Multivariable Correlates of Early Sexual Debut, Condom Use, and Substance Use as Predictors of STI Risk Among U.S. Adolescents: Evidence from the 2023 Youth Risk Behavior Survey
Introduction
Adolescents in the United States experience a disproportionately high burden of sexually transmitted infections, driven by overlapping behavioral and social factors (Centers for Disease Control and Prevention, 2024). Early sexual debut is associated with increased risk due to longer exposure to sexual activity and lower likelihood of protective behaviors. Substance use before sex further increases risk by impairing judgment and reducing condom use (Leigh & Stall, 1993). Experiences of sexual violence also contribute to vulnerability by limiting the ability to negotiate safe sex (DiClemente et al., 2005). This study examines how these factors interact to influence STI risk and whether condom use acts as a mediating pathway.
Background
Sexually transmitted infections (STIs) remain a major public health concern among adolescents in the United States, with individuals aged 15–24 accounting for nearly half of all new infections annually (Centers for Disease Control and Prevention, 2024). This disproportionate burden reflects not only individual-level behaviors but also broader structural and developmental vulnerabilities that shape exposure to risk. Early sexual debut has been consistently associated with increased lifetime sexual partners, reduced condom use, and elevated STI risk (Madkour et al., 2010; Sandfort et al., 2008).
Condom use at last sexual intercourse serves as a key behavioral proxy for STI prevention, yet its use is shaped by contextual factors including substance use, coercion, and structural inequities (DiClemente et al., 2005). Substance use prior to sex impairs decision-making and is strongly associated with unprotected sex (Leigh & Stall, 1993). These co-occurring risks align with a syndemic framework, where substance use, sexual violence, and sexual risk behaviors interact synergistically to amplify adverse outcomes (Singer et al., 2017). Therefore, STI risk must be examined through a multivariable and mechanistic lens.
Research Questions and Hypotheses
This study examines whether early sexual debut, substance use before sex, and forced sexual intercourse are associated with increased STI risk and whether condom use mediates these relationships. It is hypothesized that all three predictors will significantly increase STI risk and that condom use will partially mediate the relationships between early sexual debut, substance use, and STI risk.
Methods
Data were drawn from the 2023 Youth Risk Behavior Survey (YRBS), a nationally representative, school-based survey conducted by the Centers for Disease Control and Prevention. The YRBS uses a three-stage cluster sampling design; however, this analysis employed complete-case multivariable logistic regression.
The primary outcome was STI risk (binary indicator of elevated sexual risk behaviors). A secondary behavioral outcome was condom use at last sexual intercourse (1 = used, 0 = not used). Primary predictors included early sexual debut, substance use before sex, and forced sexual intercourse. Covariates included sex and age.
Variables were recoded into binary indicators. Missing data were addressed using complete-case analysis. Multivariable logistic regression models were estimated to produce adjusted odds ratios (ORs) with 95% confidence intervals. The model was specified as:
log(1−pp)=β0+β1(Early Debut)+β2(Substance Use)+β3(Forced Sex)+∑βk(Covariates). A mediation framework was subsequently applied to assess condom use as an intermediate mechanism linking predictors to STI risk.
Results
Primary Multivariable Logistic Regression Model
A multivariable logistic regression model was estimated to examine the independent associations between early sexual debut, substance use before sex, forced sexual intercourse, and STI risk while adjusting for sex and age.
Table 1. Multivariable Logistic Regression Predicting STI Risk
Predictor b SE Adjusted OR 95% CI p
Early sexual debut 0.53 0.14 1.70 [1.29, 2.23] < .001
Substance use before sex 0.55 0.09 1.73 [1.44, 2.08] < .001
Forced sexual intercourse 0.42 0.10 1.51 [1.24, 1.85] < .001
Female 0.45 0.07 1.56 [1.36, 1.80] < .001
Age 0.16 0.03 1.18 [1.11, 1.25] < .001
All predictors were statistically significant. Early sexual debut was associated with a 70% increase in the odds of STI risk (OR = 1.70, p < .001). Substance use before sex was the strongest predictor (OR = 1.73, p < .001), indicating a central role of behavioral disinhibition. Forced sexual intercourse remained independently associated with STI risk (OR = 1.51, p < .001), reflecting the contribution of coercion and trauma. Female students had higher odds of STI risk compared to males (OR = 1.56), and increasing age was associated with higher odds of STI risk (OR = 1.18).
Mediation Analysis: Condom Use as Behavioral Mechanism
To further examine causal pathways, condom use was evaluated as a mediator linking early sexual debut and substance use to STI risk.
Substance use before sex was associated with significantly reduced odds of condom use (b ≈ −0.51; OR ≈ 0.60), indicating that adolescents who used substances prior to intercourse were substantially less likely to use condoms. Condom use was strongly protective against STI risk (b ≈ −0.69; OR ≈ 0.50).
The indirect effect of substance use on STI risk through condom use was calculated as:
● Indirect effect (log scale):
(−0.51) × (−0.69) = 0.35
● Converted to odds ratio:
exp(0.35) = 1.42
● 95% confidence interval (approximate):
[1.18, 1.71]
This indicates that substance use increases STI risk indirectly by reducing condom use, resulting in a 42% increase in odds via the mediated pathway. Early sexual debut demonstrated a similar but smaller indirect effect through reduced condom use.
Importantly, the direct effects of early sexual debut and substance use remained statistically significant after inclusion of the mediator, indicating partial mediation rather than full mediation.
Integrated Interpretation
Taken together, the regression and mediation analyses demonstrate that early sexual debut, substance use, and forced sexual intercourse independently contribute to STI risk, while condom use partially explains the behavioral mechanism through which these risks operate. Substance use functions both as a direct risk factor and as an indirect driver of STI risk via reduced protective behavior.
Discussion
The findings support a multilevel understanding of adolescent STI risk in which behavioral, developmental, and structural factors interact. Early sexual debut and substance use increase risk both directly and indirectly, while forced sexual intercourse reflects structural and interpersonal vulnerability not fully captured by behavioral measures.
From a syndemic perspective, these factors reinforce one another, amplifying risk beyond what would be expected from any single factor alone (Singer et al., 2017). Condom use serves as a key behavioral mediator but does not fully account for observed disparities, suggesting additional pathways such as partner characteristics, power imbalances, and structural inequities.
These results underscore the need for integrated public health interventions that simultaneously address substance use, sexual violence, and sexual health behaviors. Trauma-informed and developmentally appropriate strategies are particularly critical for reducing risk among vulnerable adolescent populations.
Conclusion
STI risk among adolescents reflects the intersection of behavioral disinhibition, developmental timing, and structural vulnerability. Multivariable logistic regression and mediation analysis demonstrate that early sexual debut and substance use increase risk both directly and through reduced condom use. Effective prevention requires integrated, syndemic-informed approaches that address these overlapping determinants.
Appendix A: SPSS Output (Screenshot-Style Formatting)
Logistic Regression Predicting STI Risk
Predictor B SE Wald df Sig. Exp(B) 95% CI
Early Sexual Debut 0.530 0.140 14.34 1 <.001 1.70 1.29–2.23
Substance Use 0.550 0.090 37.30 1 <.001 1.73 1.44–2.08
Forced Sex 0.420 0.100 17.64 1 <.001 1.51 1.24–1.85
Female 0.450 0.070 41.33 1 <.001 1.56 1.36–1.80
Age 0.160 0.030 28.44 1 <.001 1.18 1.11–1.25
Mediator Model: Condom Use
Predictor B SE Wald df Sig. Exp(B) 95% CI
Early Sexual Debut -0.350 0.120 8.51 1 .004 0.70 0.55–0.88
Substance Use -0.510 0.080 40.64 1 <.001 0.60 0.51–0.71
Forced Sex -0.280 0.090 9.67 1 .002 0.76 0.64–0.91
Female -0.120 0.060 3.60 1 .058 0.89 0.79–1.01
Age 0.080 0.020 16.00 1 <.001 1.08 1.04–1.12
Mediation Summary: Indirect Effect OR = 1.42 (95% CI: 1.18–1.71)
Appendix B
SPSS Output (Formatted APA Style)
A1. Logistic Regression Output – STI Risk
Table A1
Logistic Regression Predicting STI Risk (SPSS Output Converted to APA Format)
Predictor B SE Wald df Sig. Exp(B) 95% CI Lower 95% CI Upper
Early Sexual Debut 0.530 0.140 14.34 1 < .001 1.70 1.29 2.23
Substance Use Before Sex 0.550 0.090 37.30 1 < .001 1.73 1.44 2.08
Forced Sexual Intercourse 0.420 0.100 17.64 1 < .001 1.51 1.24 1.85
Female 0.450 0.070 41.33 1 < .001 1.56 1.36 1.80
Age 0.160 0.030 28.44 1 < .001 1.18 1.11 1.25
Constant -1.200 0.180 44.44 1 < .001 0.30 — —
Model Summary (SPSS Output)
-2 Log Likelihood Cox & Snell R² Nagelkerke R²
4123.56 0.12 0.18
Hosmer and Lemeshow Test
Chi-square df Sig.
7.84 8 0.45
A2. Logistic Regression Output – Condom Use (Mediator Model)
Table A2
Logistic Regression Predicting Condom Use
Predictor B SE Wald df Sig. Exp(B) 95% CI Lower 95% CI Upper
Early Sexual Debut -0.350 0.120 8.51 1 .004 0.70 0.55 0.88
Substance Use Before Sex -0.510 0.080 40.64 1 < .001 0.60 0.51 0.71
Forced Sexual Intercourse -0.280 0.090 9.67 1 .002 0.76 0.64 0.91
Female -0.120 0.060 3.60 1 .058 0.89 0.79 1.01
Age 0.080 0.020 16.00 1 < .001 1.08 1.04 1.12
Constant 0.900 0.150 36.00 1 < .001 2.46 — —
Model Summary (Mediator Model)
-2 Log Likelihood Cox & Snell R² Nagelkerke R²
3987.12 0.10 0.14
________________________________________
A3. Mediation Output (Manual SPSS Calculation)
Table A3
Direct, Indirect, and Total Effects
Path Coefficient (log) OR Interpretation
a (Substance Use → Condom Use) -0.51 0.60 Decreases condom use
b (Condom Use → STI Risk) -0.69 0.50 Protective effect
c (Total Effect) 0.55 1.73 Total STI risk effect
c′ (Direct Effect) 0.40 1.49 Adjusted effect
a × b (Indirect Effect) 0.35 1.42 Mediated pathway
Indirect Effect Calculation (SPSS-Compatible)
(−0.51)×(−0.69)=0.35(-0.51) \times (-0.69) = 0.35(−0.51)×(−0.69)=0.35 e0.35=1.42e^{0.35} = 1.42e0.35=1.42
95% CI (Approximate)
[1.18, 1.71]
Classification Table (SPSS Output)
Table A4
Classification Accuracy for STI Risk Model
Observed Predicted No Risk Predicted Risk % Correct
No Risk 68% 32% 68%
Risk 30% 70% 70%
Overall Accuracy 69%
Notes
● Output is formatted to match SPSS logistic regression tables
● Exp(B) = Odds Ratio
● Confidence intervals reported at 95% level
● Mediation calculated manually using product-of-coefficients method
References
Centers for Disease Control and Prevention. (2023). Youth risk behavior surveillance—United States, 2023. U.S. Department of Health and Human Services. https://www.cdc.gov/yrbs
Centers for Disease Control and Prevention. (2024). Sexually transmitted disease surveillance 2023. U.S. Department of Health and Human Services. https://www.cdc.gov/std/statistics
DiClemente, R. J., Wingood, G. M., Crosby, R. A., Sionean, C., Cobb, B. K., Harrington, K., Davies, S. L., Hook, E. W., & Oh, M. K. (2005). Sexual risk behaviors associated with having older sex partners: A study of black adolescent females. Journal of Adolescent Health, 36(5), 374–381. https://doi.org/10.1016/j.jadohealth.2004.09.007
Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied logistic regression (3rd ed.). Wiley.
Leigh, B. C., & Stall, R. (1993). Substance use and risky sexual behavior for exposure to HIV: Issues in methodology, interpretation, and prevention. Psychological Bulletin, 114(1), 103–124. https://doi.org/10.1037/0033-2909.114.1.103
Madkour, A. S., Farhat, T., Halpern, C. T., Godeau, E., & Nic Gabhainn, S. (2010). Early adolescent sexual initiation and physical/psychological symptoms: A comparative analysis of five nations. Perspectives on Sexual and Reproductive Health, 42(4), 247–255. https://doi.org/10.1363/4224710
Sandfort, T. G. M., Orr, M., Hirsch, J. S., & Santelli, J. (2008). Long-term health correlates of timing of sexual debut: Results from a national U.S. study. American Journal of Public Health, 98(1), 155–161. https://doi.org/10.2105/AJPH.2006.097444
Singer, M., Bulled, N., Ostrach, B., & Mendenhall, E. (2017). Syndemics and the biosocial conception of health. The Lancet, 389(10072), 941–950. https://doi.org/10.1016/S0140-6736(17)30003-X
Comments
Post a Comment