Knowledge, attitudes and human papillomavirus vaccine recommendation practices among primary care physicians in Bahrain: a cross-sectional study

Technical Summary: Knowledge, Attitudes, and HPV Vaccine Recommendation Practices Among Primary Care Physicians in Bahrain

Problem Statement
Despite the World Health Organization's target for 90% HPV vaccination coverage among girls by age 15, global progress remains slow, with population-weighted coverage estimated at only 15% as of mid-2020. In the Gulf Cooperation Council (GCC) region, evidence regarding HPV vaccine uptake is limited and uneven, with most settings reporting uptake below 10%. While physician recommendation is a primary determinant of vaccine uptake, little is known about the frequency of such recommendations by primary care physicians in Bahrain. Although Bahrain introduced a national two-dose HPV vaccination program for adolescents (ages 12–13) via school campaigns in October 2023, primary care physicians are not the primary vaccinators. Instead, their role is largely consultative, particularly for parents who are hesitant. The study addresses the gap in understanding how often these physicians recommend the vaccine, their knowledge and attitudes toward it, and the factors influencing their recommendation practices.

Methodology
This was a cross-sectional study conducted between July and September 2025 across governmental primary health care centers in all four governorates of Bahrain.

• Participants: The study surveyed 205 physicians (response rate 90%), including family medicine consultants, specialists, residents, and general practitioners. The sample was proportionally distributed across governorates.
• Instrument: A self-administered, printed questionnaire in English was used. It comprised:
  • Demographics: Age, gender, experience, position, marital status, and children.
  • Knowledge: 24 items covering HPV infection, associated cancers, vaccine specifics, and Bahrain's national program (Kuder–Richardson 20 = 0.61). Scores were classified as poor (<60%), fair (60–79%), or good (≥80%).
  • Attitudes: Six items on a 5-point Likert scale, analyzed as two subscales: "Personal Endorsement" (safety, efficacy, personal willingness) and "Perceived Acceptance" (parental and community acceptance).
  • Practice: Frequency of recommendation in the past six months (Never, Rarely, Sometimes, Often, Always).
• Outcome: The primary outcome was "regular recommendation," defined as recommending the vaccine sometimes, often, or always.
• Analysis: Bivariate analyses (chi-square, Fisher's exact, t-tests) were followed by multivariable logistic regression to identify independent factors associated with regular recommendation. Sensitivity analyses included cluster-robust standard errors, ordinal logistic modeling, and the addition of confidence scores.

Key Results

• Knowledge: The mean knowledge score was 19.1/24 (79.5%). While nearly all physicians knew HPV causes cervical cancer (99.5%), significant gaps existed regarding practical program details: only 33.7% knew the specific vaccine type (nonavalent) used, 70.7% knew the target age group, and 76.1% knew the dose schedule. Knowledge was not associated with recommendation frequency.
• Attitudes: Physicians held favorable views of the vaccine itself (81.0% deemed it safe; 88.8% effective; 77.6% supported routine inclusion). However, perceived acceptance was low: only 28.8% believed parents usually accept the vaccine, and 36.1% believed the community would.
• Practice: In the previous six months, 35.6% of physicians never recommended the vaccine, 28.3% did so rarely (once or twice), and only 36.1% recommended it regularly.
• Predictors of Recommendation: Multivariable logistic regression identified two significant independent factors associated with regular recommendation:
  1. Clinical Experience: Physicians with 5–15 years of experience (aOR 5.34) and >15 years (aOR 5.15) were approximately five times more likely to recommend regularly compared to those with ≤4 years.
  2. Perceived Acceptance: Each one-point increase in the perceived parental/community acceptance score was associated with a 33% increase in the odds of regular recommendation (aOR 1.33).
  • Non-factors: Knowledge score, gender, marital status, and attendance at Continuing Medical Education (CME) were not significantly associated with regular recommendation in the adjusted model.

Significance and Claims
The authors claim this is the first study to examine HPV vaccine recommendation practices among primary care physicians in Bahrain. The study highlights a disconnect between physician knowledge/attitudes and actual practice. Specifically, the authors conclude that:

1. Experience and Perception Drive Practice: Recommendation behavior is driven more by clinical experience and the physician's expectation of parental acceptance than by their factual knowledge of the vaccine.
2. Pessimistic Bias: Physicians' expectations of parental refusal appear more pessimistic than reality. Companion data suggests 80.6% of Bahraini parents would be encouraged by a recommendation, and 70% of those offered the vaccine accepted it.
3. Structural Context: Because the vaccine is delivered via school campaigns, physicians have few routine opportunities to recommend it; their role is primarily counseling hesitant parents.
4. Implications: The authors suggest that educational interventions focusing solely on general knowledge are unlikely to change practice. Instead, they propose targeted training for early-career physicians (who are less likely to recommend), updates on specific national program details, and feedback mechanisms to correct physicians' pessimistic expectations regarding parental acceptance. They also note that presumptive recommendation strategies and system-level prompts (e.g., electronic health record reminders) could be effective.

The study acknowledges limitations, including its cross-sectional design (precluding causal inference), self-reported data, and the exclusion of private practice physicians. However, the consistency of findings across sensitivity analyses and the inclusion of all four governorates support the robustness of the observed associations.