Introduction

The persistence of poliovirus in Pakistan is undermining the efforts of the Global Polio Eradication Initiative.[1] Significant barriers responsible for the persistence of polio in Pakistan are vaccine hesitancy and refusals by parents.[2,3] Vaccine refusals by families were responsible for 40% of the Wild Polio Virus cases occurring in 2014.[4] Negative attitudes toward polio vaccines are not only driving these refusals[5] but they are also leading to violent attacks on polio workers.[6]

An incident in 2019 demonstrates the risk posed by anti-polio vaccination attitudes in Pakistan. In the Khyber Pakhtunkhwa (KPK) province, an angry mob burnt down a healthcare unit after a viral hoax stating that children were rushed to hospital with various medical complaints following polio vaccination.[7] Such serious implications underscore the need for assessing vaccine related attitudes. Empirical assessment of vaccine related attitudes can also help in predicting the likelihood of getting vaccinated because attitudes exert influence on behavior.[8] In the long run, this assessment can facilitate the development of more tailored eradication interventions.[5]

The first step in this endeavor requires the utilization of a psychometrically sound instrument for measuring attitudes toward polio vaccination as they exist in Pakistan. There are three similar measures of attitude toward polio immunization available.[5,9,10] However, the internal reliabilities of these instruments are low to moderate and they have been developed for specific samples (i.e., health workers, religious scholars, and residents of Quetta/Peshawar). The validation procedures for these scales only involved expert reviews and try-outs with small samples.

There are also several measures[11,12] for measuring general vaccination attitudes. However, the tool measuring attitudes toward polio vaccination can benefit from being more indigenized as these attitudes are shaped by the unique socio-political context of Pakistan.[13] The nature of this particular context has been explored by previous studies.

Polio vaccine refusals in Pakistanis, more specifically, Pakhtuns were significantly increased when the U.S. killed Osama bin Laden, the former Al-Qaeda chief, in a military operation in Pakistan by reportedly using a fake hepatitis vaccination campaign to trace bin Laden.[13] These refusals have been more prevalent among ethnic groups that inhabit regions like KPK and Balochistan, where conflict and insecurity is common.[3] These conditions exacerbate uncertainty which can stem from the immediate situation or the overall social environment.[14]

People perceive getting vaccinated to be riskier than not getting vaccinated under conditions of uncertainty.[15] Uncertainty about vaccination itself can lead to vaccine hesitancy and mistrust.[16] Furthermore, uncertainty about one aspect of life can influence other types of uncertainty.[17] Therefore, it is expected that such uncertainties influence self-uncertainty-uncertainty about one’s attitudes, beliefs, values, and place in the world which produces such aversive feelings that one is motivated to reduce them.[17]

One way of reducing self-uncertainty is becoming attached to a group or group identification,[18] a process which makes extremist propaganda more persuasive[19] and increases beliefs in conspiracy theories because people are only worried about conspiracy theories when their group identification is strong and they fear hostility from another group.[20] Conspiracy theories on their own can combat uncertainty by giving meaning to events beyond comprehension.[21] This is evidenced by the fact that people, throughout history, have shown more belief in conspiracy theories during times of conflict and crises like natural disasters, societal change, and wars.[22] Thus, it is likely that anti-vaccination conspiracy theories become more persuasive under conditions of uncertainty. This assumption holds serious implications for the polio vaccination campaign in Pakistan, a country where the tendency to believe in general conspiracy theories is prevalent.[23]

This general tendency to believe in conspiracy theories significantly increases anti-vaccination attitudes.[24] Belief in specific anti-vaccination theories also predicts anti-vaccination attitudes.[25] Since belief in specific conspiracy theories depends on the general tendency to hold such views[26] - a general conspiracy mentality.[27] Hence, it can be predicted that a conspiracy mentality negatively influences attitudes toward polio vaccination.

In Pakistan, anti-vaccination propaganda widely spreads through social media[28] and poses a significant hurdle in the eradication of polio by peddling various conspiracy theories.[13] These conspiracy theories are usually about the content, efficacy, and harmful effects of the polio vaccine[2,29,30] such as rumors that the vaccine contains pig fat, which Muslims are forbidden to consume, or that, polio vaccines cause infertility.

Statically testing the links of these variables with attitudes towards polio vaccination can help form efficacious intervention designs. However, to the best of the author’s knowledge, almost all studies about polio vaccination attitudes in Pakistan have been qualitative or descriptive.[1,2,5,30-34] Another avenue worth exploring is the difference between attitudes of men and women toward polio vaccination because these studies have not been able to find significant gender differences.[5]

Due to the significance of attitudes toward polio vaccination and the paucity of research on their psychological influencers, the present research has taken the initiative of exploring the dimensions of these attitudes. This research aims to develop an attitude toward polio vaccination scale (ATPVS) and establish its psychometric properties. Due to evidence of their link with attitudes toward vaccination uncertainty and conspiracy mentality have been explored as correlates and indicators of convergent validity.

Methods

The research commenced after approval was granted by the ethical review committee (IRB: F.No.D-107-02(02)M.Phil/Fall-2018-Admin Permission granted December 11, 2019) at the National Institute of Psychology, Quaid i Azam University, Islamabad, Pakistan. The development of ATPVS was undertaken in two phases; phase 1 was about scale development and phase 2 consisted of scale validation procedures including exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and testing of convergent validity.

Results

This section contains results of various validation procedures carried out for the ATPVS.

EFA

Further analysis was completed on the remaining 26 items to determine the underlying dimensions by means of EFA. First half of the sample was used for this analysis. Initially, the Kaiser-Meyers-Olkin (KMO) measure of sampling adequacy and Bartlett’s Test of Sphericity were computed to check the sampling adequacy (n = 310). KMO = 0.83 illustrated that correlations were compact enough to generate distinct and reliable factors with Bartlett test of Sphericity χ^² (325) = 2120.73 significant at P < 0.001 showing that the data was good enough for carrying out EFA. After the convergence of factor solution in 50 iterations, seven factors with Eigenvalues <1 were suggested. Since these seven factors were neither comprehensive nor interpretable, the Varimax method of rotation was used for the purpose of extracting meaningful factors. Subsequently, three more factor structures consisting of 2, 3, and 4 factors were extracted. SMEs were recruited to assess the meaningfulness of all these factor structures.

The 4-factor solution was discarded after the consensus of the majority of SMEs (n = 7) around 2 and 3-factor structures being more meaningful. Upon testing the alpha reliabilities of the three factors, reliability coefficients turned out to be 0.84, 0.74, and 0.40. Weak alpha value of the third factor did not improve even after omitting a few items. Therefore, the three-factor solution was also discarded. Finally, a two-factor solution was tested whose factor loadings are presented in Table 1.

Table 1 Factor loadings of 26 items through principal component analysis using the varimax method (n=310)

In the two-factor solution presented in Table 1, only items that had factor loadings of 0.30 were retained because an item loading of about 0.30 or more is considered significant for a sample size similar to the present study.[45] Items 1 and 11 were discarded for having low factor loadings. Whereas, items 7 and 24 were discarded for cross-loading on both factors I and II. The final solution containing Factor I and Factor II were not only qualitatively meaningful, but the alpha coefficients of the two factors were above 0.70.

The ATPVS consisting of 22 items was finalized as consisting of two dimensions or factors. Factor I was named Scepticism while Factor II was named Advocacy, as per the suggestion of SMEs. Items of Scepticism reflected doubts about the safety of polio vaccines, concern about their harmful effects, and belief in conspiracy theories about polio vaccines. Items of Advocacy indicated belief in the effectiveness of polio vaccines and support for the vaccination campaign. Reliability analysis revealed alpha coefficients of 0.83 and 0.73 for Scepticism and Advocacy, respectively. The inter-subscale correlation value showed a significant association of −0.27 between the two factors.

Confirming the factor structure of ATPVS

To assess the psychometric meaningfulness of two factors of ATPVS and to test its validity for use with a heterogeneous sample, the second half of the sample, containing 310 adult participants was used. AMOS 22 statistical package was used to conduct CFA with maximum likelihood. The various model fit indices computed to evaluate this model are presented in Table 2.

Table 2 Confirmatory factor analysis (indices of model fit) for attitude toward polio vaccination scale (n=310)

Initially in the default model, hypothesized CFA for ATPVS yielded unsatisfactory model fit indices. For a model to be considered acceptable, the NFI, GFI, and CFI should be <0.90, 0.90, and 0.93 respectively.[46] Whereas, the RMSEA value should be >0.08[47] and χ^²/df value should be >2 or 3.[48] To improve these indices, error covariance’s were drawn between all possible error terms suggested by modification indices [Figure 1].

Close

Figure 1

Measurement model of ATPVS

Export to PPT

After adding these error covariance’s, all model indices fell within the acceptable range except for NFI. The standardized factor loadings for Scepticism and Advocacy are given in Table 3.

Table 3 Factor loadings of CFA for attitude toward polio vaccination scale (n=310)

It is evident from Table 3 that all items possess acceptable factor loadings, that is, >0.40,[49] on the two subscales of ATPVS, Scepticism, and Advocacy. Thus, the factor structure of ATPVS, as proposed by EFA, has been confirmed through a CFA on an independent sample. Furthermore, Cronbach’s alphas of Scepticism and Advocacy calculated for this sample showed α-values of 0.91 and 0.82, respectively.

Frequency analysis for ATPVS items

The data of the total rom two previous samples (i.e., on-field and online) were combined for frequency analysis of responses to ATPVS [Table 4]. This data of 620 individuals consists of Pakistani adults who possess English literacy.

Table 4 Frequencies and percentages of Likert-type responses on attitude toward polio vaccination scale (n=620)

Convergent validity

Pearson Product Moment was used to assess the links of self-uncertainty, conspiracy mentality, and the two components of attitude toward polio vaccination, scepticism, and advocacy. The correlations coefficients of these variables along with the descriptive properties and alpha reliabilities for their measures have been tabulated [Table 5].

Table 5 Descriptive statistics and alpha reliability coefficients of scales (n=620)

The correlation matrix suggests that scepticism and advocacy have a significantly negative relationship between themselves. Conspiracy mentality has a positively significant relationship with both components of ATPVS; therefore, an increase in conspiracy mentality is associated with increases in both scepticism and advocacy of polio vaccination. Moreover, increase in self-uncertainty is associated with an increase in scepticism and a decrease in advocacy.

Gender Differences in attitudes towards polio vaccination

Independent-sample t-test was utilized to assess gender differences in attitude towards polio vaccination, self-uncertainty, and conspiracy mentality [Table 6].

Table 6 Gender differences on scepticism toward polio vaccination, advocacy of polio vaccination (n=620)

It is evident that gender differences in attitude toward polio vaccination are present but with a small effect size.[44] Women are showing less scepticism and more advocacy of polio vaccination as compared to men as evident from the group mean scores.

Discussion

There is a dearth of instruments that adequately address attitudes towards polio vaccination. Therefore, this research took the initiative of developing ATPVS to accurately reflect indigenous attitudes toward polio vaccination. An EFA and a subsequent CFA resulted in the most significant finding of this research: Attitudes toward polio vaccination in Pakistan are reflected by two dimensions; skepticism and advocacy. Skepticism covers doubts about the safety of polio vaccines, concern about their harmful effects, and belief in conspiracy theories about polio vaccines. Such concerns have also been identified by previous studies (Closser et al. 2015; Khan et al. 2020).[2,31]

The factor, advocacy, covers support for the vaccination campaign. This aspect is unique in the sense that it covers advocacy of polio vaccination on a country-wide level (e.g., ”polio workers should be facilitated during polio vaccination campaigns” and “outreach of polio vaccines should be increased all over the country). This reflects the social welfare aspect of vaccination[52] something that goes beyond personal concern. Thus, an effective campaign can benefit from a two-pronged approach: Tackling misinformation about polio vaccination and appealing to concern for social welfare.

Next, it was observed through a frequency analysis of responses to ATPVS by the combined sample that most participants tended to agree with items of the advocacy subscale, thereby, indicating their support for polio vaccination campaign and their willingness to participate in it. This finding could have occurred due to certain sample characteristics – being young, single and mostly hailing from Punjab – associated with positive attitudes towards polio vaccination.[51]

For the Skepticism scale, it was observed that participants disagreed with statements of most items, thereby, indicating their positive attitudes. However, participants showed agreement to two items: Item 1 that states that the problem of polio has been exaggerated, item 7 that states that the vaccine process is repeated so it can provide financial benefit to the vaccine developers. This finding provides proof for the two-factor structure of ATPVS in which skepticism and advocacy are not simply opposite constructs. Thus, participants who might be showing more advocacy of polio vaccination can also hold certain skeptical attitudes. However, it is also likely that these two items may be representing factual statements that individuals might be agreeing with, irrespective of their positive attitudes. Further analyses of these items are required to confirm this possibility. Nevertheless, even a few such doubts can indicate vaccine hesitancy which can soon turn into refusals.[52] Vaccine refusals from even a few individuals can result in the emergence of unvaccinated clusters where a disease outbreak might occur.[53]

Converge validity was established by exploring the relationship of ATPVS with Self-Uncertainty Scale[41] and CMQ.[42] Results revealed that self-uncertainty was positively related with skepticism and negatively related with advocacy, thereby, supporting the hypothesis that uncertainty is associated with negative vaccination attitudes. It should be noted that self-uncertainty can bleed into other uncertainties and vice-versa.[17] Hence, if uncertainty related to vaccination can lead to vaccine hesitancy and mistrust[17] then self-uncertainty is likely to have a negative effect as well. The findings of this study regarding self-uncertainty are also in line with the proposition that uncertainty can strengthen belief in conspiracy theories.[20]

The need to assess vaccine decision-making in the presence of uncertainty has significant implications for Pakistan because uncertainty prevails in low-income emerging economies[54] and this need has now become even more salient in the conditions where global uncertainty is at record high levels due to Coronavirus.[55]

Next, results reveal that conspiracy mentality is associated with an increase in skepticism. This finding is in line with the hypothesis linking beliefs in conspiracy theories and anti-vaccination attitudes.[24,25,56] This construct holds implications for vaccination attitudes in general and also for polio vaccines. Negative propaganda regularly spreads online and promotes conspiracy theories[13] which often lead to mass panic[57] and acts of violence.[34] However, dealing with individual characteristics that make people susceptible to negative attitudes should be a greater priority for the government than curbing few factions peddling anti-polio vaccination propaganda.[58]

Coming to advocacy, its positive relationship with conspiracy mentality seems to go against the hypothesis that conspiracy mentality is associated with negative attitudes. However, there are arguments that conspiracy theories are not always associated with negative outcomes.[21] For example, conspiracy theories encourage protest movements[27] and bolster support for democratic principles.[59] This tendency also represents a desire to challenge the status quo to in turn seek truth and social advancement.[60] This social concern is already linked with decision-making about vaccines.[50]

Finally, it has been observed that women show more advocacy and less skepticism towards polio vaccination as compared to men. These small but significant gender differences in attitudes negate previous findings which found no significant gender differences in attitude towards polio immunization.[5] Future studies should recruit a more diverse sample with the help of translated instruments to discover more differences not just along the lines of gender but also demographics such as provincial background, marital status, and socio-economic level.

Conclusion

The present research has discovered underlying dimensions of attitudes toward polio vaccination and has led to the development of a psychometrically sound instrument called the ATPVS with its two subscales: Scepticism and advocacy. Self-uncertainty and conspiracy mentality have been established as correlates of these factors. The insight gained from this study can help in identifying individuals and contexts that increase susceptibility to anti-polio vaccination propaganda. In the long run, more individualized campaigns can be designed and implemented to help in achieving the complete eradication of polio from Pakistan.

Authors’ Contributions

WA, HJ, MK, and GMK were responsible for composing the manuscript; WA and HJ were responsible for conceptualizing the research question, data collection, and statistical analysis. WA, HJ, MK, and GMK also edited and reviewed the manuscript. All authors have read and agreed to the published version of the manuscript.

Ethics Approval and Consent to Participate

All procedures carried out in this study were approved by the ethical committee at the National Institute of Psychology, Quaid-i-Azam University, Islamabad (IRB Number: F.No.D-107-02(02)M.Phil./Fall-2018-Admin Permission granted December 11, 2019). Written informed consents were taken from all participants.

Availability of Data and Material

The data that support the findings of this study will be provided by the corresponding author upon reasonable request.

Consent for Publication

All authors of this study gave their consent for publication.