Hesitancy to Return to the Pre-pandemic Routine.

The goal of the present study was to explore COVID-19 related hesitancy, which represents the inability of people to return to previous levels of functioning after a major medical crisis like the current pandemic. A new questionnaire was developed to evaluate participants' hesitancy. The study was conducted online in November, 2020, using convenience sampling. A total of 538 individuals from the general Greek population participated in the study and completed the Hesitancy Questionnaire, the Generalized Anxiety Disorder 2-item (GAD-2), the Short Health Anxiety Inventory and a COVID-19-related worry question. The Hesitancy Questionnaire proved to have adequate psychometric properties. Correlation with anxiety as assessed by GAD-2 proved to be significant but low, indicating that the two scales are measuring two different concepts. The greatest hesitancy was observed in older adults for both genders (males, M = 40.86, SD = 15.24; females, M = 49.34, SD = 14.74). Women in general appeared more hesitant than men scoring higher (males, M = 36.13, SD = 15.25; females, M = 42.63, SD = 17.31) with a statistically significant difference [t(536) = - 3.706, p = .001). This study provided a tool to informed understanding on how citizens perceive the new normality after the COVID-19 pandemic. If not appropriately addressed, hesitancy may increase stress levels and result in mental health or socialization problems.

Introduction

Researchers all over the world are exploring the immediate aftermath of the COVID-19 pandemic on mental health. Fear, insomnia, anxiety, depression, and posttraumatic symptomatology are well investigated topics (Blekas et al., 2020; Skapinakis et al., 2020; Tsipropoulou et al., 2020; Voitsidis et al., 2020a). As new measures are introduced in several countries, the levels of loneliness (Parlapani et al., 2020a), harmful alcohol and drug use (Panagiotidis et al., 2020), and self-harm or suicidality (Hawton et al., 2020) are also expected to rise.

In Greece, two national lockdowns were imposed by the government, as the COVID-19 pandemic is still ongoing. The prolonged medical crisis affected both individually and collectively the citizens of Greece. At the first wave, adherence of young adults to the measures was satisfactory (Parlapani et al., 2020b; Sakka et al., 2020), whereas, after lifting the lockdown measures, just before the summer months, adherence was questionable. A significant amount of confirmed cases and deaths, resulting to a fatality rate of 1.9% in November, further increased fear and uncertainty. At a social level, polarization between those who were attentive and those who were careless started to grow, as continuous reminders of the magnitude of the health crisis were projected across the media span (Fountoulakis et al., 2020). In addition, worry about the financial aftermath was significant among professionals in the private sector, whereas the public sector had minimal or no pay cuts (Economides et al., 2020).

Studying the public health response and its consequences at the individual level (Blekas et al., 2020; Parlapani et al., 2020b; Voitsidis et al., 2020b), it became clear that prolonged precariousness during this pandemic made people doubtful about what is correct and what is not, as well as more hesitant to return to previous levels of carefreeness.

To hesitate means to express uncertainty and to hold back from doing or saying something in doubt (Merriam-Webster, n.d.). In that sense, hesitancy differs from fear and anxiety because it addresses decision making. Making decisions can be hard even during ordinary periods, more so when fear or emotions are involved. Under the pressure of the current pandemic, the fear of making the wrong decision jeopardizes the basic cognitive schema of safety. Uncertainties about the accuracy of available diagnostic tests, the natural history of the disease, the effects of treatment and the healthcare efficiency create an unstable environment where it is difficult to comprehend all the elements and reach a conclusion (Hunink et al., 2014).

This fragile logical structure, with missing or incomplete elements, leads to cognitive dissonance. Cognitive dissonance is the psychological discomfort people experience when holding two or more cognitions that are psychologically inconsistent (Festinger, 1957). People adopted new habits under the imposed lockdowns to retain their safety. They have learned that being isolated and distant from others may keep them safe and healthy. The cognitive dissonance is obvious: while the first goal is the desire to socialize and return to normality, the second is the desire to stay healthy, leaving citizens hesitant to decide to snap back to work, go out, or travel. In the global COVID-19 literature, hesitancy was mainly studied to address the reluctance of some people to get vaccinated (vaccine hesitancy) (Dror et al., 2020; Dubé et al., 2013; Puri et al., 2020), but from clinical experience, it is apparent that it also influences other significant aspects of social life or decisions associated with seeking medical advice (e.g. people’s avoidance of visiting health care facilities, despite having symptoms, due to fear of contracting the virus).

In that context, the goal of the present study was to explore COVID-19-related hesitancy, which represents the inability of people to return to previous levels of functioning after a major medical crisis like the current pandemic. To achieve the goal, a new tool was created, the “Hesitancy Questionnaire” (HeQ), a self-reported questionnaire that targets severity of hesitancy. Consequently, the following objectives were addressed: (a) to provide psychometric evidence for its internal structure; (b) to provide evidence for its internal consistency; (c) to evaluate the strength of evidence from the sample and (d) to provide evidence of concurrent validity.

Methods

Development of the HeQ

At the beginning of the questionnaire construction a thorough literature review was conducted by the first author, to check if a hesitancy questionnaire had already been published, resulting to papers mainly addressing vaccine hesitancy reporting on tools identifying vaccine hesitant persons (Cella et al., 2020; Shapiro et al., 2018) and one addressing social support seeking (Gontkovsky et al., 2021). Then, a brainstorming session was carried out with the rest of the authors, where the purpose was to discuss the impact of the lockdown measures and recommend as many questions as possible. Finally, in a focused group discussion, an expert panel of five members (two clinical psychologists, one health psychologist and two psychiatrists) evaluated all the questions proposed to be included in the questionnaire, with respect to face validity, content validity, and comprehensiveness. They rated the meaning and linguistic clarity towards instructions, items and response format. A dichotomous response choice (adequate–inadequate) was used by the raters to assess clarity and content. Items that received “inadequate” by three reviewers were excluded (5 items due to lack of linguistic clarity, and 10 items due to content vagueness). Items that received “inadequate” by one or two reviewers were reassessed before final decision of exclusion or inclusion (3 items excluded). Items with similar content were merged (from 16 to 8 items). Items with ambiguous wording were rephrased. Eligible items were selected by full agreement among the panel members. Recommended changes were applied to the final version by the research team. Out of the 40 original questions, 14 were selected and formed the questionnaire to be tested in the pilot study.

Pilot Study

A pilot study was performed resulting to data from 30 participants (16 females; 14 males; Mage = 34.32 years, SD = 7.53) who completed the questionnaire twice, three weeks apart. Respondents were asked to rank their hesitancy to return back to common activities on a sliding scale of 0–10. The domains covering the most common activities were: working, socializing, traveling, health, and planning the future.

Results from the pilot study analysis helped reform the structure of the questionnaire and highlighted its reproducibility, as a test–retest correlation of 0.89 was obtained. Items with high correlations were merged. Specifically, three items (i.e. “I hesitate to visit crowded shops”; “I hesitate to visit malls”; and “I hesitate to participate in social events”) were merged to item 4 (i.e. “I hesitate to visit crowded shops, malls or social events”). Furthermore, two items (i.e. “I hesitate to travel abroad”; and “I hesitate to travel domestically” were merged to item 12 (i.e. “I hesitate to travel”). Two items that were not relevant to all participants (i.e. “I hesitate to send my children to school”; “I hesitate to visit a playground with my children”) were removed. Thus, the second draft of the HeQ consisted of 9 items.

To facilitate the cross-cultural use of the HeQ it was decided to also adopt the questionnaire in the English language. In this direction, one Greek native psychiatrist living in London and one Greek native pathologist living in Boston, with an excellent command of English, who were informed by the authors about the background and objectives of the HeQ, independently carried out forward translation of the HeQ into English. Then, the two translators discussed via a Zoom meeting and integrated the two translated versions into one. Finally, another psychiatrist, fluent in English did a back translation to Greek. The back-translated version was checked and approved by the authors of the original version. After the final proofreading, construction of the HeQ-English version was completed (“Appendix”).

Main Study

The study was conducted online in November, 2020, using convenience sampling. A total of 538 individuals from the general Greek population completed the study. All participants were at least 18 years old and fluent in Greek according to the inclusion criteria. Only the participants who provided complete values on the self-report measures were included in the analysis; missing values on the demographic characteristics were not considered an exclusion criterion. All participants were asked to complete the questionnaires taking into account the COVID-19 health crisis. Ethical approval was granted from the Papageorgiou General Hospital Scientific Board (563/2020).

Measures

A set of questions was used to obtain socio-demographic characteristics including age, gender, educational level, region, and residential area. In addition, the study included the following scales:

Generalized Anxiety Disorder 2-Item (GAD-2; Kroenke et al., 2007; Skapinakis, 2007): GAD-2 is a very brief self-report scale for assessing core anxiety symptoms during a specified period of time. The scale consists of two-items rated on a four-level Likert-like scale (“0 = not at all to 3 = nearly every day”). It is a widely used screening tool in both clinical practice and academic research. Probable cases of GAD are detected using a cutoff score of 3 or greater with good sensitivity and specificity (Plummer et al., 2016). In this study GAD-2 demonstrated an overall Cronbach’s alpha of 0.80.

Short Health Anxiety Inventory (SHAI; Leonidou & Panayiotou, 2016; Salkovskis et al., 2002): SHAI is a self-rated questionnaire widely used to assess levels of health concern and is also applicable for the identification of severe health anxiety or hypochondriasis. It contains 14 multiple-choice items related to health worries and awareness of ordinary and unusual bodily sensations, as well as a treatment process measure of four additional items that were not applicable to the present study. Each item includes four possible choice statements coded from 0 to 3. Higher mean scores signify greater severity of health anxiety. In this study the 14 items of SHAI demonstrated an overall Cronbach’s alpha of 0.86.

COVID-19-Related Worry was investigated using a single-item question i.e. “I worry a lot about coronavirus-19” scored on a 10-point Likert-type scale. The higher score was considered indicative of elevated worry toward COVID-19. Single item measures, beyond their simplicity and time-effectiveness, provide reliable and valid information and have been successfully used for assessing anxiety, depression (Turon et al., 2019), health status and overall quality of life (Bowling, 2005). The question selection was based on previous research evaluations regarding the perceived negativity during the pandemic (Blekas et al., 2020).

Data Analysis

Descriptive statistics were used to report the sample characteristics. Internal consistency was assessed by Cronbach alpha coefficients (α), inter-item correlations and corrected item-total correlations. Concurrent validity was assessed by comparing the Pearson correlations between the HeQ and GAD-2. Exploratory analysis was used to explore the factorial structure. Confirmatory Factor Analysis (CFA) with maximum-likelihood (ML) estimation was performed to further assess the goodness of fit of the factor structure. Model fit was assessed using the following indices: Satorra–Bentler scaled χ2 (SBχ2), robust versions of Comparative Fit Index (CFI), Bentler–Bonnett Non-Normed Fit Index (NNFI), Root Mean Square Error of Approximation (RMSEA) and Standardized Root Mean Square Residual (SRMR). Data were analyzed using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, N.Y., USA) and CFA was conducted by SPSS-AMOS v.23 (Arbuckle, 2014).

Results

The total sample (n = 538) consisted of 419 females (77.9%) and 119 females (22.1%). The age of the respondents ranged from 18 to 75 years (M = 34.63, SD = 4.20). More details on demographics are presented in Table 1.

Table 1

Demographic characteristics

Participants’ characteristics	Frequency	%
Gender
Male	119	22.1
Female	419	77.9
Age
18–30	130	24.2
31–45	186	34.6
46–60	184	34.2
61–75	38	7.1
Education
High school	4	.8
Lyceum	52	9.7
2-year graduate school	65	12.1
University	208	38.7
MSc	160	29.7
PhD	49	9.1
Relationship status
Married	285	53.0
Serious relationship	79	14.7
Single	119	22.1
Divorced/separated	45	8.4
Widower	10	1.9
Region
Aegean Islands	15	2.8
Crete	117	21.7
Epirus	24	4.5
Ionian Islands	19	3.5
Central Macedonia	159	29.6
Outside Greece	19	3.5
Peloponnese	6	1.1
Sterea Ellada	121	22.5
Thessaly	34	6.3
Thrace	24	4.5
Residence area
Urban area	476	88.5
Small city	30	5.6
Rural area	32	5.9

Factor Analysis

Data were randomly split into two subsamples of equal size to explore the structure of HeQ. Item analysis was carried out to provide evidence of reliability estimates. Principal Component Analysis (PCA) was performed on the first subsample (validation sample, n = 269) to identify the underlying factor structure of HeQ. Kaiser–Meyer–Olkin (KMO) and Bartlett’s Sphericity (BTS) tests showed that data were suitable for factors analysis [KMO = 0.935; BTS: χ2 (91) = 2209.13; p < 0.001]. All variables had at least one correlation coefficient greater than 0.30 and appeared suitable for factor analysis. The value of the determinant for the correlation matrix was 0.0066, indicating that there was no multicollinearity in the data. The participant to item ratio for this analysis was approximately 26 to 1, where sample size was 269 and the number of variables included was 10. This indicated that the given sample size was sufficient to produce reliable results. Based on the scree plot and the Kaiser criterion for determining the number of significant factors, a unifactorial model was selected.

PCA showed that only one factor had eigenvalue greater than 1.00. The extracted factor accounted for 59.39% of the total variance with an eigenvalue of 5.94. The analysis yielded strong factor loadings, ranging from 0.60 (item 18) to 0.84 (item 4). The final items along with item means, standard deviations, PCA and CFA loadings are presented in Table 2. The highest mean pointing to highest hesitancy were found in items 3 (“I hesitate to shake hands or hug people”), 10 (“I hesitate to participate in indoor events”) and 13 (“I hesitate to travel”).

Table 2

Hesitancy questionnaire, descriptives, PCA and CFA loadings

Item	M	SD	h2	riF	h2	riF
1. I hesitate to go to work	3.60	2.78	0.55	0.67	0.47	0.73
2. I hesitate to shake hands or hug people	6.70	2.90	0.51	0.69	0.54	0.66
3. I hesitate to meet people who belong to vulnerable groups	6.61	2.96	0.58	0.65	0.62	0.69
4. I hesitate to visit crowded shops, malls or social events	5.06	3.10	0.69	0.84	0.61	0.79
10. I hesitate to participate in indoor events	6.20	3.23	0.71	0.73	0.64	0.76
12. I hesitate to travel	7.28	1.99	0.55	0.76	0.49	0.71
13. I hesitate to plan for future activities or social events (time off, doctor’s visits, etc.)	5.48	3.00	0.49	0.72	0.51	0.69
17. I hesitate to participate in meetings with people I don’t know	5.13	3.18	0.61	0.75	0.70	0.70
18. I hesitate to spend money	4.17	2.75	0.42	0.60	0.41	0.69

M mean, SD standard deviation, h explained variance, r factor loadings, PCA principal component analysis, CFA confirmatory factor analysis

Confirmatory Factor Analysis

CFA was run to confirm whether the factor structure emerged from the PCA replicated on the second subsample (calibration sample, n = 269). Although the results of the Chi-square goodness of fit test were significant, χ2(20) = 51.87, p < 0.001, the RMSEA index [RMSEA = 0.07, 90% CI = (0.04, 0.09)], the CFI (CFI = 0.97), the TLI (TLI = 0.96), and the SRMR (SRMR = 0.04), were all indicative of a good model fit. CFA loadings are presented in Table 2. Based on the results of PCA and CFA, a unifactorial model was elected for HeQ.

Cronbach’s alpha coefficient of the scale was α = 0.89, indicating good reliability. Results of item analysis are presented in Table 3. Inter-correlations of the scale ranged from r = 0.290 to r = 0.676, and were positive and statistically significant (p < 0.001).

Table 3

Item-total statistics

	Scale mean if item deleted	Scale variance if item deleted	Corrected item-total correlation	Cronbach’s alpha if item deleted
1. I hesitate to go to work	43.55	315.980	.540	.87
2. I hesitate to shake hands or hug people	41.75	295.477	.675	.86
3. I hesitate to meet people who belong to vulnerable groups	40.51	310.878	.567	.87
4. I hesitate to visit crowded shops, malls or social events	42.26	291.105	.712	.85
10. I hesitate to participate in indoor events	41.49	295.430	.679	.86
12. I hesitate to travel	40.10	304.386	.610	.86
13. I hesitate to plan for future activities or social events (time off, doctor’s visits, etc.)	41.89	297.667	.672	.86
17. I hesitate to participate in meetings with people I don’t know	42.43	285.448	.781	.85
18. I hesitate to spend money	42.94	332.336	.568	.88

Concurrent validity was supported by the significant correlation with GAD-2 (r = 0.244, p < 0.001). Significant correlations were also presented after the correlation with COVID-related worry (r = 0.623, p < 0.001) and SHAI (r = 0.428, p < 0.001). Discriminant validity was explored by examining the mean differences between participants with low anxiety and high anxiety. Participants with high anxiety and high health anxiety symptomatology had statistically significant higher scores in hesitancy than those who reported low anxiety or health anxiety (Table 4).

Table 4

Hesitancy means and standard deviations among anxious and non-anxious participants

	HeQ mean	SD	Statistic, p value
Low anxiety (n = 430)	47.54	17.40	t(536) = 4.39, p = .001
High anxiety (n = 108)	69.60	16.64
Low health anxiety (n = 516)	40.68	16.85	t(536) = 3.43, p = .001
High health anxiety (n = 22)	53.31	18.24

Finally, a general linear model was run to explore the associations between the basic variables. The significance values of the main effects, for the variables “gender” (Male, B = − 3.589, p = 0.011), “COVID-19-related worry” (B = 3.465, p = 0.001) and “SHAI total score” (B = 0.426, p = 0.011), were less than 0.05, indicating that their effects contributed to the model, with partial η2 indicating that the effect of “COVID-19-related worry” accounted for a greater amount of variation  (Table 5).

Table 5

General linear model for hesitancy

Source	Type III sum of squares	df	Mean Square	F	p value	Partial η2
Corrected model	66,565.419a	13	5120.417	29.779	.000	.426
Intercept	4768.245	1	4768.245	27.731	.000	.050
Gender	1140.007	1	1140.007	6.630	.010	.013
Age	474.132	3	158.044	.919	.431	.005
Education	2019.962	6	336.660	1.958	.070	.022
COVID-19-related worry	34,465.315	1	34,465.315	200.441	.000	.277
GAD-2 total	50.435	1	50.435	.293	.588	.001
SHAI total	3117.289	1	3117.289	18.129	.000	.034
Error	89,756.579	522	171.947
Total	1,068,277.000	536
Corrected Total	156,321.998	535

aR2 = .426 (Adjusted R2 = .412)

The greatest hesitancy was observed in older adults for both genders (males, M = 40.86, SD = 15.24; females, M = 49.34, SD = 14.74). Women in general appeared more hesitant than men scoring higher (males, M = 36.13, SD = 15.25; females, M = 42.63, SD = 17.31) with a statistically significant difference [t(536) = − 3.706, p = 0.001).

Discussion

The HeQ proved to have adequate psychometric properties. The final version contains 9 items and has a unifactorial structure in the current sample. Total scores are calculated by averaging the scores of all items; higher scores indicate greater hesitancy (“Appendix”). Correlation with anxiety as assessed by GAD-2 proved to be significant but low, implying that the two scales measure two different concepts. Significant correlations were also presented with COVID-19-related worry and SHAI, suggesting that hesitancy during the current period is influenced by health concerns and not general anxiety.

More hesitant participants proved to be females and older adults. In accordance with relevant research, women over-report symptoms during health crises (Mayor, 2015). In this particular health crisis, women are usually responsible for caring for the elderly and the children, something that increases their burden (Power, 2020). As expected, older adults were more hesitant, as greater age is associated with a higher prevalence of underlying risk factors for severe COVID-19 disease and increased mortality (Ioannidis et al, 2020; Parlapani et al., 2020a, 2020b).

This new tool could add to the COVID-19 research field as it is crucial to develop structured and validated instruments to address the psychosocial impact of the pandemic and to contribute to the assessment and management of hesitant citizens (Cortez et al., 2020; Das et al., 2020). Any identified concerns need more in-depth understanding of the nature of hesitancy as well as of the personality traits of those who are hesitating (Larson et al., 2015).

As a result of the current medical crisis, people formed new habits, adhering to several measures to protect themselves and their loved ones (Mertens et al., 2020). They also had to adapt to new working environments and new forms of communication. While this new reality could lead to resilience and post-crisis growth (Polizzi et al., 2020; Tamiolaki & Kalaitzaki, 2020), there are some people that appear to be stuck in their indecisiveness and hesitancy.

Indecision concerns the rational dialectic, whereas hesitation concerns the psychosocial dialectic. Cognitive dissonance -the basis of hesitancy- forces us to justify a choice and dismiss any alternative, as an effort to alleviate the tension caused by indecisiveness (Festinger, 1957). Internal factors like personality characteristics, e.g. risk-taker vs risk-averse (Doob, 1990; Lauriola et al., 2005), vulnerability of getting ill from SARS-CoV-2, and preexisting psychopathology, e.g. health anxiety, may contribute to increased hesitancy. External factors, such as the financial situation, or the impact of coronavirus on one’s home region, could also have an effect on hesitancy.

The basis of decision making is the evaluation of data and information (Gold & Shadlen, 2007; Stansfield et al., 2006). During previous pandemics, daily life was not impacted as much as today, since information did not spread instantly and constantly. Misinformation complicates the decision-making process, as it has various sources such as rumors and fiction, politicians, and the media (Lewandowsky et al., 2012). Those are all factors that influence the way people perceive the likelihood and severity of a disease and affect their decision-making process towards their reactions, adherence to the measures and return to pre-pandemic activities.

Even more so, as trying to lessen the discomfort they feel by being hesitant, people tend to interpret any incoming information in a convenient way to support their decision (Garcia-Alamino, 2020), getting trapped in confirmation bias. In the absence of a vaccine and with pending questions hovering over herd immunity, the best way to avoid contracting the virus is by distancing themselves from other people as much as possible. Over-focusing on details about cases and deaths supports the belief that home is the absolute safe place, making it difficult to adapt to other environments. Overconfidence in a belief detracts from reasoning, because logical errors are intuitive and seem reasonable (Digdon, 2020). This underlying mechanism impacts on communication messages, triggering the propagation of misinformation that diffuses from one person to another, resulting either to lack of adherence or to the development of inconsistent theories.

If contributing factors lead to realistic thinking without cognitive distortions, it is easier to overcome hesitancy and develop new strategies to address any concerns, without avoiding behaviors. If this is not the case and contributing factors lead to distorted thinking, then prolonged hesitancy could lead to impaired mental health and anxiety symptoms (Fig. 1).

Fig. 1

COVID-19 hesitancy

This study provided a tool to informed understanding of how citizens perceive the new normality after the COVID-19 pandemic. If not appropriately addressed, hesitancy may increase stress levels and result in mental health or socialization issues.

The HeQ could be of value in mental health screening as it could help identifying the underlying mechanisms of anxiety such as health, social, generalized and non-specific anxiety, conditions characterized by dysfunctional decision-making processes.

Similarly, the study rational could apply to several aspects of health care since hesitancy is intervening in health decision-making in a different manner. For instance, there are patients that are hesitant to disclose medical information (e.g. substance use), visit doctors despite developing symptoms (e.g. urine incontinence), receive a medical regimen or a vaccine, (e.g. COVID-19 vaccine), undertake a medical procedure (e.g. colonoscopy) or even return to normal habits after a medical event (e.g. bone fracture, or coronary artery bypass grafting).

Altogether, the HeQ is an instrument that could be used in its current form as a screening tool considering that it can reveal a common psychological difficulty that may be affecting one’s psychological well-being. Prospectively, disease specific versions could be developed to unveil similar underlying mechanisms impacting physical health. Such versions could contribute to early detection of hesitant individuals and to tailored communication strategies to understand their concerns and address their needs.

Therefore, more research is needed to explore hesitancy and to accordingly provide the means to overcome it, because even when the first vaccines and therapeutics become widely available, no one really knows how long it will be before resuming our social and work life without discomfort.

A basic limitation of the current study was the recruitment of a convenience sample from an online survey. An additional limitation is the study’s cross-sectional design and lack of data on non-respondents. Lastly, women were overrepresented, something that may have influenced gender comparisons.

The results of this study highlighted the validity of HeQ, a short and simply scored self-administered questionnaire targeting hesitancy to return to normality. The questionnaire was developed in the COVID-19 era, but it could also be used in relation with other medical crises, both individual and social. Future research could compare the HeQ with COVID-19-related anxiety scales such as the COVID Stress Scales (CSS; Taylor et al., 2020) or the Obsessive–Compulsive Inventory-12 items (OCI-12; Abramovitch et al., 2021) to explore if increased hesitancy leads to increased COVID-19-related anxiety or health anxiety.