Disparities in collaborative patient-provider communication about HPV vaccination Jennifer L. Moss, PhD 1,2, Melissa B. Gilkey, PhD 3, Barbara K. Rimer, DrPH 1, Noel T. Brewer, PhD 1 1 University of North Carolina 2 National Cancer Institute 3 Harvard Medical School May 20, 2016
Disclosures JLM, MBG, and BKR have no potential conflicts to report. NTB has served on paid advisory boards or received research grants from Merck, GSK, CDC, and FDA, and now serves as chair of the CDC-funded National HPV Vaccination Roundtable. This research was supported by a grant from National Cancer Institute (F31 CA189411; PI: Moss) and data support from the National Center for Health Statistics. Opinions do not necessarily reflect the official positions of the NCI or the NCHS. 2
Introduction: HPV Infection can lead to Genital warts (>300k cases/yr in US) Cancer (26k cases/yr in US) High-priority populations Racial/ethnic minorities Un/underinsured Vaccine Lack of health disparities in uptake Suboptimal uptake overall U.S. HPV vaccine initiation, 2014 Girls 60% Boys 42% 3
Introduction: Patient-provider interactions Provider recommendation Communication style? Paternalistic v. collaborative Association with HPV vaccination: Mixed findings Differences in communication style by patient characteristics 4
Methods Collaborative communication Demographics HPV vaccine initiation National Immunization Survey-Teen, 2010 n=4,124 parents of girls Measures Collaborative communication (Charles et al., 1997, 1999, 2006) HPV vaccine initiation (verified by healthcare providers) Formal test of mediation (Iacobucci, 2012; Sobel, 1982) 5
Results: Participant characteristics (n=4,124) Non-Hispanic white 61% Privately insured 62% Mother with >HS degree 64% At/above poverty 78% Urban 85% 6
% initiating HPV vaccine Results: Communication and vaccination Overall HPV vaccine initiation: 49% 100% 80% *** OR=4.70 (95% CI=4.12-5.36) 60% 40% 20% 0% Non-collaborative communication Collaborative communication ***p<.001 7
NH White v Hispanic Private v public insur Mom > HS v mom HS At or above v below pov Urb/suburb v rural % reporting collaborative communication Results: Collaborative communication Overall prevalence = 53% 100% ** *** * ** ** 80% 60% 40% 20% 0% *p<.05 **p<.01 ***p<.001 8
Results: Mediation (I) Ethnicity Collaborative v non-collaborative -0.69* 1.07 Hispanic v non-hispanic white 0.56* (0.38) Vaccination v no vaccination Mediation z = -2.94* Controlling for provider recommendation for vaccination. *p<.05 9
Results: Mediation (II) Health insurance Collaborative v non-collaborative -0.35* 1.05* Public v private insurance 0.36* (0.28*) Vaccination v no vaccination Mediation z = -2.28* Controlling for provider recommendation for vaccination. *p<.05 10
Summary a = - Collaborative communication b = + Underserved groups c = ++ (c = +) HPV vaccine initiation Underserved group z -score for mediation Hispanic ethnicity -2.94* Public insurance -2.28* Below federal poverty level 1.83 - Spanish speaking -1.40 - Lower maternal education -1.14. p<.10 *p<.05 11
Strengths and limitations Strengths Nationally-representative sample Provider-verified vaccination status Limitations Cross-sectional analysis Parental self-report of communication 12
Conclusions HPV vaccination Higher when parents reported collaborative communication Equal or higher in high-priority groups Evident disparities in collaborative patient-provider communication Lower in high-priority groups Disparities in communication suppressed differences in vaccination Equalizing communication styles across groups could result in more favorable vaccination outcomes overall and for high-priority groups. 13
Jennifer L. Moss, PhD Cancer Prevention Fellow National Cancer Institute jennifer.moss@nih.gov @Jennifer_L_Moss Moss, J. L., Gilkey, M. B., Rimer, B. K., & Brewer, N. T. (in press). Disparities in collaborative patient-provider communication about human papillomavirus (HPV) vaccination. Human Vaccines & Immunotherapeutics. 14
cpfp.cancer.gov www.cancer.gov
Mediation results (III): Poverty Below v above poverty level Collaborative v non-collaborative -0.42* 1.55* 0.53* (0.31) Vaccination v no vaccination Mediation z = -2.59* *p<.01 16
Mediation results (IV): Rurality Collaborative v non-collaborative 0.36* 1.55* Rural vs urban 0.08 (0.18) Vaccination v no vaccination Mediation z = 2.87* *p<.01
Percent initiating HPV vaccination Results (V.a.): Moderation 80 +46% +33% 70 60 50 40 30 20 Non-collaborative communication Yes Collaborative communication 10 0 Rural Moderation p=.02 Urban/suburban
Percent initiating HPV vaccination Results (V.b.): Moderation 80 +44% +16% +40% +18% 70 60 50 40 30 20 Non-collab. communication Yes Collaborative communication 10 0 Non-Hispanic white Non-Hispanic black Hispanic Other Moderation p=.01
Odds ratio of HPV vaccination Step 1: Demographics & vaccination Dem. Comm. Vacc. 2.5 Hispanic (v. non-hispanic white ref.) OR=1.30 Public ins. (v. private insurance ref.) OR=1.11 Below pov. level (v. at/above poverty ref.) OR=1.36 2.0 1.5 1.0 0.5 Well-served Underserved 0.0 Race/ethnicity Insurance Poverty 20
Odds ratio of reporting collab. comm. Step 2: Demographics & communication Dem. Comm. Vacc. 2.5 Hispanic (v. non-hispanic white ref.) OR=0.54* Public ins. (v. private insurance ref.) OR=0.63** Below pov. level (v. at/above poverty ref.) OR=0.66* 2.0 1.5 1.0 0.5 0.0 Race/ethnicity Insurance Poverty Well-served Underserved *p<.01 **p<.001 21
Step 3: Communication & vaccination Dem. Comm. Vacc. HPV vaccination: 49% HPV vax OR (95% CI) DID NOT report collaborative communication 31% ref. DID report collaborative communication 66% 4.70 (4.12-5.36) (OR is controlling for provider recommendation for vaccination) 22
Step 4: Demographics, communication, and vaccination a = - Collaborative communication b = + Demographics (underserved groups) c = ++ (c = +) HPV vaccination 23
Odds ratio of HPV vaccination Differences in vaccination odds 3.0 Hispanics (v. non-hispanic white ref.) z = -3.04* Public ins. (v. private insurance ref.) z = -3.67* Below pov. (v. at/above poverty ref.) z = -2.59* 2.5 2.0 1.5 1.0 0.5 0.0 c c' c c' c c' *p<.01