Causal Inference Methodology Research Articles

Impact of addressing modifiable risk factors on the 10-year risk of cardiovascular disease in individuals with familial hypercholesterolemia: a causal analysis utilising UK biobank

Abstract Background Familial hypercholesterolemia (FH) is a prevalent autosomal dominant disorder characterised by elevated low-density lipoprotein cholesterol (LDL-C) levels from birth. These elevated LDL-C levels significantly amplify the lifetime risk of cardiovascular disease (CVD). While the impact of early and effective initiation of lipid-lowering medication (LLM) on CVD risk is well documented in genetically confirmed FH, the additional impact of addressing modifiable lifestyle risk factors (MLRF) in that patient population remains largely unquantified. Methods This study quantified the potential effects of successfully managing 4 MLRF (smoking, obesity, alcohol consumption, and low physical activity) on the 10-year incidence of a first CVD diagnosis in adults with genetically confirmed FH treated with LLM within the UK Biobank. The synthesised causal estimates of these 4 MLF were obtained using two recent causal analysis frameworks, namely the doWhy causal calculus framework and the Causal Forest Double Machine Learning Method. Results In 660 FH individuals (mean age 56±8 years; 60.15% women), smoking cessation and reducing body mass index below 30 kg/m2 delivered the most substantial risk reductions (pooled estimates, adjusted for the 3 other MLRF: -6.5% and -3.2%, respectively; both p<0.001), followed by initiating moderate or high physical activity (-1.7%; p<0.001; Figure). Notably, cessation of alcohol consumption significantly increased CVD risk (+2.5%; p<0.001), aligning with existing literature. Conclusion In over 10 years, reductions in CVD risk ranging from 1.7% to 6.5% were achievable through smoking cessation, BMI reduction, and enhanced physical activity in FH individuals. Leveraging causal inference methodologies facilitates the discernment of which risk factors offer the most pronounced additional benefits beyond lipid-lowering interventions in this high-risk population.Forest Plot of Causal Effect Estimates

A Causal Inference Methodology to Support Research on Osteopenia for Breast Cancer Patients

A Causal Inference Methodology to Support Research on Osteopenia for Breast Cancer Patients

Evaluating attitudes on health-seeking behavior among a network of people who inject drugs

People who inject drugs (PWID) are often members of HIV/AIDS risk networks, where individuals engage in sexual and injection risk behavior. Engagement in HIV care is important for this population. Low socioeconomic status, stigmatization, and lack of access to medical care often complicate successful engagement in the HIV cascade of care for PWID. This study investigates how individual's attitudes about how much control they have over HIV/AIDS risk in their life (i.e., locus of control and self-blame) affect health-seeking behavior in PWID participants and their community members. We applied causal inference methodology to PWID HIV risk networks ascertained from the Social Factors and HIV Risk Study (SFHR) conducted between 1991 and 1993 in Bushwick, Brooklyn, New York. We estimated protective disseminated effects of attitudes toward HIV/AIDS on the health-seeking behaviors of others in the PWID community. In other words, a positive attitude toward controlling HIV/AIDS can improve the health-seeking behavior of other members of the community who report a pessimistic attitude toward HIV/AIDS control. Given this finding, we also discuss potential network interventions to improve health-seeking behavior among both PWID individuals who receive the intervention and others in the PWID network informed by our analysis of disseminated effects.

Disease intervention specialist-delivered interventions and other partner services for HIV and sexually transmitted infections: A systematic review.

Disease intervention specialists (DIS) are critical for delivering partner services programs that provide partner notification, counseling, referral, and other services for HIV, sexually transmitted infections (STIs), and other infections. This systematic review of partner services and other DIS-delivered interventions for HIV and STIs was conducted to summarize the effectiveness of these programs and identify evidence gaps. A systematic literature review was conducted with a narrative synthesis. Articles were located using keyword searches in MEDLINE, Web of Science, CINAHL, and ProQuest through December 2022 and analyzed in 2023-2024. Included studies addressed an intervention of partner services or other DIS-delivered services for HIV or STIs; a United States setting; primary data collection; and an external comparison group or pre-post design. 1,915 unique records were screened for eligibility, with 30 studies included. Overall, DIS-delivered interventions improved clinical outcomes among index patients and population outcomes. Many studies focused on program process measures rather than population-level epidemiologic outcomes. All but one studies were scored as having low or medium strength of evidence. The evidence could be strengthened by establishing a streamlined set of core metrics, assessing impact using rigorous causal inference methodologies, linking program and clinical data systems, and supplementing impact evaluations with evidence on implementation strategies.

Comparing traditional and causal inference methodologies for evaluating impacts of long-term air pollution exposure on hospitalization with Alzheimer's disease and related dementias.

Alzheimer's disease and related dementias (ADRD) present a growing public health burden in the United States. One actionable risk factor for ADRD is air pollution: multiple studies have found associations between air pollution and exacerbation of ADRD. Our study builds on previous studies by applying modern statistical causal inference methodologies-generalized propensity score (GPS) weighting and matching-on a large, longitudinal dataset. We follow 50 million Medicare enrollees to investigate impacts of three air pollutants-fine particular matter (PM${}_{2.5}$), nitrogen dioxide (NO${}_2$), and summer ozone (O${}_3$)-on elderly patients' rate of first hospitalization with ADRD diagnosis. Similar to previous studies using traditional statistical models, our results found increased hospitalization risks due to increased PM${}_{2.5}$ and NO${}_2$ exposure, with less conclusive results for O${}_3$. In particular, our GPS weighting analysis finds IQR increases in PM${}_{2.5}$, NO${}_2$, or O${}_3$ exposure results in hazard ratios of 1.108 (95% CI: 1.097-1.119), 1.058 (1.049-1.067), or 1.045 (1.036-1.054), respectively. GPS matching results are similar for PM${}_{2.5}$ and NO${}_2$ with attenuated effects for O${}_3$. Our results strengthen arguments that long-term PM${}_{2.5}$ and NO${}_2$ exposure increases risk of hospitalization with ADRD diagnosis. Additionally, we highlight strengths and limitations of causal inference methodologies in observational studies with continuous treatments. Keywords: Alzheimer's disease and related dementias, air pollution, Medicare, causal inference, generalized propensity score.

Brief of Amici Curiae economists in support of respondents in Dobbs v. Jackson Women's Health Organization.

A pillar of Mississippi's argument in Dobbs v. Jackson Women's Health was that there is no evidence of "societal reliance" on abortion, meaning no reason to believe that access to abortion impacts the ability of women to participate in the economic and social life of the nation. Led by economist Caitlin Myers and attorney Anjali Srinivasan, more than 150 economists filed an amicus brief seeking to assist the Court in understanding that this assertion is erroneous. The economists describe developments in causal inference methodologies over the last three decades, and the ways in which these tools have been used to isolate the measure of the effects of abortion legalization in the 1970s and of abortion policies and access over the ensuing decades. The economists argue that there is a substantial body of well-developed and credible research that shows that abortion access has had and continues to have a significant effect on birth rates as well as broad downstream social and economic effects, including on women's educational attainment and job opportunities. What follows is a reprint of this brief.

Observation weights matching approach for causal inference

This study introduces a novel method integrating pattern recognition models with causal inference methodologies to adeptly identify and manage overlapping regions between treatment and control groups. Our approach, Observation Weights Matching (OWM), addresses the intrinsic challenges in observational studies—specifically, the fixed sample size and the lack of complete overlap in pretreatment variables. Through ensemble learning, OWM effectively retains examples within these critical overlapping regions, systematically generating weighted data distributions that aid in the precise identification of these instances. By prioritizing hard-to-classify observations and employing a novel metric of critical values for matched samples, our approach optimizes matching performance and provides greater robustness in causal analysis. Through empirical and simulation studies, we demonstrate OWM's notable advantage over traditional matching methods, enhancing causal inference in observational research. Furthermore, we show that OWM provides richer balance scores than propensity scores, ensuring unbiased estimations and advancing the field significantly.

Causal inference of Seoul bike sharing demand

The global surge in environmental consciousness has significantly boosted the demand for rental bikes, particularly in metropolitan areas such as Seoul. This study delves into the causal relationships affecting this demand using a dataset from Seoul’s bike-sharing system. Unlike previous research focusing predominantly on predictive analytics, this work innovatively applies multiple linear regression models to uncover causal inferences, offering insights that extend beyond mere forecasting. The challenges addressed include dealing with non-linear relationships and heteroscedasticity by employing the logarithmic transformation of rental counts. This approach not only aids in normalizing the data but also enhances the interpretability of the regression outcomes, emphasizing the changes in demand as a function of various environmental and temporal variables. Recent developments in causal inference methodologies have allowed for more robust and detailed analysis, paving the way for this study’s contribution to the field. The findings underscore the significant influence of factors such as hour of the day, humidity, and seasonal changes on bike rental volumes, which can inform policy-making and operational strategies in urban transport planning.

Estimating the Impact of Rurality in Disparities in Cancer Mortality.

Estimation of the independent effect of rurality on cancer mortality requires causal inference methodology and consideration of area-level socioeconomic status and rural designations. Using SEER data, we identified key incident cancers diagnosed between 2000 and 2016 at age ≥20 years (N = 3,788,273), examining a 20% random sample (n = 757,655). Standardized competing risk and survival models estimated the association between rural residence, defined by Rural-Urban Continuum Codes, and cancer-specific and all-cause mortality, controlling for age at cancer diagnosis, sex, race/ethnicity, year of diagnosis, and Area Deprivation Index (ADI). We estimated the attributable fraction (AF) of rurality and high ADI (ADI > median) to the probability of mortality. Finally, we examined county measurement issues contributing to mortality rates discordant from hypothesized rates. The 5-year standardized failure probability for cancer mortality for rural patients was 33.9% versus 31.56% for urban. The AF for rural residence was 1.04% at year 1 (0.89% by year 5), the highest among local stage disease (Y1 2.1% to Y5 1.9%). The AF for high ADI was 3.33% in Y1 (2.87% in Y5), while the joint effect of rural residence and high ADI was 4.28% in Y1 (3.71% in Y5). Twenty-two percent of urban counties and 30% of rural were discordant. Among discordant urban counties, 30% were only considered urban because of adjacency to metro area. High ADI was associated with urban discordance and low ADI with rural discordance. Rural residence independently contributes to cancer mortality. The rural impact is the greatest among those with localized disease and in high deprivation areas. Rural-urban county designations may mask high-need urban counties, limiting eligibility to state and federal resources dedicated to rural areas.

A Bayesian Analysis of Two-Stage Randomized Experiments in the Presence of Interference, Treatment Nonadherence, and Missing Outcomes

Three critical issues for causal inference that often occur in modern, complicated experiments are interference, treatment nonadherence, and missing outcomes. A great deal of research efforts has been dedicated to developing causal inferential methodologies that address these issues separately. However, methodologies that can address these issues simultaneously are lacking. We propose a Bayesian causal inference methodology to address this gap. Our methodology extends existing causal frameworks and methods, specifically, two-staged randomized experiments and the principal stratification framework. In contrast to existing methods that invoke strong structural assumptions to identify principal causal effects, our Bayesian approach uses flexible distributional models that can accommodate the complexities of interference and missing outcomes, and that ensure that principal causal effects are weakly identifiable. We illustrate our methodology via simulation studies and a re-analysis of real-life data from an evaluation of India’s National Health Insurance Program. Our methodology enables us to identify new active causal effects that were not identified in past analyses. Ultimately, our simulation studies and case study demonstrate how our methodology can yield more informative analyses in modern experiments with interference, treatment nonadherence, missing outcomes, and complicated outcome generation mechanisms.

Evaluation of a Split Flow Model for the Emergency Department

Problem definition: Split flow models, in which a physician rather than a nurse performs triage, are increasingly being used in hospital emergency departments (EDs) to improve patient flow. Before deciding whether such interventions should be adopted, it is important to understand how split flows causally impact patient flow and outcomes. Methodology/results: We employ causal inference methodology to estimate average causal effects of a split flow model on time to be roomed, time to disposition after being roomed, admission decisions, and ED revisits at a large tertiary teaching hospital that uses a split flow model during certain hours each day. We propose a regression discontinuity design to identify average causal effects, which we formalize with causal diagrams. Using electronic health records data (n = 21,570), we estimate that split flow increases average time to be roomed by about 4.6 minutes (95% confidence interval (95% CI): 2.9, 6.2 minutes) but decreases average time to disposition by 14.4 minutes (95% CI: 4.1, 24.7 minutes), leading to an overall reduction in length of stay. Split flow is also found to decrease admission rates by 5.9% (95% CI: 2.3%, 9.4%) but not at the expense of a significant change in revisit rates. Lastly, we find that the split flow model is especially effective at reducing length of stay during low congestion levels, which mediation analysis partly attributes to early task initiation by the physician assigned to triage. Managerial implications: A split flow model can improve flow and may have downstream effects on admissions but not revisits. Funding: This work was supported by the National Institutes of Health [Grants KL2TR002374 and UL1TR002373]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2022.0003

Causal microeconometrics in accounting research

Purpose: This paper reviews accounting research that employed causal inference methodology, with a focus on methods associated with causal microeconometrics (quasi-experimental). The reviewed papers were published in five leading accounting journals from 2017 through 2021. Methodology/approach: The research approach is a literature review of studies that apply the methodology of causal microeconometrics to accounting. The main section of the paper describes five methods: the treatment effects approach, propensity score matching, natural experiment, difference-in-differences estimators, and regression discontinuity design. The assumptions and limitations of each method are discussed, and selected examples of causal inference published in five leading accounting journals are provided. Findings: The study confirms the increasing frequency of the use of causal inference methodologies in accounting research. Sometimes referred to as quasi-experimental or causal microeconometric, these methods can provide a base for finding evidence of causality. However, there are limitations associated with each method. Practical implications: Statistical-econometric methodology in accounting research based on regression is rarely able to demonstrate causal relationships. This paper presents the pros and cons of applying causal inference methodologies in accounting. Originality/value: The paper’s value lies in: (1) introducing to the research community the growing presence of quasi-experimental causal methodologies in accounting, (2) presenting causal research in accounting using causal microeconometric methods, (3) identifying papers using these methods that were published in five leading accounting journals between 2017 and 2021, and (4) highlighting the challenges and the need for caution and due consideration in applying these methods.

Abstract C009: Estimating the true contribution of rural residence and area deprivation on cancer mortality

Abstract Introduction True estimation of the rural effect on cancer mortality requires causal inference methodology, consideration of area-level socioeconomic status and examination of area designations (e.g., measurement of rural vs urban). Methods Utilizing Surveillance, Epidemiology and End Results data we identified key incident cancers diagnosed between 2000 and 2016 at age ≥20 (n=3,788,273). Cancer types were selected based on prior evidence of rural mortality disparities. A 20% random sample was examined (n=757,655). Rurality was defined using the Rural Urban Continuum Codes (RUCC). County indicators were linked with the Area Deprivation Index (ADI, 0 least deprived to 100 highest deprivation). Standardized competing risk and survival models estimated causal impact of rural residence on cancer-specific mortality. All models controlled for age at diagnosis, sex, race/ethnicity, year of diagnosis, and ADI. We estimated rural residence and ADI’s attributable fraction to the probability of mortality. Models were run overall and by stage. Finally, we identified county measurement issues contributing to county rates discordant from hypothesized mortality rates (i.e., urban counties with mortality>median or rural counties with mortality≤median). Results Approximately 45% were diagnosed ≥65y, 51% were female, 71% non-Hispanic white, and 12% non-Hispanic Black. Cancer diagnoses included those with screening recommendations (70%), obesity related (57%), and tobacco-related cancers (21%). The majority of patients were diagnosed with in situ or localized disease (53%). The 5y standardized failure probability for cancer-specific mortality for rural patients was 33.9% vs. 31.6% among urban. Rural + high ADI mortality was 37.9% vs. 30.7% for urban + low ADI. Rural residence attributable fraction to the probability of cancer mortality was 1.04% at Year 1, declining to 0.9% by Year 5. The rural fraction was highest among local stage disease (Y1 2.1% to Y5 1.9%). Attributable fraction of high ADI was 3.3% in Y1 to 2.9% in Y5, while the joint effect of rural+high ADI ranged from 4.3% in Y1 to 3.7% in Y5. Discordance between hypothesized and actual mortality was greater with urban/rural definitions vs ADI cut-offs (i.e., ADI distinguishes between high and low mortality better). Among 256 urban counties, 21.7% (n=89) had a higher mortality rate than hypothesized (mortality rate > median). Of those, 27 counties had no metropolitan area, but were “attached” to a nearby metro-area according to US Economic Research Service. When these 27 counties were reclassified as rural, the rural fraction was 1.4% at Y1 and 1.2% at Y5. This was a 35% and 26% relative increase, respectively. Among 349 rural counties, 15% had mortality rates higher than expected. Conclusion: Rural residence holds a small independent contribution to cancer mortality; higher among localized disease and in areas with greater deprivation. Urban/rural county designations using standard metrics may mask high need urban counties, limiting eligibility to state or federal resources dedicated to rural areas. Citation Format: Kelly Kenzik, Elizabeth Davis, Jeffrey Franks, Smita Bhatia. Estimating the true contribution of rural residence and area deprivation on cancer mortality [abstract]. In: Proceedings of the 16th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2023 Sep 29-Oct 2;Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2023;32(12 Suppl):Abstract nr C009.

Determining the Causal Effect of Statins on Reducing the Incidence of Venous Thromboembolism after Ankle Fractures

Category: Ankle; Other Introduction/Purpose: It can be challenging to decide when to give preventative medication for venous thromboembolism (VTE). It is challenging to identify VTE in its early phases, and surgeons disagree on when prophylaxis should be given and when to suspect VTE. Additionally, patients who are not at a high risk of VTE are advised not to take prophylaxis due to the potential risk of bleeding adverse events and the evidence does not fully support providing prophylaxis in isolated foot and ankle procedures (BAE). The effectiveness of prevention, particularly in isolated foot and ankle fractures, is debatable, and research into how patients react to statins is necessary. By utilizing causal inference techniques, this research simulates a randomized control trial (RCT) from observational data. Methods: Out of a total of 1,175 patients, 238 had confirmed VTE 180 days after the incidence of ankle fracture (Case group, n=238). The inclusion criteria were 1- ankle fracture diagnosed by a physician and confirmed radiologically via X-ray or CT scan; 2- Age of 18 years or older; 3- Symptomatic VTE confirmed by a clinician and through radiologic (Duplex ultrasound, CT angiography, and/or angiography). To infer causal effects, we took the three following steps: first, the causal diagram must be created. Second, the set of variables necessary for causal inference must be identified. Third, the average treatment effect should be estimated for different treatment regimens. Results: Table 1 shows the incidence of VTE among patients who had VTE chemoprophylaxis. A non-significant 2.9% increase in VTE incidence was found, 95% CI [-1%, 7%]. The results state that, if a patient is already taking Statins, the incidence of VTE is reduced by 1%, 95% CI [-4.2%, 2.8%] as compared to patients who are not taking the drug which is not statistically significant. This indicates that there is no significant difference between administering VTE prophylaxis or not, in patients that are already consuming Statins. On the contrary, for patients in the sample that are not already taking Statins, the VTE incidence increases by 6%, 95% CI [1.6%, 10%] compared to no treatment administration which is significant since the confidence interval does not contain 0. Conclusion: Our findings provide more evidence that VTE prophylaxis may not, on average, be successful in lowering the incidence of VTE after ankle fracture, particularly in individuals who do not take Statins. To put it another way, doctors may decide not to give extra prophylaxis to avoid VTE in isolated ankle fractures if a patient is on Statins. Here, we established the causal inference methodology that can help us reproduce the results of an expensive RCT.

When the Data Are Out: Measuring Behavioral Changes Following a Data Breach

This paper investigates users’ reactions to a massive and highly sensitive data breach via an array of causal inference methodologies.

The Impact of COVID-19 Pandemic on Clinical Findings in Medical Imaging Exams: An Observational Study in a Nationwide Israeli Health Organization.

The outbreak of the COVID-19 pandemic had a major effect on consumption of healthcare services. Changes in the use of routine diagnostic exams, increased incidences of post-acute COVID-19 syndrome (PCS), and other pandemic-related factors, may have influenced detected clinical conditions. The study aimed to analyze the impact of COVID-19 on the use of outpatient medical imaging services and clinical findings therein, specifically focusing on the time period after the launch of the Israeli COVID-19 vaccination campaign. In addition, the study tested whether the observed gains in abnormal findings may be linked to PCS or COVID-19 vaccination. Our dataset included 572,480 ambulatory medical imaging patients in a national health organization, from January 1, 2019 to August 31, 2021. We compared different measures of medical imaging utilization and clinical findings therein, before and after the surge of the pandemic, to identify significant changes. We also inspected the changes in the rate of abnormal findings during the pandemic after adjusting for changes in medical imaging utilization. Finally, for imaging classes that showed increased rates of abnormal findings, we measured the causal associations between COVID-19 infection, hospitalization (indicative of COVID-19 complications), and vaccination and future risk for abnormal finding. To allow adjustment for a multitude of confounding factors, we used causal inference methodologies. After the initial drop in the utilization of routine medical imaging due to the first COVID-19 wave, the number of these exams has increased, but with lower proportions of older patients, patients with comorbidities, women, and vaccine-hesitant patients. Furthermore, we observed significant gains in the rate of abnormal findings, specifically in musculoskeletal magnetic resonance (MR-MSK) and brain computed tomography (CT-brain) exams. These results also persisted after adjusting for the changes in medical imaging utilization. Demonstrated causal associations included: COVID-19 infection increasing the risk for an abnormal finding in a CT-brain exams (odds ratio [OR] of 1.4, with 95% confidence interval [CI] 1.1 to 1.7); and COVID-19-related hospitalization increasing the risk for abnormal findings in an MR-MSK exam (OR 3.1, 95% CI 1.9 to 5.3). COVID-19 impacted the use of ambulatory imaging exams, with greater avoidance among patients at higher risk for COVID-19 complications: older patients, patients with comorbidities, and non-vaccinated patients. Causal analysis results imply that PCS may have contributed to the observed gains in abnormal findings in MR-MSK and CT-brain exams, respectively.

Effect of Antifibrotic Therapy on Survival in Patients With Idiopathic Pulmonary Fibrosis

Real-world studies have reported reduced mortality in patients with idiopathic pulmonary fibrosis (IPF) treated with antifibrotic therapy; however, the initiation or discontinuation of therapy during these studies may have introduced bias. This study investigated the effect of antifibrotic therapy on mortality and other outcomes in patients with IPF using causal inference methodology. Data from a multicenter US registry of patients with IPF were used to assess the effect of antifibrotic therapy (nintedanib or pirfenidone) on death, death or lung transplant, respiratory-related hospitalization, and acute worsening of IPF (defined as any health care encounter deemed due to acute worsening of IPF). This study used the Gran method, which accounts for differences in patient characteristics and for treatment initiations and discontinuations during follow-up. The analysis cohort was limited to patients who started antifibrotic therapy on or after the day of enrollment or had never taken it. Among the 499 patients analyzed, 352 (70.5%) received antifibrotic therapy. Estimated event rates of death at 1 year were 6.6% (95% CI, 6.1-7.1) for treated patients and 10.2% (95% CI, 9.5-10.9) for control patients. There was a numerical reduction in the risk of death (hazard ratio [HR], 0.53; 95% CI, 0.28-1.03; P=0.060) but numerical increases in risks of respiratory-related hospitalization (HR, 1.88; 95% CI, 0.90-3.92; P=0.091) and acute worsening of IPF (HR, 1.71; 95% CI, 0.36-8.09; P=0.496) in treated versus control patients. Analyses based on causal inference methodology suggest that patients with IPF who receive antifibrotic therapy have improved survival.

Randomized controlled trials of biomarker targets.

Randomized controlled trials are used to estimate the causal effect of a treatment on a health outcome of interest in a patient population. Often the specified treatment in a randomized controlled trial is a medical intervention-such as a drug or procedure-experienced directly by the patient. Sometimes the "treatment" in a randomized controlled trial is a target-such as a goal biomarker measurement-that the patient's physician attempts to reach using available medications or procedures. Large randomized controlled trials of biomarker targets are common in clinical research, and trials have been conducted to compare targets in the management of hypertension, diabetes, anemia, and acute respiratory distress syndrome. However, different randomized controlled trials intended to evaluate the same biomarker targets have produced conflicting recommendations, and meta-analyses that aggregate results of trials of biomarker targets have been inconclusive. We use causal reasoning to explain why randomized controlled trials of biomarker targets can arrive at conflicting or misleading conclusions. We describe four key threats to the validity of trials of targets: (1) intention-to-treat analysis can be misleading when a direct effect of target assignment on the outcome exists due to lack of blinding; (2) incomparability in results across trials of targets; (3) time-varying adaptive treatment strategies; and (4) Goodhart's law, "when a measure becomes a target, it ceases to be a good measure." We illustrate these findings using evidence from 15 randomized controlled trials of blood pressure targets for management of hypertension. Randomized trials of blood pressure targets exhibit substantial variation in the trial patient populations and antihypertensives used to achieve the blood pressure targets assigned in the trials. The trials did not compare or account for time-varying treatment strategies used to reach the randomized targets. Possible "off-target" effects of antihypertensive medications needed to reach lower blood pressure targets may explain the absence of a clear benefit from intensive blood pressure control. Researchers should critically assess meta-analyses of trials of targets for variation in the types, distributions, and off-target effects of therapies studied. Trial investigators should release detailed information about the biomarker targets compared in new randomized trials, as well as confounders, treatments delivered, and outcomes. New randomized controlled trials should experimentally compare treatment algorithms incorporating biomarkers, rather than targets alone. Causal inference methodology that adjusts for time-varying confounding should be used to compare time-varying treatment strategies in observational settings.

Marginal Structural Models to Estimate Causal Effects of Right-to-Carry Laws on Crime

Right-to-carry (RTC) laws allow the legal carrying of concealed firearms for defense, in certain states in the USA. I used modern causal inference methodology from epidemiology to examine the effect of RTC laws on crime over a period from 1959 up to 2016. I fitted marginal structural models (MSMs), using inverse probability weighting (IPW) to correct for criminological, economic, political and demographic confounders. Results indicate that RTC laws significantly increase violent crime by 7.5% and property crime by 6.1%. RTC laws significantly increase murder and manslaughter, robbery, aggravated assault, burglary, larceny theft and motor vehicle theft rates. Applying this method to this topic for the first time addresses methodological shortcomings in previous studies such as conditioning away the effect, overfit and the inappropriate use of county level measurements. Data and analysis code for this article are available online.

Maternal hemodynamics and neonatal birth weight in pregnancies complicated by gestational diabetes: new insights from novel causal inference analysis modeling.

ObjectiveNormal pregnancy is characterized by significant changes in maternal hemodynamics that are associated with fetal growth. Pregnancies complicated by gestational diabetes mellitus (GDM) are associated with large‐for‐gestational age and macrosomia, but the relationship between maternal hemodynamic parameters and birth weight (BW) among women with GDM has not been established. Our objective was to investigate the influence of maternal hemodynamics on neonatal BW in healthy pregnancies and in those complicated by GDM.MethodsThis was a prospective, cross‐sectional case–control study of women aged ≥ 16 years with a singleton viable pregnancy, recruited between January 2016 and February 2021 at Leicester Royal Infirmary, Leicester, UK. GDM was defined as a fasting glucose level ≥ 5.3 mmol/L and/or serum glucose level ≥ 7.8 mmol/L, 2 h following a 75‐g oral glucose load. We collected data on maternal characteristics and pregnancy outcome, including body mass index (BMI) at booking and BW centile adjusted for gestational age at delivery. Maternal hemodynamic parameters were assessed at 34–42 weeks' gestation using the Arteriograph® and bioreactance techniques. Graphical causal inference methodology was used to identify causal effects of the measured variables on neonatal BW centile.ResultsIncluded in the analysis were 141 women with GDM and 136 normotensive non‐diabetic pregnant controls. 62% of the women with GDM were managed pharmacologically, with metformin and/or insulin. Variables included in the final model were cardiac output (CO), mean arterial pressure (MAP), total peripheral resistance (TPR), aortic augmentation index (AIx), aortic pulse wave velocity (PWV) and BMI at booking. Among the controls, maternal BMI, CO and aortic PWV were significantly associated with neonatal BW. Each SD increase in booking BMI produced an increase of 8.4 BW centiles (P = 0.002), in CO produced an increase of 9.4 BW centiles (P = 0.008) and in aortic PWV produced an increase of 7.1 BW centiles (P = 0.017). We found no significant relationship between MAP, TPR or aortic AIx and neonatal BW. Maternal hemodynamics influenced neonatal BW among the women with GDM in a similar manner to that in the control group, but only the relationship between maternal BMI and neonatal BW reached statistical significance, with a 1‐SD increase in BMI producing an increase of 6.1 BW centiles (P = 0.019).ConclusionsMaternal BMI, CO and PWV were determinants of BW in our control group. The relationship between maternal hemodynamics and neonatal BW was similar between women with GDM and healthy controls. Our findings therefore suggest that fetal growth restriction in pregnancies complicated by GDM may indicate maternal cardiovascular dysfunction. © 2022 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.