Effects of vaccination against COVID-19 on the emotional health of older adults

Context

Peru is a country with a fragmented and heterogeneous health system. Among them, the Social Health Insurance of Peru (EsSalud) is one of the most important health systems in the country and is managed by the Ministry of Labor and Employment Promotion.³⁰ EsSalud gives medical attention to formal workers, retirees, and their families. The executive function of EsSalud is divided through the 29 healthcare networks, representative of each region in the country. EsSalud provides health coverage to almost a third of Peruvians, including more than 1 890 000 older adults, of which 100 000 of them are users of the Centers for the Elderly (CAM, in Spanish) present in each healthcare network in the country.³¹ The CAMs provide health services that seek to improve the functional, mental, and social capacity of people aged 60 or older affiliated with EsSalud.³²

The vaccination process against COVID-19 in Peru, organized by the Ministry of Health, was carried out by age groups, and began on April 16, 2021, with adults over 80 years old. Vaccination started on April 28, 2021, for adults over 70 years old, and on May 27, 2021, for adults over 60 years old. During that period, the BNT162b2 (BioNTech, Pfizer), ChAdOx1-S (Oxford, AstraZeneca), and BBIBP-CorV (Sinopharm) vaccines were available in Peru. However, the Peruvian Ministry of Health indicated that vaccination should be prioritized with BNT162b2 (BioNTech, Pfizer) for older adults with 60 years or older (https://cdn.www.gob.pe/uploads/document/file/1893194/Directiva%20%20Sanitaria%20N%C2%B0%20133-MINSA-2021-DGIESP%20.pdf). As a result, as of December 29, 2021, 80.8% of older adults who received at least one dose of the COVID-19 vaccine received BNT162b2 (BioNTech, Pfizer), while 10.5% and 8.7% had received the ChAdOx1-S vaccine (Oxford, AstraZeneca) and BBIBP-CorV (Sinopharm), respectively (https://www.minsa.gob.pe/reunis/data/vacunas-covid19.asp). However, the vaccination of older adults occurred during a political and health scandal in the country; Inoculation of the BBIBP-CorV (Sinopharm) vaccine candidate, outside the clinical trial, to 470 people in Peru, including health personnel and politicians.^33,34 For several months, this caused a credibility crisis for COVID-19 vaccines, particularly BBIBP-CorV (Sinopharm).³⁵

Study design

A prospective cohort study was conducted that aimed to estimate the effect of the COVID-19 vaccine on the following emotional health outcomes of older adults affiliated to EsSalud CAMs: (a) Perception of fear, (b) anxiety, and (c) worry about COVID-19, (d) general depression, and (e) general anxiety.

Population

We identified non-hospitalized adults aged 60 to 79 years affiliated to EsSalud and registered in the CAM available database at national level. The database represented 0.5% of all older adults affiliated to EsSalud. We included those older adults who had already been vaccinated (with any of the available vaccines) or who had planned to be vaccinated against COVID-19 according to the Peruvian Ministry of Health vaccination schedule at a vaccination site in Peru. We excluded adults aged 80 years or older because more than one month had elapsed since the start of vaccination in this age group by the initiation of the recruitment. We also excluded those who had some impediment to adequate communication with the interviewer via the telephone call, were diagnosed with COVID-19 in the last three months, had symptoms related to COVID-19, or refused to participate during the interview. The recruitment period was from May 27 to June 30, 2021.

Sample size calculation

Based on the recommendation of Cohen et al.,^36,37 we consider an effect size of 0.20 standard deviations for the smallest effect³⁶ estimated between any of the outcomes (perception score for fear, anxiety, and worry about COVID-19, general depression, and anxiety) and the exposure factor (unvaccinated vs. vaccinated with only one dose or vaccinated with two doses) in older adults. Assuming a significance level of 5%, statistical power of 80%, and equal variances between groups, we calculated a minimum sample of 788 older adults. Then, we corrected this value by a factor of 1.2, following the methodology proposed by Vititingghoff E. et al.,³⁸ considering the primary analysis with adjustment for potential confounding variables. Thus, we obtained a minimum sample size of 946 older adults. Finally, we considered a rejection rate of 10% and a loss to follow-up rate of 10%; then, we planned to invite 1168 participants to the study. Further details are found in the extended data (Supplementary methods: Sample size calculation).³⁹

Sampling

After excluding those who did not meet the selection criteria or did not have identification or contact data, we took 7 685 older adults from the registered CAMs into the database as a sample frame. From them, we chose a randomized and stratified sample for each of the 30 healthcare networks (n = 1 686). In addition, the sampling was carried out independently for two age subpopulations: 60 to 69 years (n = 846) and 70 to 79 (n = 840) years. For each subpopulation, we chose half of the calculated total sample size (allocation ratio 1:1). We decided to choose two age subpopulations because the recruitment period was during the start of vaccination of adults older than 60 to 69 years and one month after the initiation of vaccinations to adults older than 70 to 79 years (https://cdn.www.gob.pe/uploads/document/file/1893194/Directiva%20%20Sanitaria%20N%C2%B0%20133-MINSA-2021-DGIESP%20.pdf). The characteristics of the eligible population and selected sample are found in the extended data (Supplementary methods: Characteristics of the eligible population).³⁹ To reduce non-response bias, we adjusted sample weights to account for non-response using weighting class adjustment.^40,41 Further details are found in the extended data (Supplementary: Sampling weights calculations).³⁹

Outcomes

We assessed five outcomes in emotional health: fear of COVID-19, anxiety about COVID-19, worry about COVID-19, general anxiety, and general depression, perceived by the respondents during the last two weeks before responding to the survey. Fear of COVID-19 was measured with the Spanish version of the Fear of COVID-19 Scale (FCV-19S) with seven items that are answered on a Likert scale from 1 (strongly disagree) to 5 (strongly agree).⁴² This scale measures the emotional and somatic fear response to COVID-19. It has adequate internal consistency (Cronbach’s α = 0.88) and convergent validity with other mental health covariates in two non-probabilistic samples of adults and older adults from Lima, Peru.^42,43 The anxiety about COVID-19 was measured using the Spanish version of the Coronavirus Anxiety Scale (CAS), which measures persistent and excessive concern about COVID-19 that is accompanied by physical symptoms⁴⁴ and has five items that are answered on a Likert scale from 0 (not at all) to 4 (almost every day).⁴⁵ It is a unidimensional scale with adequate internal consistency (Cronbach’s α = 0.89 and 0.91), convergent validity with anxiety, and adjustment rates between females and males, and between older adults aged 60 to 65 years and 66 to 86 years. However, estimations in samples of adults and older adults from Lima, Peru, using Item Response Theory models, suggest that the instrument is more reliable in those with high anxiety levels to COVID-19.^46,47 The outcome ‘worry about COVID-19’ was measured with the scale to measure worry about contagion of COVID-19 (PRE-COVID-19, in Spanish), which contains six items that are answered on a Likert scale of 1 (never or rarely) to 4 (almost all the time). This scale measures the degree of worry about possible COVID-19 infection in the respondent and how this concern affects their state of mind and their ability to carry out their daily activities. It was developed in a non-probabilistic sample of adults between 18 and 50 years of age from Lima and Callao, Peru, and it proved to be unidimensional, to have adequate content validity, internal consistency (ω coefficient = 0.90), and convergent validity with other mental health covariates.¹² A higher total score on each scale means a higher respondent’s outcome level. Then, we dichotomized the total scores for each outcome and considered the upper quartile as a high outcome level.

The general anxiety outcome was measured with the Spanish version of the Generalized Anxiety Disorder (GAD-2), which contains two items that are answered on a Likert scale from 0 (never) to 3 (almost every day).⁴⁸ A total score of two or more points can diagnose clinically relevant anxiety in older adults, with 67% sensitivity and 90% specificity.⁴⁹ The general depression outcome was measured with the Spanish version of the Patient Health Questionnaire (PHQ-2), which contains two items that are answered on a Likert scale from 0 (no day) to 3 (almost every day).⁵⁰ A total score of three or more points can diagnose clinically relevant depression in older adults without cognitive impairment with 79% sensitivity and 82% specificity.⁵¹

Exposure variable

We asked about self-reported COVID-19 vaccination status (no dose, only first dose, and two doses of vaccine) and the time in days since they received each dose of the vaccine.

Covariables

In addition, we asked about sociodemographic variables, previous mental health diagnosis and treatment, and personal and family history of COVID-19. Also, we asked about the preference between BNT162b2 (BioNTech, Pfizer), ChAdOx1-S (Oxford, AstraZeneca), and BBIBP-CorV (Sinopharm) vaccine. Additionally, we assessed the comorbidity with the Geriatric Comorbidity Index. This index measures the severity degree of 15 clinical conditions, classifying them from 0 to 4 each (0: no disease, 1: asymptomatic disease, 2: asymptomatic disease with treatment, 3: uncontrolled disease despite treatment, and 4: very serious or life-threatening disease). After presenting the severity degree classification and giving simple and standardized examples about each clinical condition, the interviewers asked the responders to identify their current situation for each clinical condition. According to these scores, the degree of comorbidity was grouped into classes: without comorbidity (all conditions absent), class I (one or more conditions with a severity degree of 1 or less), class II (one or more conditions with a severity degree of 2), class III (one condition with a severity degree of 3), and class IV (two or more conditions with a severity degree of 3, or at least one condition with a severity degree of 4).⁵²

Data collection

We conducted a pilot with trained interviewers for data collection for one day using a random sample from the sampling frame of approximately 120 participants. During the pilot, we evaluated the data collection capacity of the interviewers and the availability to participate of the selected older adults. Then, we identified and corrected deficiencies for formal data collection.

After sampling, we assign an identification code to each selected older adult to facilitate recognition and monitoring within the program. Then, the interviewers contacted the selected participants through telephone calls, using the telephone numbers registered in the CAM database. Previously trained interviewers made the calls and collected data. In case of not answering two calls on two different days, the older adult was excluded from the study. During the call, the interviewers identified the older adult by asking them for their identity document number. Then, the older adult was invited to participate in the study by requesting their verbal informed consent. If the older adult agreed to participate in the survey, the interviewer asked about the selection criteria. Then, the interviewers collected the baseline data. The data collection interview took approximately 20 minutes.

The follow-up calls were made between July 1st and July 27th, 2021, considering the date of the baseline interview. During these calls, the interviewers followed the same procedure mentioned above. Again, we considered a lost record if the older adult did not answer the call twice on two days. Emotional health outcomes were asked directly using the questionnaires, but in a different order than before, with items in different places. At the end of the interview, they asked about vaccination against COVID-19.

Trained interviewers registered the collected information through the Google Form platform (Questionnaire in Spanish and English in the extended data: Supplementary methods³⁹), using the identification code of each participant. The principal investigator monitored this database every two days, looking for errors during data collection. In case of suspecting a wrong registration, we coordinated with the responsible interviewer to evaluate the need to re-register said entry.

Data analysis

All the information collected was automatically recorded in a Microsoft Excel 2021 sheet (Microsoft, WA, United States). Before the analysis, we joined the baseline database with the follow-up database considering the identification code of each older adult interviewed. We reviewed the database for inconsistencies in responses about vaccination. We considered the response in the baseline measurement as the valid one if it was inconsistent with the responses during the follow-up measurement. Then, we performed the descriptive analysis of the results in the total number of recruits and separately according to the vaccination status against COVID-19. We describe the relative and absolute frequencies of the qualitative variables and the mean ± standard deviation of the quantitative variables. The baseline prevalence of altered emotional health outcomes and their 95% confidence intervals (95% CI) were plotted for the total sample and separated according to COVID-19 vaccination status. We calculated the 95% CI with the logit adjust method for the design degrees of freedom.⁵³

We compared the frequency of altered emotional health outcomes with the Chi-squared test.

We assessed, in the baseline, the association between the vaccination situation and the outcomes in emotional health. We estimated the odds ratios (aOR) and 95% CI for having adjusted altered emotional health outcomes by sex, age, time in days since receiving the last dose of vaccine (unvaccinated were assigned with zero), living with someone, comorbidity, vaccine preference, history of emotional health, and history of COVID-19, using a logistic regression model.

Finally, we assessed the association between vaccination status and altered emotional health outcomes, performing multilevel logistic regression models with mixed effects with a three-level structure. In addition to the first individual level of each measurement, we used random intercepts for the healthcare network and each individual as the other two levels. This analysis allows us to calculate the effects considering the stratum of the care network and the correlations of the two responses over time within the same individual, thus allowing a longitudinal and stratified assessment. In addition, we adjusted the regression model for the confounding variables: time in days since receiving their last dose of vaccine, sex, age, living with someone, comorbidity, vaccine preference, history of mental health disease, history of COVID-19, and time in days since the basal measurement (The basal measurement records had a value of 0).

In all the previously mentioned analyses, we consider the study’s sample weights and the stratum using the svyset command. A value of p < 0.05 was considered statistically significant to reject the null hypothesis in all statistical tests. Statistical software STATA MP v17 (StataCorp, Texas, USA) was used for the analysis.

Ethics

Each participant gave verbal informed consent before being included in the study. The anonymity of the interviewees was always maintained, assigning each one an identification code. Therefore, the survey didn’t collect personally identifiable information. In cases where a participant had an acute event in her mental health, the interviewer immediately referred the participant to a psychiatrist free of charge who managed the event by telephone. The protocol is registered in the PRISA repository of the Peruvian National Institute of Health (ID code: EI00000001999), and it was approved by the Institutional Review Board of the Instituto Nacional del Corazón - EsSalud (Certificate of approval 25/2021-CEI).

Recruitment and baseline measurement

We randomly selected and invited 1,686 older adults to the study. A total of 51.1% (n = 861) of them met the selection criteria for baseline measurement. Among the reasons for not participating in the study were not responding to the call (n = 695), having been diagnosed with COVID-19 in the last three months (n = 57), refusing to participate (n = 51), not wanting to be vaccinated (n = 11), have symptoms related to COVID-19 at the time of the interview (n = 8) and have been vaccinated in another country (n = 3). Then, 20.8% (n = 179) refused to participate or did not respond to the call for follow-up measurement (Figure 2). The frequency distribution of gender and mean age were similar between those who did not participate in the baseline measurement or during the follow-up measurement compared to the recruited patients (Table S1 in the underlying data: Supplementary results³⁹).

Figure 2. Participant’s selection flowchart.

The main characteristics of older adults aged 60 to 79 years affiliated with EsSalud’s CAMs are in Table 1. 54.5% of widowers and divorcees older adults had been in that status for less than a year. Since the pandemic’s start, 20.8% have been hospitalized among those ever diagnosed with COVID-19. On the other hand, among older adults who had at least one family member, whom they live, with a diagnosis of COVID-19, since the start of the pandemic, 17.2% of them had at least one family member who died due to COVID-19. We observed that the age, geriatric comorbidity index, and previous COVID-19 personally or family diagnosis were associated with the vaccination status.

Regarding the vaccination situation, at the baseline measurement, 43.9% of the respondents received only one dose with an average time of 15.6 days from the date they received the vaccine, and 49.1% received the two doses with an average time of 16.9 days from the date they received the vaccine. On the other hand, 5.4% of older adults reported having a previous diagnosis of a mental health disorder. However, we found the prevalence of general anxiety, assessed with GAD-2, to be 16.4% (95% CI: 14.1 to 19.1), and the prevalence of general depression, estimated by PHQ-2, was 8.0% (95% CI: 6.3 to 10.0). In addition, we found that those who had two doses of the vaccine had less likely to have fear of COVID-19 (p<0.001), anxiety about COVID-19 (p<0.001), worry about COVID-19 (p<0.001), and general anxiety (p<0.001) compared to those who were unvaccinated or those who had one dose of the vaccine. However, we didn’t observe a trend in the case of the outcome of general depression (p=0.099) (Figure 3).

Figure 3. Baseline prevalence of emotional health outcomes in total and according to vaccination status in older adults recruited from May 27th to June 30th, 2021 (n = 861).

Association between vaccination status and emotional health outcomes at baseline

During the baseline measurement, we observed that older adults who had two doses of the vaccine had less fear of COVID-19 (aOR: 0.27; 95% CI: 0.13 to 0.56) than those who were unvaccinated. Similarly, we observed a lower anxiety about COVID-19, worry about COVID-19, and general anxiety in those who had two doses of the vaccine, but without statistical significance. Meanwhile, those who had one or two doses of the vaccine were more likely to have general depression than those who were unvaccinated, although this was not statistically significant. However, those who had two doses of the vaccine were significantly less likely to had fear of COVID-19, anxiety about COVID-19, worry about COVID-19, and general anxiety, compared to those with only one dose (Table 2).

Association between vaccination and mental health problems at one month of follow-up

The mean follow-up time for 661 older adults was 31.4 ± 0.14 days. Considering the one-month follow-up period, we observed that older adults with two doses of the COVID-19 vaccine had less fear of COVID-19 (aOR: 0.19; 95% CI: 0.07 to 0.53) and less anxiety about COVID-19 (aOR: 0.45; 95% CI: 0.22 to 0.89), compared to those who were unvaccinated. In addition, those with two doses of the vaccine were significantly less likely to had fear of COVID-19, anxiety about COVID-19, and worry about COVID-19, compared to those who had only one dose. We observed similar results in the outcomes of worry about COVID-19 and general anxiety; however, there is high uncertainty about these estimates (Table 3).

Summary of results

We hypothesize that vaccination has a causal effect in reducing fear, anxiety, and worry about COVID-19 and, also in general anxiety and depression. Our study, conducted in a cohort from a nationally representative sample of older adults affiliated to EsSalud, partially confirmed our hypothesis. We found evidence that those older adults with two doses of the COVID-19 vaccine, compared with unvaccinated and with only one dose, had less likelihood of high levels of fear and anxiety about COVID-19. However, we were unable to confirm these findings for any outcome in those who had received only one dose compared with unvaccinated. To our knowledge, this is the first study that assesses the effect of vaccination on emotional health in a representative sample of older adults, using novel COVID-19 perception outcomes.

Effect of vaccination against COVID-19 in the emotional health of older adults.

High fear and anxiety about COVID-19 were significantly less likely in those who had two doses of the COVID-19 vaccine than those who were unvaccinated or had only one dose. Interestingly, the scales that measured both constructs (FCV-19S and CAS, respectively) focus mainly on the emotional and physical reaction to thoughts related to COVID-19 and its possible contagion.^54,55 However, the PRE-COVID-19 scale, which measured the worry about COVID-19, focused on daily dysfunction caused by thoughts about the possibility of getting COVID-19.¹² This difference is relevant, as it would mean that vaccination could affect older adults by improving their mental well-being and reducing more intense psychosomatic symptoms of stress related to the pandemic (fear/anxiety of COVID-19)⁵⁶; but without reducing daily thoughts and behaviors associated with the possibility of contagion (COVID-19 concern).

On the other hand, although we observed a slight decrease in general anxiety, the effect of vaccination on general depression and anxiety outcomes, measured with PHQ-2 and GAD-2, respectively, was not significant. However, previous studies conducted in adults from the United States,²⁴ Turkey,⁵⁷ Argentina,⁵⁸ and China,⁵⁹ and health professionals from Turkey²⁷ reported that those with at least one dose of the vaccine against COVID-19 have lower scores on the depression scales, measured with the PHQ-4, PHQ-9, and the Beck Depression Inventory, and on anxiety scales, measured with the GAD-7 and the Beck Anxiety Inventory. The mechanisms and causes of depression and anxiety in older adults are related to psychosocial factors of loneliness and loss, and neuroendocrine and vascular disorders.^60,61 Previous mentioned studies included adults in general with low representation of older adults, so, according to our results, the effect of vaccination would not be sufficient to significantly reduce these outcomes in emotional health in older adults, since they would respond to other intrinsic and extrinsic factors that were not measured in the present study.

On the other hand, the evaluation of the effect of vaccination against COVID-19 on emotional health could be affected by the perception of the vaccine’s effectiveness or worries about adverse events.⁶² Older adults are particularly susceptible to fake news or misinformation,⁶³ which could influence their perception of vaccination against COVID-19 and diminish its effect on their mental health. This effect should be evaluated in future studies. Similarly, the perception of effectiveness and vaccination intention could also be affected by the type of vaccine manufacturer, with the BNT162b2 vaccine (BioNTech, Pfizer) being the most preferred, and the ChAdOx1-S vaccine (Oxford, AstraZeneca) the least preferred in developed countries.⁶⁴ Thus, even though most older adults were vaccinated with the BNT162b2 vaccine (BioNTech, Pfizer), the political scandal in Peru regarding the BBIBP-CorV (Sinopharm) vaccine³³ may have partially affected the effect of vaccination on the emotional health of this population.

Second dose of vaccine against COVID-19

We observed the effect on emotional health from the second dose and not in those who had only one dose of the vaccine. This result is different from previous studies where all those who had at least one dose were included in the vaccinated group, regardless of whether they had both doses or not.^24,27,57,59 In addition, one study from Sweden report lower anxiety and depressive symptoms after a short-time period after first and second dose.⁶⁵ Clinical effectiveness studies have shown the need for a second dose of the vaccine to have greater effectiveness in preventing mortality and severe disease from COVID-19 in older adults.^15,66 This information was communicated promptly to the population, making most people aware of the need for a second dose, especially those willing to be vaccinated.^67,68 This may explain that in the context where the communicational emphasis was placed on the need for the second dose of the COVID-19 vaccine, the effect on emotional health could mainly be observed in those who received two doses of the vaccine. Considering the high proportion of older adults who have had two doses of the vaccine,⁶⁹ the effect of vaccination reported in other studies may have been carried by those who had both doses, compared to those who had only one dose of the vaccine against COVID-19.

However, the presence of new variants of concern, such as B.1.1.529, could affect the population’s mental health^70,71 due to their greater infectivity and immune escape from vaccination.⁷² Given this, the need for a third,⁷³ or even a fourth,⁷⁴ COVID-19 vaccine booster is currently under discussion. So, considering that the perception of the vaccine’s effectiveness correlates with the level of concern about the new variants,⁷⁵ it is important to continue monitoring mental health in the most vulnerable populations such as the elderly, and its evolution during future vaccination policies against COVID-19.

Public health relevance

As of January 2022, the two-dose vaccination rate in adults aged 60 years and older in different countries was around 80% (https://ourworldindata.org/grapher/covid-fully-vaccinated-by-age?country). Among the reasons for older adults to decide to be vaccinated is the fear of developing the disease and the perception of the vaccine’s effectiveness to prevent the disease.^76,77 However, the lack of reliable information, the fear of possible adverse effects, and the limited access to receive the vaccines mean that many older adults do not get vaccinated or do not have the opportunity to get vaccinated.⁷⁶ In this sense, the communication strategy to promote vaccination against COVID-19 could be complemented with the message of reducing fear and anxiety about being infected with COVID-19. Thus, integrating with other elements necessary to have an adequate vaccination rate, such as the empowerment of the first level of care and the availability and access to vaccines,⁷⁸ it could improve citizen confidence in the vaccination process.

Currently, the COVID-19 pandemic has a significant impact on mental health, which will continue in the medium and long term.⁷⁹ From the point of view of positive epidemiology, our results propose vaccination against COVID-19 as a positive determinant of mental health in older adults.²⁹ So, vaccination against COVID-19 could contribute to the partial improvement of the emotional health of older adults. However, we must consider that the mental health of older adults depends on various intrinsic and extrinsic factors that not only respond to the COVID-19 pandemic.^60,61

Limitations and strengths

The interpretation of the results of this study must consider the following limitations. First, we were unable to obtain the planned sample size for the primary analysis, for the unvaccinated group. This made the statistical power of our results insufficient to find statistically significant results. Thus, we do not rule out the effect of the vaccine on worry about COVID-19 and general anxiety, which should be evaluated in future studies. Second, the low representativeness of the older adults affiliated with EsSalud registered in the CAM database, the high refusal to participate in the study, and the loss during follow-up could have caused selection bias. Thus, it is likely that older adults in CAMs have greater access to receiving the vaccine and to activities that improve their mental health, so the effect that we measured in the study may be overestimated. Third, the way to determine the vaccination status was by self-report, so the measurement of this variable could have been overestimated due to the social desirability bias. Fourth, even though we did not conduct clinical interviews to evaluate the emotional health in the present study, we used different specific psychometric instruments for the perception of COVID-19 and general anxiety and depression. These tools have robust evidence of psychometric validity in our population of interest, making the constructs that we measured reliable. Fifthly, we used short and general questionnaires to measure anxiety and depressive symptoms (i.e. PHQ-2 and GAD-2), which are not specific to vaccination settings. Therefore, it is possible that significant differences would have been found if instruments specifically designed for these settings had been used. Similarly, it is possible that the use of more comprehensive versions measuring depressive and anxiety symptoms, such as the PHQ-9 and GAD-7, which include emotional and somatic indicators, might have increased the variability of the measures and found significant results. However, we believe that using the PHQ-2 and GAD-2 captures the core symptoms of anxiety and depression. Sixth, it is possible that the change in outcomes such as anxiety or depressive symptoms in older adults was not solely dependent on COVID-19 vaccination. Social determinants of health or other factors may be more influential and not accounted for in our study (i.e., family support, economic status, or quality of life). Therefore, we invite other researchers to design future studies with greater methodological rigour, taking into account the considerations mentioned in our limitations.

Underlying data

Figshare: Underlying data for ‘Effects of vaccination against COVID-19 on the emotional health of older adults’, https://doi.org/10.6084/m9.figshare.20134994.⁸⁰

This project contains the following underlying data:

Data file 1: VacMentHe_DataBase.xlsx (The database has been anonymized and it has not distorted the scientific meaning.)

Extended data

Figshare: Extended data for ‘Effects of vaccination against COVID-19 on the emotional health of older adults’, https://doi.org/10.6084/m9.figshare.20135000.³⁹

This project contains the following extended data: