Assessment of the E-value in the presence of bias amplification: a simulation study.

BACKGROUND: The E-value, a measure that has received recent attention in the comparative effectiveness literature, reports the minimum strength of association between an unmeasured confounder and the treatment and outcome that would explain away the estimated treatment effect. This study contributes to the literature on the applications and interpretations of E-values by examining how the E-value is impacted by data with varying levels of association of unobserved covariates with the treatment and outcome measure when covariate adjustment is applied. We calculate the E-value after using regression and propensity score methods (PSMs) to adjust for differences in observed covariates. Propensity score methods are a common observational research method used to balance observed covariates between treatment groups. In practice, researchers may assume propensity score methods that balance treatment groups across observed characteristics will extend to balance of unobserved characteristics. However, that assumption is not testable and has been shown to not hold in realistic data settings. We assess the E-value when covariate adjustment affects the imbalance in unobserved covariates. METHODS: Our study uses Monte Carlo simulations to evaluate the impact of unobserved confounders on the treatment effect estimates and to evaluate the performance of the E-Value sensitivity test with the application of regression and propensity score methods under varying levels of unobserved confounding. Specifically, we compare observed and unobserved confounder balance, odds ratios of treatment vs. control, and E-Value sensitivity test statistics from generalized linear model (GLM) regression models, inverse-probability weighted models, and propensity score matching models, over correlations of increasing strength between observed and unobserved confounders. RESULTS: We confirm previous findings that propensity score methods - matching or weighting - may increase the imbalance in unobserved confounders. The magnitude of the effect depends on the strength of correlation between the confounder, treatment, and outcomes. We find that E-values calculated after applying propensity score methods tend to be larger when unobserved confounders result in more biased treatment effect estimates. CONCLUSIONS: The E-Value may misrepresent the size of the unobserved effect needed to change the magnitude of the association between treatment and outcome when propensity score methods are used. Thus, caution is warranted when interpreting the E-Value in the context of propensity score methods.

Associations of fine particulate matter with incident cardiovascular disease; comparing models using ZIP code-level and individual-level fine particulate matter and confounders.

BACKGROUND: PM2.5 has been positively associated with cardiovascular disease (CVD) incidence. Most evidence has come from cohorts and administrative databases. Cohorts typically have extensive information on potential confounders and residential-level exposures. Administrative databases are usually more representative but typically lack information on potential confounders and often only have exposures at coarser geographies (e.g., ZIP code). The weaknesses in both types of studies have been criticized for potentially jeopardizing the validity of their findings for regulatory purposes. METHODS: We followed 101,870 participants from the US-based Nurses' Health Study (2000-2016) and linked residential-level PM2.5 and individual-level confounders, and ZIP code-level PM2.5 and confounders. We used time-varying Cox proportional hazards models to examine associations with CVD incidence. We specified basic models (adjusted for individual-level age, race and calendar year), individual-level confounder models, and ZIP code-level confounder models. RESULTS: Residential- and ZIP code-level PM2.5 were strongly correlated (Pearson r = 0.88). For residential-level PM2.5, the hazard ratio (HR, 95 % confidence interval) per 5 µg/m3 increase was 1.06 (1.01, 1.11) in the basic and 1.04 (0.99, 1.10) in the individual-level confounder model. For ZIP code-level PM2.5, the HR per 5 µg/m3 was 1.04 (0.99, 1.08) in the basic and 1.02 (0.97, 1.08) in the ZIP code-level confounder model. CONCLUSION: We observed suggestive positive, but not statistically significant, associations between long-term PM2.5 and CVD incidence, regardless of the exposure or confounding model. Although differences were small, associations from models with individual-level confounders and residential-level PM2.5 were slightly stronger than associations from models with ZIP code-level confounders and PM2.5.

Sweden's excess mortality in 2020-2022 and reporting in the media.

AIM: The aim was to scrutinize the report in March 2023 that Sweden's excess mortality was lowest in 2020-2022 compared with other European Union and Nordic countries, a report that received great national and international attention. STUDY DESIGN: Comparison of excess mortality in Sweden and Norway. METHODS: Excess mortality for 2020-2022 was calculated for Sweden and Norway, the country with per-capita excess mortality closest to Sweden's, compared with the average mortality for 2017-2019 in the respective country, following the definitions by Statistics Sweden reported in a daily newspaper. RESULTS: Excess mortality is a measure with low misclassification compared with other pandemic outcome measures. Following the definitions, total excess mortality for the years 2020-2022 was 11,897 individuals in Sweden and 6089 in Norway. However, the distributions of excess mortality across the 3 years strongly differed. In Sweden, 60% of excess mortality was observed in 2020, 8% in 2021 and 32% in 2022. In sharp contrast, 0% of excess mortality was observed in Norway in 2020, 20% in 2021 and 80% in 2022. If the relative distribution of excess mortality in Sweden had been the same as in Norway in 2020-2022, approximately 7000 individuals who died in 2020 would instead have died as excess mortality in 2022, saving approximately 14,000 person-years in Sweden. CONCLUSIONS: The report disregards residual confounding due to the broad definition of the period 2020-2022. Mass media should avoid one-sided reporting.

The implementation of target trial emulation for causal inference: a scoping review.

OBJECTIVES: We aim to investigate the implementation of Target Trial Emulation (TTE) for causal inference, involving research topics, frequently used strategies, and issues indicating the need for future improvements. STUDY DESIGN AND SETTING: We performed a scoping review by following the Joanna Briggs Institute (JBI) guidance and Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist. A health research-focused librarian searched multiple medical databases, and two independent reviewers completed screening and extraction within covidence review management software. RESULTS: Our search resulted in 1,240 papers, of which 96 papers were eligible for data extraction. Results show a significant increase in the use of TTE in 2018 and 2021. The study topics varied and focused primarily on cancer, cardiovascular and cerebrovascular diseases, and infectious diseases. However, not all papers specified well all three critical components for generating robust causal evidence: time-zero, random assignment simulation, and comparison strategy. Some common issues were observed from retrieved papers, and key limitations include residual confounding, limited generalizability, and a lack of reporting guidance that need to be improved. CONCLUSION: Uneven adherence to the TTE framework exists, and future improvements are needed to progress applications using causal inference with observational data.

Testing Biased Randomization Assumptions and Quantifying Imperfect Matching and Residual Confounding in Matched Observational Studies.

One central goal of design of observational studies is to embed non-experimental data into an approximate randomized controlled trial using statistical matching. Despite empirical researchers' best intention and effort to create high-quality matched samples, residual imbalance due to observed covariates not being well matched often persists. Although statistical tests have been developed to test the randomization assumption and its implications, few provide a means to quantify the level of residual confounding due to observed covariates not being well matched in matched samples. In this article, we develop two generic classes of exact statistical tests for a biased randomization assumption. One important by-product of our testing framework is a quantity called residual sensitivity value (RSV), which provides a means to quantify the level of residual confounding due to imperfect matching of observed covariates in a matched sample. We advocate taking into account RSV in the downstream primary analysis. The proposed methodology is illustrated by re-examining a famous observational study concerning the effect of right heart catheterization (RHC) in the initial care of critically ill patients. Code implementing the method can be found in the supplementary materials.

Body mass index and body shape before treatment and nasopharyngeal carcinoma prognosis: A population-based patient cohort study in southern China.

A concern of reverse causation exists about the association between nasopharyngeal carcinoma (NPC) prognosis and body mass index (BMI) at diagnosis, while the prognostic impact of BMI measured years before diagnosis is unknown. Therefore, we investigated associations of prediagnosis and pretreatment BMI and body shape on NPC mortality. From a population-based patient cohort in southern China between 2010 and 2013, we included 2526 incident NPC cases with prospective follow-up through 2018. We assessed the associations of BMI and body shape at age 20 years, 10 years before diagnosis, and at diagnosis with NPC mortality, combining strategies of stratification and statistical adjustment to minimize reverse causation. We observed 25% lower all-cause mortality (hazard ratio [HR] 0.75, 95% confidence interval [CI]: 0.64-0.89) and 25% lower NPC-specific mortality (HR 0.75, 95% CI: 0.61-0.91) among overweight vs normal-weight NPC cases at diagnosis. Lean body shapes 1 and 2 at diagnosis were associated with 68% and 23% higher all-cause mortality, respectively, compared to normal body shape 3. No effect modification by cancer stage was detected for associations with all-cause or NPC-specific mortality. Associations with BMI and body shape 10 years before diagnosis were similar but attenuated, while body size and shape at age 20 were not associated with mortality. Being overweight at diagnosis decreased mortality, and thinner body shape increased mortality, compared to normal weight/body shape. These associations may be due to poorer nutrition and treatment intolerance, resulting in treatment discontinuation and worse survival outcomes.

Pragmatic considerations for negative control outcome studies to guide non-randomized comparative analyses: A narrative review.

PURPOSE: This narrative review describes the application of negative control outcome (NCO) methods to assess potential bias due to unmeasured or mismeasured confounders in non-randomized comparisons of drug effectiveness and safety. An NCO is assumed to have no causal relationship with a treatment under study while subject to the same confounding structure as the treatment and outcome of interest; an association between treatment and NCO then reflects the potential for uncontrolled confounding between treatment and outcome. METHODS: We focus on two recently completed NCO studies that assessed the comparability of outcome risk for patients initiating different osteoporosis medications and lipid-lowering therapies, illustrating several ways in which confounding may result. In these studies, NCO methods were implemented in claims-based data sources, with the results used to guide the decision to proceed with comparative effectiveness or safety analyses. RESULTS: Based on this research, we provide recommendations for future NCO studies, including considerations for the identification of confounding mechanisms in the target patient population, the selection of NCOs expected to satisfy required assumptions, the interpretation of NCO effect estimates, and the mitigation of uncontrolled confounding detected in NCO analyses. We propose the use of NCO studies prior to initiating comparative effectiveness or safety research, providing information on the potential presence of uncontrolled confounding in those comparative analyses. CONCLUSIONS: Given the increasing use of non-randomized designs for regulatory decision-making, the application of NCO methods will strengthen study design, analysis, and interpretation of real-world data and the credibility of the resulting real-world evidence.

Indirect covariate balance and residual confounding: An applied comparison of propensity score matching and cardinality matching.

PURPOSE: Propensity score matching (PSM) is subject to limitations associated with limited degrees of freedom and covariate overlap. Cardinality matching (CM), an optimization algorithm, overcomes these limitations by matching directly on the marginal distribution of covariates. This study compared the performance of PSM and CM. METHODS: Comparative cohort study of new users of angiotensin-converting enzyme inhibitor (ACEI) and ß-blocker monotherapy identified from a large U.S. administrative claims database. One-to-one matching was conducted through PSM using nearest-neighbor matching (caliper = 0.15) and CM permitting a maximum standardized mean difference (SMD) of 0, 0.01, 0.05, and 0.10 between comparison groups. Matching covariates included 37 patient demographic and clinical characteristics. Observed covariates included patient demographics, and all observed prior conditions, drug exposures, and procedures. Residual confounding was assessed based on the expected absolute systematic error of negative control outcome experiments. PSM and CM were compared in terms of post-match patient retention, matching and observed covariate balance, and residual confounding within a 10%, 1%, 0.25% and 0.125% sample group. RESULTS: The eligible study population included 182 235 (ACEI: 129363; ß-blocker: 56872) patients. CM achieved superior patient retention and matching covariate balance in all analyses. After PSM, 1.6% and 28.2% of matching covariates were imbalanced in the 10% and 0.125% sample groups, respectively. No significant difference in observed covariate balance was observed between matching techniques. CM permitting a maximum SMD <0.05 was associated with improved residual bias as compared to PSM. CONCLUSION: We recommend CM with more stringent balance criteria as an alternative to PSM when matching on a set of clinically relevant covariates.

Use of negative control outcomes to assess the comparability of patients initiating lipid-lowering therapies.

PURPOSE: Clinical trials have demonstrated efficacy of proprotein convertase subtilisin/kexin type 9 inhibitors (PCSK9i) in reducing risk of cardiovascular disease events, but effectiveness in routine clinical care has not been well-studied. We used negative control outcomes to assess potential confounding in an observational study of PCSK9i versus ezetimibe or high-intensity statin. METHODS: Using commercial claims, we identified U.S. adults initiating PCSK9i, ezetimibe, or high-intensity statin in 2015-2018, with other lipid-lowering therapy (LLT) use in the year prior (LLT cohort) or atherosclerotic cardiovascular disease (ASCVD) in the past 90 days (ASCVD cohort). We compared initiators of PCSK9i to ezetimibe and high-intensity statin by estimating one-year risks of negative control outcomes influenced by frailty or health-seeking behaviors. Inverse probability of treatment and censoring weighted estimators of risk differences (RDs) were used to evaluate residual confounding after controlling for covariates. RESULTS: PCSK9i initiators had lower one-year risks of negative control outcomes associated with frailty, such as decubitus ulcer in the ASCVD cohort (PCSK9i vs. high-intensity statin RD = -3.5%, 95% confidence interval (CI): -4.6%, -2.5%; PCSK9i vs. ezetimibe RD = -1.3%, 95% CI: -2.1%, -0.6%), with similar but attenuated associations in the LLT cohort. Lower risks of accidents and fractures were also observed for PCSK9i, varying by cohort. Risks were similar for outcomes associated with health-seeking behaviors, although trended higher for PCSK9i in the ASCVD cohort. CONCLUSIONS: Observed associations suggest lower frailty and potentially greater health-seeking behaviors among PCSK9i initiators, particularly those with a recent ASCVD diagnosis, with the potential to bias real-world analyses of treatment effectiveness.

Dangers of residual confounding: a cautionary tale featuring cognitive ability, socioeconomic background, and education.

BACKGROUND: Cognitive ability and socioeconomic background (SEB) have been previously identified as determinants of achieved level of education. According to a "discrimination hypothesis", higher cognitive ability is required from those with lower SEB in order to achieve the same level of education as those with higher SEB. Support for this hypothesis has been claimed from the observation of a positive association between SEB and achieved level of education when adjusting for cognitive ability. We propose a competing hypothesis that the observed association is due to residual confounding. METHODS: To adjudicate between the discrimination and the residual confounding hypotheses, data from the 1997 National Longitudinal Survey of Youth (NLSY97, N = 8984) was utilized, including a check of the logic where we switched predictor and outcome variables. RESULTS: The expected positive association between SEB and achieved level of education when adjusting for cognitive ability (predicted by both hypotheses) was found, but a positive association between cognitive ability and SEB when adjusting for level of education (predicted only by the residual confounding hypothesis) was also observed. CONCLUSIONS: These results highlight the potential use of reversing predictors and outcomes to test the logic of hypothesis testing, and support a residual confounding hypothesis over a discrimination hypothesis in explaining associations between SEB, cognitive ability, and educational outcome.

Robust Q-learning.

Q-learning is a regression-based approach that is widely used to formalize the development of an optimal dynamic treatment strategy. Finite dimensional working models are typically used to estimate certain nuisance parameters, and misspecification of these working models can result in residual confounding and/or efficiency loss. We propose a robust Q-learning approach which allows estimating such nuisance parameters using data-adaptive techniques. We study the asymptotic behavior of our estimators and provide simulation studies that highlight the need for and usefulness of the proposed method in practice. We use the data from the "Extending Treatment Effectiveness of Naltrexone" multi-stage randomized trial to illustrate our proposed methods.

Impact of the Improper Adjustment for Age in Research on Age-Related Macular Degeneration: An Example Using Data from the Canadian Longitudinal Study on Aging.

Purpose: Confounding is an important problem in observational research. Improper modeling of the confounder will lead to residual confounding that may distort results and impact inferences. An example of this will be presented from research on age-related macular degeneration and depression. Methods: A 3-year prospective cohort study was performed using data from the Canadian Longitudinal Study on Aging consisting of 30,097 individuals aged 45-85 years. Incident depression was assessed using the Center for Epidemiologic Studies Depression scale. Participants were asked if they had ever had a physician diagnosis of age-related macular degeneration (AMD). Multivariable Poisson regression was used. Age was modeled in four ways including as a linear term, as a 4-category variable, as a spline, and as a polynomial. Models were compared using the Akaike's Information Criteria (AIC) with lower scores indicating better performance. Results: The point estimates and inferences differed depending on how age was modeled. Age had a J-shape relationship with the incidence of depression. The model with the lowest AIC was when age was entered as a categorical variable. When age was modeled in this way, AMD was not significantly associated with the incidence of depression (relative risk (RR) = 1.21, 95% Confidence Interval (CI) 0.97, 1.53). By contrast, when age was modeled as a linear term, AMD was significantly associated with the incidence of depression (RR = 1.28, 95% CI 1.02, 1.61). Conclusions: Researchers should clearly report their adjustment strategies and should be cautious when modeling the relationship between age and depression in order to minimize residual confounding.

Compared to randomized studies, observational studies may overestimate the effectiveness of DOACs: a metaepidemiological approach.

BACKGROUND AND OBJECTIVES: Randomized controlled trials (RCTs) are criticized for including patients who are overselected. Health authorities consequently encourage "real-world" postmarketing cohort studies. Our objective was to determine the differences between RCTs and observational studies as regards their populations and efficacy/safety results. METHODS: A systematic review was conducted to identify RCTs and observational studies including patients with venous thromboembolism receiving direct oral anticoagulants or conventional treatment. Ratios of hazard ratio (RHR) comparing epidemiological studies (prospective and retrospective cohort studies and studies using living databases) with RCTs were computed. RESULTS: Six RCTs (27,121 patients) and twenty observational studies (248,971 patients) were identified and analyzed. Prospective cohort studies seemed to recruit patients who were no less selected than those of RCTs whereas other types of observational studies may reflect the population treated in real life. Among observational studies, prospective cohort studies yielded the most favorable estimates of treatment effect compared with RCTs. These studies were associated with a nonsignificant 33% increase in efficacy estimate (RHR 0.67, [95% CI, 0.39-1.18]) but no effect on safety estimate. Studies using living databases were associated with nonsignificant trends toward a greater effect on efficacy (RHR 0.82, [0.66-1.01]) and a smaller effect on safety (RHR 1.33, [0.96-1.84]). DISCUSSION: Overall, in this clinical setting, an exaggeration of the treatment efficacy estimate was seen with observational studies compared with RCTs. CONCLUSIONS: As the presence of residual confounding cannot be excluded, these results should be interpreted cautiously.

Using negative control outcomes to assess the comparability of treatment groups among women with osteoporosis in the United States.

PURPOSE: In contrast to randomized clinical trials, comparative safety and effectiveness assessments of osteoporosis medications in clinical practice may be subject to confounding by indication. We used negative control outcomes to detect residual confounding when comparing osteoporosis medications. METHODS: Using MarketScan Commercial and Supplemental claims, we identified women aged ≥55 years who initiated an oral bisphosphonate (BP) (risedronate, alendronate, or ibandronate), denosumab (an injected biologic), or intravenous zoledronic acid (ZA) from October 1, 2010 to September 30, 2015. Women with Paget's disease or cancer were excluded. We compared individual oral BPs to each other, denosumab to ZA, denosumab to oral BPs, and ZA to oral BPs, with respect to 11 negative control outcomes identified by subject matter experts. We estimated the 12-month cumulative risk difference (RD) using inverse probability of treatment and censoring weights. RESULTS: Among 148 587 women, most initiated alendronate (57%), followed by ibandronate (12%), ZA (11%), risedronate (10%), and denosumab (10%). Compared with denosumab, patients initiating ZA had similar risks of all negative control outcomes. Compared with oral BPs, patients initiating denosumab had a higher risk of a wellness visit (RD = 1.2%, 95% CI: 0.4, 1.9) and a lower risk of receiving herpes zoster vaccine (RD = -0.6%, 95% CI: -1.1, -0.2). Comparing ZA with oral BP initiators resulted in two outcomes with positive associations. CONCLUSIONS: Caution is warranted when comparing injectable vs oral osteoporosis medications, given the potential for unmeasured confounding. Evaluating negative control outcomes could be a standard validity check prior to conducting comparative studies.

Implications of nonlinearity, confounding, and interactions for estimating exposure concentration-response functions in quantitative risk analysis.

Recent advances in understanding of biological mechanisms and adverse outcome pathways for many exposure-related diseases show that certain common mechanisms involve thresholds and nonlinearities in biological exposure concentration-response (C-R) functions. These range from ultrasensitive molecular switches in signaling pathways, to assembly and activation of inflammasomes, to rupture of lysosomes and pyroptosis of cells. Realistic dose-response modeling and risk analysis must confront the reality of nonlinear C-R functions. This paper reviews several challenges for traditional statistical regression modeling of C-R functions with thresholds and nonlinearities, together with methods for overcoming them. Statistically significantly positive exposure-response regression coefficients can arise from many non-causal sources such as model specification errors, incompletely controlled confounding, exposure estimation errors, attribution of interactions to factors, associations among explanatory variables, or coincident historical trends. If so, the unadjusted regression coefficients do not necessarily predict how or whether reducing exposure would reduce risk. We discuss statistical options for controlling for such threats, and advocate causal Bayesian networks and dynamic simulation models as potentially valuable complements to nonparametric regression modeling for assessing causally interpretable nonlinear C-R functions and understanding how time patterns of exposures affect risk. We conclude that these approaches are promising for extending the great advances made in statistical C-R modeling methods in recent decades to clarify how to design regulations that are more causally effective in protecting human health.

Contemporary Use and Outcomes of Arterial Closure Devices After Percutaneous Coronary Intervention: Insights From the Veterans Affairs Clinical Assessment, Reporting, and Tracking Program.

Background Arterial closure devices reduce the length of bedrest after invasive cardiac procedures via the femoral approach, but there are conflicting data on their association with major bleeding and vascular complications. We thus sought to evaluate the contemporary use of femoral arterial closure devices and their association with major bleeding among patients undergoing percutaneous coronary intervention. Methods and Results We identified patients undergoing percutaneous intervention via the femoral approach within the Veterans Affairs Healthcare System from December 2004 through September 2018. The association between arterial closure device use and major bleeding was evaluated using both propensity matching and instrumental variable analyses, incorporating contrast-induced nephropathy as a falsification end point. We identified 132 373 percutaneous coronary interventions performed by 681 operators, with closure device use increasing 1.2% each year (linear trend P<0.001). In a propensity-matched cohort, closure devices were associated with a 1.1% reduction in periprocedural bleeding (95% CI, -1.5% to -0.6%). Closure devices were also associated with a numerical decrease in contrast-inducted nephropathy that did not reach statistical significance (-0.6%; 95% CI, -1.3% to 0.1%). In an instrumental variable analysis of closure device use, there was no difference in the bleeding rate between those who received a closure device and those who did not (0.2%; 95% CI, -0.9% to 1.2%). Conclusions Arterial closure devices are associated with a reduction in major bleeding within a propensity-matched cohort. This association dissipates in an instrumental variable analysis, highlighting some of the methodologic limitations of comparative effectiveness research in observational analyses.

Comparative effectiveness of high dose versus adjuvanted influenza vaccine: A retrospective cohort study.

BACKGROUND: Adults 65 years and older (seniors) experience more complications following influenza infection than younger adults. We estimated the relative vaccine effectiveness (rVE) of a trivalent high dose (HD-IIV3) versus an adjuvanted trivalent influenza vaccine (aIIV3) in seniors for respiratory-related hospitalizations. METHODS: We conducted a retrospective cohort study using claims data from Optum's Clinformatics® Data Mart to compare outcome rates between seniors who received HD-IIV3 versus aIIV3 during the 2016/17 and 2017/18, predominantly A/H3N2 respiratory seasons. Rates were adjusted for demographic characteristics, comorbid conditions, previous influenza vaccination, and geography. We used the previous event rate ratio (PERR) approach to address bias by time-fixed unmeasured confounders. RESULTS: We identified 842,282 HD-IIV3 and 34,157 aIIV3 recipients for the 2016/17 season and 1,058,638 HD-IIV3 and 189,636 aIIV3 recipients for the 2017/18 season. The pooled rVE of HD-IIV3 versus aIIV3 for respiratory-related hospitalizations over both seasons was 12% (95% confidence interval: 3.3%-20%); 13% (-6.4% to 32%) for the 2016/17 season and 12% (2.1%-21%) for the 2017/18 season. CONCLUSIONS: Pooled over two predominantly A/H3N2 respiratory seasons, HD-IIV3 was associated with fewer respiratory hospital admissions than aIIV3 in senior members of large national managed health care company in the U.S.

The role of sexual networks in studies of how BV and STIs increase the risk of subsequent reinfection.

Prior studies have demonstrated that both bacterial vaginosis (BV) and sexually transmitted infections (STIs) are strong independent risk factors for subsequent STI. In observational studies of this biological enhancement (BE) hypothesis, it is important to adjust for the risk of STI exposure so that the independent effect of BE can be assessed. We sought to model if two markers of local sexual network (partner concurrency and cumulative number of STIs) represented residual confounding in the models of risk for subsequent infection in a study that screened 3620 women for STIs every 3 months for a year. Mixed-effects logistic regression was used to calculate the odds ratios for an incident diagnosis of Chlamydia trachomatis, Neisseria gonorrhoeae, Trichomonas vaginalis and BV following a diagnosis of any of these four at the prior visit, controlling for the cumulative number of STIs and partner concurrency variables. We found that partner concurrency and cumulative number of STIs were each associated with incident infection, and in general, controlling for these variables reduced the strength of the association between prior and incident infections. We conclude that the frequently found association between prior and incident STIs is associated with both BE and sexual network structure.