Early Risk Factors for Long COVID That Causes Disability

Since the onset of the COVID‐19 pandemic, the global health community has increasingly turned its attention to the long‐term consequences of SARS‐CoV‐2 infection beyond the acute phase of illness [1]. One of the most pressing concerns is the emergence of long COVID, also known as post‐COVID‐19 condition, which is now recognised as a constellation of symptoms persisting for more than 12 weeks after initial infection in the absence of an alternative diagnosis [1]. These symptoms range from fatigue, breathlessness and cognitive impairment to musculoskeletal pain, gastrointestinal issues and psychological distress [1, 2]. Although the precise mechanisms remain unclear, long COVID is believed to result from a combination of viral persistence, immune dysregulation and autonomic dysfunction [1].

While many individuals report mild or gradually improving symptoms, a substantial subset experience debilitating impairments [3, 4]. Recent data estimate that 3.3% of people living in private households in England and Scotland, approximately two million individuals, self‐report long COVID. In nearly three‐quarters of these cases, day‐to‐day activities are adversely affected, with roughly one in five reporting that their activities are ‘limited a lot’ [2]. This condition, here referred to as disabling long COVID, is increasingly regarded as a significant public health concern, with implications for workforce participation, health utilisation and long‐term disability support systems [5, 6]. Despite growing policy interest, systematic evidence on who is most at risk of developing disabling long COVID remains limited.

Much of the current understanding of long COVID stems from clinical cohorts, electronic health records or convenience samples, which, while valuable, are often restricted to individuals who seek care and may underrepresent marginalised groups or those with less access to health care [1]. Moreover, many studies fail to distinguish between the presence of persistent symptoms and their functional impact—a distinction critical for understanding health inequalities and guiding resource allocation [7, 8]. Long COVID symptoms, such as fatigue and cognitive dysfunction, can be widespread, but their capacity to limit activities of daily living varies considerably [9]. Consequently, the UK Office for National Statistics (ONS) and NHS England have begun to differentiate between ‘non‐disabling’ and ‘disabling’ forms of long COVID, the latter being characterised by a self‐reported reduction in daily functioning [10]. This functional framing aligns with the broader conceptualisation of disability within global health and human rights frameworks, including the International Classification of Functioning, Disability and Health (ICF) [11]. Yet, few studies have investigated disabling long COVID using prospective data that links prepandemic baseline information with postinfection outcomes.

This study aims to identify prepandemic risk factors associated with disabling long COVID in a large population‐based cohort of UK adults. Using longitudinal data, this study estimates the association between factors measured before the pandemic and the development of disabling long COVID at follow‐up. It further examines whether these associations differ by sex.

Of the 12,033 participants included in the analytic sample, 690 (5.7%) reported disabling long COVID and 11,343 (94.3%) did not. Participants with disabling symptoms differed significantly from those without across multiple characteristics.

Descriptive characteristics varied significantly by disabling long COVID status (Table 1). Disabling symptoms were more common among adults aged 30–69 years, women and non‐White participants (all p < 0.01). Indicators of social and health disadvantage, including low income satisfaction, current smoking and pre‐existing health conditions, were more prevalent among those with disabling long COVID. Poor self‐rated health, high psychological distress and very bad sleep were also markedly more common in this group (all p < 0.001). In contrast, area of residence (urban vs. rural) did not differ significantly between groups (p = 0.739).

Figure 1 illustrates the multisystem symptom profile of the 690 individuals with disabling long COVID. Fatigue was most common (79.6%), followed by muscle aches (67.4%), shortness of breath (61.6%), headaches (58.5%) and persistent cough (55.5%). Cognitive and systemic symptoms were also frequent: loss of concentration (49.9%), fever (47.2%) and loss of smell/taste (44.4%). Over 40% reported upper respiratory symptoms (runny nose and sore throat), sleep disturbance and memory problems. Other symptoms, such as dizziness, appetite loss and chills, clustered around 30%, while chest pain, digestive issues and numbness affected 19%–24%. Only 4.6% cited ‘other’ symptoms.

After adjusting for all covariates, many associations persisted, though attenuated. Middle‐aged adults remained at elevated risk (30–49: RR 1.38, 95% CI 1.10–1.73, p = 0.006; 50–69: RR 1.28, 95% CI 1.01–1.62, p = 0.040), while the protective effect in the oldest group weakened (≥ 70: RR 0.68, 95% CI 0.46–1.00, p = 0.053). Female sex was still a significant predictor (RR 1.26, 95% CI 1.08–1.48, p = 0.003), and White ethnicity conferred lower risk versus non‐White (RR 0.75, 95% CI 0.61–0.92, p = 0.005). Compared with low‐income satisfaction, moderate (RR 0.76, 95% CI 0.62–0.93, p = 0.004) and high satisfaction (RR 0.81, 95% CI 0.67–0.98, p = 0.034) were protective. A pre‐existing health condition increased adjusted risk by 31% (RR 1.31, 95% CI 1.10–1.56, p = 0.002), and fair/poor self‐rated health was associated with a nearly 80% higher risk (RR 1.78, 95% CI 1.40–2.25, p < 0.001). Psychological distress (GHQ‐12 ≥ 4) raised risk by 44% (RR 1.44, 95% CI 1.21–1.72, p < 0.001), and ‘fairly bad’ or ‘very bad’ baseline sleep each nearly doubled risk (RR 1.92, 95% CI 1.45–2.56 and RR 1.96, 95% CI 1.37–2.81; both p < 0.001). Neither current smoking (RR 1.09, 95% CI 0.88–1.35, p = 0.428) nor rural residence (RR 1.16, 95% CI 0.97–1.37, p = 0.097) remained significant (Figure 2).

This study identifies key prepandemic risk factors associated with the development of disabling long COVID, defined as persistent symptoms lasting more than twelve weeks that limit daily activities. Adults aged 30 to 69 years were more likely to experience disabling symptoms than those aged 16 to 29 years. Women were also more affected than men, consistent with previous observations of higher long COVID prevalence among females [14–16]. In addition to age and sex, several prepandemic factors were independently associated with greater risk, including lower satisfaction with income, non‐White ethnicity, pre‐existing long‐term health conditions, poor self‐rated health, psychological distress and poor sleep quality. These findings suggest that disabling long COVID does not affect all individuals equally and that socioeconomic and mental health vulnerabilities present before the pandemic may shape who experiences the most severe and functionally limiting outcomes.

The analysis revealed clear differences in risk profiles between men and women. Among men, middle age (30–69 years) was associated with significantly higher risk compared to younger men, whereas women in the same age groups showed elevated but nonsignificant estimates. In contrast, being aged 70 years or older was protective for women but not for men. Living in a rural area was associated with increased risk only among women, and moderate income satisfaction was protective in women but not in men, with both factors showing statistically significant sex interactions. Weaker evidence also suggested possible sex differences in the associations with age and self‐rated health. These sex‐specific patterns underline the need for tailored public health responses. Overall, the findings highlight that disabling long COVID is shaped not only by clinical factors but also by socioeconomic and structural determinants established before the pandemic.

A growing body of research has established that long COVID disproportionately affects certain demographic and clinical groups. Women, older adults and individuals with pre‐existing physical or mental health conditions have consistently been identified as being at higher risk of developing persistent symptoms following SARS‐CoV‐2 infection [17]. Several studies have also linked long COVID to socioeconomic disadvantage and elevated psychological stress [18]. However, most existing research has focused primarily on the presence or persistence of symptoms, rather than on their severity or impact on functional ability. This limits the ability to assess how long COVID interferes with individuals’ capacity to work, care for others or participate in everyday life—factors that are increasingly recognised as critical in public health and disability frameworks.

Furthermore, the majority of evidence to date comes from clinical cohorts, electronic health records or online surveys. These sources are valuable but often capture only those who seek care or have the means to participate in research, which may lead to selection bias and underrepresentation of marginalised populations. Few studies have used nationally representative data or been able to draw on comprehensive baseline information collected before the pandemic began [19]. As a result, it has been difficult to determine whether observed associations reflect true predisposing risk factors or are confounded by experiences during or after infection. This gap highlights the need for prospective, population‐based studies that can disentangle pre‐existing vulnerabilities from postinfection consequences and better inform prevention and support strategies.

This study makes a significant contribution by focusing specifically on disabling long COVID, defined as persistent symptoms that interfere with daily functioning, rather than on the broader and more heterogeneous category of long COVID symptoms alone. This distinction is important because not all individuals with lingering symptoms experience a loss of functional capacity [20, 21]. By operationalising disability in alignment with public health frameworks, such as those used by the UK ONS and the WHO’s ICF, this study offers a more meaningful measure for informing social policy, workplace accommodation and long‐term healthcare planning. It moves beyond simply counting symptoms and shifts attention towards the consequences of long COVID on people’s lives.

The symptom profile observed in this study is broadly consistent with existing literature, with fatigue, muscle aches and breathlessness predominating. However, it is worth noting that these findings differ somewhat from the 2024 RECOVER‐Adult Long COVID Research Index (LCRI), which assigns a relatively high diagnostic weight to changes in smell or taste [22]. In the LCRI, symptom weights were derived using LASSO regression to identify symptoms that best discriminate between SARS‐CoV‐2‐infected and uninfected participants, rather than reflecting raw symptom prevalence. Thus, the high weight assigned to smell/taste changes in the LCRI reflects their specificity to SARS‐CoV‐2 infection, not necessarily their frequency among those with long COVID. In this sample, loss of smell or taste was reported by 44.4% of individuals with disabling long COVID, ranking below fatigue, muscle aches, breathlessness, headaches and persistent cough. This difference likely also reflects the predominance of Omicron‐era infections in the follow‐up period (2022‐23), during which olfactory symptoms became less common compared with earlier variants [23]. Additionally, because the UKHLS only collected detailed symptom data from participants who reported long COVID lasting more than 12 weeks with activity limitation, it was not possible to compare symptom profiles between those with disabling long COVID, those with nondisabling long COVID and those without long COVID. Future research with symptom data across the full spectrum of postinfection outcomes would be valuable for identifying whether specific symptoms, such as loss of smell or taste, may distinguish disabling from nondisabling forms of long COVID.

Unlike many previous studies that rely on clinical or convenience samples, this analysis is based on a large cohort with rich, prepandemic baseline data. This design allows for a clearer identification of risk factors rather than consequences. While prior research has suggested associations between long COVID and mental health, socioeconomic status or health behaviours [16–21], few studies have been able to confirm whether these factors pre‐dated infection or emerged afterwards. By prospectively linking data from before the pandemic with outcomes measured three to 4 years later, this study provides stronger evidence that psychological distress, sleep problems and financial dissatisfaction were already present and independently predicted disabling long COVID. This is a key methodological advancement, enabling the identification of individuals who may be more vulnerable to prolonged and disabling recovery trajectories.

Importantly, this study also offers new insights into sex‐specific risk patterns. While women overall had a higher risk of disabling long COVID, stratified analyses revealed that the role of age, rural or urban residence and health perceptions varied by sex. Older age appeared to be protective only among women, while rural residence increased risk only for women, suggesting that social roles, caregiving responsibilities or healthcare access may interact with sex to shape long COVID outcomes. These nuanced findings contrast with many earlier studies that report pooled estimates, potentially obscuring meaningful subgroup differences [17–19]. By highlighting these interactions, this study underlines the need for tailored interventions and policies that consider the intersection of sex, health and social determinants in shaping the long‐term impact of the pandemic.

This study identifies several prepandemic factors associated with increased risk of developing disabling long COVID, including being aged 30 to 69 years, female sex, non‐White ethnicity, low satisfaction with income, pre‐existing health conditions, poorer self‐rated health, psychological distress and poor sleep quality. By using large prospective data, it offers stronger evidence that these vulnerabilities existed before infection and may contribute to prolonged functional impairment. The findings highlight the importance of distinguishing between symptom presence and disability in understanding the true burden of long COVID. They also reveal sex‐specific differences in risk, pointing to the need for more targeted and equitable public health responses. Overall, addressing the social and health inequalities that shape susceptibility to disabling long COVID should be a central priority in ongoing recovery planning.

Yusuff Adebayo Adebisi was supported by a UKRI Economic and Social Research Council (ESRC) (ES/P000681/1).

UKHLS has received ethics clearance for all waves from the University of Essex Ethics Committee and, for nurse/biomarker modules, the National Health Service Research Ethics Committee. This analysis uses fully anonymised, publicly available data (UK Data Service, SN 2000053); no additional institutional review board approval was required.

The author declares no conflicts of interest.

Additional supporting information is available in the Supporting File, including Supporting Table S1, which presents the sensitivity analysis comparing the mutually adjusted model with a mediator‐stripped model.

Adebisi, Yusuff Adebayo , Prepandemic Risk Factors for Disabling Long COVID: A Prospective Cohort Analysis, Journal of Tropical Medicine, 2026, 9396282, 9 pages, 2026. 10.1155/jotm/9396282

Yusuff Adebayo Adebisi, Email: y.adebisi.1@research.gla.ac.uk.

Rajib Chowdhury, Email: rajib478@yahoo.com.

The data analysed in this study are publicly available from the University of Essex, Institute for Social and Economic Research, via the UK Data Service. Understanding Society (data series), 13th release, 2024. SN: 2000053. https://doi.org/10.5255/UKDA-Series-2000053↗.