Mixed-Methods Research in HIV Primary Prevention: Mixed-Methods Research in HIV Primary Prevention Eric G. Benotsch, Ph.D. Dept. of Psychology University of Colorado at Denver & Health Sciences Center, Denver, CO USA


Slide2: Acknowledgements Hanoi School of Public Health Steve Koester, Ph.D. Steve Pinkerton, Ph.D. David Seal, Ph.D. John Mikytuck, B.S. Chris Nettles, M.A. Jean Redmann, B.A. Salvatore Seeley, M.S.W. Felicia Wong, MPH Kathleen Ragsdale, Ph.D.


Training Background: Training Background Clinical Psychology doctorate training program at the University of Iowa Quantitative Tradition (research) Qualitative Methods (clinical work) Postdoctoral training at the Medical College of Wisconsin Qualitative Methods Mixed-Methods


Quantitative Inquiry: Quantitative Inquiry Uses numerical data to understand human experience Collected via surveys or questionnaires Counts Relative Scores Well-described analytic tradition ANOVA Regression


Qualitative Inquiry: Qualitative Inquiry Uses narrative data to understand human experience Collected through face-to-face interviews, focus groups, or observation Moderately well-described analytic tradition Coding (based on a priori questions, emergent ideas) Examination of themes and patterns Examination of exceptions to typical patterns


Mixed-Methods Inquiry: Mixed-Methods Inquiry Intelligently and strategically combines quantitative and qualitative methodology Recognizes that no single research approach is superior overall Each has advantages and disadvantages Research should be question-driven not method driven (pragmatist philosophical approach) Tools


Qualitative Methods -- Advantages: Qualitative Methods -- Advantages Richness of data Can be used when little is known (may be especially useful early in research for understanding a complex phenomenon) Useful for developing rapport with a community Can be used when want to understand how members of a community interpret results Compelling for some


Qualitative Methods -- Disadvantages: Qualitative Methods -- Disadvantages Labor-intensive Therefore, typically smaller samples Purposively selected – lower generalizability Time-consuming – dissemination of results may take longer than quantitative Not compelling for some


Quantitative Methods -- Advantages: Quantitative Methods -- Advantages Less labor intensive Facilitates collecting larger numbers Greater generalizability More rapid dissemination of results May be especially useful early in research for documenting a problem Compelling for some


Quantitative Methods -- Disadvantages: Quantitative Methods -- Disadvantages Requires good understanding to design appropriate measures Complexity of ideas transmitted is reduced Little give-and-take Does not capture richness of detail Not compelling for some


Slide11: Stigma Research "If you are in the general community [and are HIV-positive], then I would say that you are probably going to be shunned, you are probably going to be ridiculed, people are not going to touch you, they are not going to want you in their restaurants, they will not hire you to work for them…all the issues are there and they are very obvious. Now, if you are providing services for these people as a professional, it is very different. People come here to this agency…just to get a smile or a hug, or for somebody to talk to because they don’t have anyone. In our community people are being disowned, people have no support systems once they are diagnosed as being positive…and that goes from family to church to wherever they are.”


Slide12: Stigma Research Overall, 36% of NGO directors related that HIV-positive persons experience some form of ostracism or are not viewed as valued or complete members of their communities. In a second study with a random sample of American households, almost 30% of respondents reported that they would avoid a grocery store if the owner had AIDS, and 26% indicated that they would be uncomfortable wearing a sweater that had previously been worn by someone with AIDS (Herek, Capitanio, and Widaman; 2002).


Slide13: Mixed-Methods Inquiry Intelligently and strategically combines quantitative and qualitative methodology Complementary strengths and non-overlapping weaknesses Discovering Truth Combining research strategies allows us to broaden our understanding and obtain a more complete picture May require multiple investigators


Slide14: Mixed-Methods Inquiry Particularly compelling in Health Research Health Behaviors multiply determined Rich individual and contextual factors Example, Taking HIV medication Biological factors Social factors Individual factors


Slide15: Mixed-Methods Inquiry Key Point: Use right tools for the purpose Must meet the methodological requirements of each type of research


Practical Strategies for Combining Methods: Conversion of Data: Practical Strategies for Combining Methods: Conversion of Data Transferability between quantitative and qualitative methods Data analysis can be independent of data collection Making quantitative data qualitative – patterns (e.g., diagnostic information, categorical information)


Slide17: Practical Strategies for Combining Methods: Conversion of Data Making qualitative data quantitative – patterns (percents of individuals identifying key themes) Can subject this data to traditional quantitative analyses


Slide18: Example Many NGO directors (60%) indicated that stigmatizing attitudes were more prevalent in some segments of the community than in others. Respondents were especially likely to report more stigma in strongly religious subpopulations (mentioned by 25% of respondents) and in rural areas (18%). Consistent with NGO directors' sense that urban environments are more tolerant of HIV-positive persons, we found a negative correlation between population density of NGO cities and stigma directed at HIV-positive persons (rho = -0.25, p < .05). HIV-positive individuals also apparently fared better in states with lower rates of self-identified religious affiliation (rho = -0.23, p < .05).


Basic Research Designs: Basic Research Designs Mixed-Methods Research Quantitative and Qualitative methods used within the same study Data are collected, analyzed, and interpreted together To answer a single primary research question Often collected at the same time (although not always)


Slide20: Basic Research Designs Mixed-Models Research Quantitative and Qualitative methods used within the same program of research Data are analyzed and interpreted separately To answer multiple research questions Often collected sequentially, with one phase informing the data collection at the next phase


Mixed-Methods Research in HIV Primary Prevention: Mixed-Methods Research in HIV Primary Prevention A Case Example: Gay and Bisexual Men Engaged in Recreational Travel


Iterative Process: Iterative Process One research method followed by the next, depending on the research question being asked. Quantitative  Qualitative  Quantitative  Qualitative


Background: Background Studies with heterosexual young adults Apostolopoulos et al.: 30% of men and 31% of women reported having sex with a new partner during a spring break holiday Bellis et al.: 26% of men and 14% of women reported more than 1 sexual partner during a relatively brief vacation


Question One:: Question One: Is there a problem? Can we document it?


Study One: Key West: Study One: Key West Participants recruited in venues where traveling MSM are found. Brief, anonymous, self-report survey Assessed: Demographic information, zip code, HIV status Number of days in Key West Sexual behavior Sexual partner characteristics Disclosure of HIV status Substance use


Study One: Key West: Study One: Key West N = 219 eligible participants Mean age = 39 years Mean years of education = 14.8 years Diverse geographic representation Average of 4.1 days in Key West at that point of trip


Sexual Behavior: Sexual Behavior Majority reported no anal sex 26% reported sex with one partner 22% reported sex with multiple partners # of partners


Sexual Behavior: Sexual Behavior HIV-negative men significantly less likely to report multiple partners (18%), relative to HIV-positive men or men who did not know their HIV status (45%; p <.01) Most (51%) reported sex with a partner of unknown status Majority (77%) did not disclose HIV status to all of their sex partners


Sexual Behavior: Sexual Behavior Among sexually-active men, average # of partners was 1.95 (SD = 1.56) Most sexually-active men reported at least one instance of unprotected sex Majority of partners (71%) were first met on current trip; 45% of partners were from Key West


Sexual Behavior and Substance Use: Sexual Behavior and Substance Use 37% reported sex after having “too much” to drink 19% reported having sex after using drugs Substance use correlated with total # of partners (rho = 0.46, p <.001) and unprotected acts (rho = 0.29, p <.01)


Slide31: Mathematical Modeling Probability of acquiring HIV from a particular partner depends on number and type of sex acts, whether partner was insertive or receptive partner, partner HIV status, and per-act transmission probabilities for unprotected and protected intercourse.


Slide32: Mathematical Modeling Per-Act Probability Estimates for Unprotected Sexual Acts .001 receptive vaginal intercourse .0006 insertive vaginal intercourse .02 receptive anal intercourse .0006 insertive anal intercourse Estimated effectiveness of condoms = 90% Katz & Gerberding, 1997 Mastro & deVincenzi, 1996


Slide33: Mathematical Modeling Probability of acquiring HIV is the cumulative probability across all sexual acts and partner types Estimates of HIV status Key West Partners = 31.4% Non Key West Partners = 18.3%


Mathematical Modeling Results: Mathematical Modeling Results 1 out of every 196 sexually-active, at-risk MSM visitors to Key West would acquire HIV during a one-week stay in Key West 1 out of every 407 total MSM visitors Overall, approximately 200 new infections per year Significant economic impact


Conclusions: Conclusions Significant levels of risk behavior among MSM travelers in the United States Evidence for disease transmission


Study Two: Key West and Rehoboth Beach: Study Two: Key West and Rehoboth Beach Participants recruited in venues where traveling MSM are found. Brief, anonymous, self-report survey Assessed: Demographic information, zip code, HIV status Number of days in tourist area Sexual behavior More detailed assessment of substance use


Slide37: Study Two Participants N = 268 eligible participants Mean age = 38 years 90% reporting at least some college Predominantly white (83%), Latino (11%), African American (3%), Asian American (1%), Other (2%) Diverse geographic representation Average of 3.6 days in destination at that point of trip


Slide38: Substance Use Assessment Yes/No questions asking if they had used substances during their stay in the tourist area Marijuana, cocaine, ecstasy, LSD, methamphetamine, poppers, ketamine, rophynol, GHB, alcohol (to the point of intoxication)


Results: Results High rates of substance use Substance use associated with higher rates of HIV risk behavior


Substance Use and Sexual Risk Behavior: Substance Use and Sexual Risk Behavior % % Reporting Substance Use


Conclusions: Conclusions Significant risk Substance use as a risk factor for high-risk sexual behavior Potential problems with polysubstance use


Limitations of Existing Research: Limitations of Existing Research Convenience Samples No clear demonstration of increased risk during travel Very brief assessments of complex constructs Insufficient information for planning interventions Insufficient data collected to understand problem


Costs of Pilot Studies: Costs of Pilot Studies Study One: $800 Study Two: $600


Strengths of Pilot Studies: Strengths of Pilot Studies Clear evidence of risk – have successfully documented problem Identify some factors associated with risk Information can be used to guide development of qualitative and quantitative strategies Demonstrate team’s ability to recruit MSM travelers


Using Pilot Data: Using Pilot Data Pilot data used in support of grant application Collect Data in 3 Study Sites Proposed improved methodology Qualitative work More representative samples Larger samples Assessment in the tourist area and at home Collect information needed for intervention development


Research Questions Asked: Research Questions Asked Context of risk Motivations for travel, motivations to engage in risk behavior Information related to travel How do men reduce their risk during travel? (behavioral skills) Where are MSM travelers found in this community? What types of interventions might be used to reduce risk? How acceptable would HIV risk-reduction interventions be?


Initial Findings: Initial Findings Many men report greater disinhibition during travel “You know, it’s like I come out partly to party and have a good time. Sometimes I get carried away and do things I wouldn’t do in [home city]. The last time I was here, I had this fling with a guy I had just met and we didn’t use condoms…I was pretty worried after that.”


Slide49: Initial Findings Some men are open to risk-reduction interventions, within certain guidelines “People come here to have a good time. I don’t think anyone wants to get HIV when they are here, but they don’t want to be preached at either….I think people would be open to talking [to an outreach worker] but the message can’t be to not have fun.”


How data will be used: How data will be used Development of quantitative questionnaire Further development of intervention ideas Will be used for identification of venues for venue surveillance


Slide53: Research Questions Asked Demographic information, substance use, sexual risk Constructs identified during previous qualitative work Reactions to possible interventions Acceptability Likelihood of success


How data will be used: How data will be used Determining if environment factors (versus person characteristics) are key determinants of risk Identification of most promising intervention strategies – subsequent development


Research Questions Asked: Research Questions Asked Impressions of findings from first two phases Reactions to possible intervention strategies Data will be used for intervention development, seeking additional funding


Mixed-Methods Research in HIV Primary Prevention: Mixed-Methods Research in HIV Primary Prevention An Introduction


Slide58: Relevant Publications Benotsch, E.G., deRoon Cassini, T., Wright, V., Seal, D.W., Prosen, H., Kelly, J.A., Bogart, L.M., & Galletly, C. (under review). Stigma and AIDS: Challenges faced by community organizations conducting HIV prevention programs in the United States. Benotsch, E.G., Nettles, C.D., Wong, F., Redmann, J., Boschini, J., Pinkerton, S.D., Ragsdale, K., & Mikytuck, J.J. (in press). Sexual risk behavior and substance use in men attending Mardi Gras celebrations in New Orleans, Louisiana. Journal of Community Health. Benotsch, E.G., Seeley, S., Mikytuck, J., Pinkerton, S.D., Nettles, C.D., & Ragsdale, K. (in press). Substance use, medications for sexual facilitation, and sexual risk behavior among traveling men who have sex with men. Sexually Transmitted Diseases. Benotsch, E.G., Mikytuck, J., Ragsdale, K., & Pinkerton, S.D. (2006). Sexual risk and HIV acquisition among MSM travelers to Key West, Florida: A mathematical modeling analysis. AIDS Patient Care STDs, 20, 549-556. Benotsch, E.G., & Kalichman, S.C. (2002). Preventing HIV and AIDS. In D. Glenwick & L. Jason (Eds.). Innovative Strategies for Promoting Health and Mental Health across the Lifespan. (pp. 205-226). New York: Springer.