Exploring social determinants of health and their impacts on self-reported quality of life in long COVID-19 patients Authors: Anh Pham, Julia Smith, Kiffer G Card, Kaylee A Byers, Esther Khor Published: 2024 ABSTRACT AbstractThis study explores the health-related quality of life (HRQoL) experienced by patients with Long COVD-19 using data from British Columbia’s post-COVID-19 Recovery Clinics. A retrospective cohort of 3463 patients was analyzed to assess HRQoL through the EQ-5D-5L questionnaire which includes five dimensions (mobility, self-care, usual activities, physical health, and mental health) administered to patients; responses were analyzed using the Visual Analogue Score (VAS). Notably, 95% of participants reported HRQoL scores below 90, with 50% scoring under 60, indicating significant impacts on their well-being. The analysis revealed that HRQoL is significantly influenced by various social determinants of health (SDoH), including age, sex, employment status, and ethnicity, each showing distinct correlations with HRQoL dimensions and overall VAS scores. Specifically, older age was associated with decreased mobility and increased pain/discomfort but less anxiety and depression, highlighting varying impacts across the age spectrum. The study highlights the multifaceted impacts of Long COVID on the lives of patients and underscores the necessity of targeted strategies to improve HRQoL among diverse groups, considering specific SDoH. Such a comprehensive approach could lead to more equitable health outcomes and support the development of tailored public health policies aimed at the recovery and rehabilitation of Long COVID sufferers.Keywords: Long COVID, Post-COVID syndrome, Quality of Life, Social determinants of health, Gender, Ethnicity, AgeSubject terms: Quality of life, Health policy Introduction Long COVID, a term first developed by patients1, is a condition wherein a variety of symptoms continue or develop following an acute COVID infection2. Symptoms often include respiratory, cardiovascular, neurological and cognitive impairments, though more than 200 symptoms have been documented3. Data from different countries consistently estimate that 10–20% of SARS-CoV-2 infected individuals experience Long COVID symptoms to varying degrees and for varying durations4. In Canada, those diagnosed with Long COVID receive standard care for medically defined symptoms (e.g. respiratory impairments), but care for medically undefined symptoms, such as chronic fatigue or brain fog, is generally limited to rehabilitation therapy5. Socioeconomic impacts While understanding the symptoms of Long COVID is crucial for addressing the immediate health concerns of affected individuals, it is equally important to be aware that Long COVID has a significant socioeconomic impact on individuals and communities. Many of those living with Long COVID have to give up or reduce both productive and reproductive activities, experience stigma and medical gaslighting (it is when healthcare professionals dismiss or minimize patients’ symptoms, causing them to doubt their own experiences), and suffer emotional and mental distress4,6,7. Given that Long COVID affects multiple aspects of patients’ lives and that these impacts can last months to years without standardized treatment options, it is not surprising that a number of studies have found Long COVID to negatively impact health related quality of life (HRQoL). Liska et al. find that patients with Long COVID had a lower quality of life (QoL) compared to a healthy control group8. A study of patients with mild initial COVID infections, but ongoing Long COVID symptoms, also finds lower scores in the physical HRQoL health domain compared to the general population9. A qualitative study of adults with Long COVID similarly finds notable impacts on HRQoL related to ability to carry out regular activities and physical health10. While providing a general overview, few of these studies consider how effects on HRQoL might differ based on patient demographics and social determinants of health (SDoH). Social determinants of health SDoH are non-medical factors that influence health outcomes11. SDoH frameworks, such as those published by the World Health Organization12, help illuminate the drivers of health inequities, including social positions related to sex, gender, race, and other demographics and structural factors related to employment, housing, and security. While extensive research demonstrates relationships between SDoH, such as living conditions (security of, if alone or with others, etc.) and employment status (particularly whether employment is precarious or not) and acute COVID infections, there are very few studies on SDoH and Long COVID. A handful of studies demonstrate that female sex is associated with risk of Long COVID, and while there are biological explanations for some of these differences, studies also point to social factors such as over representation of females in sectors with increased risk of repeat infection, such as healthcare, retail, food services, and education, among others sectors13; female sex also is associated with higher prevalence of pre-existing, non-medical conditions, such as poor sleep14. There is less conclusive information about race and ethnicity related to Long COVID, but there are numerous studies indicating that racialized and ethnic minority communities in North America experienced greater risk of acute COVID-19 infection, suggesting increased risk of Long COVID as well as barriers accessing care due to pre-existing inequities15. However, even this literature on SDoH and Long COVID does not focus on HRQoL. One of the only studies that considers HRQoL, Long COVID and race—a study of Latino people with Long COVID in Texas—finds this population had slightly worse HRQoL scores, particularly related to anxiety, compared to the general population, likely due to lower socio-economic positions and barriers accessing healthcare16. Overall there is very little analysis of how SDoH interact with experiences of Long COVID11in relation to HRQoL. To advance research on the effects of SDoH on HRQoL in individuals with Long COVID, we examine the characteristics of patients referred to the Post-COVID-19 Interdisciplinary Clinical Care Network (PC-ICCN) at the Provincial Health Services Authority (PHSA) throughout British Columbia (BC), Canada. PC-ICCN, in collaboration with the regional health authorities, provided education from clinical specialists and collected standardized questionnaires at Post-COVID Recovery Clinics. The completed questionnaires included HRQoL which we assessed through a SDOH lens. We analyze two self-assessments of HRQoL from patients, looking across different dimensions to better understand a nuanced perspective of how individuals are diversly impacted by Long COVID, and then assess how these scores differ across select SDoH (sex, age, ethnicity). In doing so we aim to add nuanced analysis of how HRQoL may vary among patient groups, which can contribute to informing equity-based interventions. Methods Study design and settings This is cross-sectional study which secondarily uses data collected from a survey given to patients who experienced long COVID symptoms and had received health services at the post-COVID-19 Recovery Clinics (PCRCs) in BC17. Patients were asked to complete the survey when referred to PC-ICCN and repeated in intervals while the patient receives health services at the PCRCs and had their results recorded in the Patient Records Outcomes and Information System, a provincial-wide information system at PHSA. We included all patients who completed the questionnaire at the first admission between July 2020 and November 2023. A complete questionnaire was included in Supplementary S1. A portion of the survey used for this analysis including questions about their social demographics, behaviors, COVID-19 history, and quality of life. Explanatory variables Demographic characteristics collected by the survey include patients’ sex, age, ethnicity (Caucasian, Indigenous, and visible minorities), living arrangement (living alone or with others), employment status (employed or unemployed), number of regularly working hours (full-time, part-time with more or less than 50% of working hours, and unable to work due to COVID-19). The questionnaire also asked for their history of smoking (including cigarette, marijuana, and vaping), history of COVID-19 symptoms, medical history, and comorbidities. In relation to Long COVID symptoms, the questionnaire included level of cough and shortness of breath. Based on the SDoH framework12, we selected the following variable for analysis: sex (gender was not collected), age, ethnicity, living arrangement, employment status, and number of work hours. Outcome variables HRQoL was defined as “those aspects of self-perceived well-being that are related to or affected by the presence of disease or treatment”18. It was measured using the EQ-5D-5L questionnaire, which includes five dimensions: the ability to walk (mobility), to wash and dress (self-care), to do daily usual activities (work, study, care, recreation, etc.), as well as states of physical health (pain or discomfort) and mental health (anxiety or depression). Patients identified themselves using five-level Likert scales, from having no problem (1) to having extreme problem (5), for each dimension19. The EQ-5D-5L is widely regarded as a valid and reliable tool for measuring health-related quality of life (HRQoL) due to its comprehensive assessment of various health dimensions and consistent performance across diverse populations20,21. Participants were also asked to indicate their general EQ-5D-VAS score, which is a visual scale from 0 (the worst) to 100 (the best HRQoL). Statistical analysis All analyses were conducted using Python 3.022 and STATA 1523. We performed a descriptive analysis, in which, percentage, mean and standard deviation (SD) were reported for continuous variables, while median and interquartile range (IQR) were reported for categorical variables. Individual components of HRQoL were reported using Likert scales presented in the questionnaire, while self-identified VAS score was reported as mean and SD. We used univariable and multivariable regression (Ordinary Least Square with robust standard errors where assumptions of homoscedasticity and/or normality of errors were violated) to investigate the association between predictors (demographics, EQ’s dimensions) and the outcome (VAS). Purposeful selection methods, with significance cut-off of 0.2, was used to select relevant variables for the multivariable regression24. A significance level of 0.2 allows for a more inclusive selection of variables in the early stages of model building25. We also used partial correlation coefficients to evaluate the strength of correlation between variables and outcomes26. Partial correlation coefficients measure the strength and direction of the relationship between two variables while controlling for the influence of one or more additional variables26; it allowed us to access correlation of each variable at a time while controlling for confounders. We considered the p-value < 0.05 threshold for statistical significance. Results Demographic characteristics A total of 3,613 patients completed the questionnaire. After removing those who did not respond to questions about age, sex, and EQ-VAS, we included 3,463 participants in this analysis. The mean age was 50 (SD: 15) with 48% aged 40–59. Of these respondents, 2,168 (62%) were females; 2,077 (59%) were Caucasian, 796 (28%) were Asian, and the fewer number of participants identified as Indigenous (3%), Latin American (2%), and Black (1%). Given differences in sample size across groups, for the purpose of this analysis we have regrouped participants as Caucasian, Indigenous, and Visible Minorities (the term used by Statistics Canada to define non-Caucasian populations)27. Most of the participants lived with others in the same household (79%), were employed (68%)—though only one-third worked full-time—and 16% were unable to work following COVID-19 illness (Table 1). Table 1Demographic characteristics of studied cohort.Characteristicsn = 3463n (%*)Age (mean (SD))50 (15)Sex Males1322 (38) Females2141 (62)Ethnicity Caucasian2065 (60) Indigenous117 (3) Asian780 (23) Latin American86 (2) Black25 (0.7)Living setting Alone653 (19) With others2738 (79)Employment status Employed2347 (68) Unemployed1024 (30)Work arrangement Full-time1225 (36) Part-time (> 50% regular hours)265 (8) Part-time (< 50% regular hours)257 (7) Unable to work after COVID570 (16)*These percentages might not add up to 100% as there were patients who did not respond to specific question(s). Demographic characteristics of studied cohort. *These percentages might not add up to 100% as there were patients who did not respond to specific question(s). Impacts of long COVID on HRQoL The mean EQ-5D-VAS score is 56 (SD: 21) with half of the cohort scored below 60 and a quarter of them scored below 40. In general, females score higher than males (59 vs 54), with the youngest females have the highest average score (mean: 68, SD: 22). Though these differences are not statistically different across groups (Fig. 1). Mean EQ-5D-VAS score by age group and sex. Among the five dimensions of HRQoL, four-over-five patients experience difficulties doing usual activities (82%) or some pain/discomfort (81%), three quarters of patients report mental health problems (76%), 60% have problems in walking around, while the majority of them find themselves having no problem related to self-care (i.e. dress and wash by themselves). Only 5% of the cohort experience no impact from Long COVID, the mean of EQ-5D-VAS score for this group is 83 (SD: 19). The differences in having problems with the five measured dimensions across age groups are not statistically significant (Fig. 2). Percentages of participants (n = 3463) reporting having problems (slight or more) for each dimension of the EQ-5D-5L questionnaire by age group. EQ-5D-5L dimensions and EQ-5D-VAS We explore the associations between various demographic factors and the dimensions of HRQoL as measured by the EQ-5D-VAS. A complete result from multivariable regression is presented in Supplementary S2. The findings reveal predominantly negative coefficients across different levels of the five dimensions, suggesting that more severe symptoms are associated with lower quality of life scores. Additionally, the cumulative effect of deteriorating levels across all five dimensions shows a statistically significant negative impact on the VAS, with each incremental level of difficulty reducing the EQ-5D-VAS score by approximately 3. However, there is an exception in the dimension of self-care: individuals unable to dress or wash themselves exhibited a coefficient of 10.8, unexpectedly indicating a higher EQ-5D-VAS score compared to those facing fewer challenges in self-care. This unusual finding is attributed to two outliers who, despite severe self-care limitations, rated their quality of life very highly (scores of 70 and 90 out of 100). When these outliers are excluded, the association between the most severe level of self-care difficulty and EQ-5D-VAS score becomes statistically insignificant. Select social determinants of health Otherwise, ethnicity presents significant associations; Indigenous and visible minority groups reported higher VAS scores relative to Caucasians, with coefficients of 3.92 and 2.82, respectively, both statistically significant. Other demographic status like age, sex, living arrangement, employment status do not have a consistent trend when accessed individually; in the multivariable regression model, they do not statistically associate with EQ-5D-VAS. Findings from partial correlation coefficients indicate that different characteristics affects differently to each dimension. A complete table of the findings from this analysis can be found in Table 2 of Supplementary S2. A positive coefficient shows an increment in levels of dimension (the higher the level, the less desirable), while a negative coefficient shows a decrement in levels (the better the patients can perform). These reported coefficients reveal a range of effects that various characteristics have on different dimensions of HRQoL. Age and gender Aging is associated with slightly higher levels of mobility, as indicated by positive coefficients of 0.09 and 0.10 for people aged 40–59 and 60 + , respectively. Conversely, older age correlates with a reduction in anxiety and depression, as shown by a negative coefficient of − 0.08 and − 0.17. Gender also plays a significant role; being female is linked to higher levels in all dimensions, particularly in usual activities and pain/discomfort, where the coefficients are 0.11 and 0.10, respectively. Living arrangement and employment Living with others generally results in better outcomes in all examined dimensions, with coefficients ranging from − 0.06 to − 0.03. Being unemployment shows significant impact on mobility and self-care ability with coefficients of 0.04 and 0.07, respectively. Working full-time, compared to part-time, with and near-zero coefficients, suggesting little to no effect on HRQoL dimensions; however, being unable to work following an acute infection of COVID links to worse state of health for all five dimensions, which has the highest coefficients ranging from 0.11 to 0.27. Ethnicity Belonging to an Indigenous group or being a visible minority correlates with variations in HRQoL, in which Indigenous people indicate worse states for four dimensions, except usual activities; while people of visible minorities experienced more difficulties with self-care, and better states of usual activities, pain/discomfort and anxiety/depression. These relationships highlight the complex interplay between social determinants of health and various aspects of HRQoL. QoL score When partial correlations were done for EQ-5D-VAS, the score was statistically higher for middle-age (39-60y) people, females, people who live with others, and for those of visible minorities (coefficient was 0.08, 0.10, and 0.13, respectively); it decreased among people who were unable to work after COVID in comparison to full-time workers (coefficient − 0.19); and did not statistically associate with employment status or Indigenous people. Discussion As of December 2023, BC had approximately 400,000 COVID-19 cases reported28, and approximately 65,000 BC adults living with Long COVID29. This is a significant portion of the Canadian population and highlights the importance of better understanding the impacts of this relatively new illness on the health and well-being of patients. Overall, we find that patients reported impacts of Long COVID on their quality of life, but that these impacts vary by age and ethnicity. Differences across identities demonstrate different impacts related to SDoH and underscore the need to better understand these dynamics and to ensure that supports are tailored to diverse populations. We find that the majority of participants in our sample are older (50% aged 50 or older) and female (62%). This aligns with national trends in reporting of Long COVID whereby these number reported by Canadian Census 2021 were 40% and 51%, respectively29. In comparison to the 2021 census30, participants in our cohort are more likely to live with other relatives or non-relatives (79% vs 40%), to work full-time (35% vs 32%), though only 14% of our cohort worked part-time (vs 28%), and 16% lost their jobs due to COVID-19. Overall, our cohort can be bias toward those who would normally be more concerned about their health and more likely to seek for health services, for example, older females who lived with their caregivers31. The prolonged impacts of Long COVD-19 signs and symptoms can hugely affect people’s health and quality of life32. Our study assessed quality of life using two tools. HRQoL scores based on five dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) and a self-reported EQ-5D-VAS. The EQ-5D-VAS was predictable from the EQ-5D-5L states with people with more difficulties ranking themselves with lower VAS. The low HRQoL scores (mean 55.8) from self-assessments indicates profound impacts on HRQoL. In comparison to the mean EQ-VAS population norm of 82.3 for Canadians33, 89% of patients with Long COVID in this cohort had HRQoL scores lower than the norm, while 50% had a score lower than 60. Our findings are similar to other studies on quality of life and Long COVID, including those that compare patients with Long COVID to control groups32,34. The EQ-5D-VAS scores from our sample are similar to those for diseases of the blood, digestive and nervous systems, as well as those with congenital malformations35, suggesting Long COVID be recognized as a severe health condition requiring multiple supports to improve patient HRQoL. It also aligns with reports on the chronic and multi-systemic impacts, specifically on pulmonary and nervous system, of Long COVID36. Survey responses indicate that Long-COVID particularly affects ability to do usual activities (e.g. work, study, housework, family, or leisure activities), causes pain or discomfort and mental health problems across age groups. Other studies using the HRQoL dimensions have similarly documented significant impacts in terms of pain, anxiety and depression37. Our SDoH analysis reveals that older people (aged 60 +) and those of visible minorities or Indigenous population reported significantly higher VAS; though the score did not statistically associate with those who worked full-time, lived alone, or Indigenous people. Age and sex Female sex and older age are associated with worse outcomes across all five dimensions suggesting that females with Long COVID experience more challenges in their daily life and health conditions than males. In our study, they rated their quality of life higher than males, though the difference was not statistically significant. This finding aligns with other research which find female sex to be associated with lower HRQoL among Long COVID patients38. Lower HRQoL among women, in general, has been associated with their lower average socioeconomic status, than men, as well as more frequent dismal of symptoms (i.e., medical gaslighting)39. Future research identifying older females with Long COVID as a targeted group with special needs can look into ways to provide more appropriate support for this population. Ethnicity Indigenous identity was associated with HRQoL impacts related to higher levels of difficulties in mobility and self-care, as well as pain/discomfort and anxiety/depression; while being a member of a visible minority was also associated with more difficulties in self-care, but better scores in terms of usual activities, pain/discomfort scopes, and anxiety/depression. While our data does not provide explanations for these differences, results should be understood within a context of colonization, structural racism (including within the Canadian health system), income inequality and lack of culturally sensitive care—factors that have been found to be related to health inequities experienced by Indigenous Peoples and visual minorities in Canada40. Given that our cohort was a relatively good representative sample for Indigenous and Asian people (3% and 28% in comparison to 5% and 12% of the Canadian population)30, these results also reflect a need to prioritize culturally sensitive interventions to improve quality of life among Indigenous peoples and visible minorities living with Long COVID. It is worth noting that we did not have a sizeable sample of people from other visible minorities, which might be a result of the difference between BC’s population structure and other Canadian provinces30, but it could also be a result of unknown reasons for not getting access to appropriate care for Long COVID. Living arrangement The association of living alone with worsened states across all HRQoL dimensions in our study highlights the significant impact of social isolation as a SDoH41. Individuals who live alone may experience reduced emotional support and access to care, which are crucial for managing chronic illnesses like Long COVID42. This lack of support can exacerbate symptoms, particularly in mental health domains such as anxiety and depression43. Future research should investigate interventions that enhance social connectivity and support for this demographic, and public health policies need to incorporate these findings to improve support systems and services for those living alone, ensuring a holistic approach to health and wellbeing. Working conditions Being employed is associated with lower effects on mobility and self-care, and working full-time is associated with better states across all five dimensions when compares to people who work less then 50% number of regular hour and people who are unable to work. This may present a fact that people with better health condition are more likely to be in the workforce, particularly as full-time workers. However, given that they are not completely in a healthy state, this may point to the challenges people with Long COVID face in managing an episodic health condition and full-time employment demands44,45. Results can also be considered in relation to the growing literature on precarious employment as a SDoH. It maybe the precarious work has compounding effect on HRQoL for those with Long COVID46. Since even before the pandemic, a discussion about the flexibility in work arrangement to maintain a work-life balance was already a major concern of the workforce worldwide47. The Canadian government has called for public’s contribution to a formal discussion about the issue48. Other studies note those with Long COVID benefit from flexible work arrangements49. Multiple modifications to the Canada Labour Standards Regulations have been made since then. However, the policies for flexible work arrangement are still on a case-by-case basis, which can add unnecessary burdens to workers who are living with or recovering from Long COVID. Furthermore, Long COVID patients have reported challenges accessing employment benefits, which may restrict ability to leave full-time employment even when symptoms are debilitating50,51. Taken together, such results point to flexible employment arrangements to maintain one’s economic stability and increased access to benefits as means to improve HRQoL for Long COVID patients52. Limitations Our study using data collected provincially from the Post-COVID Interdisciplinary Clinical Network (PC-ICCN) has its limitation of a secondary research, in which we could not control for variable selection and the quality of collected data. Here we analyzed select SDoH to better understand the effects of Long COVID on HRQoL based on availability of data. As such our analysis was limited to what the initial survey conducted and we could not analyze for gender as well as sex, or other important SDoH that could influence HRQoL, for instance, factors like community support, neighborhood characteristics, or access to healthcare services, which might significantly impact HRQoL, were not included. This is specifically true for people of prioritized populations, such as Indigenous people or people of visible minorities, where the connection between gender, culture, religious or faith, etc. and health access and HRQoL can be complicated and require a thorough approach. Secondly, because the sample was drawn from primary care referrals to PC-ICCN, the findings might not be generalizable to all individuals with Long COVID, particularly those who do not seek or have access to clinic-based care. Thirdly, as it is a self-report questionnaire, we solely relied on patients and their personal experience; hence, potential biases in reporting (e.g., recall bias, social desirability bias) may occurred and could affect the accuracy of the data related to HRQoL and the severity of symptoms. And finally, without a comparable control group, we are unable to investigate causal effects between SDoH and HRQoL. Future research Despite the limitations discussed above, there are several promising directions for future research that can significantly enhance our understanding of Long COVID and its impacts. (1) Longitudinal studies are essential for examining the progression and long-term effects of Long COVID, particularly in understanding causal relationships between SDoH and HRQoL. (2) Analyzing HRQoL along SDoH axes deepens understanding of how the experience of living with Long COVID differs among population groups. Such analysis demonstrates that HRQoL is not simply determined by the diagnosed medical condition, but also non-medical factors such as demographics (age, sex, etc..) and social positions (ethnicity, employment status). (3) Deeper understanding of the relationship between SDoH and HRQoL among Long COVID patients can inform interventions designed to improve quality of life for specific population groups and address the underlying determinants of poor health outcomes. This study initiates the exploration of social determinants of health (SDoH) by focusing on selected factors available through secondary data. Further research is needed to examine additional SDoH, such as access to healthcare, education level, and social support networks. This should include investigating how various SDoH, including structural determinants like colonization, impact the health outcomes of Indigenous peoples and visible minorities, particularly in relation to health-related quality of life (HRQoL) among individuals with Long COVID. SDoH frameworks may also shed light on relationships between HRQoL, access to health care and health outcomes for those with Long COVID. Further research should also expand the scope of SDoH analyzed and incorporate diverse population samples from various geographic and healthcare settings to improve the generalizability of findings. Including control groups and employing qualitative research methods can provide deeper insights into the unique challenges and personal experiences of Long COVID patients. Additionally, exploring the biological and psychological mechanisms of Long COVID, alongside the effectiveness of technological and innovative healthcare solutions, can drive the development of targeted interventions and inform public health strategies. This comprehensive approach will not only elucidate the complexities of Long COVID but also enhance healthcare responses to this and similar public health challenges. Conclusion In conclusion, this study underscores the profound impact of Long COVID on health-related quality of life, revealing significant correlations with various social determinants of health. These findings demonstrate the need for equity-based policy responses to Long COVID that take into consideration the needs of those whose HRQoL is most effected. This might include: The development of culturally competent Long COVID care for Indigenous Peoples, The development of culturally competent Long COVID care for Indigenous Peoples, Education for clinicians on unconscious bias, sex and gender and medical gaslighting, Education for clinicians on unconscious bias, sex and gender and medical gaslighting, Promoting flexible work arrangements for those with Long COVID, and Promoting flexible work arrangements for those with Long COVID, and Social prescribing for Long COVID patients, particularly those that live alone. Social prescribing for Long COVID patients, particularly those that live alone. Our findings further highlight the need for multifaceted approaches to research incorporating longitudinal studies, broader SDoH analysis, and diverse population samples. Future research should also explore the specific biological and psychological mechanisms underlying Long COVID, as well as the potential benefits of technological innovations in healthcare. By deepening our understanding and expanding our methodological approaches, we can better tailor interventions and inform policies to effectively address the challenges posed by Long COVID and improve patient outcomes. Supplementary Information Supplementary Information 1. Supplementary Information 2. Acknowledgements This study was funded through the support of Health Research BC and the Post-COVID Interdisciplinary Care Network. KGC, JS and KB are supported by a Michael Smith Health Research BC Scholar Award. Author contributions JS, KB, KC: conceptualization, methodology, manuscript reviewing and editing; AP: methodology, data analysis, original draft manuscript preparation; EK: conceptualization, manuscript reviewing and editing. All authors have read, revised, and approved of the final manuscript. Data availability The dataset can be available upon reasonable request with REB and approved agreements in place to Esther Khor, Post-COVID-19 Interdisciplinary Clinical Care Network, Provincial Health Services Authority, Vancouver, Canada. Declarations Competing interests The authors declare no competing interests. Ethics declaration This study has received approval from the Research Ethics Board at the Simon Fraser University (#30001542) and adheres to all relevant guidelines and regulations for research involving de-identified health data (e.g. CHREB, Tri-Council Policy Statement on the Ethical Conduct for Research Involving Humans [TCPS2]). Consent to participate/consent to publish Not applicable. According to the Freedom of Information and Protection of Privacy Act—British Columbia Law (RSBC 1996, c 165, “https://canlii.ca/t/566l3” retrieved on 2024-07-22), aggregated data from medical records can be used for research purposes at health organizations; therefore, informed consent is not required. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Supplementary Information The online version contains supplementary material available at 10.1038/s41598-024-81275-4. References Associated Data Supplementary Materials Supplementary Information 1. Supplementary Information 2. Data Availability Statement The dataset can be available upon reasonable request with REB and approved agreements in place to Esther Khor, Post-COVID-19 Interdisciplinary Clinical Care Network, Provincial Health Services Authority, Vancouver, Canada.