By measuring parents’ attitudes regarding disease salience and community benefit, high scores on a four-item scale was associated with fivefold greater likelihood of not fully vaccinating their children, according to a study published in the.
Mohammad Tahir Yousafzai, MPH, of Aga Khan University, Karachi, Pakistan, and colleagues developed a larger 14-item scale to measure parental attitudes and surveyed 901 households in the Sindh province of Pakistan during 2014. Part of this scale was a short 4-item subscale focusing on disease salience and community benefit, whereas the remaining 10 items form another subscale that measures parents’ perceptions and concerns regarding vaccines directly. The items are presented as 1-5 Likert scales, and scoring higher represents holding more negative attitudes regarding vaccines.
Of the 901 households surveyed, 25% of children were fully vaccinated, which meant children received all primary vaccines up to 14 weeks of age, and 54% were partially vaccinated, which meant at least one of those primary vaccines had been missed. The remaining 21% were unvaccinated.
High scores on the full 14-item scale showed some correlation with no vaccination versus partial or full vaccination (odds ratio, 3.05; 95% confidence interval, 1.75-5.31); the association disappeared after adjustment for children’s age and gender. The subscales performed better after adjustment: The adjusted ORs were 1.52 (95% CI, 1.05-2.21) for the longer subscale and 5.21 (95% CI, 3.60-7.55) for the shorter subscale. The data also showed high association between high scores on the shorter subscale and the likelihood of no or only partial vaccination (aOR, 9.65; 95% CI, 4.81-19.37).
The researchers noted that most similar scales used in developed countries are longer (10 or more items) and have lower internal consistency. This four-item scale, on the other hand, may be especially useful among lower-income populations and those in developing areas that have lower literacy rates.
One of the limitations of the study was that it was conducted only in Pakistan and not multiple developing countries; the researchers acknowledged this could limit generalizability. Another limitation is that the study size was too small for subdomain analysis; the researchers wrote that, although a lack of such analysis is unfortunate, it shouldn’t hamper the shorter subscale’s usability. A further limitation is that there was little variability across Likert scores – mostly answers at the extreme ends rather than a mix. This suggests that interviewees may not have understood how Likert scales work and therefore may not have answered accurately, noted the researchers, who employed a visual chart, trained interviewers, and field monitoring to mitigate this possibility.
“Measurement of the parental attitudes toward childhood vaccination is very important for the appropriate planning of strategies for increasing vaccine coverage and for monitoring,” they wrote.
The study was sponsored by Gavi, The Vaccine Alliance. The authors had no funding or conflicts of interest to disclose.
SOURCE: Yousafzai MT et al. .