COVID-19 vaccine acceptance and hesitancy among the general population of Pakistan: a population-based survey.

OBJECTIVES: This study aims to determine the COVID-19 vaccination coverage and the factors associated with vaccine acceptance and hesitancy in the general population of Pakistan.
SETTING: This population-based study covers all major areas of Pakistan, including Sindh, Punjab, Khyber Pakhtunkhwa and Baluchistan provinces and the capital Islamabad. PARTICIPANTS: A total of 541 male and female Pakistani adults above 18 years were interviewed to determine the COVID-19 vaccination coverage and understand the factors associated with vaccine acceptance and hesitancy. OUTCOME: The outcome was COVID-19 vaccination status (not vaccinated or vaccinated).
RESULTS: Of 541 participants, 227 (41.96%) were non-vaccinated and 314 (58.04%) were vaccinated. Two-thirds of the participants from both the non-vaccinated and vaccinated groups (185 (81.50%) vs 236 (75.16%), p=0.008) reside in Sindh. Nearly one-third of participants from both groups were ever infected with COVID-19 (77 (33.92%) and 90 (28.66%)). The odds of COVID-19 vaccination among the age group 34-42 years were 1.75 times higher (95% CI 1.35 to 2.09, p=0.008) than the other age groups. The odds of COVID-19 vaccination among those with COVID-19 ever-infected family members were 1.87 times higher (95% CI 1.56 to 2.34, p=0.032) than those with uninfected family members.
CONCLUSIONS: Targeted interventions for subsets of populations reluctant to vaccination can improve vaccine coverage. Moreover, advocacy and explaining the public health benefits of vaccination can enhance the coverage in Pakistan. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

This study, unlike others in Pakistan, has collected data from the general population instead of only focusing on healthcare workers.

The sample size was large and data were collected from all over Pakistan, making the results generalisable.

This was self-reported vaccination status, which might have caused recall bias.

Introduction

COVID-19 has impacted the overall health status of populations for more than 2 years. It quickly spread across the globe with a high transmissibility rate.1 Over time, numerous viral variants have emerged, posing another challenge to healthcare professionals and health scientists.2 3 In such a situation, the most effective action that can be taken to control the spread of infection among the masses and improve their health status is vaccination and herd immunity.4

Across the globe, various vaccine trials have been underway for the past 2 years, some of which have been successfully scaled up to larger populations.5 However, the availability of COVID-19 vaccines has not resulted in 100% vaccination rates in any population. One important reason for this is vaccine hesitancy.6 The WHO defines vaccine hesitancy as ‘refusal of vaccines despite availability of vaccine services’.7 This results in people not wanting to get vaccinated themselves and those around them.8 Recent studies found vaccine hesitancy among the general population and even healthcare workers of countries in Asia, Africa and South America. Harapan et al9 reported that 77.6% of their study population from 10 countries were found to be vaccine-hesitant. This ultimately reduces the chance of herd immunity in various countries, resulting in worsening health status of the populations.8 10–12 Such trends are even seen in healthcare workers,11–15 suggesting that knowledge may not change these perceptions.

Pakistan has also suffered from the significant burden of COVID-19 cases and deaths. About 1.6 million individuals in Pakistan were infected by COVID-19, with 30 000 deaths, during the last 2 years (https://COVID-19.gov.pk/).13 In Pakistan, COVID-19 vaccination was introduced in February 2021. The Sinopharm vaccine was initially introduced and later Sinovac, Pfizer, AstraZeneca and Moderna. Some clinical trials on vaccination were also done in Pakistan.14 The officials did certain public health messages and campaigns to inform the safety and efficacy of COVID-19 vaccination. After so much effort, the vaccination rate was still low.15

Vaccine hesitancy has affected the success of the COVID-19 vaccination programme, thus affecting the chances of achieving herd immunity.16 This is especially true in Pakistan, where the healthcare system has barely been keeping up with the pressure of COVID-19. Despite this, due to limited evidence, the level of vaccine acceptance and hesitancy in Pakistan is unclear, as are the specific factors that affect these.17 This study aims to determine the COVID-19 vaccination coverage and the factors associated with vaccine acceptance and hesitancy in the general population of Pakistan.

Methods

We conducted a population-based survey to determine the COVID-19 vaccination coverage and to understand the factors associated with vaccine acceptance and hesitancy among the general population of Pakistan. This study was conducted from August to December 2021 using a structured questionnaire for data collection. All individuals, male and female, above 18 years of age, and residing in Pakistan, including in Sindh, Punjab, Khyber Pakhtunkhwa, Baluchistan and Islamabad, were included in the study. The tool was taken from a research study/global survey conducted in 20 countries.18 The tool has a Cronbach’s alpha of 0.90. The original tool in English was translated to local languages by language experts and the content validity index of the translated tool was 0.88.

Multistage cluster randomised sampling technique was used to enrol participants in the study. First, we selected households from the list taken from the Pakistan Population Welfare Department. The selection of households in each district was randomised through a computer-generated list of all district households. Once the households were selected, data collectors visited the families chosen for data collection. A single person/adult from each house was selected randomly for data collection.

The sample size was calculated using the prevalence/coverage of COVID-19 vaccination in Pakistan, which is 53%, as reported by Chaudhary et al.19 Considering 95% CI and 80% power, the calculated sample size was 530. When accounting for a 10% non-respondent rate, the final sample size was 583.

The formula and calculation are as follows:

n=[Z2 a/2 (pq)]+B2×design effect

where Z2 (a/2)=(1.96 at 95% CI); p=proportion of vaccinated: 0.53; q=proportion of unvaccinated: 1−0.53=0.47; B2=bound on error (0.05)2; and design effect=2.

The final sample size adjusting to the cluster effect came to be: (265×2)=530. With the inclusion of 10% non-response and missing data, the final sample size came to be 583.

The outcome variable was COVID-19 vaccination status (not vaccinated or vaccinated). Being vaccinated meant that the person has received at least one dose of the COVID-19 vaccine. Not being vaccinated meant not receiving even a single dose of the COVID-19 vaccine. Covariates included sociodemographic details of the participants, such as age, sex, location, education and occupation. Variables related to COVID-19 infection, vaccination type and factors for COVID-19 vaccination acceptance and hesitancy were also reported.

Data on the outcome (not vaccinated or vaccinated) were stratified and analysed. Frequencies and percentages were calculated for categorical variables. The outcome was binary; hence, multivariable binary logistic regression was performed to check the association between COVID-19 vaccination status and independent variables using a stepwise model-building technique. A p value ≤0.05 was considered significant. Akaike’s information criterion and Bayesian information criterion statistics were used to assess the model’s goodness of fit by calculating the difference in the nested and fit model. Analysis was performed on R V.4.1 statistical software.

Patient and public involvement

This study did not involve patients. The study findings are publicly available to all participants and the general public. Informed written consent was taken from all participants before the data collection. Participants were informed about the data collection processes and the risk and benefits of enrolling in the study. After participants agree, they were informed about the data collection tool and the outcomes of the study. Moreover, during and after the data collection, participants’ privacy and confidentiality were maintained.

Results

A total of 541 responses were received, with a response rate of 92.8%. Of the 541 participants, 227 (41.96%) were non-vaccinated and 314 (58.04%) were vaccinated. Majority of the participants in the non-vaccinated group (68, 29.96%) were 18–28 years, while those in the vaccinated group were mostly above 42 years. Most of the participants in the non-vaccinated group were female (n=121, 53.30%), while most in the vaccinated group were male (n=167, 53.18%). Two-thirds of the participants in both the non-vaccinated and vaccinated groups (185 (81.50%) vs 236 (75.16%), p=0.008) reside in Sindh. Most of the participants in the non-vaccinated and vaccinated groups (125 (55.07%) vs 152 (48.41%), p<0.001) were graduates. Among the non-vaccinated, majority (n=80, 35.24%) were self-employed, and among the vaccinated majority (n=110, 35.03%) were healthcare workers (p<0.001) (table 1).

Table 1

Baseline characteristics of participants based on COVID-19 vaccination status

Characteristics	COVID-19 vaccination status, n (%)		P value
	Non-vaccinated, n=227 (41.96)	Vaccinated, n=314 (58.04)
Age (years)
18–28	68 (29.96)	80 (25.48)	0.064
29–33	51 (22.47)	79 (26.16)
34–42	59 (25.99)	65 (20.70)
>42	49 (21.59)	90 (28.66)
Gender
Male	106 (46.70)	167 (53.18)	0.136
Female	121 (53.30)	147 (46.82)
Resident of
Sindh	185 (81.50)	236 (75.16)	0.008
Punjab	24 (10.57)	44 (14.01)
Khyber Pakhtunkhwa	8 (3.52)	25 (7.96)
Baluchistan	1 (0.44)	6 (1.91)
Islamabad	9 (3.96)	3 (0.96)
Education
Postgraduate	71 (31.28)	60 (19.11)	<0.001
Graduate	125 (55.07)	152 (48.41)
SSC/HSSC	11 (4.85)	33 (10.51)
Middle school	15 (6.61)	45 (14.33)
No schooling	5 (2.20)	24 (7.64)
Occupation
Business	31 (13.66)	38 (12.10)	<0.001
Educationist	11 (4.85)	26 (8.28)
Healthcare worker	8 (3.52)	110 (35.03)
Other	57 (25.11)	88 (28.03)
Self-employed	80 (35.24)	46 (14.65)
Unemployed	40 (17.62)	6 (1.91)

HSSC, Higher Secondary School Certificate; SSC, Secondary School Certificate.

Nearly one-third of participants from both groups were ever infected with COVID-19 (77 (33.92%) and 90 (28.66%)), and almost half of participants’ family members ever had COVID-19 infection (112 (49.34%) and 140 (44.59%)). Majority of the participants in both groups had, at some point, taken care of an infected family member (77 (68.75%) and 90 (64.75%)). Among the vaccinated participants, 149 (47.45%) received Sinopharm and 147 (46.81%) received Sinovac COVID-19 vaccines (table 2).

Table 2

COVID-19 infection and vaccination-related details among the general population (N=541)

Characteristics	COVID-19 vaccination status, n (%)		P value
	Non-vaccinated, n=227 (41.96)	Vaccinated, n=314 (58.04)
Infected ever with COVID-19
No	150 (66.08)	224 (71.34)	0.191
Yes	77 (33.92)	90 (28.66)
Infected family member ever
No	115 (50.66)	174 (55.41)	0.274
Yes	112 (49.34)	140 (44.59)
Taken care of an infected family member ever
No	35 (31.25)	49 (35.25)	0.504
Yes	77 (68.75)	90 (64.75)
Vaccination received
Sinopharm	–	149 (47.45)	–
Sinovac	–	147 (46.81)
AstraZeneca	–	5 (1.59)
Other	–	13 (4.14)

When factors for vaccine hesitancy were investigated, non-vaccinated individuals reported lack of belief in vaccination (n=92, 40.53%), followed by waiting for a better vaccine (n=47, 20.70%). In contrast, factors for vaccination acceptance showed that a significant number of people getting vaccinated wanted protection against COVID-19 (n=87, 27.71%), while many (n=114, 36.31%) mentioned family safety as the prime reason for accepting COVID-19 vaccination (table 3).

Table 3

Factors for COVID-19 vaccine acceptance and hesitancy among the general population (N=541)

Factors	COVID-19 vaccination status, n (%)		P value
	Non-vaccinated, n=227 (41.96)	Vaccinated, n=314 (58.04)
I do not believe the vaccine will work	92 (40.53)	–	–
I am waiting for a better vaccine	47 (20.70)	–
Fear of COVID-19 spread	26 (11.45)	–
Worried about adverse events	30 (13.22)	–
I do not think so COVID-19 exists	32 (14.10)	–
Self-protection against COVID-19	–	87 (27.71)	–
Family safety	–	114 (36.31)
Satisfied with previous vaccines	–	65 (20.70)
Mandatory at office	–	26 (8.28)
Mandatory for travel	–	22 (7.01)

The multivariable analysis reported that age (years), education, occupation and infected family members were significantly associated with COVID-19 vaccination status. The age group 34–42 years had greater odds of getting vaccinated. The odds of COVID-19 vaccination among the age group 34–42 years were 1.75 times higher (CI 1.35 to 2.09, p=0.008) compared with the age groups 18–28, 29–33 and >42 years. Similarly, the odds of COVID-19 vaccination among people with postgraduate education were 2.24 times higher (CI 1.80 to 2.73, p<0.001) compared with people who graduated and those with schooling and no schooling. The odds of COVID-19 vaccination among healthcare workers were 1.88 times greater (CI 1.30 to 2.05) and 1.69 times higher (CI 1.27 to 1.91) among educationists and 1.90 times higher (CI 1.43 to 2.11) in self-employed individuals compared with business people, unemployed and other occupations (p<0.001). The odds of COVID-19 vaccination among those with COVID-19 ever-infected family members were 1.87 times higher (CI 1.56 to 2.34, p=0.032) compared with those with uninfected family members ever (table 4).

Table 4

Univariate and multivariable analyses reporting crude and adjusted OR for COVID-19 vaccination status (N=541)

Characteristics	Crude OR (95% CI)	P value	Adjusted OR (95% CI)	P value
Age (years)
18–28	1	0.065*	1	0.008†
29–33	1.76 (1.24 to 1.99)		1.59 (1.24 to 1.89)
34–42	1.91 (1.49 to 2.25)		1.75 (1.35 to 2.09)
>42	1.54 (1.33 to 1.87)		1.67 (1.29 to 197)
Gender
Female	1
Male	1.29 (1.22 to 1.82)	0.136*	–	–
Education
No schooling	1	<0.001*		<0.001†
Postgraduate	2.96 (1.57 to 3.33)		2.24 (1.80 to 2.73)
Graduate	1.71 (1.18 to 2.00)		1.96 (1.42 to 2.23)
SSC/HSC	1.38 (1.05 to 1.86)		1.70 (1.29 to 1.90)
Middle school	1.43 (1.15 to 1.96)		1.40 (1.12 to 1.67)
Occupation
Unemployed	1	<0.001*	1	<0.001†
Business	1.36 (1.02 to 1.95)		1.22 (1.06 to 1.65)
Educationist	1.80 (1.20 to 2.06)		1.69 (1.27 to 1.91)
Healthcare worker	2.42 (1.43 to 3.11)		1.88 (1.30 to 2.05)
Other	1.13 (1.08 to 1.91)		1.24 (1.17 to 1.87)
Self-employed	1.81 (1.16 to 2.04)		1.90 (1.43 to 2.11)
Infected ever with COVID-19
No	1	0.665
Yes	1.65 (0.87 to 1.88)		–	–
Infected family member ever
No	1	0.211*	1	0.032†
Yes	1.23 (1.04 to 1.56)		1.87 (1.56 to 2.34)
Taken care of an infected family member ever
No	1	0.359		–
Yes	0.93 (0.67 to 1.45)		–

*Significant at the univariate level.

†Significant at the multivariable level.

Discussion

This study finds that about 58% of the individuals were vaccinated against COVID-19 vaccination, with most youngsters, compared with other studies, reporting lower coverage in Pakistan. Education level and previous infection played a significant role in vaccine acceptance, while not believing in vaccination was the primary reason for hesitancy.

Most individuals opting out of getting vaccinated against COVID-19 did not believe the vaccine would work and were waiting for a ‘better’ vaccine. Those opting to receive the vaccine were primarily driven by securing their family’s safety and protection. Being middle-aged, obtaining higher education, being self-employed or a healthcare worker, and caring for an infected family member were significantly associated with COVID-19 vaccination status.

The ‘3C’ model has historically explained vaccine hesitancy. This model explores three significant factors for hesitancy: complacency, confidence and convenience.20 In this study, we see that complacency and confidence played a role in the choice of vaccination made in our sample. Those concerned about their and their family’s protection were vaccinated (complacency). Similarly, poor trust in the vaccine was seen in those not vaccinated (confidence). Pakistan has a history of commonplace misconceptions and widespread distrust in vaccinations; the classic example is the polio vaccine.21 22 A complex interplay of religious, social and media influence and logistical issues made herd immunity challenging to achieve. The COVID-19 vaccines, which are new vaccines developed by foreign countries and have never been tested on large populations, are also suspected to play a role in vaccine hesitancy.

Our results showed that 42% of our sample were not vaccinated. This is contrast to the stats by government of Pakistan which reported the coverage of 38%, with 110 million individuals immunised with the first dose, as the government of Pakistan reported in the first week of February (https://covid.gov.pk/). This is similar to the surrounding region, with Umakanthan et al12 in a study from India reporting that nearly half of all Indians were planning on getting vaccinated. In this study, the most typical reason for refusing vaccination was poor confidence in vaccination, with fear of side effects.12 A survey of 15 countries also found that, in lower-middle-income countries, protection from COVID-19 disease and fear of vaccine side effects were the two significant factors for accepting or rejecting vaccines.23 Similarly, in a study conducted in Saudi Arabia, an Islamic country, Al-Mohaithef et al24 found that older individuals (older than 45 years) and those with a postgraduate degree were more likely to accept the COVID-19 vaccine. This suggests that older age group and education affect vaccine perceptions positively.

This study is foundational and future research can help explore our population’s behaviour towards vaccine hesitancy. Identifying people who are less likely to get vaccinated and identifying reasons for their vaccination decisions can help healthcare and government systems address poor vaccine coverage. Targeted interventions in such populations,10 such as dispelling myths and tracking down and regulating sources of false information fuelling negative perceptions about vaccines, can help increase vaccine acceptance. This will increase vaccine coverage, a crucial ingredient of herd immunity.24 25

Strength and limitations

This study reported vaccine hesitancy and acceptance in the general population of Pakistan. Also, this study is among the studies that have catered to both urban and rural populations, including all occupational categories. Most of the surveys in Pakistan have focused on healthcare workers. Our sample population was not biased towards any specific institution or age group. The vaccinated and non-vaccinated groups had comparable demographics. We translated our data collection tool into various local languages. Thus, our sample is more diverse due to fewer language barriers. Moreover, the results of this study are generalisable to the whole population.

This study has some limitations. Although the survey has a large sample size, our numbers cannot compare with many other studies due to lack of resources. Moreover, these data were gathered from self-reported information. We did not check vaccination records, which might have caused recall bias. Thus, our sample may not fully represent the general population of Pakistan; however, our results function as preliminary findings on which further studies can be established.

Conclusion

Targeted interventions for subsets of people more likely to refuse vaccination can improve vaccine coverage. Future research must explore the factors affecting COVID-19 vaccine hesitancy in the Pakistani population.