Reporting of covariate selection and balance assessment in propensity score analysis is suboptimal: a systematic review

Increased risk of arrhythmias, heart failure, and thrombosis in SARS-CoV-2 positive individuals persists at one year post-infection

Risk of cardiovascular events is increased after COVID-19. However, information on cardiovascular risk trends after COVID-19 infection is lacking and estimates by sex are inconsistent. Our aim was to examine cardiovascular outcomes and mortality in a large cohort (164,346 participants) of SARS-CoV-2 positive individuals compared to non-positive individuals, stratified by sex. Data were obtained from the Spanish Health System's electronic medical records. Selected individuals were ≥ 45 years old with/without a positive SARS-CoV-2 test in the period March-May 2020. Follow-up was obtained until January 31, 2021, for cardiovascular events (angina/myocardial infarction, arrhythmias, bypass/revascularization, heart failure, peripheral artery disease, stroke/transient ischemic attack, and thrombosis), and until March 31, 2021, for mortality. Individuals were matched by propensity score. Incidence of cardiovascular events and mortality was compared with accelerated failure time models. The effect of matching and of COVID-19 severity was assessed with sensitivity analyses. In the first 3 months of follow-up, SARS-CoV-2 positive individuals had a higher risk of mortality and of all cardiovascular events. From 4-12 months, there was increased risk of mortality in SARS-CoV-2 positive individuals overall, of heart failure in SARS-CoV-2 positive females (HR= 1.26 [1.11-1.42]), and of arrhythmias and thrombosis in SARS-CoV-2 positive males (HR= 1.29 [1.14-1.47] and HR= 1.35 [1.03-1.77], respectively). When COVID-19 patients admitted to the ICU were excluded, incidence of thrombosis was similar in males regardless of positive/non-positive SARS-CoV-2 status. In the full year of follow-up, increased incidence of heart failure and of arrhythmias and thrombosis was observed in SARS-CoV-2 positive females and males, respectively.

Association between periodontal disease and malignant soft tissue sarcomas

OBJECTIVE

Malignant soft tissue sarcoma (MSTS) is a rare disease, but is seen in patients undergoing orthopedic surgery. Although the association of periodontal disease with various cancers occurring in the oral cavity, gastrointestinal tract, lungs, and prostate, has been reported, the association between periodontal disease and MSTS remains unclear. This study investigated the association between periodontal disease and MSTS in patients undergoing orthopedic surgery.

SUBJECTS AND METHODS

One hundred fifteen patients who underwent orthopedic surgery between 2017 and 2021 were retrospectively enrolled (mean age = 66.8 ± 10.7 years). The patient background was adjusted by the propensity score (PS). Subsequently, the association of periodontal disease with MSTS was analyzed using PS inverse probability of treatment weighting (IPTW). Periodontal status was determined by evaluating the periodontal inflamed surface area, which was calculated by measuring the periodontal probing pocket depth and detecting bleeding on probing.

RESULTS

Multivariate logistic regression analysis after adjustment by the PS showed that severe periodontitis was significantly associated with MSTS (odds ratio 2.81, p = 0.04). Furthermore, IPTW showed that severe periodontitis was significantly associated with MSTS (odds ratio 3.21, p = 0.01).

CONCLUSION

The results indicate an association between periodontal inflammation and MSTS.

Risk of carpal tunnel syndrome among patients with osteoarthritis: a US population-based study

BACKGROUND

Carpal tunnel syndrome (CTS), an entrapment neuropathy caused by pressure of the median nerve, is a progressive condition that can lead to a decreased quality of life. Studies suggest an association between CTS and arthritis; however, previous studies examining osteoarthritis (OA) and CTS are limited in number, scope and study design. This study estimated the incidence and risk of CTS among patients with OA, both overall and by specific joints, in a large population-based cohort in the United States.

METHODS

Patients from the Optum claims database aged ≥ 45 years and diagnosed with OA between January 1, 2018, and December 31, 2022, were eligible for the OA cohort. The non-OA cohort included those without a diagnosis of OA at the index date and no history of OA for 12 months pre-index. Baseline characteristics were balanced using propensity score matching. The risk of CTS in the OA and non-OA cohort were evaluated using incidence rates and adjusted hazard ratios that were estimated using Cox regression.

RESULTS

After applying the inclusion/exclusion criteria, 3,610,240 of the 6,023,384 adults with a diagnosis of OA remained in the OA cohort. After propensity-score matching, each cohort included 1,033,439 individuals. The incidence rates for CTS per 1000 person-years were 7.35 (95% confidence interval [CI] 7.21-7.49) in the OA cohort and 1.44 (95% CI 1.38-1.50) in the non-OA cohort. The risk of developing CTS in patients with OA was ~ 4 times that of patients without (hazard ratio = 3.80; 95% CI 3.54-4.07). This increased risk was found across all OA joint types, with OA of the hand/wrist having the highest risk for CTS. Additionally, multiple OA joints presented a higher risk compared with a single affected joint.

CONCLUSIONS

OA increases the risk of CTS, but this is not limited to patients with hand/wrist OA, suggesting a systemic impact of OA on CTS. While the risk appears highest for patients with hand/wrist OA, patients with more distant affected joints like knee or hip also have an increased risk of CTS.

Use and reporting of inverse-probability-of-treatment weighting for multicategory treatments in medical research: a systematic review

OBJECTIVES

Causal inference methods for observational data represent an alternative to randomised controlled trials when they are not feasible or when real-world evidence is sought. Inverse-probability-of-treatment weighting (IPTW) is one of the most popular approaches to account for confounding in observational studies. In medical research, IPTW is mainly applied to estimate the causal effect of a binary treatment, even when the treatment has in fact multiple categories, despite the availability of IPTW estimators for multiple treatment categories. This raises questions about the appropriateness of the use of IPTW in this context. Therefore, we conducted a systematic review of medical publications reporting the use of IPTW in the presence of a multi-category treatment. Our objectives were to investigate the frequency of use and the implementation of these methods in practice, and to assess the quality of their reporting.

STUDY DESIGN AND SETTING

Using Pubmed, Embase and Web of Science, we screened 5660 articles and retained 106 articles in the final analysis that were from 17 different medical areas. This systematic review is registered on PROSPERO (CRD42022352669).

RESULTS

The number of treatment groups varied between 3 and 9, with a large majority of articles (90 [84.9%]) including 3 or 4 groups. The most commonly used method for estimating the weights was multinomial regression (51 [48.1%]) and generalized boosted models (48 [45.3%]). The covariates of the weight model were reported in 91 articles (85.9 %). Twenty-six articles (24.5 %) did not discuss the balance of covariates after weighting, and only 16 articles (15.1 %) referred to the assumptions needed to obtain correct inferences.

CONCLUSION

The results of this systematic review illustrate that medical publications scarcely use IPTW methods for more than two treatment categories. Among the publications that did, the quality of reporting was suboptimal, in particular in regard to the assumptions and model building. IPTW for multi-category treatments could be applied more broadly in medical research, and the application of the proposed guidelines in this context will help researchers to report their results and to ensure reproducibility of their research.

Combined oral contraceptive utilization and uterine fibroid incidence: A prospective study in a cohort of African-American women

Published associations between combined oral contraceptive use and uterine fibroid development have lacked prospective imaging with ultrasound to distinguish between incident and prevalent fibroids. The Study of Environment, Lifestyle, and Fibroids prospectively followed fibroid-free, African-American women (the group with the highest disease burden in the U.S.) to identify incident cases. We examined associations between combined oral contraceptive use and the 40-month cumulative risk of fibroids. History of hormonal contraceptive use was collected via telephone interview at enrollment. Fibroid identification was performed using transvaginal ultrasonography at enrollment, and at 20 and 40-months of follow-up. Inverse probability weights for exposures and censoring were used to construct weighted risk ratios (wRR) and weighted risk different (wRD) estimators which control for differences in fibroid risk factors between exposure groups. In addition, unweighted fully adjusted log-binomial regression models (aRR) were run for comparison. Of the 1,308 participants in the analysis sample, 70% had used combined oral contraceptives and 17% developed fibroids by 40 months. We observed an inverse association between ever use of combined oral contraceptives and cumulative fibroid incidence (wRR: 0.78; 95% Confidence Interval (CI): 0.60, 1.00; wRD: -0.05, 95% CI: -0.11, 0; aRR: 0.76, 95% CI: 0.60, 0.98). Fibroid incidence was greater in participants who started using combined oral contraceptives after age 17 years than among younger initiators, though the restriction to ever-users made this estimate less precise (wRR: 1.25; 95% CI: 0.89, 1.76; wRD: 0.04, 95% CI: -0.02, 0.10). No consistent patterns of fibroid incidence were seen among ever-users for duration of, or years since, last combined oral contraceptives use.

Influence of initial misdiagnosis on mortality in patients with bacteraemia: propensity score matching and propensity score weighting analyses

BACKGROUND

The diagnostic process is a key element of medicine but it is complex and prone to errors. Infectious diseases are one of the three categories of diseases in which diagnostic errors can be most harmful to patients. In this study we aimed to estimate the effect of initial misdiagnosis of the source of infection in patients with bacteraemia on 14 day mortality using propensity score methods to adjust for confounding.

METHODS

Data from a previously described longitudinal cohort of patients diagnosed with monobacterial bloodstream infection (BSI) at the Leiden University Medical Centre (LUMC) between 2013 and 2015 were used. Propensity score matching and inversed probability of treatment weighting (IPTW) were applied to correct for confounding. The average treatment effect on the treated (ATT), which in this study was the average effect of initial misdiagnosis on the misdiagnosed (AEMM), was estimated. Methodological issues that were encountered when applying propensity score methods were addressed by performing additional sensitivity analyses. Sensitivity analyses consisted of varying caliper in propensity score matching and using different truncated weights in inversed probability of treatment weighting.

RESULTS

Data of 887 patients were included in the study. Propensity scores ranged between 0.015 and 0.999 and 80 patients (9.9%) had a propensity score > 0.95. In the matched analyses, 35 of the 171 misdiagnosed patients died within 14 days (20.5%), versus 10 of the 171 correctly diagnosed patients (5.8%), yielding a difference of 14.6% (7.6%; 21.6%). In the total group of patients, the observed percentage of patients with an incorrect initial diagnosis that died within 14 days was 19.8% while propensity score reweighting estimated that their probability of dying would have been 6.5%, if they had been correctly diagnosed (difference 13.3% (95% CI 6.9%;19.6%)). After adjustment for all variables that showed disbalance in the propensity score a difference of 13.7% (7.4%; 19.9%) was estimated. Sensitivity analyses yielded similar results. However, performing weighted analyses without truncation yielded unstable results.

CONCLUSION

Thus, we observed a substantial increase of 14 day mortality in initially misdiagnosed patients. Furthermore, several patients received propensity scores extremely close to one and were almost sure to be initially misdiagnosed.

Covariate Balancing Methods for Randomized Controlled Trials Are Not Adversarially Robust

The first step toward investigating the effectiveness of a treatment via a randomized trial is to split the population into control and treatment groups then compare the average response of the treatment group receiving the treatment to the control group receiving the placebo. To ensure that the difference between the two groups is caused only by the treatment, it is crucial that the control and the treatment groups have similar statistics. Indeed, the validity and reliability of a trial are determined by the similarity of two groups' statistics. Covariate balancing methods increase the similarity between the distributions of the two groups' covariates. However, often in practice, there are not enough samples to accurately estimate the groups' covariate distributions. In this article, we empirically show that covariate balancing with the standardized means difference (SMD) covariate balancing measure, as well as Pocock and Simon's sequential treatment assignment method, are susceptible to worst case treatment assignments. Worst case treatment assignments are those admitted by the covariate balance measure, but result in highest possible ATE estimation errors. We developed an adversarial attack to find adversarial treatment assignment for any given trial. Then, we provide an index to measure how close the given trial is to the worst case. To this end, we provide an optimization-based algorithm, namely adversarial treatment assignment in treatment effect trials (ATASTREET), to find the adversarial treatment assignments.

Effectiveness of residential treatment services for veterans with substance use disorders: A propensity score matching evaluation

BACKGROUND

Prior reviews of substance use disorder (SUD) treatment have found mixed support for residential level of care but are limited by methodology problems and the ethical concerns of randomizing patients with severe SUD to lower levels of care.

METHODS

The present study is the first to use a large archival SUD residential sample with a matched comparison group and one-year follow-up period to examine the benefits of residential treatment provided to adults clinically assessed as warranting SUD residential care. We used propensity score matching in our sample (N = 6177) of veterans with a SUD who were screened and accepted for Veterans Affairs (VA) SUD residential treatment between January 1st, 2019 and June 30th, 2019.

RESULTS

We found evidence that VA SUD residential treatment saves veteran lives with an average 66% all-cause mortality risk reduction during the study period (b = -1.09, exp(b) = 0.34, p <0.001). Medium-to-large residential pre- to post-treatment self-reported mental health and SUD symptom improvements (|SMDrobust| = 0.54-0.93) were sustained by one-year post-screening. These residential treatment improvements were significantly larger than estimated counterfactual outcomes across self-reported SUD and stress disorder symptoms at one-year post-screening (ps <0.001). We found mixed behavioral, service utilization, and other self-reported mental health outcomes.

CONCLUSIONS

We conclude that VA SUD residential treatment is an effective level of care for veterans warranting residential care particularly for SUD symptom improvements and reductions in mortality risk.

A comparison of full model specification and backward elimination of potential confounders when estimating marginal and conditional causal effects on binary outcomes from observational data

A common view in epidemiology is that automated confounder selection methods, such as backward elimination, should be avoided as they can lead to biased effect estimates and underestimation of their variance. Nevertheless, backward elimination remains regularly applied. We investigated if and under which conditions causal effect estimation in observational studies can improve by using backward elimination on a prespecified set of potential confounders. An expression was derived that quantifies how variable omission relates to bias and variance of effect estimators. Additionally, 3960 scenarios were defined and investigated by simulations comparing bias and mean squared error (MSE) of the conditional log odds ratio, log(cOR), and the marginal log risk ratio, log(mRR), between full models including all prespecified covariates and backward elimination of these covariates. Applying backward elimination resulted in a mean bias of 0.03 for log(cOR) and 0.02 for log(mRR), compared to 0.56 and 0.52 for log(cOR) and log(mRR), respectively, for a model without any covariate adjustment, and no bias for the full model. In less than 3% of the scenarios considered, the MSE of the log(cOR) or log(mRR) was slightly lower (max 3%) when backward elimination was used compared to the full model. When an initial set of potential confounders can be specified based on background knowledge, there is minimal added value of backward elimination. We advise not to use it and otherwise to provide ample arguments supporting its use.

The impact of sufentanil versus remifentanil on surgical site wound healing in caesarean section primiparas undergoing epidural anaesthesia: A systematic meta-analysis

Caesarean section (C-section) is a prevalent obstetric surgical procedure, with the choice of analgesic agents playing a pivotal role in postoperative recovery. This systematic meta-analysis aimed to compare the effects of sufentanil (ST) and remifentanil (RT) on postoperative wound healing in caesarean section primiparas undergoing epidural anaesthesia. A comprehensive search was conducted across multiple databases, adhering to PRISMA guidelines, yielding eight randomized controlled trials (RCTs) for inclusion. The primary outcome was wound healing assessment using the REEDA (redness, edema, ecchymosis, discharge, approximation) scale on the third, fifth and tenth postoperative days. The meta-analysis encompassed 691 primiparas. A significant difference in wound healing was observed between ST and RT on the third (I2  = 99%; Random: SMD: 6.75, 95% CIs: 3.11-10.39, p < 0.01) and fifth days (I2  = 99%; Random: SMD: 3.63, 95% CIs: 1.56-5.70, p < 0.01) postcaesarean section. However, no significant difference was noted on the tenth day (I2  = 5%; Random: SMD: 0.00, 95% CIs: -0.45-0.45, p = 0.35). Sufentanil and remifentanil exhibit differential effects on early postoperative wound healing in caesarean section primiparas undergoing epidural anaesthesia. While both opioids are effective analgesics, sufentanil demonstrates a more pronounced impact on wound healing during the immediate postoperative days. Clinicians should consider these findings when selecting an opioid for pain management in this patient population.

Propensity Scores in Health Disparities Research: The Example of Cognitive Aging and the Hispanic Paradox

BACKGROUND

Individuals of Mexican ancestry in the United States experience substantial socioeconomic disadvantages compared with non-Hispanic white individuals; however, some studies show these groups have similar dementia risk. Evaluating whether migration selection factors (e.g., education) associated with risk of Alzheimer disease and related dementia (ADRD) explain this paradoxical finding presents statistical challenges. Intercorrelation of risk factors, common with social determinants, could make certain covariate patterns very likely or unlikely to occur for particular groups, which complicates their comparison. Propensity score (PS) methods could be leveraged here to diagnose nonoverlap and help balance exposure groups.

METHODS

We compare conventional and PS-based methods to examine differences in cognitive trajectories between foreign-born Mexican American, US-born Mexican American, and US-born non-Hispanic white individuals in the Health and Retirement Study (1994-2018). We examined cognition using a global measure. We estimated trajectories of cognitive decline from linear mixed models adjusted for migration selection factors also associated with ADRD risk conventionally or with inverse probability weighting. We also employed PS trimming and match weighting.

RESULTS

In the full sample, where PS overlap was poor, unadjusted analyses showed both Mexican ancestry groups had worse baseline cognitive scores but similar or slower rates of decline compared with non-Hispanic white adults; adjusted findings were similar, regardless of method. Focusing analyses on populations where PS overlap was improved (PS trimming and match weighting) did not alter conclusions.

CONCLUSIONS

Attempting to equalize groups on migration selection and ADRD risk factors did not explain paradoxical findings for Mexican ancestry groups in our study.

Causal evidence that herpes zoster vaccination prevents a proportion of dementia cases

The root causes of dementia are still largely unclear, and the medical community lacks highly effective preventive and therapeutic pharmaceutical agents for dementia despite large investments into their development. There is growing interest in the question if infectious agents play a role in the development of dementia, with herpesviruses attracting particular attention. To provide causal as opposed to merely correlational evidence on this question, we take advantage of the fact that in Wales eligibility for the herpes zoster vaccine (Zostavax) for shingles prevention was determined based on an individual's exact date of birth. Those born before September 2 1933 were ineligible and remained ineligible for life, while those born on or after September 2 1933 were eligible to receive the vaccine. By using country-wide data on all vaccinations received, primary and secondary care encounters, death certificates, and patients' date of birth in weeks, we first show that the percentage of adults who received the vaccine increased from 0.01% among patients who were merely one week too old to be eligible, to 47.2% among those who were just one week younger. Apart from this large difference in the probability of ever receiving the herpes zoster vaccine, there is no plausible reason why those born just one week prior to September 2 1933 should differ systematically from those born one week later. We demonstrate this empirically by showing that there were no systematic differences (e.g., in pre-existing conditions or uptake of other preventive interventions) between adults across the date-of-birth eligibility cutoff, and that there were no other interventions that used the exact same date-of-birth eligibility cutoff as was used for the herpes zoster vaccine program. This unique natural randomization, thus, allows for robust causal, rather than correlational, effect estimation. We first replicate the vaccine's known effect from clinical trials of reducing the occurrence of shingles. We then show that receiving the herpes zoster vaccine reduced the probability of a new dementia diagnosis over a follow-up period of seven years by 3.5 percentage points (95% CI: 0.6 - 7.1, p=0.019), corresponding to a 19.9% relative reduction in the occurrence of dementia. Besides preventing shingles and dementia, the herpes zoster vaccine had no effects on any other common causes of morbidity and mortality. In exploratory analyses, we find that the protective effects from the vaccine for dementia are far stronger among women than men. Randomized trials are needed to determine the optimal population groups and time interval for administration of the herpes zoster vaccine to prevent or delay dementia, as well as to quantify the magnitude of the causal effect when more precise measures of cognition are used. Our findings strongly suggest an important role of the varicella zoster virus in the etiology of dementia.

Effect of perioperative oral management on postoperative bloodstream infection in heart valve surgery patients

OBJECTIVE

There is a well-known relationship between oral hygiene and infective endocarditis. Epidemiological evidence regarding perioperative oral management (POM) for cancer surgery has been accumulated, but this evidence is not sufficient for cardiac surgery. Therefore, our purpose was to investigate whether POM can prevent postoperative complications in patients undergoing heart valve surgery.

SUBJECTS AND METHODS

Using single-arm medical information, we retrospectively enrolled 301 patients who underwent heart valve surgery between April 2010 and March 2019. The patient background was adjusted by the propensity score (PS). We then analyzed the impact of POM on postoperative bloodstream infection (PBSI), postoperative pneumonia, and mortality using PS inverse probability of treatment weighting (IPTW).

RESULTS

IPTW revealed that the POM group had a lower incidence of PBSI than the control group, with an odds ratio of 0.316 (p = 0.003). The mortality in the POM group was significantly lower than that in the control group (p = 0.023). Fourteen patients died in the present study and 6 of them were infection-related.

CONCLUSIONS

POM was significantly associated with decreased incidence of PBSI and mortality. The results suggest that POM is beneficial for the prevention of PBSI and mortality in patients undergoing heart valve surgery.

Perceptions of and subjective difficulties with social cognition in schizophrenia from an internet survey: Knowledge, clinical experiences, and awareness of association with social functioning

AIM

Social cognition impacts social functioning in schizophrenia; however, little is known about how patients with schizophrenia themselves perceive social cognition. This study, using an internet survey, aimed to investigate their knowledge of social cognition, clinical experiences related to social cognition, awareness of social cognition's role in one's social life, and the relationships between subjective difficulties with social cognition and social functioning.

METHODS

Data from 232 outpatients with schizophrenia (SZ) and 494 healthy controls (HC) were obtained through an internet survey and were weighted according to the demographics of the national population. A newly developed survey questionnaire on knowledge, experience, and awareness of social cognition was administered. Subjective difficulties were evaluated using the Self-Assessment of Social Cognition Impairments and the Observable Social Cognition Rating Scale.

RESULTS

Less than a quarter of both groups were familiar with the term or concept of social cognition. Less than 5% of both groups had experienced being assessed or treated for social cognition. More than half of both groups were aware of the relationship between social cognition and social functioning. The SZ group had higher levels of subjective difficulties than the HC group across all social cognitive domains. The attributional bias domain of subjective difficulties was negatively associated with social functioning.

CONCLUSION

Patients with schizophrenia had substantial subjective difficulties in social cognition, which they perceived as being related to social functioning. However, their knowledge of social cognition was limited, and the assessment and treatment might not be widespread in regular clinical practice.

The impact of OHIP+ pharmacare on use and costs of public drug plans among children and youth in Ontario: a time-series analysis

BACKGROUND

In 2018, Ontario implemented a pharmacare program (Ontario Health Insurance Plan Plus [OHIP+]) to provide children and youth younger than 25 years with full coverage for prescription medications in the provincial formulary. We aimed to assess the use of public drug plans and costs of publicly covered prescriptions before and after the program's implementation and modification.

METHODS

We conducted a population-based, interrupted time-series analysis using data on prescription drug claims, from the Canadian Institute for Health Information's National Prescription Drug Utilization Information System, for people younger than 25 years from January 2016 to October 2019 in Ontario, using British Columbia as the control. We assessed changes in the level and trend of publicly covered prescriptions and expenditures after the introduction of OHIP+ in January 2018 and after program modifications in April 2019. We also assessed plan use and expenditures for publicly covered prescriptions for diabetes and asthma.

RESULTS

Publicly covered prescriptions in Ontario increased by 290%, from 756 per 1000 people before OHIP+ to 2952 per 1000 (p < 0.001) after its implementation. After program modification, prescriptions decreased by 52% to 1421 per 1000 (p < 0.001). Similarly, total public drug expenditures increased by 254%, from $379 million in 2017 to $839 million in 2018, then reduced by 49% to $204 million in 2019. Monthly public plan expenditures increased by $115.94 (95% confidence interval [CI] $100.93 to $130.94) post-OHIP+ implementation and decreased by $99.97 (95% CI -$119.79 to -$80.15) per person per month after April 2019.

INTERPRETATION

Adopting OHIP+ increased use of public drug plans and expenditures for publicly funded prescription medicines, and the program modification was associated with decreases in both outcomes. This study's findings can inform the national pharmacare debate; future research should investigate associations with health outcomes.

Variable inclusion strategies through directed acyclic graphs to adjust health surveys subject to selection bias for producing national estimates

Along with the rapid emergence of web surveys to address time-sensitive priority topics, various propensity score (PS)-based adjustment methods have been developed to improve population representativeness for nonprobability- or probability-sampled web surveys subject to selection bias. Conventional PS-based methods construct pseudo-weights for web samples using a higher-quality reference probability sample. The bias reduction, however, depends on the outcome and variables collected in both web and reference samples. A central issue is identifying variables for inclusion in PS-adjustment. In this paper, directed acyclic graph (DAG), a common graphical tool for causal studies but largely under-utilized in survey research, is used to examine and elucidate how different types of variables in the causal pathways impact the performance of PS-adjustment. While past literature generally recommends including all variables, our research demonstrates that only certain types of variables are needed in PS-adjustment. Our research is illustrated by NCHS' Research and Development Survey, a probability-sampled web survey with potential selection bias, PS-adjusted to the National Health Interview Survey, to estimate U.S. asthma prevalence. Findings in this paper can be used by National Statistics Offices to design questionnaires with variables that improve web-samples' population representativeness and to release more timely and accurate estimates for priority topics.

Consolidative chemotherapy after definitive concurrent chemoradiotherapy for esophageal squamous cell carcinoma patients: a population based cohort study

BACKGROUND

The role of consolidative chemotherapy (CCT) for locally advanced esophageal squamous cell carcinoma (LA-ESCC) patients treated with definitive concurrent chemoradiotherapy (dCCRT) is unclear. We aimed to compare the overall survival (OS) of those treated with vs without CCT via a population based approach.

METHODS

Eligible LA-ESCC patients diagnosed between 2011 and 2017 were identified via the Taiwan Cancer Registry. We used propensity score (PS) weighting to balance observable potential confounders between groups. The hazard ratio (HR) of death and incidence of esophageal cancer mortality (IECM) were compared between those with vs without CCT. We also evaluated the OS in supplementary analyses via alternative approaches.

RESULTS

Our primary analysis consisted of 368 patients in whom covariates were well balanced after PS weighting. The HR of death when CCT was compared to without was 0.67 (95% confidence interval 0.52-0.86, P = 0.002). The HR of IECM was 0.66 (P = 0.04). The HR of OS remained similarly in favor of CCT in supplementary analyses.

CONCLUSIONS

We found that CCT was associated with significantly improved OS for LA-ESCC patients treated with dCCRT. Randomized controlled trials were needed to confirm this finding.

Recommendations for the use of propensity score methods in multiple sclerosis research

BACKGROUND

With many disease-modifying therapies currently approved for the management of multiple sclerosis, there is a growing need to evaluate the comparative effectiveness and safety of those therapies from real-world data sources. Propensity score methods have recently gained popularity in multiple sclerosis research to generate real-world evidence. Recent evidence suggests, however, that the conduct and reporting of propensity score analyses are often suboptimal in multiple sclerosis studies.

OBJECTIVES

To provide practical guidance to clinicians and researchers on the use of propensity score methods within the context of multiple sclerosis research.

METHODS

We summarize recommendations on the use of propensity score matching and weighting based on the current methodological literature, and provide examples of good practice.

RESULTS

Step-by-step recommendations are presented, starting with covariate selection and propensity score estimation, followed by guidance on the assessment of covariate balance and implementation of propensity score matching and weighting. Finally, we focus on treatment effect estimation and sensitivity analyses.

CONCLUSION

This comprehensive set of recommendations highlights key elements that require careful attention when using propensity score methods.

Impact of Sevoflurane and Thiopental Used Over the Course of Electroconvulsive Therapy: Propensity Score Matching Analysis

Objective

Although anesthetics play an important role in electroconvulsive therapy (ECT), the clinical efficacy and seizure adequacy of sevoflurane in the course of ECT remain unclear. The purpose of this study was to examine the clinical efficacy and seizure adequacy of sevoflurane, compared with those of thiopental, in the course of ECT in patients with mood disorders.

Methods

We conducted a retrospective chart review. Patients who underwent a course of ECT and received sevoflurane (n = 26) or thiopental (n = 26) were included. Factors associated with ECT and treatment outcomes were compared between the two groups using propensity score (PS) matching. Between-group differences were examined using an independent t-test for continuous variables and a χ²-test for categorical variables.

Results

Patients who received sevoflurane needed more stimulations (sevoflurane: 13.2 ± 4 times, thiopental: 10.0 ± 2.5 times, df = 51, p = 0.001) and sessions (sevoflurane: 10.0 ± 2.1 times, thiopental: 8.4 ± 2.1 times, df = 51, p = 0.01) and had more inadequate seizures (sevoflurane: 5 ± 3.9 times, thiopental: 2.7 ± 2.7 times, df = 51, p = 0.015). Remission and response rates were similar in both groups.

Conclusion

The present findings indicate that sevoflurane should be used with caution in ECT and only when the clinical rationale is clear.

Biological Disease-Modifying Antirheumatic Drugs and Osteoporotic Fracture Risk in Patients with Rheumatoid Arthritis: A Danish Cohort Study

OBJECTIVES

Clinical trials have shown a beneficial effect from biological disease-modifying antirheumatic drugs (bDMARDs) on hand or axial bone loss in patients with rheumatoid arthritis; however, it is unclear if this translates to a reduced fracture risk. We investigated the effect of bDMARDs on osteoporotic fracture risk compared to no biological treatment in rheumatoid arthritis.

METHODS

A cohort of patients with rheumatoid arthritis aged 18+ from DANBIO was linked to population-based health registries in Denmark (2006-2016). Adopting a prevalent new-user design, we matched bDMARD users to bDMARD-naïve patients using time-conditional propensity scores. The risk of incident osteoporotic fractures (including hip, vertebrae, humerus, and forearm) was estimated among the matched patients by Cox proportional hazards models.

RESULTS

Out of 24,678 patients with rheumatoid arthritis, 4265 bDMARD users were matched to the same number of bDMARD-naïve patients (mean age 56.2 years, 74% female). During follow-up, 229 osteoporotic fractures occurred among bDMARD users and 205 fractures among bDMARD-naïve patients (incidence rates 12.1 and 13.0 per 1000 person-years, respectively). The use of bDMARDs was not associated with a reduced risk of osteoporotic fractures among patients with rheumatoid arthritis (hazard ratio 0.97, 95% confidence interval 0.78-1.20), compared with no biological treatment. The risk estimates were similar for all osteoporotic fracture sites.

CONCLUSION

We found no independent beneficial effect from using bDMARDs on reducing the risk of osteoporotic fractures in patients with rheumatoid arthritis.

BioFACTS: biomarkers of rhabdomyolysis in the diagnosis of acute compartment syndrome - protocol for a prospective multinational, multicentre study involving patients with tibial fractures

INTRODUCTION

The ischaemic pain of acute compartment syndrome (ACS) can be difficult to discriminate from the pain linked to an associated fracture. Lacking objective measures, the decision to perform fasciotomy is based on clinical findings and performed at a low level of suspicion. Biomarkers of muscle cell damage may help to identify and monitor patients at risk, similar to current routines for patients with acute myocardial infarction. This study will test the hypothesis that biomarkers of muscle cell damage can predict ACS in patients with tibial fractures.

METHODS AND ANALYSIS

Patients aged 15-65 years who have suffered a tibial fracture will be included. Plasma (P)-myoglobin and P-creatine phosphokinase will be analysed at 6-hourly intervals after admission to the hospital (for 48 hours) and-if applicable-after surgical fixation or fasciotomy (for 24 hours). In addition, if ACS is suspected at any other point in time, blood samples will be collected at 6-hourly intervals. An independent expert panel will assess the study data and will classify those patients who had undergone fasciotomy into those with ACS and those without ACS. All primary comparisons will be performed between fracture patients with and without ACS. The area under the receiver operator characteristics curves will be used to identify the success of the biomarkers in discriminating between fracture patients who develop ACS and those who do not. Logistic regression analyses will be used to assess the discriminative abilities of the biomarkers to predict ACS corrected for prespecified covariates.

ETHICS AND DISSEMINATION

The study has been approved by the Regional Ethical Review Boards in Linköping (2017/514-31) and Helsinki/Uusimaa (HUS/2500/2000). The BioFACTS study will be reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology recommendations.

TRIAL REGISTRATION NUMBER

NCT04674592.

Effectiveness of Systemic Corticosteroids Therapy for Nonsevere Patients With COVID-19: A Multicenter, Retrospective, Longitudinal Cohort Study

OBJECTIVES

Corticosteroids were clinically used in the treatment of nonsevere patients with COVID-19, but the efficacy of such treatment lacked sufficient clinical evidence, and the impact of dose had never been studied. This study aimed to evaluate the effect of systemic corticosteroid use (SCU) in nonsevere patients with COVID-19.

METHODS

We conducted a multicenter retrospective cohort study in Hubei Province. A total of 1726 patients admitted with nonsevere type COVID-19 were included. Mixed-effect Cox model, mixed-effect Cox model with time-varying exposure, multiple linear regression, and propensity score analysis (inverse probability of treatment weight and propensity score matching) were used to explore the association between SCU and progression into severe type, all-cause mortality, and length of stay.

RESULTS

During the follow-up of 30 days, 29.8% of nonsevere patients with COVID-19 received treatment with systemic corticosteroids. The use of systemic corticosteroids was associated with higher probability of developing severe type (adjusted hazard ratio 1.81; 95% confidence interval 1.47-2.21), all-cause mortality (adjusted hazard ratio 2.92; 95% confidence interval 1.39-6.15) in time-varying Cox analysis, and prolonged hospitalization (β 4.14; P < .001) in multiple linear regression. Analysis with 2 propensity score cohorts displayed similar results. Besides, increased corticosteroid dose was significantly associated with elevated probability of developing severe type (P < .001) and prolonged hospitalization (P < .001).

CONCLUSIONS

Corticosteroid treatment against nonsevere patients with COVID-19 was significantly associated with worse clinical outcomes. The higher dose was significantly associated with elevated risk of poor disease progression. We recommend that SCU should be avoided unless necessary among nonsevere patients with COVID-19.

Quality of conduct and reporting of propensity score methods in studies investigating the effectiveness of antimicrobial therapy

Background

Propensity score methods are becoming increasingly popular in infectious disease medicine to correct for confounding in observational studies. However, applying and reporting propensity score techniques correctly requires substantial knowledge of these methods. The quality of conduct and reporting of propensity score methods in studies investigating the effectiveness of antimicrobial therapy is yet undetermined.

Methods

A systematic review was performed to provide an overview of studies (2005–2020) on the effectiveness of antimicrobial therapy that used propensity score methods. A quality assessment tool and a standardized quality score were developed to evaluate a subset of studies in which antibacterial therapy was investigated in detail. The scale of this standardized score ranges between 0 (lowest quality) and 100 (excellent).

Results

A total of 437 studies were included. The absolute number of studies that investigated the effectiveness of antimicrobial therapy and that used propensity score methods increased 15-fold between the periods 2005–2009 and 2015–2019. Propensity score matching was the most frequently applied technique (65%), followed by propensity score–adjusted multivariable regression (25%). A subset of 108 studies was evaluated in detail. The median standardized quality score per year ranged between 53 and 61 (overall range: 33–88) and remained constant over the years.

Conclusions

The quality of conduct and reporting of propensity score methods in research on the effectiveness of antimicrobial therapy needs substantial improvement. The quality assessment instrument that was developed in this study may serve to help investigators improve the conduct and reporting of propensity score methods.

Comparison of estimated treatment effects between randomized controlled trials, case-matched, and cohort studies on laparoscopic versus open distal gastrectomy for advanced gastric cancer: a systematic review and meta-analysis

PURPOSE

In actual surgical research, case-matched studies are frequently conducted as an alternative to randomized controlled trials (RCTs). However, it is still unclear what differences there are between RCTs and case-matched studies in upper gastrointestinal surgery, and clarifying them is a very important clinical issue. Thus, the purpose of this study was to investigate estimated treatment effects between RCTs, case-matched studies, and cohort studies regarding laparoscopic distal gastrectomy (LDG) for advanced gastric cancer (AGC).

METHODS

We searched the PubMed, Cochrane Central Register of Controlled Trials, and Web of Science databases for studies that compared LDG versus open distal gastrectomy for AGC published from the inception of the databases until July 2021. A meta-analysis was performed using the Review Manager version 5.3 software program from the Cochrane Collaboration, and six short-term outcomes and three long-term outcomes were assessed.

RESULTS

Twenty-three studies with 13698 patients were included. There was no difference in estimated treatment effects between RCTs and case-matched studies for all outcomes except for the number of retrieved lymph nodes and postoperative complications. In terms of intraoperative blood loss, postoperative hospital stay, number of retrieved lymph nodes, and recurrence, observational studies tended to overestimate the treatment effects.

CONCLUSION

The estimated treatment effects of LDG for AGC in the case-matched study were almost the same as in the RCTs. However, to assess the true magnitude of the treatment effect, the design and actual implementation of the analysis must be critically evaluated.

Transparency of high-dimensional propensity score analyses: guidance for diagnostics and reporting

Purpose

The high‐dimensional propensity score (HDPS) is a semi‐automated procedure for confounder identification, prioritisation and adjustment in large healthcare databases that requires investigators to specify data dimensions, prioritisation strategy and tuning parameters. In practice, reporting of these decisions is inconsistent and this can undermine the transparency, and reproducibility of results obtained. We illustrate reporting tools, graphical displays and sensitivity analyses to increase transparency and facilitate evaluation of the robustness of analyses involving HDPS.

Methods

Using a study from the UK Clinical Practice Research Datalink that implemented HDPS we demonstrate the application of the proposed recommendations.

Results

We identify seven considerations surrounding the implementation of HDPS, such as the identification of data dimensions, method for code prioritisation and number of variables selected. Graphical diagnostic tools include assessing the balance of key confounders before and after adjusting for empirically selected HDPS covariates and the identification of potentially influential covariates. Sensitivity analyses include varying the number of covariates selected and assessing the impact of covariates behaving empirically as instrumental variables. In our example, results were robust to both the number of covariates selected and the inclusion of potentially influential covariates. Furthermore, our HDPS models achieved good balance in key confounders.

Conclusions

The data‐adaptive approach of HDPS and the resulting benefits have led to its popularity as a method for confounder adjustment in pharmacoepidemiological studies. Reporting of HDPS analyses in practice may be improved by the considerations and tools proposed here to increase the transparency and reproducibility of study results.

Modified Frailty Index to Assess Risk in Elderly Patients Undergoing Distal Pancreatectomy: A Retrospective Single-Center Study

Background

To compare the postoperative course of elderly patients (≥70 years) submitted to minimally invasive (MIDP) versus open distal pancreatectomy (ODP) and to evaluate if the modified Frailty Index (mFI) predicts the surgical course of elderly patients submitted to DP.

Methods

Data of patients aged ≥70 who underwent DP at a single institution between March 2011 and December 2019 were retrospectively retrieved. A 2:1 propensity score matching (PSM) was used to correct for differences in baseline characteristics. Then, postoperative complications were compared between the two groups (MIDP vs. ODP). Additionally, the entire cohort of DP elderly patients was stratified according to the mFI into three groups: non-frail (mFI = 0), mildly frail (mFI = 1/2), or severely frail (mFI = 3) and then compared.

Results

A total of 204 patients were analyzed. After PSM, 40 MIDP and 80 ODP patients were identified. The complications considered stratified homogenously between the two groups, with no statistically significant differences. The severity of the postoperative course increased as mFI did among the three groups regarding any complication (p = 0.022), abdominal collection (p = 0.014), pulmonary complication (p = 0.001), postoperative confusion (p = 0.047), Clavien-Dindo severity ≥3 events (p = 0.036), and length of stay (p = 0.018).

Conclusions

Elderly patients can be safely submitted to MIDP. The mFI identifies frail elderly patients more prone to develop surgical and non-surgical complications after DP.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00268-021-06436-2.

How well do covariates perform when adjusting for sampling bias in online COVID-19 research? Insights from multiverse analyses

COVID-19 research has relied heavily on convenience-based samples, which-though often necessary-are susceptible to important sampling biases. We begin with a theoretical overview and introduction to the dynamics that underlie sampling bias. We then empirically examine sampling bias in online COVID-19 surveys and evaluate the degree to which common statistical adjustments for demographic covariates successfully attenuate such bias. This registered study analysed responses to identical questions from three convenience and three largely representative samples (total N = 13,731) collected online in Canada within the International COVID-19 Awareness and Responses Evaluation Study ( www.icarestudy.com ). We compared samples on 11 behavioural and psychological outcomes (e.g., adherence to COVID-19 prevention measures, vaccine intentions) across three time points and employed multiverse-style analyses to examine how 512 combinations of demographic covariates (e.g., sex, age, education, income, ethnicity) impacted sampling discrepancies on these outcomes. Significant discrepancies emerged between samples on 73% of outcomes. Participants in the convenience samples held more positive thoughts towards and engaged in more COVID-19 prevention behaviours. Covariates attenuated sampling differences in only 55% of cases and increased differences in 45%. No covariate performed reliably well. Our results suggest that online convenience samples may display more positive dispositions towards COVID-19 prevention behaviours being studied than would samples drawn using more representative means. Adjusting results for demographic covariates frequently increased rather than decreased bias, suggesting that researchers should be cautious when interpreting adjusted findings. Using multiverse-style analyses as extended sensitivity analyses is recommended.

OUP accepted manuscript

Aims

There is currently no consensus on whether atrial fibrillation (AF) patients at low risk for stroke (one non-sex-related CHA₂DS₂-VASc point) should be treated with an oral anticoagulant.

Methods and results

We conducted a multi-country cohort study in Sweden, Denmark, Norway, and Scotland. In total, 59 076 patients diagnosed with AF at low stroke risk were included. We assessed the rates of stroke or major bleeding during treatment with a non-vitamin K antagonist oral anticoagulant (NOAC), a vitamin K antagonist (VKA), or no treatment, using inverse probability of treatment weighted (IPTW) Cox regression. In untreated patients, the rate for ischaemic stroke was 0.70 per 100 person-years and the rate for a bleed was also 0.70 per 100 person-years. Comparing NOAC with no treatment, the stroke rate was lower [hazard ratio (HR) 0.72; 95% confidence interval (CI) 0.56–0.94], and the rate for intracranial haemorrhage (ICH) was not increased (HR 0.84; 95% CI 0.54–1.30). Comparing VKA with no treatment, the rate for stroke tended to be lower (HR 0.81; 95% CI 0.59–1.09), and the rate for ICH tended to be higher during VKA treatment (HR 1.37; 95% CI 0.88–2.14). Comparing NOAC with VKA treatment, the rate for stroke was similar (HR 0.92; 95% CI 0.70–1.22), but the rate for ICH was lower during NOAC treatment (HR 0.63; 95% CI 0.42–0.94).

Conclusion

These observational data suggest that NOAC treatment may be associated with a positive net clinical benefit compared with no treatment or VKA treatment in patients at low stroke risk, a question that can be tested through a randomized controlled trial.

Independent Supported Housing for non-homeless individuals with severe mental illness: Comparison of two effectiveness studies using a randomised controlled and an observational study design

BACKGROUND

No randomised controlled study (RCT) on the effectiveness of Independent Supported Housing (ISH) vs. housing as usual (HAU) settings for non-homeless individuals with severe mental illness (SMI) has been conducted to date because of limited feasibility. Alternative designs, such as observational studies, might be suitable for providing adequate evidence if well conducted. To test this hypothesis, this article reports on a prospective, direct comparison of the designs of two parallel studies in this field.

METHODS

A two-centre, parallel-group non-inferiority effectiveness study was conducted at two locations in Switzerland using identical instruments and clinical hypotheses. One centre applied an RCT design and the other an observational study (OS) design with propensity score methods (ClinicalTrials.gov: NCT03815604). The comparability of the two study centres was investigated in terms of participants, procedures, and outcomes. The primary outcome was social inclusion and the secondary outcomes were quality of life and psychiatric symptoms.

RESULTS

The study included 141 participants (RCT: n = 58; OS: n = 83). Within one year, 27% study dropouts occurred (RCT: 34%; OS: 22%). A similar balance of sample characteristics was achieved in the RCT and the OS using propensity score methods (inverse probability of treatment weighting). After one year, ISH was non-inferior to the control condition regarding social inclusion (mean differences [95% CI]) in the RCT (6.28 [-0.08 to 13.35]) and the OS (2.24 [-2.30 to 6.77]) and showed no significant differences in quality of life (RCT: 0.12 [-0.52 to 0.75]; OS: 0.16 [-0.26 to 0.58]) and symptoms (RCT: -0.18 [-0.75 to 0.40]; OS: 0.21 [-0.17 to 0.60]) in both study centres. However, strong and persistent preferences for ISH in the RCT control group reduced participants' willingness to participate. Because of several limitations in the RCT, the results of the RCT and the OS are not comparable.

CONCLUSION

Participants were comparable in both study sites. However, there were significant problems in conducting the RCT because of strong preferences for ISH. The OS with propensity score methods provided results of more stable groups of participants and revealed balanced samples and valid outcome analysis. Our results do not support further investment in RCTs in this field.

G-computation and doubly robust standardisation for continuous-time data: A comparison with inverse probability weighting

In time-to-event settings, g-computation and doubly robust estimators are based on discrete-time data. However, many biological processes are evolving continuously over time. In this paper, we extend the g-computation and the doubly robust standardisation procedures to a continuous-time context. We compare their performance to the well-known inverse-probability-weighting estimator for the estimation of the hazard ratio and restricted mean survival times difference, using a simulation study. Under a correct model specification, all methods are unbiased, but g-computation and the doubly robust standardisation are more efficient than inverse-probability-weighting. We also analyse two real-world datasets to illustrate the practical implementation of these approaches. We have updated the R package RISCA to facilitate the use of these methods and their dissemination.

Pulmonary function changes after thoracoscopic lobectomy versus intentional thoracoscopic segmentectomy for early-stage non-small cell lung cancer

BACKGROUND

Thoracoscopic segmentectomy is increasingly used in the surgical treatment of early-stage non-small cell lung cancer. However, it remains unclear whether pulmonary function loss after thoracoscopic lung resection is in direct proportion to the number of resected segments, and thus intentional thoracoscopic segmentectomy has the function-preserving advantage over thoracoscopic lobectomy.

METHODS

In this prospective observational study, spirometry tests were performed preoperatively and 6 months postoperatively. The observed functional loss was compared with the expected loss estimated by the segment counting method. Resection extent index was introduced as the number of resected segments to total number of segments in the corresponding lobe. Spirometry changes after thoracoscopic lobectomy and intentional thoracoscopic segmentectomy were compared using propensity score matching.

RESULTS

There were 338 thoracoscopic lobectomies and 321 thoracoscopic segmentectomies. Overall, the observed pulmonary function loss after segmentectomy was significantly less than after lobectomy. But the observed functional loss was significantly greater than the expected loss after segmentectomy. And pulmonary function loss per segment resected was almost doubled after segmentectomy comparing to lobectomy. For segmentectomies with a resection extent index less than 0.5, especially single segmentectomies, pulmonary function loss was significantly less than after corresponding lobectomies. Otherwise, no significant differences in spirometry changes between lobectomies and segmentectomies were detected.

CONCLUSIONS

Pulmonary function loss after thoracoscopic lung resection cannot be accurately evaluated by the number of resected segments. According to the resection extent index, intentional thoracoscopic segmentectomy may help preserve more pulmonary function than thoracoscopic lobectomy only when less than half of the corresponding lobe are resected.

Minimizing bias in massive multi-arm observational studies with BCAUS: balancing covariates automatically using supervision

BACKGROUND

Observational studies are increasingly being used to provide supplementary evidence in addition to Randomized Control Trials (RCTs) because they provide a scale and diversity of participants and outcomes that would be infeasible in an RCT. Additionally, they more closely reflect the settings in which the studied interventions will be applied in the future. Well-established propensity-score-based methods exist to overcome the challenges of working with observational data to estimate causal effects. These methods also provide quality assurance diagnostics to evaluate the degree to which bias has been removed and the estimates can be trusted. In large medical datasets it is common to find the same underlying health condition being treated with a variety of distinct drugs or drug combinations. Conventional methods require a manual iterative workflow, making them scale poorly to studies with many intervention arms. In such situations, automated causal inference methods that are compatible with traditional propensity-score-based workflows are highly desirable.

METHODS

We introduce an automated causal inference method BCAUS, that features a deep-neural-network-based propensity model that is trained with a loss which penalizes both the incorrect prediction of the assigned treatment as well as the degree of imbalance between the inverse probability weighted covariates. The network is trained end-to-end by dynamically adjusting the loss term for each training batch such that the relative contributions from the two loss components are held fixed. Trained BCAUS models can be used in conjunction with traditional propensity-score-based methods to estimate causal treatment effects.

RESULTS

We tested BCAUS on the semi-synthetic Infant Health & Development Program dataset with a single intervention arm, and a real-world observational study of diabetes interventions with over 100,000 individuals spread across more than a hundred intervention arms. When compared against other recently proposed automated causal inference methods, BCAUS had competitive accuracy for estimating synthetic treatment effects and provided highly concordant estimates on the real-world dataset but was an order-of-magnitude faster.

CONCLUSIONS

BCAUS is directly compatible with trusted protocols to estimate treatment effects and diagnose the quality of those estimates, while making the established approaches automatically scalable to an arbitrary number of simultaneous intervention arms without any need for manual iteration.

Mobile Health, Disease Knowledge, and Self-Care Behavior in Chronic Kidney Disease: A Prospective Cohort Study

Mobile health (mHealth) management is an emerging strategy of care for patients with chronic diseases. However, the effect of mHealth management on clinical outcomes of patients with chronic kidney disease (CKD) has not been well-studied. The aim of this study was to investigate the additional influence of mHealth on disease knowledge and self-care behavior in CKD patients who had received traditional education. We designed and developed a new healthcare mobile application, called iCKD, which has several major features, including home-based physiological signal monitoring, disease health education, nutrition analysis, medication reminder, and alarms and a warning system. Trained nurses interviewed patients with CKD using structured questionnaires of disease knowledge and self-care behavior. After propensity score matching, we analyzed 107 patients who used iCKD and traditional education, and 107 who received traditional education. The patients who used iCKD had higher disease knowledge scores than those who received traditional education. In multivariate analysis, iCKD was significantly and positively associated with disease knowledge scores. Patients with high education levels could have greater disease knowledge through using mHealth. There was no significant difference in total scores of self-care behavior between the two groups. In conclusion, mHealth can significantly increase disease knowledge in patients with CKD.

Evaluation of an intervention targeted with predictive analytics to prevent readmissions in an integrated health system: observational study

OBJECTIVES

To determine the associations between a care coordination intervention (the Transitions Program) targeted to patients after hospital discharge and 30 day readmission and mortality in a large, integrated healthcare system.

DESIGN

Observational study.

SETTING

21 hospitals operated by Kaiser Permanente Northern California.

PARTICIPANTS

1 539 285 eligible index hospital admissions corresponding to 739 040 unique patients from June 2010 to December 2018. 411 507 patients were discharged post-implementation of the Transitions Program; 80 424 (19.5%) of these patients were at medium or high predicted risk and were assigned to receive the intervention after discharge.

INTERVENTION

Patients admitted to hospital were automatically assigned to be followed by the Transitions Program in the 30 days post-discharge if their predicted risk of 30 day readmission or mortality was greater than 25% on the basis of electronic health record data.

MAIN OUTCOME MEASURES

Non-elective hospital readmissions and all cause mortality in the 30 days after hospital discharge.

RESULTS

Difference-in-differences estimates indicated that the intervention was associated with significantly reduced odds of 30 day non-elective readmission (adjusted odds ratio 0.91, 95% confidence interval 0.89 to 0.93; absolute risk reduction 95% confidence interval -2.5%, -3.1% to -2.0%) but not with the odds of 30 day post-discharge mortality (1.00, 0.95 to 1.04). Based on the regression discontinuity estimate, the association with readmission was of similar magnitude (absolute risk reduction -2.7%, -3.2% to -2.2%) among patients at medium risk near the risk threshold used for enrollment. However, the regression discontinuity estimate of the association with post-discharge mortality (-0.7% -1.4% to -0.0%) was significant and suggested benefit in this subgroup of patients.

CONCLUSIONS

In an integrated health system, the implementation of a comprehensive readmissions prevention intervention was associated with a reduction in 30 day readmission rates. Moreover, there was no association with 30 day post-discharge mortality, except among medium risk patients, where some evidence for benefit was found. Altogether, the study provides evidence to suggest the effectiveness of readmission prevention interventions in community settings, but further research might be required to confirm the findings beyond this setting.

Neonatal Abstinence Syndrome and Maternal Opioid-Related Diagnoses: Analysis of ICD-10-CM Transition, 2013-2017

BACKGROUND AND OBJECTIVES

Hospital discharge records remain a common data source for tracking the opioid crisis among pregnant women and infants. The International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) transition from the International Classification of Diseases, Ninth Revision, Clinical Modification may have affected surveillance. Our aim was to evaluate this transition on rates of neonatal abstinence syndrome (NAS), maternal opioid use disorder (OUD), and opioid-related diagnoses (OUD with ICD-10-CM codes for long-term use of opioid analgesics and unspecified opioid use).

METHODS

Data from the 2013-2017 Healthcare Cost and Utilization Project's National Inpatient Sample were used to conduct, interrupted time series analysis and log-binomial segmented regression to assess whether quarterly rates differed across the transition.

RESULTS

From 2013 to 2017, an estimated 18.8 million birth and delivery hospitalizations were represented. The ICD-10-CM transition was not associated with NAS rates (rate ratio [RR]: 0.99; 95% confidence interval [CI]: 0.90-1.08; P = .79) but was associated with 11% lower OUD rates (RR: 0.89; 95% CI: 0.80-0.98; P = .02) and a decrease in the quarterly trend (RR: 0.98; 95% CI: 0.96-1.00; P = .04). The transition was not associated with maternal OUD plus long-term use rates (RR: 0.98; 95% CI: 0.89-1.09; P = .76) but was associated with a 20% overall increase in opioid-related diagnosis rates including long-term and unspecified use (RR: 1.20; 95% CI: 1.09-1.32; P < .001).

CONCLUSIONS

The ICD-10-CM transition did not appear to affect NAS. However, coding of maternal OUD alone may not capture the same population across the transition, which confounds the interpretation of trend data spanning this time period.

Matching Methods for Confounder Adjustment: An Addition to the Epidemiologist's Toolbox

Propensity score weighting and outcome regression are popular ways to adjust for observed confounders in epidemiological research. Here, we provide an introduction to matching methods, which serve the same purpose but can offer advantages in robustness and performance. A key difference between matching and weighting methods is that matching methods do not directly rely on the propensity score and so are less sensitive to its misspecification or to the presence of extreme values. Matching methods offer many options for customization, which allow a researcher to incorporate substantive knowledge and carefully manage bias/variance trade-offs in estimating the effects of nonrandomized exposures. We review these options and their implications, providing guidance for their use, and comparison with weighting methods. Because of their potential advantages over other methods, matching methods should have their place in an epidemiologist's methodological toolbox.

The 10-year Survival Analysis of Radiofrequency Ablation for Solitary Hepatocellular Carcinoma 5 cm or Smaller: Primary versus Recurrent HCC

Background Although favorable outcomes have been reported with radiofrequency ablation (RFA) for limited hepatocellular carcinoma (HCC), the efficacy of this treatment for recurrent HCC has not been thoroughly investigated. Purpose To compare the long-term outcomes and analyze the prognostic factors for outcomes after RFA for initial HCC versus as a second-line treatment for recurrent HCC. Materials and Methods This retrospective study included 560 patients with solitary tumors 5 cm or smaller (263 initial HCCs, 297 -recurrent HCCs) who underwent percutaneous US-guided RFA from January 2005 to December 2016. Of 297 patients with -recurrent HCC, 134 had previously undergone hepatectomy, 128 had undergone transarterial chemoembolization (TACE), and 35 had undergone local ablation therapy. Overall survival (OS) between initial HCC and recurrent HCC was compared before and after propensity score matching. Prognostic factors for all patients were analyzed with the log-rank test and Cox proportional hazards model. Results A total of 560 patients (mean age, 60 years ± 12 [standard deviation]; 441 men) were evaluated. Before matching, the OS rates at 1, 3, 5, and 10 years were 92.6%, 73.9%, 59.3%, and 39.6%, respectively, in patients with recurrent HCC and 92.8%, 75.4%, 63.3%, and 44.7% in patients with initial HCC (P = .27). After matching, the OS rates at 1, 3, 5, and 10 years were 94.8%, 75.7%, 61.6%, and 47.3% in the initial HCC group and 91.9%, 71.2%, 58.7%, and 45.2% in the recurrent HCC group (P = .32). Among patients with recurrent HCC, no significant difference in mean OS was noted for local recurrence versus distant recurrence (81.6 months ± 5.1 vs 83.8 months ± 6.6, P = .82) or previous treatment modality (82.0 months ± 7.3 in the resection group, 82.7 months ± 5.3 in the TACE group, and 79.3 months ± 10.8 in the local ablation group; P = .83). Local tumor progression after previous local ablation (10 of 35 [28.6%]) was higher than that after previous hepatectomy (15 of 134 [11.2%], P = .04). Multivariable analysis demonstrated that tumor size (hazard ratio, 1.58; 95% CI: 1.06, 2.36; P = .02), portal hypertension (hazard ratio, 1.52; 95% CI: 1.03, 2.26; P = .04), Child-Pugh class (hazard ratio, 2.01; 95% CI: 1.02, 3.96; P = .045), and serum α-fetoprotein level (hazard ratio, 1.62; 95% CI: 1.10, 2.39; P = .01) were independent predictive factors for recurrent HCC outcomes. Conclusion Radiofrequency ablation provides similar long-term survival for solitary hepatocellular carcinoma of 5 cm or less, regardless of whether treatment is initial or salvage therapy. © RSNA, 2021 Online supplemental material is available for this article.

Identifying the causal effect of child marriage on unmet needs for modern contraception and unintended pregnancy in Nepal: a cross-sectional study using propensity score matching

OBJECTIVES

This study aimed to investigate whether child marriage had causal effects on unmet needs for modern contraception, and unintended pregnancy, by estimating the marginal (population-averaged) treatment effect of child marriage.

DESIGN

This study used secondary data from the Nepal Demographic and Health Survey 2016. Applying one-to-one nearest-neighbour matching with replacement within a calliper range of ±0.01, 15-49 years old women married before the age of 18 were matched with similar women who were married at 18 or above to reduce selection bias.

SETTING

Nationally representative population survey data.

PARTICIPANTS

The sample consisted of 7833 women aged 15-49 years who were married for more than 5 years.

OUTCOME MEASURES

Unmet needs for modern contraception and unintended pregnancy.

RESULTS

The matching method achieved adequate overlap in the propensity score distributions and balance in measured covariates between treatment and control groups with the same propensity score. Propensity score matching analysis showed that the risk of unmet needs for modern contraception, and unintended pregnancy among women married as children were a 14.3 percentage point (95 % CI 10.3 to 18.2) and a 10.1 percentage point (95 % CI 3.7 to 16.4) higher, respectively, than among women married as adults. Sensitivity analysis indicated that the estimated effects were robust to unmeasured covariates.

CONCLUSIONS

Child marriage appears to increase the risk of unmet needs for modern contraception and unintended pregnancy. These findings call for social development and public health programmes that promote delayed entry into marriage and childbearing to improve reproductive health and rights.

Effectiveness of Image-Guided Radiotherapy in Adjuvant Radiotherapy on Survival for Localized Breast Cancer: A Population-Based Analysis

Purpose

Image-guided radiotherapy (IGRT) is an advanced radiotherapy technique to improve the radiotherapy delivery. We aimed to compare the overall survival (OS) for localized breast cancer (LBC) patient treated with adjuvant conventional fractionated radiotherapy (CFRT) using IGRT vs those without IGRT via a population-based analysis.

Patients and Methods

Eligible LBC patients diagnosed between 2011 and 2013 were identified via the Taiwan Cancer Registry. We used propensity score (PS) weighting to balance observable potential confounders between groups. The hazard ratio (HR) of death and other outcomes were compared between IGRT and non-IGRT. We also evaluated OS in various supplementary analyses.

Results

Our primary analysis included 6490 patients in whom covariates were well balanced after PS weighing. The HR for death when IGRT was compared with non-IGRT was 1.02 (95% confidence interval 0.80–1.31, P = 0.86). There were also no significant differences in the supplementary analyses.

Conclusion

We found that OS of LBC patients treated with adjuvant CFRT was not statistically different between those treated with IGRT versus without IGRT. This was the first study in this regard to our knowledge but randomized controlled trials were needed to confirm our finding.

Safety of image-guided radiotherapy in definitive radiotherapy for localized prostate cancer: a population-based analysis

Objectives:

Image-guided radiotherapy (IGRT) is a recommended advanced radiation technique that is associated with fewer acute and chronic toxicities. However, one Phase III trial showed worse overall survival in the IGRT arm. The purpose of this observational study is to evaluate the impact of IGRT on overall survival.

Methods:

We used the Taiwan Cancer Registry Database to enroll cT1-4N0M0 prostate cancer patients who received definitive radiotherapy between 2011 and 2015. We used inverse probability treatment weighting (IPW) to construct balanced IGRT and non-IGRT groups. We compared the overall survival of those in the IGRT and non-IGRT groups. Supplementary analyses (SA) were performed with alternative covariates in propensity score (PS) models and PS approaches. The incidence rates of prostate cancer mortality (IPCM), other cancer mortality (IOCM), and cardiovascular mortality (ICVM) were also evaluated.

Results:

There were 360 patients in the IGRT arm and 476 patients in the non-IGRT arm. The median follow-up time was 50 months. The 5-year overall survival was 88% in the IGRT arm and 86% in the non-IGRT arm (adjusted hazard ratio [HR] of death = 0.93; 95% CI, 0.61–1.45; p = 0.77). The SA also showed no significant differences in the overall survival between those in the IGRT and non-IGRT arms. Both groups did not significantly differ in terms of IPCM, IOCM, and ICVM.

Conclusions:

The overall survival of localized prostate cancer patients who underwent IGRT was not inferior to those who did not.

Advances in knowledge:

We demonstrated that the overall survival for prostate cancer patients with IGRT was not worse than those who did not undergo IGRT; this important outcome comparison has not been previously examined in the general population.

Chemotherapy alone versus definitive concurrent chemoradiotherapy for cT4b esophageal squamous cell carcinoma: a population-based study

Background

The role of radiotherapy for cT4bNanyM0 esophageal squamous cell carcinoma (ESqCC) is relatively unclear, with both chemotherapy (C/T) alone and definitive concurrent chemoradiotherapy (dCCRT) being treatment options in the current guidelines. We aimed to compare the survival of dCCRT versus C/T for these patients via a population-based approach.

Methods

Eligible cT4b ESqCC patients diagnosed between 2011 and 2017 were identified via the Taiwan Cancer Registry. We used propensity score (PS) weighting to balance the observable potential confounders between groups. The hazard ratio (HR) of death and incidence of esophageal cancer mortality (IECM) were compared between dCCRT and C/T. We also evaluated OS in subgroups of either low or standard radiotherapy doses.

Results

Our primary analysis consisted of 247 patients in whom covariates were well balanced after PS weighing. The HR for death when dCCRT was compared with C/T was 0.36 (95% confidence interval 0.24–0.53, P < 0.001). Similar results were found for IECM. Statistical significance was only observed in the standard RT dose but not in the low dose in subgroup analyses.

Conclusions

In this population-based nonrandomized study of cT4bNanyM0 ESqCC patients from Asia (Taiwan), we found that the use of radiotherapy with chemotherapy was associated with better overall survival than chemotherapy alone. Further studies (especially RCTs) are needed to confirm our findings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-021-01742-4.

Greedy caliper propensity score matching can yield variable estimates of the treatment-outcome association-A simulation study

PURPOSE

Greedy caliper propensity score (PS) matching is dependent on randomness, which can ultimately affect causal estimates. We sought to investigate the variation introduced by this randomness.

METHODS

Based on a literature search to define the simulation parameters, we simulated 36 cohorts of different sizes, treatment prevalence, outcome prevalence, treatment-outcome-association. We performed 1:1 caliper and nearest neighbor (NN) caliper PS-matching and repeated this 1000 times in the same cohort, before calculating the treatment-outcome association.

RESULTS

Repeating caliper and NN caliper matching in the same cohort yielded large variations in effect estimates, in all 36 scenarios, with both types of matching. The largest variation was found in smaller cohorts, where the odds ratio (OR) ranged from 0.53 to 10.00 (IQR of ORs: 1.11-1.67). The 95% confidence interval was not consistently overlapping a neutral association after repeating the matching with both algorithms. We confirmed these findings in a noninterventional example study.

CONCLUSION

Caliper PS-matching can yield highly variable estimates of the treatment-outcome association if the analysis is repeated.

Difference in surgical outcomes of rectal cancer by study design: meta-analyses of randomized clinical trials, case-matched studies, and cohort studies

Background

RCTs are considered the standard in surgical research, whereas case-matched studies and propensity score matching studies are conducted as an alternative option. Both study designs have been used to investigate the potential superiority of robotic surgery over laparoscopic surgery for rectal cancer. However, no conclusion has been reached regarding whether there are differences in findings according to study design. This study aimed to examine similarities and differences in findings relating to robotic surgery for rectal cancer by study design.

Methods

A comprehensive literature search was conducted using PubMed, Scopus, and Cochrane CENTRAL to identify RCTs, case-matched studies, and cohort studies that compared robotic versus laparoscopic surgery for rectal cancer. Primary outcomes were incidence of postoperative overall complications, incidence of anastomotic leakage, and postoperative mortality. Meta-analyses were performed for each study design using a random-effects model.

Results

Fifty-nine articles were identified and reviewed. No differences were observed in incidence of anastomotic leakage, mortality, rate of positive circumferential resection margins, conversion rate, and duration of operation by study design. With respect to the incidence of postoperative overall complications and duration of hospital stay, the superiority of robotic surgery was most evident in cohort studies (risk ratio (RR) 0.83, 95 per cent c.i. 0.74 to 0.92, P < 0.001; mean difference (MD) –1.11 (95 per cent c.i. –1.86 to –0.36) days, P = 0.004; respectively), and least evident in RCTs (RR 1.12, 0.91 to 1.38, P = 0.27; MD –0.28 (–1.44 to 0.88) days, P = 0.64; respectively).

Conclusion

Results of case-matched studies were often similar to those of RCTs in terms of outcomes of robotic surgery for rectal cancer. However, case-matched studies occasionally overestimated the effects of interventions compared with RCTs.

Differences in the Microbial Composition of Hemodialysis Patients Treated with and without β-Blockers

β-blockers are commonly prescribed to treat cardiovascular disease in hemodialysis patients. Beyond the pharmacological effects, β-blockers have potential impacts on gut microbiota, but no study has investigated the effect in hemodialysis patients. Hence, we aim to investigate the gut microbiota composition difference between β-blocker users and nonusers in hemodialysis patients. Fecal samples collected from hemodialysis patients (83 β-blocker users and 110 nonusers) were determined by 16S ribosomal RNA amplification sequencing. Propensity score (PS) matching was performed to control confounders. The microbial composition differences were analyzed by the linear discriminant analysis effect size, random forest, and zero-inflated Gaussian fit model. The α-diversity (Simpson index) was greater in β-blocker users with a distinct β-diversity (Bray-Curtis Index) compared to nonusers in both full and PS-matched cohorts. There was a significant enrichment in the genus Flavonifractor in β-blocker users compared to nonusers in full and PS-matched cohorts. A similar finding was demonstrated in random forest analysis. In conclusion, hemodialysis patients using β-blockers had a different gut microbiota composition compared to nonusers. In particular, the Flavonifractor genus was increased with β-blocker treatment. Our findings highlight the impact of β-blockers on the gut microbiota in hemodialysis patients.

Similarities and differences between study designs in short- and long-term outcomes of laparoscopic versus open low anterior resection for rectal cancer: A systematic review and meta-analysis of randomized, case-matched, and cohort studies

AIM

Randomized controlled trials (RCT) are the gold standard in surgical research, and case-matched studies, such as studies with propensity score matching, are expected to serve as an alternative to RCT. Both study designs have been used to investigate the potential superiority of laparoscopic surgery to open surgery for rectal cancer, but it remains unclear whether there are any differences in the findings obtained using these study designs. We aimed to examine similarities and differences between findings from different study designs regarding laparoscopic surgery for rectal cancer.

METHODS

Systematic review and meta-analyses. A comprehensive literature search was conducted using PubMed, Scopus, and Cochrane. RCT, case-matched studies, and cohort studies comparing laparoscopic low anterior resection and open low anterior resection for rectal cancer were included. In total, 8 short-term outcomes and 3 long-term outcomes were assessed. Meta-analysis was conducted stratified by study design using a random-effects model.

RESULTS

Thirty-five studies were included in this review. Findings did not differ between RCT and case-matched studies for most outcomes. However, the estimated treatment effect was largest in cohort studies, intermediate in case-matched studies, and smallest in RCT for overall postoperative complications and 3-year local recurrence.

CONCLUSION

Findings from case-matched studies were similar to those from RCT in laparoscopic low anterior resection for rectal cancer. However, findings from case-matched studies were sometimes intermediate between those of RCT and unadjusted cohort studies, and case-matched studies and cohort studies have a potential to overestimate the treatment effect compared with RCT.

Optimal sessions of hyperbaric oxygen therapy in patients with carbon monoxide poisoning: A prospective observational study

INTRODUCTION

Hyperbaric oxygen (HBO) therapy may be a useful treatment to prevent the development of delayed neuropsychiatric sequelae (DNS) in patients with acute carbon monoxide (CO) poisoning. However, there is no clear consensus regarding the optimal number of HBO therapy sessions in patients with CO poisoning. Here, we compared the development of DNS after 3 and > 3 sessions of HBO therapy in patients with acute CO poisoning.

METHODS

This prospective observational study recruited 299 patients with CO poisoning. Demographic and clinical information were obtained, including comorbidities, vital signs, and symptoms. Patients were divided into two groups according to whether they received 3 or > 3 sessions of HBO therapy (3 HBO vs. >3 HBO). A propensity score-matching process was used to balance potential prognostic factors in both groups.

RESULTS

Of the 299 patients with acute CO poisoning enrolled in this study, 183 (59.0%) were included in the analysis. Patients were excluded for the following reasons: age < 18 years, not underwent HBO therapy, discharged against medical advice, and loss to follow-up. The overall rate of DNS development was 17.5%. The >3 HBO group had a higher incidence of DNS development compared to the 3 HBO group (36.3% vs. 16.3%; p = 0.09). Propensity score-matching analysis revealed similar incidences of DNS (31.3% vs. 28.1%, respectively; p > 0.99).

CONCLUSIONS

There is a critical need to determine the optimal number of HBO therapy sessions for patients with acute CO poisoning. This study showed no difference in DNS development after 3 and > 3 sessions of HBO therapy.

Comparing mortality between positive and negative blood culture results: an inverse probability of treatment weighting analysis of a multicenter cohort

Background

The association between blood culture status and mortality among sepsis patients remains controversial hence we conducted a tri-center retrospective cohort study to compare the early and late mortality of culture-negative versus culture-positive sepsis using the inverse probability of treatment weighting (IPTW) method.

Methods

Adult patients with suspected sepsis who completed the blood culture and procalcitonin tests in the emergency department or hospital floor were eligible for inclusion. Early mortality was defined as 30-day mortality, and late mortality was defined as 30- to 90-day mortality. IPTW was calculated from propensity score and was employed to create two equal-sized hypothetical cohorts with similar covariates for outcome comparison.

Results

A total of 1405 patients met the inclusion criteria, of which 216 (15.4%) yielded positive culture results and 46 (21.3%) died before hospital discharge. The propensity score model showed that diabetes mellitus, urinary tract infection, and hepatobiliary infection were independently associated with positive blood culture results. There was no significant difference in early mortality between patients with positive or negative blood culture results. However, culture-positive patients had increased late mortality as compared with culture-negative patients in the full cohort (IPTW-OR, 1.95, 95%CI: 1.14–3.32) and in patients with severe sepsis or septic shock (IPTW-OR, 1.92, 95%CI: 1.10–3.33). After excluding Staphylococcal bacteremia patients, late mortality difference became nonsignificant (IPTW-OR, 1.78, 95%CI: 0.87–3.62).

Conclusions

Culture-positive sepsis patients had comparable early mortality but worse late mortality than culture-negative sepsis patients in this cohort. Persistent Staphylococcal bacteremia may have contributed to the increased late mortality.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-021-05862-w.