School racial composition, effect modification by caring teacher/staff presence, and mid/late-life depressive symptoms: findings from the Study of Healthy Aging among African Americans

For Black students in the United States, attending schools with a higher proportion of White students is associated with worse mental and physical health outcomes in adolescence/early adulthood. No prior studies evaluate K-12 school racial composition and later-life mental health. In a cohort of Black adults ages 50+ in Northern California who retrospectively self-reported school racial composition for grades 1, 6, 9, and 12, we assessed the association between attending a school with mostly Black students vs. not and mid/late-life depressive symptoms (8-item PROMIS depression score, standardized to US adult population) using age-, sex/gender-, southern US birth-, and parental education-adjusted generalized estimating equations, and assessed effect modification by caring teacher/staff presence. Later-life depressive symptoms were lower among those who attended schools with mostly Black students in grades 1 and 6 (b=-0.12, 95% CI: -0.23, 0.00 and b=-0.11, 95% CI: -0.22, 0.00, respectively). In grade 6, this difference was larger for students without an adult at school who cared about them (b=-0.29, 95% CI: -0.51, -0.07 vs. b=-0.04, 95% CI: -0.17, 0.09). Among Black Americans, attending early school with mostly Black students may have later life mental health benefits; this protective association appears more important for students without caring teachers/staff.

Clinical Predictors and Outcomes of Invasive Anal Cancer for People With Human Immunodeficiency Virus in an Inception Cohort

BACKGROUND

Due to the heterogeneity of risk for invasive anal cancer (IAC) among people with human immunodeficiency virus (PWH), we investigated predictors of IAC and described outcomes among those with a cancer diagnosis.

METHODS

Using a longitudinal inception cohort of anal cancer screening, we evaluated risk factors and outcome probabilities for incident IAC in Cox models. Screening included anal cytology and digital anorectal examination, and, if results of either were abnormal, high-resolution anoscopy.

RESULTS

Between 30 November 2006 and 3 March 2021, a total of 8139 PWH received care at the University of California, San Diego, with 4105 individuals undergoing screening and subsequently followed up over a median of 5.5 years. Anal cancer developed in 33 of them. IAC was more likely to develop in patients with anal high-grade squamous intraepithelial lesions (aHSILs) on initial or subsequent follow-up cytology (hazard ratio, 4.54) and a nadir CD4 cell count ≤200/µL (2.99). The joint effect of aHSILs and nadir CD4 cell count ≤200/µL amplified the hazard of IAC by 9-fold compared with the absence of both. PWH with time-updated cytology aHSIL and CD4 cell counts ≤200/µL had 5- and 10-year probabilities of IAC of 3.40% and 4.27%, respectively. Twelve individuals with cancer died, 7 (21% of the total 33) due to cancer progression, and they had clinical stage IIIA or higher cancer at initial diagnosis.

CONCLUSIONS

PWH with both aHSIL and a nadir CD4 cell count ≤200/µL have the highest risk of IAC. PWH who died due to IAC progression had clinical stage IIIA cancer or higher at diagnosis, highlighting the importance of early diagnosis through high-resolution anoscopic screening.

Impact of recipient age on mortality among Cytomegalovirus (CMV)-seronegative lung transplant recipients with CMV-seropositive donors

BACKGROUND

Cytomegalovirus (CMV)-seronegative lung transplant recipients (LTRs) with seropositive donors (CMV D+/R-) have the highest mortality of all CMV serostatuses. Due to immunosenescence and other factors, we hypothesized CMV D+/R- status might disproportionately impact older LTRs. Thus, we investigated whether recipient age modified the relationship between donor CMV status and mortality among CMV-seronegative LTRs.

METHODS

Adult, CMV-seronegative first-time lung-only recipients were identified through the Scientific Registry of Transplant Recipients between May 2005 and December 2019. We used adjusted multivariable Cox regression to assess the relationship of donor CMV status and death. Interaction between recipient age and donor CMV was assessed via likelihood ratio testing of nested Cox models and by the relative excess risk due to interaction (RERI) and attributable proportion (AP) of joint effects.

RESULTS

We identified 11,136 CMV-seronegative LTRs. The median age was 59 years; 65.2% were male, with leading transplant indication of idiopathic pulmonary fibrosis (35.6%); and 60.8% were CMV D+/R-. In multivariable modeling, CMV D+/R- status was associated with 27% increased hazard of death (adjusted hazard ratio: 1.27, 95% confidence interval: 1.21-1.34) compared to CMV D-/R-. Recipient age ≥60 years significantly modified the relationship between donor CMV-seropositive status and mortality on the additive scale, including RERI 0.24 and AP 11.4% (p = 0.001), that is, the interaction increased hazard of death by 0.24 and explained 11.4% of mortality in older CMV D+ recipients.

CONCLUSIONS

Among CMV-seronegative LTRs, donor CMV-seropositive status confers higher risk of posttransplant mortality, which is amplified in older recipients. Future studies should define optimal strategies for CMV prevention and management in older D+/R- LTRs.

Social interaction level modulates the impact of frailty on cognitive decline: a longitudinal study

OBJECTIVES

This study evaluates whether social interaction level modifies the association between frailty and cognitive decline in older adults.

METHODS

A total of 2701 adults aged ≥65 years enrolled in the Progetto Veneto Anziani (Pro.V.A.), participated in the study. At baseline, participants were classified as having low, moderate, or high social interactions based on: cohabiting status, frequency of contacts with relatives/friends or involvement in childcare activities, and frequency of participation in social or community activities. Baseline frailty was defined as the presence of ≥3 criteria among: weight loss, weakness, slowness, low physical activity, and exhaustion. Cognitive function was assessed through the Mini-Mental State Examination (MMSE) at baseline and after 4.4 and 7 years. The association between frailty and MMSE changes over time was evaluated through linear mixed models. Interaction and subgroup analyses explored the modifying effect of social interaction level on the above association.

RESULTS

The mean age of participants was 76.1 years, and 59.1% were women. Frail individuals had a steeper annual MMSE decline than their non-frail counterparts (ß=-0.40, 95%CI: -0.59, -0.20). When stratifying participants by social interaction level, we found that the association between frailty and MMSE changes over time was stronger in those with low social interactions (ß=-0.74, 95%CI: -1.33, -0.15) while attenuated in those with moderate (ß=-0.42, 95%CI: -0.74, -0.11) or high social interaction level (ß=-0.29, 95%CI: -0.58, 0.01).

CONCLUSION

Maintaining frequent social interactions might mitigate the negative impact of frailty on older people's cognitive functions.

Does Self-Reported BMI Modify the Association Between Stroke and Depressive Symptoms?

BACKGROUND

Depressive symptoms are common in stroke survivors. While obesity has been associated with stroke and depression, its influence on the association between stroke and depressive symptoms is unknown.

METHODS

Cross-sectional data from 2015 to 2016 Canadian Community Health Survey was used. History of stroke was self-reported and our outcome of interest was depressive symptoms in the prior 2 weeks, measured using the 9-item Patient Health Questionnaire. Self-reported body mass index (BMI) was modeled as cubic spline terms to allow for nonlinear associations. We used multivariable logistic regression to evaluate the association between stroke and depressive symptoms and added an interaction term to evaluate the modifying effect of BMI.

RESULTS

Of the 47,521 participants, 694 (1.0%) had a stroke and 3314 (6.5%) had depressive symptoms. Those with stroke had a higher odds of depressive symptoms than those without (aOR = 3.13, 95% CI 2.48, 3.93). BMI did not modify the stroke-depressive symptoms association (P interaction = 0.242) despite the observed variation in stroke-depressive symptoms association across BMI categories,: normal BMI [18.5-25 kg/m2] (aOR† = 3.91, 95% CI 2.45, 6.11), overweight [25-30 kg/m2] (aOR† = 2.63, 95% CI 1.58, 4.20), and obese [>30 kg/m2] (aOR† = 2.76, 95% CI 1.92, 3.94). Similar results were found when depressive symptoms were modeled as a continuous measure.

CONCLUSION

The association between stroke and depressive symptoms is not modified by BMI, needing additional work to understand the role of obesity on depression after stroke.

Do dental anxiety and satisfaction with dental professionals modify the association between affordability and dental service use? A population-based longitudinal study of Australian adults

OBJECTIVES

This study aimed to assess the association between affordability in terms of difficulty paying dental bills in Australian dollars and dental service use in the presence of sociodemographic confounders, and to assess the role of dental anxiety and satisfaction with dental professionals as mediators. The second aim was to investigate how dental anxiety and satisfaction with dental professionals modify the association between affordability and use of dental services in Australian adults.

METHODS

Longitudinal data from the Australian National Study of Adult Oral Health (2004-06 and 2017-18) was used. Poisson regression and path analysis were conducted to determine the association between affordability and frequency of use of dental services. Effect measure modification (EMM) analysis was performed by stratification of dental anxiety and satisfaction with dental professionals.

RESULTS

The study included 1698 Australian adults and identified that the prevalence of low frequency of dental visits was 20% more for those who had difficulty paying dental bills. Adults with dental anxiety (prevalence ratio [PR] = 1.14) and those who were dissatisfied with dental professionals (PR = 1.17) had a higher prevalence of low frequency of dental visits in the presence of difficulty paying dental bills. This indicated that dental anxiety and dissatisfaction with dental professionals were effect modifiers on this pathway.

CONCLUSIONS

Adults who experience dental anxiety and dissatisfaction with dental professionals are more likely to avoid dental visits when faced with difficulty paying dental bills. However, it is important to note that these associations do not necessarily imply a causal relationship.

Unveiling causal connections: Long-term particulate matter exposure and type 2 diabetes mellitus mortality in Southern China

Evidence of the potential causal links between long-term exposure to particulate matters (PM, i.e., PM1, PM2.5, and PM1-2.5) and T2DM mortality based on large cohorts is limited. In contrast, the existing evidence usually suffers from inherent bias with the traditional association assessment. A prospective cohort of 580,757 participants in the southern region of China were recruited during 2009 and 2015 and followed up through December 2020. PM exposure at each residential address was estimated by linking to the well-established high-resolution simulation dataset. Hazard ratios (HRs) were calculated using time-varying marginal structural Cox models, an established causal inference approach, after adjusting for potential confounders. During follow-up, a total of 717 subjects died from T2DM. For every 1 μg/m3 increase in PM2.5, the adjusted HRs and 95% confidence interval (CI) for T2DM mortality was 1.036 (1.019-1.053). Similarly, for every 1 μg/m3 increase in PM1 and PM1-2.5, the adjusted HRs and 95% CIs were 1.032 (1.003-1.062) and 1.085 (1.054-1.116), respectively. Additionally, we observed a generally more pronounced impact among individuals with lower levels of education or lower residential greenness which as measured by the Normalized Difference Vegetation Index (NDVI). We identified substantial interactions between NDVI and PM1 (P-interaction = 0.003), NDVI and PM2.5 (P-interaction = 0.019), as well as education levels and PM1 (P-interaction = 0.049). The study emphasizes the need to consider environmental and socio-economic factors in strategies to reduce T2DM mortality. We found that PM1, PM2.5, and PM1-2.5 heighten the peril of T2DM mortality, with education and green space exposure roles in modifying it.

Post-traumatic stress and future substance use outcomes: leveraging antecedent factors to stratify risk

Background

Post-traumatic stress disorder (PTSD) and substance use (tobacco, alcohol, and cannabis) are highly comorbid. Many factors affect this relationship, including sociodemographic and psychosocial characteristics, other prior traumas, and physical health. However, few prior studies have investigated this prospectively, examining new substance use and the extent to which a wide range of factors may modify the relationship to PTSD.

Methods

The Advancing Understanding of RecOvery afteR traumA (AURORA) study is a prospective cohort of adults presenting at emergency departments (N = 2,943). Participants self-reported PTSD symptoms and the frequency and quantity of tobacco, alcohol, and cannabis use at six total timepoints. We assessed the associations of PTSD and future substance use, lagged by one timepoint, using the Poisson generalized estimating equations. We also stratified by incident and prevalent substance use and generated causal forests to identify the most important effect modifiers of this relationship out of 128 potential variables.

Results

At baseline, 37.3% (N = 1,099) of participants reported likely PTSD. PTSD was associated with tobacco frequency (incidence rate ratio (IRR): 1.003, 95% CI: 1.00, 1.01, p = 0.02) and quantity (IRR: 1.01, 95% CI: 1.001, 1.01, p = 0.01), and alcohol frequency (IRR: 1.002, 95% CI: 1.00, 1.004, p = 0.03) and quantity (IRR: 1.003, 95% CI: 1.001, 1.01, p = 0.001), but not with cannabis use. There were slight differences in incident compared to prevalent tobacco frequency and quantity of use; prevalent tobacco frequency and quantity were associated with PTSD symptoms, while incident tobacco frequency and quantity were not. Using causal forests, lifetime worst use of cigarettes, overall self-rated physical health, and prior childhood trauma were major moderators of the relationship between PTSD symptoms and the three substances investigated.

Conclusion

PTSD symptoms were highly associated with tobacco and alcohol use, while the association with prospective cannabis use is not clear. Findings suggest that understanding the different risk stratification that occurs can aid in tailoring interventions to populations at greatest risk to best mitigate the comorbidity between PTSD symptoms and future substance use outcomes. We demonstrate that this is particularly salient for tobacco use and, to some extent, alcohol use, while cannabis is less likely to be impacted by PTSD symptoms across the strata.

Exploring the Modifying Role of GDP and Greenness on the Short Effect of Air Pollutants on Respiratory Hospitalization in Beijing

It is unclear whether Gross Domestic Product (GDP) and greenness have additional modifying effects on the association between air pollution and respiratory system disease. Utilizing a time-stratified case-crossover design with a distributed lag linear model, we analyzed the association between six pollutants (PM2.5, PM10, NO2, SO2, O3, and CO) and 555,498 respiratory hospital admissions in Beijing from 1st January 2016 to 31st December 2019. We employed conditional logistic regression, adjusting for meteorological conditions, holidays and influenza, to calculate percent change of hospitalization risk. Subsequently, we performed subgroup analysis to investigate potential effect modifications using a two-sample z test. Every 10 μg/m3 increase in PM2.5, PM10, NO2, SO2, and O3 led to increases of 0.26% (95%CI: 0.17%, 0.35%), 0.15% (95%CI: 0.09%, 0.22%), 0.61% (95%CI: 0.44%, 0.77%), 1.72% (95%CI: 1.24%, 2.21%), and 0.32% (95%CI: 0.20%, 0.43%) in admissions, respectively. Also, a 1 mg/m3 increase in CO levels resulted in a 2.50% (95%CI: 1.96%, 3.04%) rise in admissions. The links with NO2 (p < 0.001), SO2 (p < 0.001), O3 (during the warm season, p < 0.001), and CO (p < 0.001) were significantly weaker among patients residing in areas with higher levels of greenness. No significant modifying role of GDP was observed. Greenness can help mitigate the effects of air pollutants, while the role of GDP needs further investigation.

Planning stepped wedge cluster randomized trials to detect treatment effect heterogeneity

Stepped wedge design is a popular research design that enables a rigorous evaluation of candidate interventions by using a staggered cluster randomization strategy. While analytical methods were developed for designing stepped wedge trials, the prior focus has been solely on testing for the average treatment effect. With a growing interest on formal evaluation of the heterogeneity of treatment effects across patient subpopulations, trial planning efforts need appropriate methods to accurately identify sample sizes or design configurations that can generate evidence for both the average treatment effect and variations in subgroup treatment effects. To fill in that important gap, this article derives novel variance formulas for confirmatory analyses of treatment effect heterogeneity, that are applicable to both cross-sectional and closed-cohort stepped wedge designs. We additionally point out that the same framework can be used for more efficient average treatment effect analyses via covariate adjustment, and allows the use of familiar power formulas for average treatment effect analyses to proceed. Our results further sheds light on optimal design allocations of clusters to maximize the weighted precision for assessing both the average and heterogeneous treatment effects. We apply the new methods to the Lumbar Imaging with Reporting of Epidemiology Trial, and carry out a simulation study to validate our new methods.

Potential causal links of long-term air pollution with lung cancer incidence: From the perspectives of mortality and hospital admission in a large cohort study in southern China

Evidence on the potential causal links of long-term air pollution exposure with lung cancer incidence (reflected by mortality and hospital admission) was limited, especially based on large cohorts. We examined the relationship between lung cancer and long-term exposure to particulate matter (PM, including PM2.5 , PM10 and PM10-2.5 ) and nitrogen dioxide (NO2 ) among a large cohort of general Chinese adults using causal inference approaches. The study included 575 592 participants who were followed up for an average of 8.2 years. The yearly exposure of PM and NO2 was estimated through satellite-based random forest approaches and the ordinary kriging method, respectively. Marginal structural Cox models were used to examine hazard ratios (HRs) of mortality and hospital admission due to lung cancer following air pollution exposure, adjusting for potential confounders. The HRs of mortality due to lung cancer were 1.042 (95% confidence interval [CI]: 1.033-1.052), 1.032 (95% CI:1.024-1.041) and 1.052 (95% CI:1.041-1.063) for each 1 μg/m3 increase in PM2.5 , PM10 and NO2 , respectively. In addition, we observed statistically significant effects of PMs on hospital admission due to lung cancer. The HRs (95%CI) were 1.110 (1.027-1.201), 1.067 (1.020-1.115) and 1.079 (1.010-1.153) for every 1 μg/m3 increase in PM2.5 , PM10 , PM10-2.5 , respectively. Furthermore, we found larger effect estimates among the elderly and those who exercised more frequently. We provided the most comprehensive evidence of the potential causal links between two outcomes of lung cancer and long-term air pollution exposure. Relevant policies should be developed, with special attention to protecting the vulnerable groups of the population.

Racial Disparities in Infliximab Efficacy for Ulcerative Colitis: Evidence Synthesis and Effect Modification Assessment

An increasing amount of research explores the role of race in clinical phenotypes and outcomes in ulcerative colitis (UC). We aimed to investigate racial differences in infliximab (IFX) treatment efficacy in UC. We used aggregate data from IFX trials and evidence synthesis methods to generate race-specific efficacy estimates. Then, we tested the effect modification by race by comparing the race-specific estimates derived from independent evidence syntheses. We computed ratios of relative risks (RRRs) and performed tests of statistical interaction. We analyzed data from five randomized, placebo-controlled trials evaluating IFX as induction and maintenance therapy for adults with moderate-to-severe UC (875 participants; 45% Asians). We found no substantial evidence of racial differences concerning the efficacy of IFX in inducing clinical response (RRR = 0.89, 95% CI: 0.66-1.20; p = 0.44), clinical remission (RRR = 0.58, 95% CI: 0.24-1.44; p = 0.24), and mucosal healing (RRR = 0.99, 95% CI: 0.69-1.41; p = 0.95), or maintaining clinical remission (RRR = 0.81, 95% CI: 0.46-1.42; p = 0.45) and mucosal healing (RRR = 0.84, 95% CI: 0.48-1.46; p = 0.53), between Asian and Caucasian populations. Future clinical studies should expand the participation of racial minorities to comprehensively assess potential racial differences in the effectiveness of advanced therapies, including IFX, in the context of treating UC.

Do respiratory virus infections modify associations of asthma exacerbation with aeroallergens or fine particulate matter? A time series study in Philadelphia PA

Respiratory virus infections are related to over 80% of childhood asthma exacerbations. They enhance pro-inflammatory mediator release, especially for sensitized individuals exposed to pollens/molds. Using a time-series study design, we investigated possible effect modification by respiratory virus infections of the associations between aeroallergens/PM2.5 and asthma exacerbation rates. Outpatient, emergency department (ED), and inpatient visits for asthma exacerbation among children with asthma (28,540/24,444 [warm/cold season]), as well as viral infection counts were obtained from electronic health records of the Children's Hospital of Philadelphia from 2011 to 2016. Rate ratios (RRs, 90th percentile vs. 0) for late-season grass pollen were 1.00 (0.85-1.17), 1.04 (0.95-1.15), and 1.12 (0.96-1.32), respectively, for respiratory syncytial virus (RSV) counts within each tertile. However, similar trends were not observed for weed pollens/molds or PM2.5. Overall, our study provides little evidence supporting effect modification by respiratory viral infections.

Do psychotic experiences act as effect modifiers in youths with common mental health problems allocated to transdiagnostic cognitive behavioural therapy versus management as usual? Secondary analyses of the Mind-My-Mind randomized trial

BACKGROUND

Psychotic experiences (PEs) are common in help-seeking youths with non-psychotic mental health problems, yet the clinical importance of PEs as potential effect modifiers of psychotherapy interventions has been scarcely examined. We examined if PEs were associated with a differential response to transdiagnostic cognitive behavioural therapy (CBT) aimed at common emotional and behavioural problems.

METHODS

We present secondary analyses from the Mind My Mind (MMM) trial that randomized 396, 6-16-year-old youths to either 9-13 sessions of transdiagnostic modular community-based CBT (MMM) or community-based management as usual (MAU). MMM was superior to MAU in reducing parent-reported impact of mental health problems according to the Strengths and Difficulties Questionnaire (SDQ). PEs were assessed by semi-structured screening interviews at baseline. The contrast between subgroups (presence/absence of PEs) was calculated to test if PEs are potential effect modifiers regarding the change in parent-reported SDQ-impact (primary outcome, rated 0[low]-10[high]) and other SDQ-related outcomes.

RESULTS

Baseline PEs were present in 74 (19%) of youths. The superior effect of MMM on changes in SDQ-impact from baseline to week 18 was not effect modified by the presence of PEs (PEs[yes] -0.89 [95%CI -1.77;-0.01] vs. PEs[no] -1.10 [95%CI -1.52;-0.68], p-value for interaction .68). For secondary outcomes similar patterns were observed. Limitations Statistical power was limited to show if PEs modified treatment response. Replication and meta-analytic evidence are needed.

CONCLUSIONS

The beneficial effects of MMM transdiagnostic CBT did not differ by PE-status, indicating that youths with emotional and behavioural problems could be offered such psychotherapy irrespective of co-occurring PEs.

Use of multiple covariates in assessing treatment-effect modifiers: A methodological review of individual participant data meta-analyses

Individual participant data (IPD) meta-analyses of randomised trials are considered a reliable way to assess participant-level treatment effect modifiers but may not make the best use of the available data. Traditionally, effect modifiers are explored one covariate at a time, which gives rise to the possibility that evidence of treatment-covariate interaction may be due to confounding from a different, related covariate. We aimed to evaluate current practice when estimating treatment-covariate interactions in IPD meta-analysis, specifically focusing on involvement of additional covariates in the models. We reviewed 100 IPD meta-analyses of randomised trials, published between 2015 and 2020, that assessed at least one treatment-covariate interaction. We identified four approaches to handling additional covariates: (1) Single interaction model (unadjusted): No additional covariates included (57/100 IPD meta-analyses); (2) Single interaction model (adjusted): Adjustment for the main effect of at least one additional covariate (35/100); (3) Multiple interactions model: Adjustment for at least one two-way interaction between treatment and an additional covariate (3/100); and (4) Three-way interaction model: Three-way interaction formed between treatment, the additional covariate and the potential effect modifier (5/100). IPD is not being utilised to its fullest extent. In an exemplar dataset, we demonstrate how these approaches lead to different conclusions. Researchers should adjust for additional covariates when estimating interactions in IPD meta-analysis providing they adjust their main effects, which is already widely recommended. Further, they should consider whether more complex approaches could provide better information on who might benefit most from treatments, improving patient choice and treatment policy and practice.

"We adjusted for race": now what? A systematic review of utilization and reporting of race in American Journal of Epidemiology and Epidemiology, 2020-2021

Race is a social construct, commonly used in epidemiologic research to adjust for confounding. However, adjustment of race may mask racial disparities, thereby perpetuating structural racism. We conducted a systematic review of articles published in Epidemiology and American Journal of Epidemiology between 2020 and 2021 to (1) understand how race, ethnicity, and similar social constructs were operationalized, used, and reported; and (2) characterize good and poor practices of utilization and reporting of race data on the basis of the extent to which they reveal or mask systemic racism. Original research articles were considered for full review and data extraction if race data were used in the study analysis. We extracted how race was categorized, used-as a descriptor, confounder, or for effect measure modification (EMM)-and reported if the authors discussed racial disparities and systemic bias-related mechanisms responsible for perpetuating the disparities. Of the 561 articles, 299 had race data available and 192 (34.2%) used race data in analyses. Among the 160 US-based studies, 81 different racial categorizations were used. Race was most often used as a confounder (52%), followed by effect measure modifier (33%), and descriptive variable (12%). Fewer than 1 in 4 articles (22.9%) exhibited good practices (EMM along with discussing disparities and mechanisms), 63.5% of the articles exhibited poor practices (confounding only or not discussing mechanisms), and 13.5% were considered neither poor nor good practices. We discuss implications and provide 13 recommendations for operationalization, utilization, and reporting of race in epidemiologic and public health research.

Sample size requirements for testing treatment effect heterogeneity in cluster randomized trials with binary outcomes

Cluster randomized trials (CRTs) refer to a popular class of experiments in which randomization is carried out at the group level. While methods have been developed for planning CRTs to study the average treatment effect, and more recently, to study the heterogeneous treatment effect, the development for the latter objective has currently been limited to a continuous outcome. Despite the prevalence of binary outcomes in CRTs, determining the necessary sample size and statistical power for detecting differential treatment effects in CRTs with a binary outcome remain unclear. To address this methodological gap, we develop sample size procedures for testing treatment effect heterogeneity in two-level CRTs under a generalized linear mixed model. Closed-form sample size expressions are derived for a binary effect modifier, and in addition, a computationally efficient Monte Carlo approach is developed for a continuous effect modifier. Extensions to multiple effect modifiers are also discussed. We conduct simulations to examine the accuracy of the proposed sample size methods. We present several numerical illustrations to elucidate features of the proposed formulas and to compare our method to the approximate sample size calculation under a linear mixed model. Finally, we use data from the Strategies and Opportunities to Stop Colon Cancer in Priority Populations (STOP CRC) CRT to illustrate the proposed sample size procedure for testing treatment effect heterogeneity.

Investigating a Paradox: Toward a Better Understanding of the Relationships Between Racial Group Membership, Stress, and Major Depressive Disorder

Epidemiologic studies in the United States routinely report a lower or equal prevalence of major depressive disorder (MDD) for Black people relative to White people. Within racial groups, individuals with greater life stressor exposure experience greater prevalence of MDD; however, between racial groups this pattern does not hold. Informed by theoretical and empirical literature seeking to explain this "Black-White depression paradox," we outline 2 proposed models for the relationships between racial group membership, life stressor exposure, and MDD: an effect modification model and an inconsistent mediator model. Either model could explain the paradoxical within- and between-racial group patterns of life stressor exposure and MDD. We empirically estimated associations under each of the proposed models using data from 26,960 self-identified Black and White participants in the National Epidemiologic Survey on Alcohol and Related Conditions III (United States, 2012-2013). Under the effect modification model, we estimated relative risk effect modification using parametric regression with a cross-product term, and under the inconsistent mediation model, we estimated interventional direct and indirect effects using targeted minimum loss-based estimation. We found evidence of inconsistent mediation (i.e., direct and indirect effects operating in opposite directions), suggesting a need for greater consideration of explanations for racial patterns in MDD that operate independent of life stressor exposure. This article is part of a Special Collection on Mental Health.

The Conditional Effects of Multimorbidity on Operative Versus Nonoperative Management of Emergency General Surgery Conditions: A Retrospective Observational Study Using an Instrumental Variable Analysis

OBJECTIVE

To understand how multimorbidity impacts operative versus nonoperative management of emergency general surgery (EGS) conditions.

BACKGROUND

EGS is a heterogenous field, encompassing operative and nonoperative treatment options. Decision-making is particularly complex for older patients with multimorbidity.

METHODS

Using an instrumental variable approach with near-far matching, this national, retrospective observational cohort study of Medicare beneficiaries examines the conditional effects of multimorbidity, defined using qualifying comorbidity sets, on operative versus nonoperative management of EGS conditions.

RESULTS

Of 507,667 patients with EGS conditions, 155,493 (30.6%) received an operation. Overall, 278,836 (54.9%) were multimorbid. After adjustment, multimorbidity significantly increased the risk of in-hospital mortality associated with operative management for general abdominal patients (+9.8%; P = 0.002) and upper gastrointestinal patients (+19.9%, P < 0.001) and the risk of 30-day mortality (+27.7%, P < 0.001) and nonroutine discharge (+21.8%, P = 0.007) associated with operative management for upper gastrointestinal patients. Regardless of multimorbidity status, operative management was associated with a higher risk of in-hospital mortality among colorectal patients (multimorbid: + 12%, P < 0.001; nonmultimorbid: +4%, P = 0.003), higher risk of nonroutine discharge among colorectal (multimorbid: +42.3%, P < 0.001; nonmultimorbid: +55.1%, P < 0.001) and intestinal obstruction patients (multimorbid: +14.6%, P = 0.001; nonmultimorbid: +14.8%, P = 0.001), and lower risk of nonroutine discharge (multimorbid: -11.5%, P < 0.001; nonmultimorbid: -11.9%, P < 0.001) and 30-day readmissions (multimorbid: -8.2%, P = 0.002; nonmultimorbid: -9.7%, P < 0.001) among hepatobiliary patients.

CONCLUSIONS

The effects of multimorbidity on operative versus nonoperative management varied by EGS condition category. Physicians and patients should have honest conversations about the expected risks and benefits of treatment options, and future investigations should aim to understand the optimal management of multimorbid EGS patients.

Impact of sarcopenia on clinical outcomes for patients with resected hepatocellular carcinoma: a retrospective comparison of Eastern and Western cohorts

BACKGROUND

Patient fitness is important for guiding treatment. Muscle mass, as a reflection thereof, can be objectively measured. However, the role of East-West differences remains unclear. Therefore, we compared the impact of muscle mass on clinical outcomes after liver resection for hepatocellular carcinoma (HCC) in a Dutch [the Netherlands (NL)] and Japanese [Japan (JP)] setting and evaluated the predictive performance of different cutoff values for sarcopenia.

METHOD

In this multicenter retrospective cohort study, patients with HCC undergoing liver resection were included. The skeletal muscle mass index (SMI) was determined on computed tomography scans obtained within 3 months before surgery. The primary outcome measure was overall survival (OS). Secondary outcome measures were: 90-day mortality, severe complications, length of stay, and recurrence-free survival. The predictive performance of several sarcopenia cutoff values was studied using the concordance index (C-index) and area under the curve. Interaction terms were used to study the geographic effect modification of muscle mass.

RESULTS

Demographics differed between NL and JP. Gender, age, and body mass index were associated with SMI. Significant effect modification between NL and JP was found for BMI. The predictive performance of sarcopenia for both short-term and long-term outcomes was higher in JP compared to NL (maximum C-index: 0.58 vs. 0.55, respectively). However, differences between cutoff values were small. For the association between sarcopenia and OS, a strong association was found in JP [hazard ratio (HR) 2.00, 95% CI [1.230-3.08], P =0.002], where this was not found in NL (0.76 [0.42-1.36], P =0.351). The interaction term confirmed that this difference was significant (HR 0.37, 95% CI [0.19-0.73], P =0.005).

CONCLUSIONS

The impact of sarcopenia on survival differs between the East and West. Clinical trials and treatment guidelines using sarcopenia for risk stratification should be validated in race-dependent populations prior to clinical adoption.

Instrumented difference-in-differences

Unmeasured confounding is a key threat to reliable causal inference based on observational studies. Motivated from two powerful natural experiment devices, the instrumental variables and difference-in-differences, we propose a new method called instrumented difference-in-differences that explicitly leverages exogenous randomness in an exposure trend to estimate the average and conditional average treatment effect in the presence of unmeasured confounding. We develop the identification assumptions using the potential outcomes framework. We propose a Wald estimator and a class of multiply robust and efficient semiparametric estimators, with provable consistency and asymptotic normality. In addition, we extend the instrumented difference-in-differences to a two-sample design to facilitate investigations of delayed treatment effect and provide a measure of weak identification. We demonstrate our results in simulated and real datasets.

Effect modification of the association between fine particulate air pollution during a wildfire event and respiratory health by area-level measures of socio-economic status, race/ethnicity, and smoking prevalence

Fine particulate air pollution (PM2.5) is decreasing in most areas of the United States, except for areas most affected by wildfires, where increasing trends in PM2.5 can be attributed to wildfire smoke. The frequency and duration of large wildfires and the length of the wildfire season have all increased in recent decades, partially due to climate change, and wildfire risk is projected to increase further in many regions including the western United States. Increasingly, empirical evidence suggests differential health effects from air pollution by class and race; however, few studies have investigated such differential health impacts from air pollution during a wildfire event. We investigated differential risk of respiratory health impacts during the 2008 northern California wildfires by a comprehensive list of socio-economic status (SES), race/ethnicity, and smoking prevalence variables. Regardless of SES level across nine measures of SES, we found significant associations between PM2.5 and asthma hospitalizations and emergency department (ED) visits during these wildfires. Differential respiratory health risk was found by SES for ED visits for chronic obstructive pulmonary disease where the highest risks were in ZIP codes with the lowest SES levels. Findings for differential effects by race/ethnicity were less consistent across health outcomes. We found that ZIP codes with higher prevalence of smokers had greater risk of ED visits for asthma and pneumonia. Our study suggests that public health efforts to decrease exposures to high levels of air pollution during wildfires should focus on lower SES communities.

Effect modification of age and hypertension on cancer and prevalence of self-reported stroke - A cross-sectional study

The objective of this study was to examine the effect modification of age on the relationship between cancer and prevalence of self-reported stroke. We used cross-sectional data from the 2015-2016 iteration of the Canadian Community Health Survey. A multivariable logistic regression model was used to assess the association between cancer and self-reported stroke. Covariates were assessed for effect modification using the maximum likelihood estimation method. We analyzed 86,809 subjects; the prevalence of self-reported stroke was 1.11%. The odds ratio for the association between cancer and self-reported stroke was 1.26 (95% CI 0.98-1.61) after adjusting for age, sex, dyslipidemia, hypertension, diabetes, heart disease, education, and household income. Age and hypertension were found to be effect modifiers, and the association between cancer and self-reported stroke was stronger in younger adults and in those without hypertension. These results suggest that cancer-associated strokes may have unique underlying mechanisms compared to conventional strokes.

Linear mixed models for investigating effect modification in subgroup meta-analysis

Subgroup meta-analysis can be used for comparing treatment effects between subgroups using information from multiple trials. If the effect of treatment is differential depending on subgroup, the results could enable personalization of the treatment. We propose using linear mixed models for estimating treatment effect modification in aggregate data meta-analysis. The linear mixed models capture existing subgroup meta-analysis methods while allowing for additional features such as flexibility in modeling heterogeneity, handling studies with missing subgroups and more. Reviews and simulation studies of the best suited models for estimating possible differential effect of treatment depending on subgroups have been studied mostly within individual participant data meta-analysis. While individual participant data meta-analysis in general is recommended over aggregate data meta-analysis, conducting an aggregate data subgroup meta-analysis could be valuable for exploring treatment effect modifiers before committing to an individual participant data subgroup meta-analysis. Additionally, using solely individual participant data for subgroup meta-analysis requires collecting sufficient individual participant data which may not always be possible. In this article, we compared existing methods with linear mixed models for aggregate data subgroup meta-analysis under a broad selection of scenarios using simulation and two case studies. Both the case studies and simulation studies presented here demonstrate the advantages of the linear mixed model approach in aggregate data subgroup meta-analysis.

GRADE Guidance 36: Updates to GRADE's approach to addressing inconsistency

OBJECTIVES

To update previous GRADE guidance addressing inconsistency and interpreting subgroup analyses.

STUDY DESIGN AND SETTING

Using an iterative process, we consulted with members of the GRADE working group through multiple rounds of written feedback and discussions at GRADE working group meetings.

RESULTS

The guidance complements previous guidance with clarification on two areas: (1) assessing inconsistency and (2) assessing the credibility of possible effect modifiers that might explain inconsistency. Specifically, the guidance clarifies that inconsistency refers to variability in results, not variability in study characteristics; that inconsistency assessment for binary outcomes requires consideration of both relative and absolute effects; how to decide between narrower and broader questions in systematic reviews and guidelines; that, with the same evidence, ratings of inconsistency may differ depending on the target of certainty rating; and how GRADE inconsistency ratings relate to a statistical measure of inconsistency I² depending on the context in which one views results. The second part of the guidance illustrates, based on a worked example, the use of the Instrument to assess the Credibility of Effect Modification Analyses (ICEMAN). The guidance explains the stepwise process of moving from a subgroup analysis to assessing the credibility of effect modification and, if found credible, to subgroup-specific effect estimates and GRADE certainty ratings.

CONCLUSION

This updated guidance addresses specific conceptual and practical issues that systematic review authors frequently face when considering the degree of inconsistency in estimates of treatment effects across studies.

Network meta-interpolation: Effect modification adjustment in network meta-analysis using subgroup analyses

Effect modification (EM) may cause bias in network meta-analysis (NMA). Existing population adjustment NMA methods use individual patient data to adjust for EM but disregard available subgroup information from aggregated data in the evidence network. Additionally, these methods often rely on the shared effect modification (SEM) assumption. In this paper, we propose Network Meta-Interpolation (NMI): a method using subgroup analyses to adjust for EM that does not assume SEM. NMI balances effect modifiers across studies by turning treatment effect (TE) estimates at the subgroup- and study level into TE and standard errors at EM values common to all studies. In an extensive simulation study, we simulate two evidence networks consisting of four treatments, and assess the impact of departure from the SEM assumption, variable EM correlation across trials, trial sample size and network size. NMI was compared to standard NMA, network meta-regression (NMR) and Multilevel NMR (ML-NMR) in terms of estimation accuracy and credible interval (CrI) coverage. In the base case non-SEM dataset, NMI achieved the highest estimation accuracy with root mean squared error (RMSE) of 0.228, followed by standard NMA (0.241), ML-NMR (0.447) and NMR (0.541). In the SEM dataset, NMI was again the most accurate method with RMSE of 0.222, followed by ML-NMR (0.255). CrI coverage followed a similar pattern. NMI's dominance in terms of estimation accuracy and CrI coverage appeared to be consistent across all scenarios. NMI represents an effective option for NMA in the presence of study imbalance and available subgroup data.

Study of effect modifiers of genetically predicted CETP reduction

Genetic variants in drug targets can be used to predict the long-term, on-target effect of drugs. Here, we extend this principle to assess how sex and body mass index may modify the effect of genetically predicted lower CETP levels on biomarkers and cardiovascular outcomes. We found sex and body mass index (BMI) to be modifiers of the association between genetically predicted lower CETP and lipid biomarkers in UK Biobank participants. Female sex and lower BMI were associated with higher high-density lipoprotein cholesterol and lower low-density lipoprotein cholesterol for the same genetically predicted reduction in CETP concentration. We found that sex also modulated the effect of genetically lower CETP on cholesterol efflux capacity in samples from the Montreal Heart Institute Biobank. However, these modifying effects did not extend to sex differences in cardiovascular outcomes in our data. Our results provide insight into the clinical effects of CETP inhibitors in the presence of effect modification based on genetic data. The approach can support precision medicine applications and help assess the external validity of clinical trials.

The impact of monthly air pollution exposure and its interaction with individual factors: Insight from a large cohort study of comprehensive hospitalizations in Guangzhou area

Background

Although the association between short-term air pollution exposure and certain hospitalizations has been well documented, evidence on the effect of longer-term (e. g., monthly) air pollution on a comprehensive set of outcomes is still limited.

Method

A total of 68,416 people in South China were enrolled and followed up during 2019-2020. Monthly air pollution level was estimated using a validated ordinary Kriging method and assigned to individuals. Time-dependent Cox models were developed to estimate the relationship between monthly PM₁₀ and O₃ exposures and the all-cause and cause-specific hospitalizations after adjusting for confounders. The interaction between air pollution and individual factors was also investigated.

Results

Overall, each 10 μg/m³ increase in PM₁₀ concentration was associated with a 3.1% (95%CI: 1.3%-4.9%) increment in the risk of all-cause hospitalization. The estimate was even greater following O₃ exposure (6.8%, 5.5%-8.2%). Furthermore, each 10 μg/m³ increase in PM₁₀ was associated with a 2.3%-9.1% elevation in all the cause-specific hospitalizations except for those related to respiratory and digestive diseases. The same increment in O₃ was relevant to a 4.7%-22.8% elevation in the risk except for respiratory diseases. Additionally, the older individuals tended to be more vulnerable to PM₁₀ exposure (P _interaction: 0.002), while the alcohol abused and those with an abnormal BMI were more vulnerable to the impact of O₃ (P _interaction: 0.052 and 0.011). However, the heavy smokers were less vulnerable to O₃ exposure (P _interaction: 0.032).

Conclusion

We provide comprehensive evidence on the hospitalization hazard of monthly PM₁₀ and O₃ exposure and their interaction with individual factors.

The Interplay Among Natural Menopause, Insomnia, and Cognitive Health: A Population-Based Study

PURPOSE

The interrelationships among age at menopause, sleep, and brain health have been insufficiently studied. This study sought to examine the influence of age at natural menopause and insomnia symptoms on long-term cognitive function among US women.

PATIENTS AND METHODS

Our study included a nationally representative cohort of US adults age 50+ from the Health and Retirement Study (2008-2018). We restricted this cohort to 5880 women age 50+, from a diverse racial and ethnic groups. Age at menopause was retrieved from baseline (2008) for women having natural menopause. Five questions were used to identify women with insomnia symptoms (2010 and 2012): trouble falling asleep, nighttime awakenings, early morning awakenings, feelings of nonrestorative sleep, and use of sleep aids. A battery of four neuropsychological tests was conducted biennially (years) to evaluate cognitive function. Longitudinal associations between age at natural menopause and cognitive function were estimated with mixed effects models with a random intercept. Insomnia symptoms were examined as potential mediators or modifiers in the pathway between age at menopause and cognition.

RESULTS

One year earlier in age at menopause was associated with a 0.49 lower mean in composite cognitive score, in any given survey year (adjusted p = 0.002). Earlier age at menopause was associated with higher risk of developing insomnia symptoms (eg, trouble falling asleep OR = 0.97; 95% CI: 0.96, 0.99), and insomnia symptoms were associated with worse cognitive performance (eg, trouble falling asleep, beta = -0.5, p-value = 0.02). Therefore, insomnia symptoms could potentially mediate the association between age at natural menopause and cognition.

CONCLUSION

Earlier age at menopause is associated with a lower score in cognitive performance. This association may be mediated by insomnia symptoms. Our findings spotlight that among women who experience early menopause, there is the need for studies of sleep-based interventions to mitigate cognitive decline.

Validating the Assumptions of Population Adjustment: Application of Multilevel Network Meta-regression to a Network of Treatments for Plaque Psoriasis

BACKGROUND

Network meta-analysis (NMA) and indirect comparisons combine aggregate data (AgD) from multiple studies on treatments of interest but may give biased estimates if study populations differ. Population adjustment methods such as multilevel network meta-regression (ML-NMR) aim to reduce bias by adjusting for differences in study populations using individual patient data (IPD) from 1 or more studies under the conditional constancy assumption. A shared effect modifier assumption may also be necessary for identifiability. This article aims to demonstrate how the assumptions made by ML-NMR can be assessed in practice to obtain reliable treatment effect estimates in a target population.

METHODS

We apply ML-NMR to a network of evidence on treatments for plaque psoriasis with a mix of IPD and AgD trials reporting ordered categorical outcomes. Relative treatment effects are estimated for each trial population and for 3 external target populations represented by a registry and 2 cohort studies. We examine residual heterogeneity and inconsistency and relax the shared effect modifier assumption for each covariate in turn.

RESULTS

Estimated population-average treatment effects were similar across study populations, as differences in the distributions of effect modifiers were small. Better fit was achieved with ML-NMR than with NMA, and uncertainty was reduced by explaining within- and between-study variation. We found little evidence that the conditional constancy or shared effect modifier assumptions were invalid.

CONCLUSIONS

ML-NMR extends the NMA framework and addresses issues with previous population adjustment approaches. It coherently synthesizes evidence from IPD and AgD studies in networks of any size while avoiding aggregation bias and noncollapsibility bias, allows for key assumptions to be assessed or relaxed, and can produce estimates relevant to a target population for decision-making.

HIGHLIGHTS

Multilevel network meta-regression (ML-NMR) extends the network meta-analysis framework to synthesize evidence from networks of studies providing individual patient data or aggregate data while adjusting for differences in effect modifiers between studies (population adjustment). We apply ML-NMR to a network of treatments for plaque psoriasis with ordered categorical outcomes.We demonstrate for the first time how ML-NMR allows key assumptions to be assessed. We check for violations of conditional constancy of relative effects (such as unobserved effect modifiers) through residual heterogeneity and inconsistency and the shared effect modifier assumption by relaxing this for each covariate in turn.Crucially for decision making, population-adjusted treatment effects can be produced in any relevant target population. We produce population-average estimates for 3 external target populations, represented by the PsoBest registry and the PROSPECT and Chiricozzi 2019 cohort studies.

Long-term exposure to air pollution and lung function among children in China: Association and effect modification

Background

Children are vulnerable to the respiratory effects of air pollution, and their lung function has been associated with long-term exposure to low air pollution level in developed countries. However, the impact of contemporary air pollution level in developing countries as a result of recent efforts to improve air quality on children's lung function is less understood.

Methods

We obtained a cross-sectional sample of 617 schoolchildren living in three differently polluted areas in Anhui province, China. 2-year average concentrations of air pollutants at the year of spirometry and the previous year (2017-2018) obtained from district-level air monitoring stations were used to characterize long-term exposure. Forced vital capacity (FVC), forced expiratory volume in 1 second (FEV₁), and forced expiratory flow between 25 and 75% of FVC (FEF_25-75) were determined under strict quality control. Multivariable regression was employed to evaluate the associations between air pollution level and lung function parameters, overall and by demographic characteristics, lifestyle, and vitamin D that was determined by liquid chromatography tandem mass spectrometry.

Results

Mean concentration of fine particulate matter was 44.7 μg/m³, which is slightly above the interim target 1 standard of the World Health Organization. After adjusting for confounders, FVC, FEV₁, and FEF_25-75 showed inverse trends with increasing air pollution levels, with children in high exposure group exhibiting 87.9 [95% confidence interval (CI): 9.5, 166.4] mL decrement in FEV₁ and 195.3 (95% CI: 30.5, 360.1) mL/s decrement in FEF_25-75 compared with those in low exposure group. Additionally, the above negative associations were more pronounced among those who were younger, girls, not exposed to secondhand smoke, non-overweight, physically inactive, or vitamin D deficient.

Conclusions

Our study suggests that long-term exposure to relatively high air pollution was associated with impaired lung function in children. More stringent pollution control measures and intervention strategies accounting for effect modification are needed for vulnerable populations in China and other developing countries.

Inappropriate Evaluation of Effect Modifications Based on Categorical Outcomes: A Systematic Review of Randomized Controlled Trials

Our meta-epidemiological study aimed to describe the prevalence of reporting effect modification only on relative scale outcomes and inappropriate interpretations of the coefficient of interaction terms in nonlinear models on categorical outcomes. Our study targeted articles published in the top 10 high-impact-factor journals between 1 January and 31 December 2021. We included two-arm, parallel-group, interventional superiority randomized controlled trials to evaluate the effects of modifications on categorical outcomes. The primary outcomes were the prevalence of reporting effect modifications only on relative scale outcomes and that of inappropriately interpreting the coefficient of interaction terms in nonlinear models on categorical outcomes. We included 52 articles, of which 41 (79%) used nonlinear regression to evaluate effect modifications. At least 45/52 articles (87%) reported effect modifications based only on relative scale outcomes, and at least 39/41 (95%) articles inappropriately interpreted the coefficient of interaction terms merely as indices of effect modifications. The quality of the evaluations of effect modifications in nonlinear models on categorical outcomes was relatively low, even in randomized controlled trials published in medical journals with high impact factors. Researchers should report effect modifications of both absolute and relative scale outcomes and avoid interpreting the coefficient of interaction terms in nonlinear regression analyses.

Doubly robust adaptive LASSO for effect modifier discovery

Effect modification occurs when the effect of a treatment on an outcome differsaccording to the level of some pre-treatment variable (the effect modifier). Assessing an effect modifier is not a straight-forward task even for a subject matter expert. In this paper, we propose a two-stageprocedure to automatically selecteffect modifying variables in a Marginal Structural Model (MSM) with a single time point exposure based on the two nuisance quantities (the conditionaloutcome expectation and propensity score). We highlight the performance of our proposal in a simulation study. Finally, to illustrate tractability of our proposed methods, we apply them to analyze a set of pregnancy data. We estimate the conditional expected difference in the counterfactual birth weight if all women were exposed to inhaled corticosteroids during pregnancy versus the counterfactual birthweight if all women were not, using data from asthma medications during pregnancy.

Effects of risk factors for ovarian cancer in women with and without endometriosis

OBJECTIVE

To evaluate the associations between 10 well-established ovarian cancer risk factors and risk of ovarian cancer among women with vs. without endometriosis.

DESIGN

Pooled analysis of 9 case-control studies in the Ovarian Cancer Association Consortium.

SETTING

Population-based.

PATIENT(S)

We included 8,500 women with ovarian cancer, 13,592 control women.

INTERVENTION(S)

Ten well-established ovarian cancer risk factors.

MAIN OUTCOME MEASURE(S)

Risk of ovarian cancer for women with and without endometriosis.

RESULT(S)

Most risk factor-ovarian cancer associations were similar when comparing women with and without endometriosis, and no interactions were statistically significant. However, body mass index (BMI) 25-<30 kg/m² was associated with increased ovarian cancer risk among women with endometriosis (odds ratio [OR] = 1.27, 95% confidence interval [CI] 1.00-1.60), but not associated with the risk among women without endometriosis (OR = 0.97; 95% CI, 0.91-1.05) when compared with BMI 18.5-<25 kg/m²; an increased risk was observed for a BMI ≥30 kg/m², although there was little difference comparing women with endometriosis (OR = 1.21; 95% CI, 0.94-1.57) to women without (OR = 1.13; 95% CI, 1.04-1.22) (P-interaction = .51). Genital talcum powder use and long-term menopausal estrogen-only therapy use showed increased ovarian cancer risk, but risk appeared greater for those with endometriosis vs. those without (genital talcum powder: OR = 1.38; 95% CI, 1.04-1.84 vs. OR = 1.12; 95% CI, 1.01-1.25, respectively; ≥10 years of estrogen-only therapy: OR = 1.88; 95% CI, 1.09-3.24 vs. OR = 1.42; 95% CI, 1.14-1.76, respectively); neither of these interactions were statistically significant (P-interaction = .65 and P-interaction = .96, respectively).

CONCLUSION(S)

The associations between ovarian cancer and most risk factors were similar among women with and without endometriosis. However, there was some suggestion of differences by endometriosis status for BMI, menopausal hormone therapy use, and genital talcum powder use, highlighting the complexity of ovarian cancer etiology.

Spillover benefit of pre-exposure prophylaxis for HIV prevention: evaluating the importance of effect modification using an agent-based model

We developed an agent-based model using a trial emulation approach to quantify effect measure modification of spillover effects of pre-exposure prophylaxis (PrEP) for HIV among men who have sex with men (MSM) in the Atlanta-Sandy Springs-Roswell metropolitan area, Georgia. PrEP may impact not only the individual prescribed, but also their partners and beyond, known as spillover. We simulated a two-stage randomised trial with eligible components (≥3 agents with ≥1 HIV+ agent) first randomised to intervention or control (no PrEP). Within intervention components, agents were randomised to PrEP with coverage of 70%, providing insight into a high PrEP coverage strategy. We evaluated effect modification by component-level characteristics and estimated spillover effects on HIV incidence using an extension of randomisation-based estimators. We observed an attenuation of the spillover effect when agents were in components with a higher prevalence of either drug use or bridging potential (if an agent acts as a mediator between ≥2 connected groups of agents). The estimated spillover effects were larger in magnitude among components with either higher HIV prevalence or greater density (number of existing partnerships compared to all possible partnerships). Consideration of effect modification is important when evaluating the spillover of PrEP among MSM.

Redefining effect modification

The odds ratio (OR) has been misunderstood in evidence based medicine and clinical epidemiology. Currently, "noncollapsibility" is considered a problem with interpretation of the OR and it is thought that the OR is rarely the parameter of interest for causal inference or interpretation of effect modification. The current focus on the relative risk (RR) and risk difference (RD) suffers from an important limitation: they are not solely measures of effect and vary numerically with baseline risk. In this paper, generalized linear models are examined in terms of the three binary effect measures commonly used in epidemiology to demonstrate that ORs may be the only way to interpret effect modification and have properties that should make them the parameter of interest for causal inference. We look forward to discussion, debate, and counter-views on this issue from the epidemiology community.

Association between household incense burning and executive function in Chinese children

Incense burning is common in Asia including China. Research investigating the association between household incense burning and children's neurodevelopment is scarce. We aimed to examine this association in Chinese children. In 2019, we randomly enrolled 8293 children aged 6-12 years from 5 elementary schools in Guangzhou, southern China. Information on duration and frequency of household incense burning was collected using a questionnaire. Children's executive function was evaluated using the parental report of the Behavioral Rating Inventory of Executive Function. A general linear model was used to assess the associations between incense burning and executive function. Children who occasionally and frequently exposed to incense burning exhibited worse performance on executive function. For example, frequent incense burning was associated with increases in behavioral regulation index (BRI) of 1.77 (95%CI: 0.97, 2.58) points and metacognition index (MI) of 1.40 (95%CI: 0.60, 2.20) points, compared to never incense burning group. Parental smoking and household income were significant modifiers of the associations, with the stronger associations were observed in children having smoking parent(s) and poorer household income. The findings suggest that household incense burning was associated with poorer executive function, especially in children whose parent(s) were smokers and in those with low household income.

Do Neighborhood Factors Modify the Effects of Lead Exposure on Child Behavior?

Lead exposure and neighborhoods can affect children’s behavior, but it is unclear if neighborhood characteristics modify the effects of lead on behavior. Understanding these modifications has important intervention implications. Blood lead levels (BLLs) in children (~7 years) from Montevideo, Uruguay, were categorized at 2 µg/dL. Teachers completed two behavior rating scales (n = 455). At one-year follow-up (n = 380), caregivers reported child tantrums and parenting conflicts. Multilevel generalized linear models tested associations between BLLs and behavior, with neighborhood disadvantage, normalized difference vegetation index (NDVI), and distance to nearest greenspace as effect modifiers. No effect modification was noted for neighborhood disadvantage or NDVI. Children living nearest to greenspace with BLLs < 2 µg/dL were lower on behavior problem scales compared to children with BLLs ≥ 2 µg/dL. When furthest from greenspace, children were similar on behavior problems regardless of BLL. The probability of daily tantrums and conflicts was ~20% among children with BLLs < 2 µg/dL compared to ~45% among children with BLLs ≥ 2 µg/dL when closest to greenspace. Furthest from greenspace, BLLs were not associated with tantrums and conflicts. Effect modification of BLL on child behavior by distance to greenspace suggests that interventions should consider both greenspace access and lead exposure prevention.

The moderating role of partners' education on early antenatal care in northern Ghana

BACKGROUND

Early antenatal care (ANC) is essential for improving maternal and child health outcomes. The primary aims of this study were to 1) estimate the association between partners' education attainment and early ANC, and 2) determine whether partners' level of education modified the relationship between mothers' education, mothers' age, planned pregnancy, employment status and early ANC.

METHODS

Data were obtained from a cross-sectional study conducted from April to May 2021 among 519 mothers with a live birth in the past year in the Nabdam district in the Upper East Region in northern Ghana. Generalized estimating equations were used to assess whether partners' level of education modified the relationship between mothers' education, mothers' age, planned pregnancy, employment status and early ANC. Effect modification was assessed on the additive and multiplicative scales using adjusted prevalence ratios (aPR) and corresponding 95% confidence intervals.

RESULTS

Mothers whose partners had secondary or higher education had a 26% higher prevalence of early ANC compared to mothers whose partners had less than a secondary level of education (aPR: 1.26, 95% CI: 1.05,1.51). There was evidence of effect modification by partners' education on the relationship between planned pregnancy and early ANC on both the additive (Relative excess risk due to interaction [RERI]: 0.61, 95% CI: 0.07,0.99), and multiplicative (ratio of PRs: 1.64, 95% CI: 1.01,2.70) scales. Among mothers whose partners had less than secondary education, mothers who had teenage pregnancy (i.e., aged 18-19 years old during pregnancy) were less likely to have early ANC compared to those who did not have teenage pregnancy (aPR: 0.71, 95% CI: 0.53,0.97). Among mothers whose partners had a secondary or higher education, early ANC was more prevalent among employed mothers compared to those who were unemployed (aPR: 1.27, 95% CI: 1.02,1.57).

CONCLUSIONS

Our findings suggest that whilst mothers whose partners had a secondary or higher education were more likely to initiate early ANC, supporting such women to plan their pregnancies can further increase the coverage of early ANC.

E-values for effect heterogeneity and approximations for causal interaction

Background

Estimates of effect heterogeneity (i.e. the extent to which the causal effect of one exposure varies across strata of a second exposure) can be biased if the exposure–outcome relationship is subject to uncontrolled confounding whose severity differs across strata of the second exposure.

Methods

We propose methods, analogous to the E-value for total effects, that help to assess the sensitivity of effect heterogeneity estimates to possible uncontrolled confounding. These E-value analogues characterize the severity of uncontrolled confounding strengths that would be required, hypothetically, to ‘explain away’ an estimate of multiplicative or additive effect heterogeneity in the sense that appropriately controlling for those confounder(s) would have shifted the effect heterogeneity estimate to the null, or alternatively would have shifted its confidence interval to include the null. One can also consider shifting the estimate or confidence interval to an arbitrary non-null value. All of these E-values can be obtained using the R package EValue.

Results

We illustrate applying the proposed E-value analogues to studies on: (i) effect heterogeneity by sex of the effect of educational attainment on dementia incidence and (ii) effect heterogeneity by age on the effect of obesity on all-cause mortality.

Conclusion

Reporting these proposed E-values could help characterize the robustness of effect heterogeneity estimates to potential uncontrolled confounding.

Association Between Depression and Risk of Incident Cardiovascular Diseases and Its Sex and Age Modifications: A Prospective Cohort Study in Southwest China

To examine possible associations between depression and cardiovascular disease (CVD) incidence and whether demographic factors modified those associations in the Chinese population. This prospective cohort study comprised 7,735 adults aged 18 years or older in Guizhou, China from 2010 to 2020. The Patient Health Questionnaire-9 (PHQ-9) was used to measure the prevalence of depression. Cox proportional hazard models were used to estimated hazard ratios (HRs) and 95% confidence intervals (95%CIs) of depression and incident CVD. We identified 215 CVD cases (including 28 acute myocardial infarction (AMI) and 197 stroke cases) during an average follow-up of 7.07 years. In the multivariable-adjusted model, baseline PHQ-9 score was associated with incident CVD, AMI, and stroke. The HR per 1-SD increase for PHQ-9 score was 1.14 (95%CI: 1.03, 1.26) for CVD, 1.26 (95%CI: 1.01, 1.57) for AMI, and 1.12 (95%CI: 1.01, 1.25) for stroke. Compared with participants without depression, those with any mild or more advanced depression had a higher risk of incident CVD (HR: 1.69, 95%CI: 1.08, 2.64) and AMI (HR: 3.36, 95%CI: 1.17, 10.56). Associations between depression with CVD and stroke were suggested to be even stronger among women and participants aged <65 years (P for interaction <0.05). The effect of depression on stroke tended to be preserved in non-smokers. Depression was associated with a higher risk of incident CVD, AMI, and stroke in adults of Southwest, China, particularly in women, participants aged <65 years, and non-smokers. These findings highlighted the importance and urgency of depression healthcare.

Sex Differences in Long-term Outcomes After Group B Streptococcal Infections During Infancy in Denmark and the Netherlands: National Cohort Studies of Neurodevelopmental Impairments and Mortality

BACKGROUND

Male infants have a higher incidence of invasive group B Streptococcus disease (iGBS) compared with female infants; however, data on sex differences in mortality and long-term outcomes after iGBS are lacking. We assessed whether a child's sex influences the effects of iGBS on mortality and risk of neurodevelopmental impairments (NDIs).

METHODS

We used Danish and Dutch registry data to conduct a nationwide cohort study of infants with a history of iGBS. A comparison cohort, children without a history of iGBS, was randomly selected and matched on relevant factors. Effect modification by sex was assessed on additive and multiplicative scales.

RESULTS

Our analyses included data from children with a history of iGBS in Denmark (period 1997 -2017; n = 1432) and the Netherlands (2000 -2017; n = 697) and from 21 172 children without iGBS. There was no clear evidence of between-sex heterogeneity in iGBS-associated mortality. Boys had a higher risk of NDI, with evidence for effect modification on additive scale at the age of 5 years for any NDI (relative excess risk due to interaction = 1.28; 95% confidence interval [CI], -0.53 to 3.09 in Denmark and 1.14; 95% CI, -5.13 to 7.41 in the Netherlands). A similar pattern was observed for moderate/severe NDI at age 5 years in Denmark and age 10 years in the Netherlands.

CONCLUSION

Boys are at higher risk of NDI ; our results suggest this is disproportionally increased in those who develop iGBS. Future studies should investigate mechanisms of this effect modification by sex.

The Role of Temperature in Modifying the Risk of Ozone-Attributable Mortality under Future Changes in Climate: A Proof-of-Concept Analysis

Air pollution risk assessments typically estimate ozone-attributable mortality counts using concentration-response (C-R) parameters from epidemiologic studies that treat temperature as a potential confounder. However, some recent epidemiologic studies have indicated that temperature can modify the relationship between short-term ozone exposure and mortality, which has potentially important implications when considering the impacts of climate change on public health. This proof-of-concept analysis quantifies counts of temperature-modified ozone-attributable mortality using temperature-stratified C-R parameters from a multicity study in which the pooled ozone-mortality effect coefficients change in concert with daily temperature. Meteorology downscaled from two global climate models is used with a photochemical transport model to simulate ozone concentrations over the 21st century using two emission inventories: one holding air pollutant emissions constant at 2011 levels and another accounting for reduced emissions through the year 2040. The late century climate models project increased summer season temperatures, which in turn yields larger total counts of ozone-attributable deaths in analyses using temperature-stratified C-R parameters compared to the traditional temperature confounder approach. This analysis reveals substantial heterogeneity in the magnitude and distribution of the temperature-stratified ozone-attributable mortality results, which is a function of regional variability in both the C-R relationship and the model-predicted temperature and ozone.

Vitamin B12 Supplementation and Vitamin B12 Blood Serum Levels: Evaluation of Effect Modification by Gender and Smoking Status

Research suggests that high intake of supplemental vitamin B₁₂ may be associated with increased risk of cancer, with some evidence that this association may vary by gender and smoking status. This investigation evaluates if similar patterns in association are observed for data for 11,757 adults from the National Health and Nutrition Examination Survey (1999-2006). Survey-weighted multivariable-adjusted linear regression was used to evaluate the association between regular B₁₂ supplement use and log-transformed serum B₁₂ levels. Persons taking vitamin B₁₂ through a multivitamin/multimineral (MVMM) had a median supplemental intake of 12 mcg/day (Q1: 6, Q3: 25), compared to 100 mcg/day (Q1: 22, Q3: 500) for persons reporting supplemental B₁₂ intake through a MVMM-exclusive source. MVMM users had a geometric mean serum B12 26% (95% CI: 23%-30%) higher than nonusers, whereas MVMM-exclusive users' geometric mean was 61% (95% CI: 53%-70%) higher than nonusers (p-trend < 0.001). Although a positive trend (p-trend < 0.001) was observed for both men and women, the association was stronger among women (p-interaction < 0.001). No interaction was observed for smoking status (p-interaction = 0.45). B₁₂ supplementation is associated with higher levels of serum B₁₂, with significant interaction by gender but not smoking. Further work is needed to better understand the interplay of B₁₂ and gender.

The Associations between Poor Antibiotic and Antimicrobial Resistance Knowledge and Inappropriate Antibiotic Use in the General Population Are Modified by Age

OBJECTIVES

Understanding factors influencing inappropriate antibiotic use can guide the design of interventions to improve antibiotic practices and reduce antimicrobial resistance (AMR).

METHODS

A nationally representative cross-sectional survey (N = 2004) was conducted between November 2020 and January 2021. Knowledge of antibiotic use and AMR using the World Health Organization's Multi-Country AMR Survey questionnaire, and antibiotic practices were examined. Multivariable logistic regression was performed to identify factors associated with inappropriate antibiotic use and examine effect measure modifications.

RESULTS

After adjusting for potential confounding, poor knowledge of antibiotic use was associated with a 3x increased odds of inappropriate antibiotic use in adults aged ≥50 years (aOR 3.11, 95% CI [2.24-4.32]), 5× increased odds in those aged 35-49 years (aOR 4.88, 95% CI [3.32-7.16]), and 7× increased odds in those aged 21-34 years (aOR 6.58, 95% CI [4.19-10.33]). While there was no statistically significant association in adults aged ≥50 years, poor knowledge of AMR increased the odds of inappropriate antibiotic use by 4 times in adults aged 35-49 years (aOR 3.73, 95% CI [1.53-9.11]) and 5 times in those aged 21-34 years (aOR 4.90, 95% CI [1.84-13.02]).

CONCLUSIONS

Targeted educational interventions for specific age groups are needed in conjunction with empowering the public with knowledge of antibiotic use and AMR.

The association between obstructive sleep apnea risk and cognitive disorders: a population-based study

STUDY OBJECTIVES

To examine the association between OSA risk and cognitive disorders among US adults.

METHODS

Data from the 2016 wave of the Health and Retirement Study (HRS) were utilized. Probable OSA cases were identified with survey items that resembled critical elements of a clinically validated OSA screen (STOP-Bang questionnaire). Weighted prevalence of cognitive impairment not dementia (CIND) and dementia among individuals with and without probable OSA were assessed. Cross-sectional analyses of associations between OSA risk and cognitive outcomes, along with effect modification by race and ethnicity, were estimated using imputed data.

RESULTS

Of the 20,910 HRS participants, 60% had probable OSA. CIND and dementia were more common among adults with probable OSA as compared to those without (12.7% vs. 8.0% for CIND; 3.2% vs 2.0% for dementia). Probable OSA was associated with CIND (OR=1.22, 1.08-1.37) and dementia (OR=1.27, 1.04-1.54). Race/ethnicity significantly modified the association between probable OSA and CIND, with a higher risk for CIND in Whites (OR=1.35, 1.17-1.57) as compared to non-Whites (OR=0.98, 0.81-1.19).

CONCLUSIONS

CIND and dementia are more common among older adults who are at high risk for OSA, as compared to low-risk individuals. These data highlight the importance of consideration of OSA risk in large-scale studies of OSA and cognitive disorders.

Donor Death Category Is an Effect Modifier Between Cold Ischemia Time and Post-transplant Graft Function in Deceased-Donor Kidney Transplant Recipients

Objectives: We aimed to analyze the effect of cold ischemia time (CIT) on post-transplant graft function through mixed-effect model analysis to reduce the bias caused by paired mate kidneys.

Methods: We reviewed all kidney transplantation records from 2015 to 2019 at our center. After applying the exclusion criteria, 561 cases were included for analysis. All donor characteristics, preservation and matching information, and recipient characteristics were collected. Transplant outcomes included delayed graft function (DGF) and estimated glomerular filtration rate (eGFR). Generalized linear mixed models were applied for analysis. We also explored potential effect modifiers, namely, donor death category, expanded criteria donors, and donor death causes.

Results: Among the 561 cases, 79 DGF recipients developed DGF, and 15 recipients who died after surgery were excluded from the eGFR estimation. The median stable eGFR of the 546 recipients was 60.39 (47.63, 76.97) ml/min/1.73 m². After adjusting for confounding covariates, CIT had a negative impact on DGF incidence [odds ratio = 1.149 (1.006, 1.313), P = 0.041]. In the evaluation of the impact on eGFR, the regression showed that CIT had no significant correlation with eGFR [β = −0.287 (−0.625, 0.051), P = 0.096]. When exploring potential effect modifiers, only the death category showed a significant interaction with CIT in the effect on eGFR (P_interaction = 0.027). In the donation after brain death (DBD) group, CIT had no significant effect on eGFR [β = 0.135 (−0.433, 0.702), P = 0.642]. In the donation after circulatory death/donation after brain death followed by circulatory death (DCD/DBCD) group, CIT had a significantly negative effect on eGFR [β= −0.700 (−1.196, −0.204), P = 0.006]. Compared to a CIT of 0–6 h, a CIT of 6–8 or 8–12 h did not decrease the post-transplant eGFR. CIT over 12 h (12–16 h or over 16 h) significantly decreased eGFR. With the increase in CIT, the regenerated eGFR worsened (P_trend = 0.011).

Conclusion: Considering the effect of paired mate kidneys, the risk of DGF increased with prolonged CIT. The donor death category was an effect modifier between CIT and eGFR. Prolonged CIT did not reduce the eGFR level in recipients from DBDs but significantly decreased the eGFR in recipients from DCDs/DBCDs. This result indicates the potential biological interaction between CIT and donor death category.

Association of Night Shift Work With Chronic Spontaneous Urticaria and Effect Modification by Circadian Dysfunction Among Workers

Purpose: Night shift work is common in the current working environment and is a risk factor for many diseases. The study aimed to explore the relationship between night shift work with chronic spontaneous urticaria (CSU), and the modification effect of circadian dysfunction on it. Methods: A cross-sectional survey was conducted among Chinese workers. Exposure was measured by night work history and duration. Circadian dysfunction was characterized by excessive daytime sleepiness (EDS). The diagnosis of CSU was made by dermatologists who were investigating on the spot. The effect size was expressed as odds ratios (ORs). Results: A total of 8,057 participants were recruited, and 7,411 (92%) with complete information were included in the final analyses. The prevalence rates of CSU for workers without night shift and those with night shift history were 0.73 and 1.28%, respectively. Compared with workers who never worked night shifts, the risk of CSU increased with the length of night shift work: OR = 1.55 (95% confidence interval [CI]: 0.78-3.06) for duration <5 years and OR = 1.91 (95% CI: 1.12-3.26) for duration ≥5 years. EDS s EDS has been shown to modify this combination. Among workers without EDS, there was no association between night shift and CSU (OR = 0.94; 95% CI: 0.49-1.79). Whereas, in participants with EDS, the correlation was significant (OR = 3.58; 95% CI: 1.14-11.20). However, the effect modification by sleep disturbance was not observed. Conclusions: Night shift work is a risk factor for CSU, and there is a dose-response relationship between night shift work hours and the risk of CSU. This connection may be modified by circadian dysfunction.

Why Does Influenza Hit the Poor More Than the Rich? A 1931 Social Epidemiology Article That Broke New Ground in the History of Confounding, Mediation, and Interaction

In 1931, Edgar Sydenstricker, the former statistician of the US Public Health Service, challenged the common belief that the 1918 influenza outbreak had affected "the rich and the poor alike." Using data from 112,317 participants in a 1918 US national survey, he observed that, on the contrary, both morbidity and mortality from the flu had been higher among the poor than among the rich. To explain these differences, Sydenstricker stratified the analyses by 2 measures of affluence collected in the survey: "economic status" (from "very poor" to "well-to-do") and household crowding (i.e., number of people per household room). Economic status was associated with influenza attack rates within categories of crowding, but not the opposite, suggesting that characteristics of poverty other than "household congestion" were the culprit of the poor's higher influenza burden. Attack rate ratios for influenza in infants and older adults were greater for the poor or very poor. Sydenstricker reanalyzed an already 12-year-old data set in the context of the Great Depression to build the evidence base relating poverty to ill health. For this purpose he used a stratification approach to assess confounding, mediation, and interaction before the concepts were formally named.

Association between early antenatal care and antenatal care contacts across low-and middle-income countries: effect modification by place of residence

Objectives

The relationship between early antenatal care (ANC) and having eight or more antenatal contacts (ANC8+) have not been well characterised across low-and middle-income countries (LMICs). It is also unclear whether the association between early ANC and ANC8+ is modified by a woman's place of residence. Our primary aim was to assess the relationship between early ANC and ANC8+ and assessed whether this relationship was modified by place of residence. We also estimated the coverage of ANC8+ across LMICs.

Methods

We analysed data on 207388 mothers with a live birth using multiple indicator cluster surveys (MICSs) from 30 LMICS conducted between 2017 and 2020. Modified Poisson regression with a robust variance was used to assess the relationship between early ANC and ANC8+, whilst adjusting for country, clustering, stratification and sampling weights. Effect modification by place of residence was assessed on the additive and multiplicative scales. Meta-analysis was used to pool prevalent estimates of ANC8+ across all countries.

Results

The overall prevalence of ANC8+ was 35.6% and ranged from 1.7% in Madagascar to 99.4% in Belarus. Early ANC was positively associated with ANC8+ [adjusted prevalence ratio (aPR): 2.61, 95% CI: 1.82,3.74]. There was evidence of positive effect modification on the additive [relative excess risk due to interaction (RERI): 0.39, 95% CI: 0.35, 0.44] and multiplicative (aPR: 1.78, 95% CI: 1.08, 2.95) scales.

Conclusion

Many LMICs may not have adopted the WHO 2016 guidelines on ANC8+. Women who had early ANC were more likely to have ANC8+, particularly those in rural areas.