June 2020 and March 2023, both linked and non-linked Virus Watch-estimated positivity and incidence rates showed strong global (Spearman = 0.90-0.92, P < .001) synchrony with CIS estimates, despite lower absolute values. Local synchrony was also strong across both approaches, but consistently higher with linked data (median range: 0.75-0.76) than with non-linked data alone (median range: 0.66-0.67). This alignment persisted before and after the end of free national testing, although differences increased in the post-free testing period. Hospitalisation rates from Virus Watch were lower than SARI estimates and demonstrated weaker synchrony. In Wales, estimates were more variable and less locally synchronised than those from England. These findings show that Virus Watch can reliably track COVID-19 trends at a substantially lower cost than large-scale surveillance studies such as ONS CIS. Its research approach can serve as a scalable model for surveillance systems in settings lacking national surveillance infrastructure.
Since the start of the pandemic, health disparities have been exacerbated among underserved populations, particularly those from deprived and ethnic minority groups. While there was increasing evidence of similar inequalities among migrants in other high-income countries, the effect of migration status on severe COVID-19 outcomes in the UK remained poorly understood. In Chapter 3, we analysed COVID-19related hospitalisation rates among migrants and UK-born adults across the first three pandemic waves (Wild-type, Alpha, and Delta), using secondary care data linked to the Virus Watch cohort. Among 30,276 participants, migrants (12.5% of the sample) had nearly double the crude hospitalisation incidence rate of UK-born individuals (4.6 vs. 2.7 per 1,000 person-years). After adjusting for confounders, pooled incidence rate ratios (IRRs) across all three waves indicated that migrants were 35% more likely to be hospitalised for or with COVID-19 (IRR 1.35 [95% CI 0.71-2.60]). Although the 95% CI were wide and overlapped 1, the consistent direction of effect across waves suggested that migrant populations in the UK faced an increased risk of COVID-19-related hospitalisation, highlighting the need for equitable interventions, particularly those aimed at increasing vaccination uptake in this group.
A proportion of individuals infected with COVID-19 developed post-COVID-19 condition (PCC), a syndrome characterised by persistent symptoms such as fatigue, breathlessness, and cognitive difficulties. It was also important to assess whether inequalities observed in acute and severe COVID-19 extended to PCC. Chapter 4 examined how deprivation, migration status, and ethnic minority status influenced the likelihood of developing PCC among Virus Watch participants, both across the entire cohort and among those who had ever been infected. During the pre-Omicron period, the probability of PCC was higher among more deprived groups, particularly among females (most deprived: 7.8% [95% CI 4.6-11.0%]; least deprived: 3.5% [2.5%-4.5%]). Among migrant and ethnic minority groups, females overall had a greater likelihood of PCC. However, disparities in PCC risk were less pronounced compared to males. Migrant males in the full cohort had a 6.3% (1.8-10.8%) likelihood of developing PCC compared to those UK-born (14.4% [10.5%-18.3%]. Additionally, ethnic minority males previously infected also had a higher likelihood of PCC than their white British counterparts (white British: 16.2%, 12.5–20.0%; ethnic minority: 38.8% [21.2–56.4%]). By the Omicron period, these disparities had attenuated across groups, highlighting the impact of intervention-induced inequalities. Elevated PCC risk in the entire and infected cohorts also reflected greater exposure to infection and post-infection determinants, underscoring the need for targeted policies to reduce infection risk, minimise intervention-induced inequalities, and ensure accessible healthcare for those living with PCC.
PCC can affect an individual's ability to perform daily activities, such as attending work or school and caring for themselves and others. Chapter 5 examined whether deprivation, migration status, and ethnic minority status were associated with experiencing functional limitations beyond PCC risk among those living with the condition. 776 Virus Watch participants reported persistent symptoms meeting the World Health Organization definition of PCC, with 88% experiencing limitations in at least one functional activity: work or education, concentration, self-care, caring for others, necessary activities outside the home, and engaging in enjoyable activities. We found that deprivation primarily drives these limitations. Compared to the least deprived, those in the most deprived quintile had higher odds of limitations in work or education (adjusted odds ratio (aOR) 2.30 [95% CI 1.12,4.93]), concentration (2.78 [1.42, 5.81]), self-care (2.11 [1.06, 4.16]), and performing necessary activities outside the home (2.06 [1.12, 3.90]). Participants in the second most deprived quintile also faced higher odds of limitations in work, education, and concentration. In contrast, we found no evidence of associations with migration status or ethnic minority status across any functional limitations.
Persistent symptoms and limitations in daily activities are likely to negatively affect one’s health-related quality of life (HRQoL). Chapter 6 investigated whether HRQoL loss associated with respiratory PAIS, persistent symptoms following any acute respiratory infection, also varied by deprivation level, using EuroQol-5-dimension, 5-level (EQ-5D-5L) survey data from 16,418 Virus Watch participants, among whom 1,084 (6.6%) had respiratory PAIS. Stratifying by respiratory PAIS status, we estimated the odds of experiencing any HRQoL loss, the extent of loss using disutility scores, and EQ-Visual Analogue Scale (EQ-VAS) scores across deprivation levels. Deprivation was consistently linked to poorer HRQoL regardless of PAIS status. However, the burden was greatest among individuals in the most deprived areas with PAIS. They showed higher odds of HRQoL loss (no PAIS aOR: 1.57 [95% CI 1.34, 1.85]; PAIS: 2.10 [0.83, 5.32]), greater disutility (no PAIS adjusted coefficient: 0.067 [0.053, 0.081]; PAIS: 0.094 [0.043, 0.144]), and lower EQ-VAS scores (no PAIS adjusted coefficient: -4.55 [-5.88, -3.23]; PAIS: -8.00 [-12.70, -3.28]) than the least deprived. A similar pattern was observed among those without respiratory PAIS, but the differences in HRQoL loss across deprivation levels were less pronounced. Across all deprivation levels, participants with respiratory PAIS had higher predicted mean disutility scores and lower predicted mean EQ-VAS scores than those without, indicating an added negative impact of respiratory PAIS on top of existing deprivation-related disadvantage.
The consequences of respiratory PAIS, including PCC, can perpetuate cycles of deprivation, which can further compound pre-existing health inequalities. It is therefore essential that policy responses prioritise infection prevention, early detection of respiratory PAIS, integrated care pathways that address respiratory symptoms alongside pain and mental health, and equitable access to rehabilitation and support services, particularly among deprived populations, where the impact of persistent symptoms is greatest.