Evaluation of post-COVID health status using the EuroQol-5D-5L scale

ABSTRACT

SARS-CoV-2 has had a lasting effect on the overall health of recovered patients, called ‘long COVID’. Currently, there is a lack of a validated standard questionnaire to assess post-COVID health status. A retrospective observational study involving the recovered COVID patients admitted to a secondary care hospital in India between June to December 2020 (n = 123), was conducted using the EuroQol-5D-5L scale at discharge, 4 weeks and 8 weeks post-discharge. A significant difference in anxiety/depression scores was found (χ² = 65.6, p < 0.000) among the 3 categories of time (discharge, 4 weeks and 8 weeks). The anxiety/depression dimension scores showed a significant change (p < 0.0001) between discharge and 8 weeks, using paired t-test. Age had a significant relationship with the anxiety/depression dimension at 4 weeks (OR = 5.617, 95% CI = 1.0320–30.5746, p < 0.05). A significant difference was found using Kruskal-Wallis rank-sum test on mean index scores (χ² = 60.0, p < 0.000) among the three categories of time (discharge, 4 weeks and 8 weeks). There was a statistically significant difference of time on EQ Index scores as determined by one-way repeated measures ANOVA (F(2,375) = 18.941, p = <0.00001). Our study found time to have a statistically significant impact on the mean index scores, level sum scores and dimension scores. Smoking was found to be significantly associated with usual activity scores at 4 weeks. The most remarkable changes occurred in the anxiety/depression dimension. Overall, there was a general trend of health improvement.

Introduction

In the current pandemic of SARS-CoV-2, there is an ever-changing narrative of the struggles faced by each person that has contracted the disease. This virus has not only affected our physical health, but also our mental and psychosocial well-being. The period of infection is highly critical, owing to the various complications caused by SARS-CoV-2 and post-recovery, patients are expected to resume their functions normally[1]. While early recovery rates are encouraging, long-term sequelae will need to be further investigated and there may be an increase in patients with persistent post-viral functional loss as a result of the pandemic[2]. For survivors of severe COVID-19 disease, having defeated the virus is just the beginning of an uncharted recovery path. The sequelae to the acute phase of SARS-CoV-2 infection impact all spheres of the patient’s life. Social support plays a key role in well-being and during times of crisis, it is emphasized as a coping mechanism. At the same time, one of the major preventive efforts for limiting the spread of COVID-19 includes social distancing. Hence, the impact of this disease has to be studied with due attention given to the psychosocial well-being of the recovered patient[3]. Early recognition, investigation and management of COVID-19-related neurological disease are challenging. Follow-up studies are necessary to ascertain the long-term neurological and neuropsychological consequences of this pandemic[4].

Post-acute care of patients with coronavirus (COVID-19) has become particularly relevant after having addressed the surge of infections in acute care settings. The prevalence of COVID-19 during the study period was 1.5–7.7% in the study area[5]. Majority of the asymptomatic cases were reported to be in individuals aged 16–45 years while symptomatic individuals were between 31–65 years old[6]. It is anticipated that infection with Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) may have a major impact on physical, cognitive, mental and social health status, involving even patients with mild disease.

The Karnataka Mental Health Report 2019 reported around 10–14.9 thousand cases (nearly 8%) of common mental disorders like depression and generalized anxiety disorder (GAD) in the study area with a burden of 12.5% among children and adolescents and 32.6% in the elderly[7]. Overall, India has a 10.6% prevalence for mental illnesses. [8] Depression, anxiety and stress were more prevalent among the Indian population during the lockdown periods[9]. Individuals with limited supplies were affected as family affluence in India is negatively proportional to mental distress[10]. Along with focusing on measures to contain the spread of COVID-19, the Indian government and mental health experts have been urged to heed attention to the mental health of citizens[9].

Despite the huge number of COVID-related papers that have flooded scientific journals, the clinical picture of the COVID-19 after-math is still indistinct. Clinicians can derive details of the so-called ‘long-COVID’ only from case reports and small studies, especially in the absence of large-scale observational studies. The growing number of survivors demands us to view their recuperation in a new light [11,12]. A vague idea exists about the effects of COVID-19 which linger long after recovery, but a quantifiable scale is necessary if we are to address and cope with these effects[13]. Moreover, a clear idea of the disease and what consequences the survivors and their physicians should anticipate can prepare them to deal with the same in a better manner.

The holistic recovery of the COVID-19 survivors is vital for the restoration of their quality of life. The survivorship experience may highlight the importance of collaboration between the fields of critical care and rehabilitation to optimize post-COVID-19 recovery [14,15].

A clinician’s responsibility extends beyond discharge from the hospital, into effective follow-up. For this purpose, a simple, standard and reproducible tool is required to guide physicians and public health officials. Such an instrument will help to pinpoint the red flags in all sectors of the survivors’ health and also assist the physicians in guiding patients to ease back into their routine life. Currently, there is a lack of an efficient follow-up system and a validated standard questionnaire to assess post-COVID recovery. Upon establishing a well-rounded evaluation system, the scope for rehabilitation centers arises. This will assure holistic recovery edging into tertiary prevention. It will also aid in developing screening and treatment programs to minimize the long-term cognitive consequences of COVID-19[16]. The objective of this study was to assess the post-COVID health status in recovered patients using the EuroQol-5D questionnaire.

Methodology

This was a retrospective observational study set in a secondary-care hospital in Karnataka, India. The study subjects consisted of COVID-19 positive patients admitted, treated and discharged from the hospital in the period of June-December 2020. The COVID-19 diagnosis was confirmed using reverse transcription Polymerase Chain Reaction (RT-PCR) testing of nasopharyngeal swab samples from the subjects. The data was collected via telephonic interviews. The assessments were made at discharge, 4 weeks and 8 weeks from the date of discharge. All subjects who met the study criteria were recruited into the study and a convenient sampling method was followed. Sample size was calculated by performing a pilot study where it was found that 92% of the subject population had health state 11111 (full health). Using the 4pq/L² formula i.e. (4x8x92)/25 = 118 patients was calculated to be the minimum sample size.

Inclusion criteria

• All post-COVID patients who are willing to participate in the study

• Aged over 18 years

Exclusion criteria

• Patients who are not willing to participate in this study.

• Patients who had a diagnosis of major depressive disorder/anxiety/chronic fatigue syndrome/Medically Unexplained Symptoms (MUS) before the diagnosis of COVID-19

• Patients who required assistance for Activities of Daily Life (ADL) before the diagnosis of COVID-19

• Patients who had restricted ADL before the diagnosis of COVID-19

• Patients with vocational problems due to their health status before the diagnosis of COVID-19

• Patients with diagnosed pulmonary function impairment/ muscle weakness before the diagnosis of COVID-19 (respiratory impairment due to COPD, TB, pneumonia, cystic fibrosis, or other persistent lung diseases/muscle weakness due to trauma, myasthenia gravis, poliomyelitis, or other persistent musculoskeletal disorders).

The EQ-5D-5L tool

The EuroQol five-dimension five-level (EQ-5D-5L) is a broad, standardized tool developed by the EuroQol group, which has been used to evaluate the quality of life (QoL) in various clinical settings the world. It is a descriptive, patient-reported measure and has 5 dimensions – mobility, self-care, usual activities, pain/discomfort and anxiety/depression. The questionnaire is set in a multiple-choice pattern and the patient is directed to choose a single option under each dimension that best describes their current health status. Responses in each category are coded as 1, 2, 3, 4, or 5 (going from 1, the best health option to 5, the worst health option). Putting the codes from each category together, a patient’s ‘health status profile’ or ‘health state’ can be described as a 5-digit number ranging from 11111 (no problems in any dimension) to 55555 (the most severe problem in every dimension). There are thus 3125 (=5⁵) health states that can be described by the EuroQol-5D-5L. Another way to represent the responses is to obtain an EQ-5D ‘index score’ by allotting a numerical value to each health state. Considering that health preferences can vary between populations, there are different value sets for each country. In this study, the index score was calculated with SPSS using the United States (US) Pickard value set- Version 1.1 (Updated 16/11/2020)[17]. The Indian value set is yet to be completed and hence, the US value set was used instead. The index values range from 1 (full health) to 0 (death) and lower, up to −0.59 (which indicate very poor health states considered to be worse than death on this scale).

Permission from the EuroQol Research Foundation was obtained for the use of their EQ-5D-5L for Paper Interviewer Administration in English (USA) in this study.

Statistical Analysis

Mean and standard deviations were calculated for continuous variables, frequencies and percentages for categorical variables. The associations of the sociodemographic factors and the dimension, index and level-sum scores of EQ-5D were analyzed with t-test, analysis of variance (ANOVA) and nonparametric statistics (Mann-Whitney U test or Kruskal-Wallis test). The percentage of people in each dimension was calculated and χ2 tests were performed to examine the statistical significance of the difference between groups in the percentage of reported problems. Fisher’s exact test was used when the exact theory frequency was less than 1. Logistic regression models were used with the five health dimensions considered to be dependent variables (0 = no problem, 1 = any problem). Statistical significance was set at 0.05 using two-sided tests and statistical analysis was performed (Image 1).

Image 1.

STROBE participant flow diagram.

Results

The mean age of the respondents was 45.28 years with a standard deviation of 17.76. All the active smokers and ex-smokers were male whereas passive smokers had equal gender distribution. Table 1 shows the frequencies of the respondents by sociodemographic categories.

Table 1.

Socio-demographic characteristics.

Variable	Category	Count	Percentage
Type of respondent	Patient alone	87	71%
	Patient assisted by a relative/caretaker	10	8%
	Only another person (relative/caretaker)	26	21%
Age	18–39 years	53	43.1%
	40–59 years	45	36.6%
	≥ 60 years	25	20.3%
Sex	Male	72	58.5%
	Female	51	41.5%
Smoking history	Nonsmokers	112	91.1%
	Ex-smokers	1	0.8%
	Active smokers	4	3.3%
	Passive smokers	6	4.8%

A) EuroQol Dimensions

Table 2 calculates the frequencies of item responses in each EQ-5D-5L dimension by percentage (%). The means and standard deviation, as well as changes in dimension-specific scores, were calculated for all EQ-5D dimensions. The average EuroQol dimension values for mobility, self-care, usual activities, pain/discomfort, anxiety/depression at 0, 4 and 8 weeks respectively were (1.18, 1.1, 1.15, 1.2, 1.41), (1.16, 1.08, 1.1, 1.12, 1.05) and (1.1, 1.04, 1.06, 1.09, 1.02).

Table 2.

Frequencies of item responses in each EQ-5D-5L dimension by level and time in percentage (%).

Level	MO			SC			UA			PD			AD
	D	*	‖	D	*	‖	D	*	‖	D	*	‖	D	*	‖
1 (No problem)	90.2	91.9	95.1	95.1	96.7	98.4	91.1	94.3	96.7	87.8	93.5	95.9	68.3	95.9	99.2
2 (Slight problem)	4.9	2.4	2.4	3.3	0.8	0.8	5.7	3.3	2.4	8.1	4.1	2.4	22.8	3.3	0
3 (Moderate problem)	2.4	4.1	0.8	0	0.8	0	1.6	1.6	0	2.4	0.8	0	8.1	0.8	0.8
4 (Severe problem)	1.6	0.8	0.8	0	0.8	0	0.8	0	0	0	0	0	0.8	0	0
5 (Extreme problem/inability)	0.8	0.8	0.8	1.6	0.8	0.8	0.8	0.8	0.8	1.6	1.6	1.6	0	0	0

MO = Mobility, SC = Self-care, UA = Usual Activity, PD = Pain/Discomfort, AD = Anxiety/Depression dimensions, D = Discharge, * = 4 weeks, ‖ = 8 weeks

Table 3 describes longitudinal changes in the EuroQol health dimensions over the three setpoints in time. Kruskal-Wallis rank-sum test was conducted to examine the differences of time on dimension scores. A significant difference in anxiety scores was found (χ2 = 65.6, p < 0.000) among the 3 categories of time (discharge, 4 weeks and 8 weeks). A pairwise post-hoc Dunn test with Holm correction was significant for 4 weeks – Discharge (p < 0.000) and 8 weeks – Discharge (p < 0.000).

Table 3.

Longitudinal analysis – Describing changes in health dimensions between multiple time points.

Time of assessment	Number/percentage of respondents reporting ‘any problem’ in the dimension
	Mobility	Self-care	Usual activities	Pain/ discomfort	Anxiety/ depression
At discharge	12 (9.8%)	6 (4.9%)	11 (8.9%)	15 (12.2%)	39 (31.7%)
At 4 weeks post-discharge	10 (8.1%)	4 (3.3%)	7 (5.7%)	8 (6.5%)	5 (4.07%)
At 8 weeks post-discharge	6 (4.9%)	2 (1.6%)	4 (3.3%)	5 (4.1%)	1 (0.8%)
Change between first and last assessment	−6	−4	−7	−10	−38
Percentage change between first and last assessment	−50%	−66.7%	−63.6%	−66.7%	−97.4%

Two-sided paired t-tests were used to assess the statistical significance of the change in the EQ-5D-5L dimension scores at discharge and 8 weeks. Wilcoxon signed-rank test was used to assess the strength of the paired t-tests (p-value). Cohen’s d was estimated as the standardized effect size (SES)[18]. The anxiety/depression dimension showed a significant change (p < 0.0001) between these two assessments, as seen in Table 4.

Table 4.

Comparison of EQ-5D-5L Dimension scores at discharge (baseline) and 8 weeks.

Dimension	At Discharge	At 8 weeks	p value
Mobility
Mean (SD*) Mean change (95%CI**) Standardized effect size	1.18 (0.63) 0.08(−0.07 to 0.23) 0.142044	1.1 (0.5)	0.2770
Self-care
Mean (SD) Mean change (95%CI) Standardized effect size	1.1 (0.53) 0.06 (−0.06 to 0.17) 0.131275	1.04 (0.37)	0.3383
Usual activity
Mean (SD) Mean change (95%CI) Standardized effect size	1.15 (0.55) 0.09 (−0.03 to 0.21) 0.188774	1.06 (0.39)	0.1529
Pain/discomfort
Mean (SD) Mean change (95%CI) Standardized effect size	1.2 (0.64) 0.11(−0.05 to 0.26) 0.187209	1.09 (0.53)	0.1692
Anxiety/depression
Mean (SD) Mean change (95%CI) Standardized effect size	1.41 (0.68) 0.40(0.27 to 0.53) 0.784088	1.02 (0.18)	<0.0001

*SD = Standard deviation, **CI = Confidence Interval. Dimension score was compared between two assessment times using the paired t-test.

Henceforth, ‘any problem’ will be used to refer to a deviation of the dimension score from 1. As seen in Table 3, there is an improvement in all dimensions over time. The most striking of these improvements lies in the anxiety/depression dimension, where those reporting ‘any problem’ dropped from 31.7% of those surveyed at discharge to just 0.8% at 8 weeks post-discharge. Hence, 97.4% of individuals who initially had ‘any problem’ at discharge, showed an improvement in this dimension by 8 weeks post-discharge.

Most respondents in all age groups reported anxiety/depression at the time of discharge. Furthermore, self-care and usual activity problems were reported most in those ≥ 60 years of age, mobility problems in the 40–59 years age group and pain/discomfort in the 18–39 years age group. Image 2 illustrates the distribution of respondents by dimension and time for all age groups.

Image 2.

Distribution of respondents by dimension and time for all age groups.

Each dimension of EQ-5D was dichotomized and used as the dependent variable. All the sociodemographic factors were included as independent variables and multivariate logistic regression models were conducted for each assessment period. Age had a significant relationship with the anxiety/depression dimension at 4 weeks (OR = 5.617, 95% CI = 1.0320–30.5746, p < 0.05).

Multiple linear regression was calculated to predict EQ-5D-5L dimension scores based on the sociodemographic factors.

At discharge, age was found to be a significant predictor for mobility, self-care and usual activity scores. A significant regression equation was found between age and usual activity scores (F (1,121) = 140.926, p < 0.01) with an R² of 0.538. Patients’ predicted Usual Activity score is equal to 0.383 + 0.638*(Age) where age was coded as 0 for ages less than 60 years and 1 for ages greater than 60 years. A significant regression equation was found between age and mobility scores (F (1,121) = 181.368, p < 0.01) with an R² of 0.599. Patients’ predicted mobility score is equal to 0.265 + 0.765*(Age). A significant regression equation was found between age and self-care scores (F (1,121) = 146.237, p < 0.01) with an R² of 0.547. Patients’ predicted self-care score is equal to 0.355 + 0.622*(Age).

At 4 weeks, age was found to be a significant predictor for self-care scores and smoking was a significant predictor for usual activity scores. A significant regression equation was found (F (1,121) = 146.237, p < 0.05) with an R² of 0.031. Patients’ predicted self-care score is equal to 1.071 + 0.248*(Age). A significant regression equation was found between smoking and usual activity scores (F (1,121) = 3.984, p < 0.05) with an R² of 0.032. Patients’ predicted usual activity score is equal to 1.363–0.292*(Smoking history) where nonsmokers (including ex-smokers) were coded as 0 and smokers (both active and passive) as 1.

At 8 weeks, age was found to be a significant predictor for mobility, usual activity, pain/discomfort and anxiety/depression scores. A significant regression equation was found between age and mobility scores (F (1,121) = 4.243, p < 0.05) with an R² of 0.034. Patients’ predicted mobility score is equal to 1.051 + 0.229*(Age). A significant regression equation was found between age and usual activity scores (F (1,121) = 3.984, p < 0.05) with an R² of 0.035. Patients’ predicted usual activity score is equal to 1.02 + 0.18*(Age). A significant regression equation was found between age and pain/discomfort scores (F (1,121) = 6.24, p < 0.05) with an R² of 0.049. Patients’ predicted pain/discomfort score is equal to 1.031 + 0.29*(Age). A significant regression equation was found between age and anxiety/depression scores (F (1,121) = 4.017, p < 0.05) with an R² of 0.032. Patients’ predicted anxiety/depression score is equal to 1.0 + 0.08*(Age).

B) EuroQol Health States

The health state ‘11111’ was the most prevalent at all three times of assessment, appearing in 55.3% of participants at discharge, 89.4% at 4 weeks and 94.3% at 8 weeks. The 10 most frequently observed health states and their frequencies at all 3 assessment times are described in Table 5.

Table 5.

Most frequently observed health states and their frequencies at Discharge, 4 and 8 weeks.

Top 10 Health states			Count (and Frequency/percentage (%))			Cumulative frequency (%)
At D	At *	At ‖	At D	At *	At ‖	At D	At *	At ‖
11111	11111	11111	68 (55.3)	110 (89.4)	116 (94.3)	55.3	89.4	94.3
11112	11112	11121	27 (22.0)	1 (0.8)	1 (0.8)	77.3	90.2	95.1
11113	11121	21211	7 (5.7)	1 (0.8)	1 (0.8)	83.0	91	95.9
11121	11212	21221	5 (4.1)	1 (0.8)	1 (0.8)	87.1	91.8	96.7
21111	21211	22211	3 (2.4)	1 (0.8)	1 (0.8)	89.5	92.6	97.5
11122	21222	31151	1 (0.8)	1 (0.8)	1 (0.8)	90.3	93.4	98.3
11211	22211	41121	1 (0.8)	1 (0.8)	1 (0.8)	91.1	94.2	99.1
11221	31111	55553	1 (0.8)	1 (0.8)	1 (0.8)	91.9	95.0	100
11314	31121		1 (0.8)	1 (0.8)		92.7	95.8
21211	31151		1 (0.8)	1 (0.8)		93.5	96.6
… (continued for all other health states)
55555	55555	55555	0 (0.0)	0 (0.0)	0 (0.0)	100	100	100

D = Discharge, * = 4 weeks, ‖ = 8 weeks

At discharge, there were 8 other health states seen beside the health states mentioned in the table [22221, 22333, 31221, 32231, 32251, 41121, 45433, 55553] and at 4 weeks, 5 other health states were observed [31151, 33331, 34322, 41121, 55553].

C) Paretian Classification of Health Change (PCHC) Analysis

Table 6 describes the changes in health according to the PCHC analysis. [19]

Table 6.

Changes in health for different assessment periods according to the PCHC.

Change	Discharge to 4 weeks	4 weeks to 8 weeks	Discharge to 8 weeks
Improve	52 (42.3%)	6 (4.9%)	52 (42.3%)
Worsen	1 (0.8%)	0	0
No change	70 (56.9%)	117 (95.1%)	71 (57.7%)
Mixed change	0	0	0

D) EQ-5D Level Sum Scores (LSS)

Level sum scores (LSS) are calculated simply by adding the 5 individual dimension scores. Kruskal-Wallis rank-sum test was conducted to examine the differences of time on LSS. A significant difference was found (χ2 = 60.4, p < 0.000) among the 3 categories of time (discharge, 4 weeks and 8 weeks). A pairwise post-hoc Dunn test with Holm correction was significant for 4 weeks – Discharge (p < 0.000) and 8 weeks – Discharge (p < 0.000).

E) EQ-5D Index Scores

Index scores at discharge ranged from −0.375 to 1 with a mean of 0.91, a standard deviation of 0.26 and a 95% CI of 0.87 to 0.95. The mode was 1. The same range of scores was seen at both 4 weeks and 8 weeks post-discharge. However, the mean improved to 0.94 (95%CI = 0.90 to 0.98) at 4 weeks and further to 0.95 (95%CI = 0.92 to 0.99) at 8 weeks post-discharge. The median index score for both these assessment times was 1.

Kruskal-Wallis rank-sum test was conducted to examine the differences of time on the mean index scores. A significant difference was found (χ2 = 60.0, p < 0.000) among the 3 categories of time (discharge, 4 weeks and 8 weeks). A pairwise post-hoc Dunn test with Holm correction was significant for 4 weeks – Discharge (p < 0.000) and 8 weeks – Discharge (p < 0.000).

Kruskal-Wallis rank-sum test was conducted to examine the differences of age on the mean index scores. At 8 weeks, a significant difference was found (χ2 = 9.58(2), p = 0.0083) among the 3 categories of age (18–39 years, 40–59 years and ≥ 60 years).

There was a statistically significant difference of time on EQ Index score as determined by one-way repeated measures ANOVA (F(2,375) = 18.941, p = <0.00001). The means obtained during the ANOVA test were 0.91 ± 0.23, 0.94 ± 0.21 and 0.95 ± 0.20 at discharge, 4 and 8 weeks respectively.

t-test showed no significant differences between mean index scores and gender at any assessment time. Similarly, there were no significant differences between mean index scores and smoking history as calculated by the Kruskal-Wallis test.

There were seven subjects, whose index scores did not improve to full health (Index score 1) as the rest of the subject population’s scores did. Three subjects were in the 70–80 age group and the remaining four were in the 20–40 age group. Out of these seven subjects, 3 subjects’ index scores remained the same across all three assessment times whereas, 4 subjects’ index scores improved slightly. Of these 4 subjects, 3 subjects’ index scores improved from assessment at discharge to assessment at 4 weeks but remained the same when assessed at 8 weeks. Only 1 subject showed improvement across all 3 assessment times.

Discussion

Patient-reported outcome measures have been perennially used in clinical and research settings to gauge the quality of life of the patient in convalescence. Various studies have successfully outlined its importance[20]. Current evidence denotes the importance of the quality of life assessment for patients recovering from Covid-19 [21–23], emphasizing the role of an international rehabilitation routine to direct physicians to focus on the holistic recovery of the patient [24–26]. The present study is the first of its kind to evaluate the quality of life (QoL) in post-COVID survivors in the Indian subpopulation of Dakshina Kannada. The authors identified the need for such a study, given the prevalence and impact of COVID-19. Literature indicates that multiple questionnaires have been used to explore QoL such as the Professional QoL questionnaire by Busselli et al [27] and a similar tool, PROMIS, by Jacobs et al [28]. However, there is limited literature for the use of EuroQol-5D in evaluating Post-COVID QoL [29–35]. Studies exploring QoL have been conducted widely across the globe including China, Australia, parts of Europe and India as well [36,37]. The Indian studies primarily focused on health care workers (HCWs) and patients with co-morbidities rather than a general overview of the QoL in COVID survivors. Hence, the present study is a pioneer study in elucidating the determining factors QoL in COVID-19 survivors, especially in the current study population and area.

Longitudinal and descriptive analysis, from discharge to 8 weeks, showed general trends of progress toward better health consistently. The current study found time to have a statistically significant impact on the mean index scores, level sum scores and dimension scores. The Paretian Classification of Health Change (PCHC) analysis showed maximum improvement from discharge to 4 weeks, indicating a relatively short curve for recovery to near-normal quality of life.

A drastic drop was noted in the number of people reporting anxiety/depression. Additionally, there was a statistically significant change between this dimension’s score at discharge to the one at 8 weeks. These findings are consistent with existing literature which shows a direct relationship between emotional distress and pandemics [38–42]. The anxiety/depression gradually reduces over time, as the immediate fear of mortality and morbidity wanes and the psychopathological effects of inflammatory mediators wear off[43]. Moreover, the lack of physical presence of loved ones during hospitalization and recovery leads to a feeling of isolation [44–46].

The underlying reasons for 18–39-year-olds having anxiety/depression during the pandemic, are explored by Larcher et al [47]. They reported that youngsters and adolescents had a generalized fear about their future, their family and broader society in the aftermath of the pandemic. Counterintuitively, elderly individuals (individuals older than 60 years) were found to be resilient to the mental stresses imposed on them by the pandemic[48]. However, the present logistic regression analysis found that at 4 weeks, the odds of individuals having anxiety/depression increase with age. Age was also found to have a statistically significant impact on the mean index scores at 8 weeks. Furthermore, age had a significant linear relationship with the individual dimension scores, differing across the assessment times, especially in consideration of the two categories of <60 years and >60 years of age. Evidence indicates young individuals attain a good quality of life by the end of their recovery period, whereas the older individuals were less likely to attain their pre-COVID quality of life. The average recovery time in patients older than and younger than 60 years was found to be 25 days and 21 days respectively in a study conducted by Barman et al [49]. The reason for this trend of younger individuals faring better against COVID-19 could be due to a decrease in the expression of angiotensin-converting enzyme-2 (ACE-2) with age[50]. ACE-2 is the primary target receptor for SARS-CoV-2 and has lung-protective effects by limiting angiotensin-2 mediated pulmonary capillary leak and inflammation. The rate of regeneration of alveolar epithelium is also better in younger individuals resulting in better pulmonary recovery from COVID-19.

In the present study, patients aged more than 60 years were noted to commonly report difficulty in the usual activity and self-care dimensions. A significant relationship was found between age and usual activity scores, at discharge and 8 weeks. The literature points out the association between disease severity and old age. [51,52] They also tend to have higher rates of morbidity, mortality and hospitalization. Many theories can explain this, ranging from cellular senescence and mitochondrial dysfunction to preexisting co-morbidities and inflammaging [53–55]. The most prominent reasoning is provided by the immunosenescence theory [56,57]. The preexisting comorbidities render these elderly individuals dependent, even before the COVID-19 infection [58,59]. Due to the need to isolate, they probably could not get assistance with daily activities.

Difficulty with mobility was reported most in the middle age group (40–59 years). A significant relationship was found to exist between age and mobility scores, at discharge and 8 weeks. It could be arising from myalgia and arthralgia which are common symptoms following most viral infections, including COVID-19 [60–62]. Moreover, individuals on regular pain management medication saw a significant disruption of the same due to the pandemic and especially if they contracted the disease, this leads to limited mobility[63]. The degenerative changes in weight-bearing joints begin in middle-aged individuals due to cellular senescence and oxidative damage of aging in general, which are further exacerbated by viral inflammation[64]. Likely, these were major symptoms in 69.2% of patients in a study conducted by Murat et al about pain as a clinical presentation in COVID-19 patients[65]. However some studies claim that their subjects had no symptoms of joint pain [66,67]. Friedman et al reported these symptoms to be more common in Influenza afflicted patients and rare in Covid-19 patients[68].

The current study found that the majority of the individuals reporting ‘any problem’ in the pain/discomfort dimension belonged to the age group of 18–39 years. A significant relationship was found to exist between age and pain/discomfort scores, at 8 weeks. This could be linked to the ubiquitous post-viral syndrome, viral-associated end-organ damage and new-onset pain due to increased sleep disturbance, inactivity, fear, anxiety, depression [69,70]. Post-viral pain/discomfort is also common in other infections like EBV and Coxiella burnetii, which extends into the recovery period[71]. However, its association with COVID-19 goes beyond just the pathological effect of the virus and can be explained by a biopsychosocial model[72]. Ibanez et al found an association between disruption of daily routine and elevated pain intensity and pain sensitivity in COVID-19 survivors[73]. This higher pain sensitivity coupled with their reluctance to rest or premature resumption of pre-COVID activities could explain why younger patients were affected more.

The majority of subjects in our study were nonsmokers. Smoking was also found to have a significant relationship with usual activity scores, at 4 weeks. Upon analysis, a direct relationship exists between usual activity and tobacco smoking. A smoker (active or passive) was found to be less likely to have problems with ‘usual activity. It is well-established that tobacco smoke has deleterious effects, ranging from COPD and cardiovascular disease to various forms of cancer [74,75]. However, pertaining to COVID-19, recent literature has shown a low prevalence of tobacco smokers in the COVID-19-afflicted patient population. This is not to be interpreted as tobacco smoke having a protective effect against COVID-19, as it’s merely an incidental finding. However, some studies have shown that nicotinic acetylcholine receptors could be targeted to prevent COVID-19 infection[76]. Other articles exhibit a potentially hazardous impact of tobacco smoking via over-expression of the ACE-2 receptor gene [77,78]. Current smokers had the highest expression and never-smokers had the lowest expression of ACE-2 receptors [79,80]. This relationship between smoking and COVID-19 severity is controversial and further research with a large sample size is required to forge a solid association.

Limitations

The study did not include an assessment using the Visual Analog Scale (VAS) of the EQ-5D tool. The study population was minimal due to the limited incidence of COVID-19 in the rural setting of a secondary-care hospital in Dakshina Kannada, India. The subjective nature of EuroQol 5D-5L limits the reliability of the information provided. There is a possibility of recall bias while using such retrospective tools. Sensitivity analysis was not performed.

A larger sample size along with the inclusion of the VAS and a baseline assessment of health status before COVID infection/hospitalization could provide stronger associations in a future study. Furthermore, of the 10 respondents absent from the interview, the cause cited was unavailability rather than disability. Hence, the answers provided by the subjects’ family members do not affect the validity of the study.

Conclusion

Quality of life assessment is a much-needed tool to improve the current COVID-19 recovery protocol. Our study found time to have a statistically significant impact on the mean index scores, level sum scores and dimension scores. Age was found to be a significant predictor for different dimensions, varying as per the assessment points. The most remarkable natural improvement occurred in the anxiety/depression dimension. Tobacco smoking status was found to be significantly associated with usual activity scores at 4 weeks. Overall, there was a general trend of improvement in health across all analysis. EQ-5D-5L is a useful tool to assess the health status of post-COVID patients in India.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).