People's attitudes towards the use of quarantine in the Covid-19 pandemic in Iran: Validity and reliability study.

BACKGROUND: The purpose of this study was first to translate (into Persian) and evaluate the psychometric properties of the attitudes towards the use of quarantine questionnaires (PAQ).
METHODS: This methodological study was conducted on 1000 adults' participants in Iran, using convenience sampling. Based on permission from the developer of the scale, it was back-translated. Content validity, exploratory and confirmatory factor analyses were tested. Reliability of the questionnaire was evaluated with test-retest and internal consistency.
RESULTS: The adequacy of the sample was acceptable because of the results of the Kaiser-Meyer-Olkin test (0.788), and Bartlett's test of sphericity revealed statistically significant results (P < .001). Exploratory factor analysis showed the 4-factor structure of the instrument. The reliability of the scale was 0.77 based on internal consistency. Also, the intra-class reliability of the scale was 0.9. The attitude of the maximum number of respondents towards quarantine is positive and they support a quarantine in the pandemic of COVID-19 in Iran.
CONCLUSIONS: Current study showed that the Persian version of attitudes towards the use of quarantine (PAQ) in the Iranian population including 17 items could be used as a valid and reliable instrument.

What’s known

There is few statistics approximately people’s attitudes towards the use of quarantine to control Covid‐19 in Iran, and there is a lack of preferred questionnaires to degree this.

What’s new

This study led to the development of a valid and reliable tool based on cultural differences to measure people’s attitudes towards quarantine in Iran.

INTRODUCTION

COVID‐19 is respiratory pneumonia caused by Severe Acute Respiratory Syndrome Coronavirus‐2 (SARS‐CoV‐2) that was declared as a pandemic by WHO on 11 March 2020 and has affected many countries. According to the latest WHO statistics on 9 June 2020, there are 7 019 918 confirmed cases and 4 404 396 deaths worldwide because of COVID‐19. Iran is one of the most infected countries with COVID‐19 with 173 832 confirmed cases.

Since there are no approved and definitive treatments, non‐pharmaceutical approaches such as prevention are important and essential in lowering the spread and death rate. Traditional methods for controlling this disease include isolation, quarantine and social distancing. Quarantine is to separate people who are suspected of having the disease but who do not have any symptoms from other healthy people.

With the outbreak of COVID‐19, governments in different countries ordered precautionary measures. Schools, universities, parks and unnecessary jobs were suspended in many countries and even in some cities, people were allowed to leave their homes only in emergency situations, or body temperature was checked in many public places. These measures have changed people’s lifestyles and social relationships and have had side effects on people’s lives, including psychological problems caused by increased anxiety and distance from relatives and acquaintances. Physical problems such as weight gain threaten people’s lives, but it does not mean that quarantine should not be performed, because breaking quarantine and outbreak of the disease may have worse psychological effects.

If the quarantine experience is negative, it can have long‐term consequences that will affect not only the quarantined individuals but also the healthcare system that governs quarantine and affects public health politicians and officials. However, depriving people of their freedom for the public good often leads to disagreements and must be carefully considered. If quarantine is necessary, the authorities must take all necessary measures to ensure that this experience is as tolerable as possible for the people.

One of the most important points in evaluating the attitude towards quarantine against COVID‐19 disease is the impact of psychological aspects of society towards this disease, , which seems to have an interfering and confounding effect on the results of evaluating the practice of individuals. For this purpose, the use of effective tools is necessary to obtain a more accurate assessment.

Some instruments have been used to measure public attitudes towards the use of quarantine during infectious disease outbreaks, such as the one developed by Tracy et al (2009). Because of the lack of a standard instrument for measuring such issues in Iran, we aimed to translate (into Persian) and examine the psychometric properties of the tool of Public attitudes towards quarantine.

MATERIALS AND METHODS

The current methodological study was performed in two steps: (a) translation of the public attitudes towards quarantine (Qx) to Persian, guided by six steps in the process of translation and adaptation of instruments and (b) evaluation of the psychometric properties of the P‐PAQ, consisting of reliability and validity assessments.

Participants and settings

Data were collected by the online feedback of participants during the survey through the “Convenient Sampling.” A structured questionnaire was administered online (Google Form) which also contains the covering letter explaining the purpose of the response from participants, and eventually, the hyperlink to the questionnaire was published in cyberspace and social networks. Data assortment happened between April 2020 and June 2020, with participants from 31 of Iran’s 32 provinces. The inclusion criteria included age older than 18 years, ability to read and write simple Persian language and willingness to complete questionnaires independently, a primary residence located within the study area during the COVID‐19 outbreak. The exclusion criterion was the defective completion of the questionnaire. Generally, a minimum sample size of 10 times the number of the items is needed for exploratory factor analysis (EFA), whereas a sample size of 300 is considered adequate for confirmatory factor analysis (CFA). Tabachnick considers at least 300 samples necessary for factor analysis, but Hair et al state that the sample size should be more than 100, but some other sources such as Comrey and Lee have classified that 100 is weak, 200 is relatively good, 300 is good, 500 is very good, 1000 and more are excellent. , In the next step, two independent samples were recommended for use in performing EFA and CFA when analysing the psychometric properties. A final sample of 1000 individuals was achieved through standard random sampling. Of those, a critical sample of 500 participants was entered for the EFA, and a validation sample of 500 participants was entered to conduct CFA.

MEASURES

Socio‐demographical data sheet

The participants “demographical variables information” was collected by a self‐administered questionnaire that included age, education level, marital status, work status and province they live in.

The public attitudes towards quarantine

The public attitudes towards quarantine (PAQ) were developed by Tracy et al in 2009. The data reported in this study are derived from a subset of 17 survey items specifically designed to measure public attitudes towards the use of quarantine during infectious disease outbreaks. Participants were asked to indicate their level of agreement/disagreement with each item; the response format was a 5‐point Likert‐type design (1 = “strongly disagree,” 2 = “somewhat disagree,” 3 = “neutral,” 4 = “somewhat agree” and 5 = “strongly agree”).

Step 1: Translation and cross‐cultural adaptation process

Translation of the PAQ closely followed the process of translation and adaptation of instruments by the World Health Organization in six steps. In the first step, the English version of the questionnaire was translated into Persian by two bilingual native Persian speakers who had sufficient experience and mastery in translating English texts. During the translation, an attempt was made not to change the meaning and concept of the phrases and their difficulty level, and the conceptual equivalence of words and sentences was emphasised. In the second step, the initial translations were compared and revised, and finally, the initial translations were merged. In the third step, the translated version was returned from the Persian language to the original language and revised conceptually. In the fourth step, the questionnaire was studied by 10 participants and their opinions were applied. In the fifth step, the final corrections were made to the questionnaire and the final version was prepared for psychometrics. In the sixth step, all steps were properly documented.

Step 2: PAQ psychometric testing procedure

To evaluate the content validity, the content validity ratio (CVR) and content validity index (CVI) were used. To determine the CVR, 20 experts were asked to examine each phrase based on a three‐part spectrum (necessary, useful but not necessary, not necessary). Acceptance of phrases was based on the Lawshe Table. None of the phrases was deleted in this section. In order to assess the CVI, 20 experts were asked to rate each of the questionnaire phrases in terms of “relevance.” The construct validity of the questionnaire was performed by exploratory and confirmatory factor analysis (EFA and CFA) in a sample size of 1000 participants, each method by 500 participants.

Data analysis

IBM SPSS, version 18.0 and Lisrel, version 8.80 were used for analysing psychometric properties. The construct validity: EFA was used to evaluate the basic construct of items using principal component analysis with varimax rotation. The Kaiser–Meyer–Olkin index (KMO) test was used to evaluate the quality of sampling, Bartlett’s test of sphericity (BTS) was used to examine the correlation matrix between items (P < .05), and the Kaiser index was used to estimate the number of factors. The KMO value varies from zero to one, with 0.9 to 1 excellent, 0.8 to 0.9 good, 0.7 to 0.8 satisfactory, 0.6 to 0.7 average, 0.5 to 0.6 sample size is insufficient and <0.5 is considered unacceptable. The results of adequacy sampling and analysability of data are shown in Table 1.

TABLE 1

KMO and Bartlett's test

Kaiser–Meyer–Olkin measure of sampling adequacy	0.788
Bartlett's test of sphericilty
Approx. χ 2	1432.99
df	136
Sig.	<0.001

Confirmatory factor analysis was performed to confirm the pattern obtained from EFA. To evaluate the fit of the model, Normed Fit Index (NFI), Non‐Normed Fit Index (NNFI), Comparative Fit Index (CFI), Goodness of Fit Index (GFI), Degrees of Freedom (CMIN/DF) and Root Mean Square Error (RMSEA) were used, which are commonly used in determining fitness in CFA.

The Cronbach'’s coefficient was calculated to check the internal consistency, if higher than 0.70 was considered acceptable. Intraclass correlation coefficient (ICC) was calculated to assess the validity of the retest. An ICC value >0.70 indicates sufficient stability. <0.3, weak and between 0.3 and 0.7, average and acceptable.

RESULTS

One thousand people participated in this study. Most of these people were students, employees or self‐employed. Details of the demographic characteristics of the subjects are reported in Table 2.

TABLE 2

The demographical profile and attitude of people towards quarantine

Variable	EFA sample (n = 500)	CFA sample (n = 500)	Total (n = 1000)	PAQ	P value
	n (%)	n (%)	n (%)	Mean (±SD)
Age group
18‐35 y	403 (80.6)	400 (80)	803 (80.3)	74.64 (±7.46)	.032
36‐65 y	94 (18.8)	92 (18.4)	186 (18.6)	73.61 (±7.21)
>65 y	3 (0.6)	8 (1.6)	11 (1.1)	69.64 (±10.74)
Gender
Female patients	302 (60.4)	308 (61.6)	610 (61.0)	74.48 (±7.00)	.77
Male patients	198 (39.6)	192 (38.4)	390 (39.0)	74.26 (±8.17)
Marriage
Single	340 (68)	350 (70)	690 (69.0)	74.55 (±7.33)	.5
Married	160 (32)	150 (30)	310 (31.0)	74.04 (±7.78)
Education
Diploma and sub‐diploma	213 (42.6)	210 (42)	423 (42.3)	74.99 (±7.5)	.019
Bachelor’s degree	182 (36.4)	206 (41.2)	388 (38.8)	74.2 (±7.31)
Master's degree	61 (12.2)	44 (8.8)	105 (10.5)	73.78 (±8.41)
Doctoral degree	44 (8.8)	40 (8)	84 (8.4)	73.06 (±6.63)
Job
Unemployed	136 (27.2)	82 (16.4)	218 (21.8)	74.65 (±8.5)	.24
Self‐employment	99 (19.8)	90 (18.0)	189 (18.9)	74.56 (±7.94)
Housewife	58 (11.6)	24 (4.8)	82 (8.2)	74.29 (±8.11)
Employee	52 (10.4)	60 (12.0)	112 (11.2)	74.9 (±6.72)
Student	121 (24.2)	174 (34.8)	295 (29.5)	74.35 (±6.9)
Worker	10 (2.0)	5 (1.0)	15 (1.5)	73.08 (±6.97)
Medical staff	15 (3.0)	46 (9.2)	61 (6.1)	73.7 (±5.94)
Retired	9 (1.8)	19 (3.8)	28 (2.8)	72.14 (±5.88)
Location
Tehran	102 (20.4)	106 (21.2)	208 (20.8)	77.34 (±6.54)	<.001
Golestan	143 (28.6)	45 (9)	188 (18.8)	71.63 (±7.18)
Mazandran	30 (6)	53 (10.6)	83 (8.3)	74.35 (±7.75)
28 Other states	225 (45)	296 (59.2)	521 (52.1)	73.54 (±5.74)
Was anyone in your home quarantined during Coronavirus?
No	351 (70.2)	334 (66.8)	685 (68.5)	74.27 (±7.64)	0.2
Yes, myself but nobody else in my home	20 (4)	22 (4.4)	42 (4.2)	74.65 (±7.23)
Yes, myself and someone else in my home	111 (22.2)	113 (22.6)	224 (22.4)	73.33 (±7.11)
Yes, not myself but someone else in my home	18 (3.6)	31 (6.2)	49 (4.2)	76.38 (±6.13)

As a result of the content validity, changes have been made in the form and richness of the words and no item has been omitted in this research. The CVR was 0.7 for the whole scale and 0.5‐1 for the items. The CVI in this study was also calculated to be 0.9.

EFA results showed the adequacy of the samples for data analysis with a score of 0.772 based on KMO test. In addition, the results of the BTS were statistically significant (P < .001) and led to the development of a 4‐factor solution [“accountable organisation” (six questions), “quarantine compliance” (five questions), “quarantine regulations” (four questions) and “ethics in quarantine” (two questions)] as a domain. The resulting domains accounted for 47.141% of the observed variance of PAQ with 17 items. The factor loadings of the 4‐factor solution are reported in Table 3. Fitness indicators for the PAQ questionnaire are reported in Table 4.

TABLE 3

Factor loadings of PAQ on rotated factor structure matrix

No	Factor 1	Factor 2	Factor 3	Factor 4
q1	0.596
q2	0.772
q3			0.458
q4	0.694
q5			0.764
q6			0.663
q7			0.737
q8		0.523
q9		0.717
q10	0.311	0.600
q11		0.646		0.339
q12		0.627
q13	0.333	0.390
q14	0.333
q15				0.691
q16				0.711
q17	0.606

TABLE 4

Model fit index summary

Model fit index	Admissibility	Result
ꭓ 2 P‐value (chi‐squared P‐value)	>0.05	>0.001
RMSEA (root mean square error of approximation)	<0.08 perfect fit, 0.08‐0.10 good fit, >0.10 weak fit	0.069
NFI (normed fit index)	>0.9	0.94
NNFI (non normed fit index)	>0.9	0.95
GFI (goodness of fit index)	>0.9	0.92
CFI (comparative of fit index)	>0.9	0.95
CMIN/DF (minimum discrepancy function by degree of freedom divided)	<3 good; <5 sometimes permissible	3.23

To examine the internal consistency of PAQ, Cronbach’s α was calculated. The Cronbach’s α coefficients of the four factors and the total scale were 0.69, 0.76, 0.67, 0.40 and 0.77, respectively, indicating adequate internal consistency. To assess duration, the scale was administered twice to 50 participants at 2‐week intervals. The ICC was 0.9 and the SEM was 0.73, indicating adequate stability over time.

Attitudes towards the use of quarantine

Participants’ attitudes towards the use of quarantine during the COVID‐19 pandemic is shown in Table 5. From the table above, it is clear that most respondents agree with the various items.

TABLE 5

People's attitudes towards quarantine (Qx) by a factor (N = 1000)

	Strongly agree, %	Somewhat agree, %	Neutral, %	Somewhat disagree, %	Strongly disagree, %
Quarantine compliance
Health Department should have the power to order people into Qx during outbreaks	81.1	11.1	6	6	1.2
Qx is the right way to stop the spreading of the outbreak of infectious disease	71.4	20.3	5.2	2.4	0.7
If I go into Qx, my family/friends/community will be protected from becoming infected	71.9	19.8	4.4	2.4	1.5
It is reasonable for some rights to be taken away during an infectious disease outbreak, like the freedom to go wherever you want and talk with whoever you want	63.3	19.9	9	3.2	4.6
As members of the community, people owe it to the rest of us to obey a quarantine order	83.4	11.4	3.4	9	9
Accountable organization
Health Department needs to explain to everyone why they should be allowed to use Qx.	85.4	8.3	4.5	9	9
Government should pay for doctors, nurses and counsellors to help people who are in Qx	80.7	10.9	6.5	1.1	8
Health Department should ensure that people have food and shelter while in Qx and pay for it with public money if needed	91.8	3.5	3.2	1	0.5
Government should pay for support groups so that people coming out of Qx have someone to talk to about it	75.8	13.3	8.8	1.3	0.8
People in Qx should get money from the government to pay for workdays lost or missed	81.6	8.8	7.7	1.2	0.7
Public Health should ensure that there is no discrimination in the way they quarantine people	85.1	8.8	5.2	0.6	0.3
Quarantine regulations
If someone is given a Qx order by the Public Health Department, they should follow it no matter what else is going on in their life at work or home	47.5	20.9	13.9	12.4	5.3
People who break Qx orders intentionally should face legal penalties like a fine or jail	51.2	26.2	13	5.5	4.1
Government should be able to lock up people if they fail to obey Qx orders	64.2	22	9.1	2.9	1.8
Health Department should use electronic bracelets and in‐home surveillance cameras for people who disobey Qx orders	26.3	17	20.2	13.4	23.1
Ethics in quarantine
For people who disagree with their Qx, the order should request a further review to have an early release	29.6	22.2	27.2	10.1	10.9
Being in quarantine is private health information—so doctors, hospitals and public health departments SHOULD NOT TELL ANYONE ELSE the names of people who are in quarantine	51.3	13.9	17.1	9.5	8.2

The total mean score of attitude towards the use of quarantine with a variety between 32 and 85 was 74.39 ± 7.47. The mean scores for the four questionnaire domains: “quarantine compliance,” “accountable organization,” “quarantine regulations” and “ethics in quarantine” were 22.97 ± 2.71, 28.40 ± 2.74, 15.61 ± 3.44 and 7.40 ± 2.09, respectively. The adjusted mean score based on the number of questions in each domain was 4.59 ± 0.54, 4.73 ± 0.45, 3.90 ± 0.86 and 3.70 ± 1.04 respectively.

Table 2 shows the relationship between people’s attitudes towards quarantine and demographic factors. The following is a comparison of attitudes towards quarantine in demographic subgroups with adjustments for the impact of alternative variables in Table 6.

TABLE 6

Factors related to attitudes towards people's quarantine based on demographic characteristics

Variable	PAQ	95% Confidence interval	P value
		(lower, upper)
Constant	77.88	(76.4, 79.36)	<.001
Age group (18‐35 y)
36‐65 y	−0.80	(−2.29, 0.7)	.30
>65 y	−6.13	(−10.76, −1.51)	.01
Gender (female)
Male	−0.83	(−1.84, 0.19)	.11
Marriage (single)
Married	0.74	(−0.56, 2.04)	.26
Education (diploma and sub‐diploma)
Bachelor's degree	−1.08	(−2.17, 0.01)	.052
Master's degree	−1.78	(−3.46, −0.11)	.04
Doctoral degree	−1.34	(−3.19, 0.5)	.15
Job (unemployed)
Self‐employment	0.66	(−0.83, 2.15)	.39
Housewife	−0.55	(−2.69, 1.59)	.61
Employee	1.84	(−0.07, 3.76)	.06
Student	0.89	(−0.46, 2.24)	.19
Worker	−0.05	(−4.21, 4.12)	.98
Medical staff	1.37	(−0.91, 3.64)	.24
Retired	0.31	(−2.98, 3.6)	.85
Location (Tehran)
Golestan	−6.22	(−7.74, −4.7)	<.001
Mazandran	−4.14	(−6.04, −2.24)	<.001
28 Other states	−3.24	(−4.43, −2.05)	<.001
Was anyone in your home quarantined during Coronavirus? (NO)
Yes, myself but nobody else in my home	0.29	(−0.84, 1.41)	.62
Yes, myself and someone else in my home	−0.95	(−3.09, 1.19)	.39
Yes, not myself but someone else in my home	1.82	(−0.47, 4.1)	.12

DISCUSSION

The Persian version’s reliability of PAQ was 0.77, the acceptable value of which is estimated based on studies >0.70. , The reliability of the instrument was not stated in the original version, but in a study conducted in India, the reliability of the questionnaire was reported to be 0.9.

The current study was performed with 1000 participants. It was found that the sample size in the current study is adequate based on the results of KMO test. Regarding the size of the participants, Comrey et al , explained that if items are <40, a sample size of 200 is sufficient for most of the factors’ analysis, but 500 is very good and 1000 is excellent. Mukherjee et al (2020) used 150 participants in a similar study that psychometric the PAQ questionnaire. Many rules regarding the size of samples with the validity of the analysis factor have been stated by researchers in various studies, which ultimately help reduce the difficulty and increase the quality of studies.

In methodological studies, CFA is used to examine the validity of the internal structure of scales and CFI is more appropriate than other fit indicators because it is less affected by sample size. Its acceptable value is also value >0.9. All fit indicators were within an acceptable threshold, indicating confidence in the validity of the internal structure. We could not compare our findings with the main study of the questionnaire because internal validity was not assessed in the original study and also no translation study of the instrument was available.

Finally, the attitude of 1000 respondents to the use of quarantine during the COVID‐19 pandemic was positive and they support a qquarantine in this pandemic of COVID‐19 in Iran. Our findings were similar to the attitude of the Indian people towards the use of quarantine in 2020 and the attitude of the Canadian people (2009) in the SARS pandemic. , Previous studies have shown that the culture of different communities can influence people’s attitudes towards quarantine. According to the following statistics, quarantine is considered a viable solution to the epidemic of infectious diseases: 65% in India, 94% Canada, 89% in Singapore, 95% in Taiwan, 76% in the United States and 92% in China. These results support the findings of this study, 91% of the participants believed that quarantine is the best way to prevent infectious diseases. In the current study, age and education had a significant relationship with attitudes towards quarantine, in order that with increasing age and level of education of participants, their attitudes towards quarantine decreased. Tracy et al (2009) and Song et al (2020) noted a significant relationship between age and gender with an attitude towards quarantine, with female patients scoring significantly higher than male patients. This score decreased with ageing. However, there are differences in age, Tracy et al stated that their participants have a more positive attitude towards quarantine with ageing, while in the Song et al study, similar to the findings of the present study, reported a lower attitude with increasing age. ,

One of the strengths of the current study is the high sample size and collection of samples from all over the country. Our study had some limitations too. Because we are translating and psychometric PAQ for the first time, and no questionnaire in Iran examines people’s attitudes towards quarantine, the conditions were very difficult to compare. We suggest that researchers identify the factors that influence people’s attitudes towards using quarantine in future studies.

CONCLUSION

Based on our assessment of the structure and content of the PAQ questionnaire, this tool has an acceptable structure and format for evaluating the attitude towards quarantine among the Iranian population. Based on the performed factor analysis, the obtained 4‐factor format with fine reliability and validity will be able to evaluate people’s attitudes. However, considering the potential effects of cultural and ethnic tendencies as well as the effects of psychological factors, it seems that the use of these items can help improve the questionnaire.

DISCLOSURES

The authors have declared no disclosures.