The impact of stress and its influencing factors among dentists during the COVID-19 pandemic in Kingdom of Bahrain

1.1 Research questions (RQ)

Q1. Is there any impact between fear of social isolation and dentists’ outcome during the COVID-19 pandemic in Kingdom of Bahrain?

Q2. Does discomfort caused by protective equipment affect the outcome of dentists during the COVID-19 pandemic in Kingdom of Bahrain?

Q3. To what extent do difficulties and anxieties related to the measures taken to control the infection, affect the dentist’s outcome during the pandemic in the Kingdom of Bahrain?

Q4. How does the burden of caring for patients affect the dentists’ performance outcome during the COVID-19 pandemic in Kingdom of Bahrain?

1.2 Research objectives (RO)

This research aims to assess the stress, influencing factors, and determinants affecting dentists’ practices outcomes during covid-19 in the Kingdom of Bahrain. The research objectives among others are as follows:

O1. To examine the impact of fear of social isolation and dentists’ outcome during the COVID-19 pandemic in Kingdom of Bahrain.

O2. To assess the effects of discomfort caused by personal protective equipment towards the dentists during the COVID-19 pandemic in the Kingdom of Bahrain.

O3. To investigate the impacts of the measures taken to control the infection that affects dentists’ outcome during the COVID-19 pandemic in Kingdom of Bahrain.

O4. To evaluate how the pandemic of COVID-19 has impacted the responsibilities of dentists in caring for patients essentially, in the Kingdom of Bahrain.

2.1 The effect of stress and influence factors among dentists

In general, stress harms both physical and mental health. But more specifically, stress harms physicians who are exposed to a high level of stress at work.⁸ Moreover, stress is not only harmful to mental and physical health but to social life as well.⁹ It could also strongly implicate the human body, such as the pathogenesis of coronary heart disease.¹⁰ Several studies have found that individuals who are reported with high levels of stress and tension are more likely to suffer from fatal diseases such as heart disease and psycho-emotional disorders, which ultimately affect their professional and social lives.^11,12 Stress and tension affect the nervous system as well. The impacts of stress can be seen in terms of reduced focus levels at work, which can lead to other serious complications.

In recent years, an increasing number of studies have mentioned the negative impact of stress on healthcare workers during COVID-19. For instance, a study conducted by Ahmed et al. (2020) has analysed the data from 30 countries. The essence of this study was on how practitioners around the globe are in a constant state of anxiety and fear while working in their respective fields due to the COVID-19 pandemic’s impact on humanity.¹³ According to Stankovska et al. (2020), healthcare workers are mostly at high risk of getting coronavirus due to direct contact with patients.¹⁴ This often causes fear and panic among the medical staff, fear of the consequences of the virus when communicating with the family members, fear of isolation, and fear of transmitting the virus to their family members.

The researcher has claimed that doctors’ frequent medical errors are mostly due to high stress levels they are experiencing.¹⁵ Moreover, Nabil Mohsen Al-Zubair et al. suggested that the most influential source contributing to stress, which received the highest vote compared to the other stress factors, was the repetition of work or boredom.

Another study led by Myers (2014) argued that other stress factors are the strict work timing and the attendance issues among dentists.¹⁶ Plus, the workload issue has led to sleep disturbances. In addition, several studies conducted in several countries showed that dentists experiencing nervousness and burnout at work led to the deterioration of their psychological state.^17,18

Pouradeli et al. (2016) mentioned that concentration at work during the COVID-19 pandemic was 66%, while time pressure was 65%. These statistics yet again highlighted some stress-causing factors amongst dentists.¹⁹ A similar study revealed that anxiety represented 35.1% in China, which was the highest factor other than sleep deprivation and depression.²⁰ Stankovska et al. (2020) and Jarvis et al. (2020) noted that everyone was anxious of the risks of catching a potentially fatal new virus and might experience mental stress, anxiety, loneliness, depression; thus, dentists were the individuals who confronted this problem the most, when looking at the nature of their job.^14,21

Regarding the stress related to sex among dentists, a study conducted by Ahmed Arafa1 et al. (2021) claimed that more women were experiencing psychological disturbances compared to men. Therefore, the study was formulated to test the impact between stress dimensions as independent variables and the dentists’ performance outcome as the dependent variable. Consequently, these hypotheses are displayed in detail to ensure they achieve the objectives of the study. Therefore, this has led to the following hypothesis:

Figure 1 shows the conceptual framework and hypotheses model for this work.

Figure 1. Conceptual framework and hypotheses model.

BCP=Burden of caring for patients; DAIC=Difficulties and anxieties related to infection control; DCPE=Discomfort caused by protective equipment; FI=Fear of social isolation. Source: Author’s own Conceptual Framework and Hypotheses Model.

3.1 Ethical consideration

Expedited ethics approval was received from the Research Crown Prince Center, BDF Hospital, Kingdom of Bahrain (approval no. BDF/R&REC/2021-625) in September 2021. All respondents were asked for written informed consent to participate before filling out the questionnaire.

3.2 Sample

Several dentists in the Kingdom of Bahrain were requested to participate in an online study regarding stress and its influencing factors among dentists during the COVID-19 pandemic. A total of 306 respondents selected from a list of 1489 registered professionals in the NHRA (National Health Regulatory Authority Bahrain). A convenience sample of dentists was contacted to participate in this research²² A sample size of n=306 was considered sufficient for this study according to the table of Krejcie and Morgan 1970.²³ In addition, a convenience sample of dentists was contacted to participate in this study.

3.3 Survey

A cross-sectional survey was designed to assess the impact of stress and its influencing factors among Bahraini dentists. An online survey was conducted to minimise face-to-face communication as well as to accommodate to dental practitioners who work in different locations of hospitals and dental clinics in Bahrain. A web-based survey was created using Google forms and emailed to the dentists in Bahrain. Data was collected between the 8^th of September 2021 and 10^th of October 2021.

Table 1 illustrates the correlation among research objectives, research questions, and related hypotheses (created by the researcher).

The quantitative approach was utilized in this study since it is more structured than the qualitative method. Hence, the data from the respondents were collected using the online questionnaire, which was more feasible during this pandemic. Both independent (BCP, DAIC,DCPE, FI) and dependent variables (Outcome) were measured using a five-point Likert scale ranging from (1) ‘Strongly Disagree’ (2) ‘Disagree’ (3) ‘Natural’ (4) ‘Agree’ (5) ‘Strongly Agree’.This research covers the respondents’ socio-demographic profile, including gender, age, education, experience, marital status, having a child. The exogenous variables have 24 questions measured by using an instrument developed by Kuo, F. L. et al. (2020),²⁴ while the endogenous variables have 6 questions measured by using an instrument developed by Al-zubair et al. (2014).

3.4 Data analysis

Descriptive analysis was done using the SPSS version 25 program to analyse the study data. The partial least squares SEM (PLS-SEM) was used to estimate complex cause-effect relationship models with the latent variables. The internal consistency (Cronbach’s alpha) of the exogenous and endogenous variables were 0.938, suggesting that the questionnaires were internally consistent. This study has five dimensions as presented in Table 2.

4.1 Background characteristics of the respondents

Figure 2 illustrates the procedure of data gathering.

Figure 2. Procedure of data gathering.

Out of 306 participants included in the survey, 253 responded, giving a response rate of 83.6%.⁴⁹ Overall, 30.8% were female, and 69.2% were male. The average mean and standard deviation in gender variable were 1.35±0.35. There were more men than women in the sample (69% vs 30.9%). Table 3 shows that the majority of the sample were aged between 45-50 years, which represented 39.9%. In contrast, 61.1% represented the rest of the age groups. The average mean and standard deviation in age were 2.72±0.96. In terms of experience, most dental practitioners who participated in the survey had over 20 years clinical experience. Over two-thirds of the surveyed dentists (69.2%) hold a bachelor’s degree. The average mean and standard deviation were 1.19 ±0.28. There were more bachelor’s degree holders than master’s degree holders in the sample (69.2.8% vs 3.8%). Regarding marital status, 81.1%. of respondents were married, and most of the participants had children (84.6%), while the rest (15.4%) did not have children. Table 4 shows the descriptive statistics for constructs. Tables 5 to 9 show the descriptive statistics for variables.

The findings reveal that the mean of all variables is greater than 2.80. This suggests that responders were aware of the impact of stress during the COVID-19 epidemic. According to Table 4, the most common stress factor was difficulties and anxieties towards infection risks. The mean score was 2.90±1.36, which was significantly higher than the rest of factors respectively, followed by fear of social isolation and the mean score was 2.87±1.14, which was the lowest standard deviation (SD) among the others. Both discomfort caused by personal protective equipment and outcome variables received nearly identical scores of 2.84±1.51 and 2.83±1.46. Finally, the burden of caring for patients had a mean of 2.81±1.72, which was the lowest mean and the highest SD among the other research dimensions.

The overall mean for all dimensions was 2.86±1.43. reflecting the existence of considerable acceptable variability within the data set. Figure 3 illustrates a simple bar graph representing mean of variable study.

Figure 3. Simple bar mean of variable study.

BCP=Burden of caring for patients; DAIC=Difficulties and anxieties related to infection control; DCPE=Discomfort caused by protective equipment; FI=Fear of social isolation; O=outcome.

4.2 Reliability and composite reliability

Reliability measurement gives internal consistency in the measurement of the variable.²⁵ The instrument’s reliability must be greater than 0.60 to be accepted.^26,27 In this study, the researcher has conducted two types of reliability tests namely Cronbach’s alpha through SPSS 25 and composite reliability (C.R). The findings showed the reliability (Cronbach’s alpha) values were ranged from 0.997 to 0.745 while composite reliability (CR) values ranged from 0.997 to 0.974. The discriminant validity was achieved as the square root of AVE (0.990) and (0.852) was higher than the correlation value (0.50) between two latent construct.²⁸ Furthermore, it implied that the model was free from any redundant items. Therefore, all values for composite reliability, Cronbach’s alpha, and AVE were significant and had met the required acceptance criteria. In general, Table 10 shows the characteristics of the three constructs used. The Cronbach’s alphas are greater than 0.7, and all composite reliability estimates higher than 0.9. This demonstrates a reasonable level of reliability thus, this finding has a good internal consistency that is fit and suitable for this study.²⁹

4.3 Assessment of measurement model

Figure 4 shows the path model result.

Figure 4. Path model result.

4.4 Convergent validity

Convergent validity refers to a level to which a set of variables converge in their measurement of a specific concept.³⁰ The establishment and confirmation of convergent validity call for meeting specific criteria, including factors loadings, composite reliability (CR) as well as average variance extracted (AVE); these were all employed in a simultaneous manner as recommended by Hair et al. 2013.³¹ In doing so, the item loadings were observed and it was evident that the entire items were over 0.50 indicating acceptable levels (Hair et al., 2010). All the factor loadings in this study are significant at the significant level of 0.1.

4.5 Discriminant validity for latent variables

The discriminant validity of the measures reveals how well items differ among constructs. Simply put, it confirms that the items that use various constructs do not overlap. As a result, all the variables have sufficient discriminant validity in this study that the AVE exceeds the correlation among variables. The results AVE for each variable have a greater correlation square of the given variable with any other variable. Table 12 below illustrates that the AVE’s square root for all variables is higher than the correlations between the variable and other variables in the model.

4.6 Coefficient of determination result R²

The R - squared value shows how much variance is explained by the independent variables for dependent variables. Consequently, a higher R² value enhances the predictive ability of the structural model. Plonsky et al. (2018) recommended that R² has to be larger than 0.75 to be deemed substantial, with adequate power above 0.25.³² In this study, the model indicated an excellent fit to the data as evidenced by the squared multiple correlations (R²) values for the dependent variables (Outcome): O (R²=0.992); as shown in Table 13 and Figure 5. Thus, the one latent variable of independents (IV) explained a substantial of variance for the outcome of dental practitioners (DV). Meanwhile, dependents variables have revealed that there were more than 98% of the variation for the impact of stress and influencing factors among dentists during the COVID-19 pandemic in Bahrain.

Figure 5. R square.

4.7 Effect size f²${\mathit{}}^{\mathbf{}}$

The main purpose of using effect size in this study is to measure the effect and relationship between variables. According to Cohen (1988), the effect size that should be less than 0.02 (0.02=small,0.15=medium, 0.35=high).³³ From Table 14, the effect size of BCP, DAIC, DCPE are small meanwhile, the effect size of FI is medium. As illustrated in Table 14 and Figure 6, the effect of stress and influencing factors of stress among dentists during the COVID-19 pandemic in the Kingdom of Bahrain were calculated by the following formula:

Figure 6. f Square (f2).

f²=above 0.35 is considered large effect size

f²=ranging from 0.15 to 0.35 is considered medium effect size

f²=between 0.02 to 0.15 is considered small effect size

f²=values less than 0.02 are considered with no effect size

f²=(R² included – R² excluded)/(1-R² included)

As shown in Table 14 and Figure 6, the effect size of the burden of caring for patients (BCP) ranged between 0.02 to 0.15, which seemed a weak support for the hypothesis with limitation of small effect size, approximately 2.6%. While difficulties and anxieties related to infection control (DAIC) ranged between 0.02 to 0.15 was considered as small effect size, represented 3.5%. Similarly, discomfort caused by protective equipment (DCPE) was considered as a small effect. Finally, fear of social isolation (FI) had the strongest effect size among other independent variables. Its medium ranged was above 0.015 and less than 0.35 (represented 20%), which was considered as a medium-size effect according to Cohen’s criterion.³⁵ In summary, all the variables of the studies have contributed considerable effects in varying proportions as the total effect size had a medium with a total average (29.8%). Therefore, there was a support for the proposed hypotheses with a high effect size rate.

4.8 The assessment of the inner model and hypotheses testing procedures

Following the confirmation of the goodness of the outer model, the researcher tested the hypothesised relationship among the constructs through Smart PLS 3.0, PLS Algorithm. R is an open-source software which could also be used. Figure 7 and Table 15 display the path coefficients and t-values that were obtained from the test.

Figure 7. Path coefficient.

BCP=Burden of caring for patients; DAIC=Difficulties and anxieties related to infection control; DCPE=Discomfort caused by protective equipment; FI=Fear of social isolation.

4.9 Path model significance results

To ensure that the path coefficients are statistically significant, the current study has chosen the bootstrapping method integrated with Smart PLS3. The main purpose of using bootstrapping was to generate t-values coupled with each path coefficient, and as a subsequence, the p-values of the hypotheses were also generated, as displayed in Table 15. As shown in Table 15, it can be concluded that the burden of caring for patients (BCP) has negatively impacted the outcome at the significant level of 0.01 (β=-0.184, t=01.560, p=0.060). Additionally, the result rejected the hypothesis due to the t-value, which was less than 1.96, while the highest SD among the others and p>0.05. Moreover, dental practitioners did not support the burden of caring for patients as an independent variable and the outcome as the dependent variable. In other words, this result indicated that their priority was to taking care of the patient, despite the fatal risk they were exposed to during the COVID-19 pandemic. Apparently, the participants did not pay much attention thus implied the variable was not their primary concern. It became unnecessary to examine the impact of (BCP) on the stated impact as suggested by Hair et al.³⁶ Furthermore, the researcher rejected alternate hypotheses and accepted null hypothesis instead, due to disapproving of the impact (BCP) on the outcome during the COVID-19.³⁷ However, the impact between difficulties and anxieties related to infection control (DAIC) and the outcome as the dependent variable was supported at the level of significance of 0.01 (p<0.05, β=0.210, t=2.118, p=0.017). Meanwhile, the discomfort caused by protective equipment was supported by predictors of dental practitioners at the significance level of 0.01 (p<0.05, β=0.212, t=2.354, p=0.009). This result was found by another study which was conducted in Taiwan in 2020, which revealed that discomfort caused by protective equipment had a moderate stressor among another dimension of stressor investigated.³⁸ Surprisingly, fear of social isolation revealed the strongest correlation with outcome among independent variables at the significance level of 0.01 (β=0.393, t=5.090, p<0.05 (0.000). Statistically, FI was considered the lowest SD among other dimensions, which enhanced researcher’s analysis. This result indicated that fear of social isolation revealed (FI) as one of the most important challenges and the most stressful for dental practitioners in Bahrain. Therefore, this result is in line with previous studies that supported healthcare workers such as dental practitioners did experience stress related to fear of social isolation, in fact they were even more stressful of the fear of having the infection spread to their family and relatives.^38,39 All the relationships mentioned earlier were examined by using the PLS.3 bootstrapping method.

5.1 Limitations

This is the first study in the Kingdom of Bahrain investigating the impact of stress and its influences on dentists during COVID-19. However, we have identified many limitations in our research. Firstly, the outcome of this study is based on data collected from the dentists on the impact of stress during COVID-19 at one point in time. This study is based on cross-sectional data, which is limited in nature and lacks follow-up. This research could be more accurate if using a longitudinal approach to explain the complex relationship of the impact over a longer period. Additionally, the low response rate, a biased selection, and the fact that this study used only two variables with no mediation or moderator variables all serve as limitations. Despite these limitations, this study represents a large-scale study with validated measurements that provides a comprehensive understanding of stress and its influencing factors among dentists in Bahrain.

5.2 Recommendations

It is recommended that further studies utilise longitudinal approaches rather than cross-sectional plus other different analysis methods such as CFA and SEM techniques. It is recommended that future studies also include other countries, such as Arab countries in general, practically in GCC (Gulf Cooperation Council) countries. Another recommendation for future studies is to increase the sample size.

Underlying data

Dryad: The impact of stress and its influencing factors among dentists during the COVID-19 pandemic. https://doi.org/10.5061/d ryad.1g1jwstzq.⁴⁹

This project contains the following underlying data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).