Jump to comment:

• Re: Trimming the estimate: Unmeasured confounding in the association between pubic hair grooming and sexually transmitted infections
  Benjamin N Breyer
  Published on: 20 April 2017
• Trimming the estimate: Unmeasured confounding in the association between pubic hair grooming and sexually transmitted infections
  Rohit P. Ojha
  Published on: 20 April 2017

• Published on: 20 April 2017

  Re: Trimming the estimate: Unmeasured confounding in the association between pubic hair grooming and sexually transmitted infections

  • Benjamin N Breyer
  • Other Contributors:
    • E Charles Osterberg, Thomas W Gaither, Mohannad A Awad, Matthew D Truesdale, Isabel Allen and Siobhan Sutcliffe

  As few studies have examined the relation between pubic hair grooming and sexually transmitted infections (STIs), we took advantage of nationally representative survey data to begin to explore this possible association and to develop hypotheses for future prospective studies. In our analysis, we observed a positive association between self-reported pubic hair grooming and STI history, which we interpreted in several poss...

  Show More

  As few studies have examined the relation between pubic hair grooming and sexually transmitted infections (STIs), we took advantage of nationally representative survey data to begin to explore this possible association and to develop hypotheses for future prospective studies. In our analysis, we observed a positive association between self-reported pubic hair grooming and STI history, which we interpreted in several possible ways given the limitations of our study (e.g., its cross- sectional study design, self-reported assessment of grooming and STIs, and lack of information on additional possible confounders, such as condom use) Interpretations of our findings included: a) grooming-associated epithelial microtears leading to increased risk of cutaneous STIs; b) grooming-related protection against lice by pubic hair removal; c) residual confounding by characteristics common to both grooming and STIs, such as risky sexual behaviors; and d) what we believe is the most likely interpretation, a combination of all of these explanations given our differing magnitudes of association by type of STI. Based on these interesting hypothesis-generating findings, we recommended that additional epidemiologic studies be conducted with prospective data collection, laboratory confirmation of STIs, and collection detailed STI risk information to elucidate the mechanisms underlying our findings.

  In response to Ojha and colleagues' concerns about confounding by gender, race, income, and sexual frequency, we adjusted our analyses for these variables: gender (male, female), race (White/Caucasian, Black/African American, Hispanic/Latino, Mixed Races, or Other), income (<$50,000, $50,000-74,999, $75,000-99,000, and >$100,000 USD), and sexual frequency (daily, weekly, monthly, every three months or less). We also repeated the analyses with and without survey weights. The association between grooming and a history of STI remained (aOR= 1.71, 95% CI 1.37-2.15). The other analyses yielded generally similar magnitudes of association and the same inferences as the results presented in our manuscript.

  Notwithstanding these largely unchanged estimates, we believe that residual confounding is still a possible interpretation of our findings and encourage more research on this topic. However, even in the presence of residual confounding, we would like to reinforce that use of grooming as a marker of high-risk sexual behaviors may still have value for STI prevention efforts to help identify individuals who would benefit most from STI prevention counseling.

  Conflict of Interest:

  None declared

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top
• Published on: 20 April 2017

  Trimming the estimate: Unmeasured confounding in the association between pubic hair grooming and sexually transmitted infections

  • Rohit P. Ojha, Director, Associate Professor

  Osterberg et al. [1] assessed the association between pubic hair grooming and sexually transmitted infections (STIs) using self-reported data from a cross-sectional survey of adults aged 18 to 65 years in the United States. The primary result was that individuals who reported ever- grooming had 1.8 times the odds (odds ratio [OR]=1.8, 95% confidence limits [CL]: 1.4, 2.2) of a history of STIs compared with individuals who...

  Show More

  Osterberg et al. [1] assessed the association between pubic hair grooming and sexually transmitted infections (STIs) using self-reported data from a cross-sectional survey of adults aged 18 to 65 years in the United States. The primary result was that individuals who reported ever- grooming had 1.8 times the odds (odds ratio [OR]=1.8, 95% confidence limits [CL]: 1.4, 2.2) of a history of STIs compared with individuals who reported never-grooming. The authors thus concluded that pubic hair grooming is associated with a history of STIs. Nevertheless, these results may be explained by confounding, selection, misclassification, and protopathic (i.e. reverse causality) biases. We focus herein on unmeasured confounding as an alternate explanation for the observed estimates.

  The authors adjusted for age and number of sexual partners, but these covariates are insufficient for adequately reducing confounding bias for the exposure-outcome association of interest. Confounding pathways include all common causes of exposure and outcome [2]. For example, gender influences pubic hair grooming habits [3] and STIs [4], but gender was not adjusted in the analysis by Osterberg et al. [1]. Therefore, gender is a source of unmeasured confounding and the consequences may be nontrivial.

  We used data reported by Osterberg et al. [1] for a sensitivity analysis of unmeasured confounding using the following formula by Ding and Vanderweele [5] to derive an adjustment factor,

  (OReu*ORud)/(OReu+ORud-1)

  where OReu is the odds ratio for the association between gender (males as reference) and any grooming (OR=2.67), and ORud is the odds ratio for the association between gender and STIs (OR=1.43). The observed OR and corresponding CL (OR=1.8, 95% CL: 1.4, 2.2) are subsequently divided by the adjustment factor (1.23), which results in an OR=1.5 (95% CL: 1.1, 1.8) for the estimate after adjustment for gender and the original covariates (age and number of sexual partners). The adjusted estimate is attenuated from the authors' reported estimate and may be even closer to the null if other relevant covariates could be adjusted such as race/ethnicity and socioeconomic status. In addition, the authors reported that the survey used sampling probability weights, but these weights did not seem to be used in the analysis. The consequence is potential bias in point estimates and overly narrow confidence limits [6], which raises further questions about the authors' interpretation.

  In summary, the interpretation by Osterberg et al. [1] may be based on biased estimates. Greater attention to unmeasured confounding and other sources of bias is warranted before attributing STIs to grooming habits.

  REFERENCES

  [1] Osterberg EC, Gaither TW, Awad MA, Truesdale MD, Allen I, Sutcliffe S, et al. Correlation between pubic hair grooming and STIs: results from a nationally representative probability sample. Sex Transm Infect. 2016. [2] Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10:37-48. [3] Butler SM, Smith NK, Collazo E, Caltabiano L, Herbenick D. Pubic hair preferences, reasons for removal, and associated genital symptoms: comparisons between men and women. J Sex Med. 2015;12:48-58. [4] Satterwhite CL, Torrone E, Meites E, Dunne EF, Mahajan R, Ocfemia MC, et al. Sexually transmitted infections among US women and men: prevalence and incidence estimates, 2008. Sex Transm Dis. 2013;40:187-93. [5] Ding P, VanderWeele TJ. Sensitivity Analysis Without Assumptions. Epidemiology. 2016;27:368-77. [6] Kreuter F, Valliant R. A survey on survey statistics: What is done and can be done in Stata. Stata Journal. 2007;7:1.

  Conflict of Interest:

  None declared

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top