Propensity score analysis for time-dependent exposure

Handling time-varying treatments in observational studies: A scoping review and recommendations

OBJECTIVE

Time-varying treatments are common in observational studies. However, when assessing treatment effects, the methodological framework has not been systematically established for handling time-varying treatments. This study aimed to examine the current methods for dealing with time-varying treatments in observational studies and developed practical recommendations.

METHODS

We searched PubMed from 2000 to 2021 for methodological articles about time-varying treatments, and qualitatively summarized the current methods for handling time-varying treatments. Subsequently, we developed practical recommendations through interactive internal group discussions and consensus by a panel of external experts.

RESULTS

Of the 36 eligible reports (22 methodological reviews, 10 original studies, 2 tutorials and 2 commentaries), most examined statistical methods for time-varying treatments, and only a few discussed the overarching methodological process. Generally, there were three methodological components to handle time-varying treatments. These included the specification of treatment which may be categorized as three scenarios (i.e., time-independent treatment, static treatment regime, or dynamic treatment regime); definition of treatment status which could involve three approaches (i.e., intention-to-treat, per-protocol, or as-treated approach); and selection of analytic methods. Based on the review results, a methodological workflow and a set of practical recommendations were proposed through two consensus meetings.

CONCLUSIONS

There is no consensus process for assessing treatment effects in observational studies with time-varying treatments. Previous efforts were dedicated to developing statistical methods. Our study proposed a stepwise workflow with practical recommendations to assist the practice.

Changes in Glomerular Filtration Rate After Switching From Tenofovir Disoproxil Fumarate to Tenofovir Alafenamide Fumarate for Human Immunodeficiency Virus Preexposure Prophylaxis

Background

Tenofovir alafenamide fumarate (TAF) was promoted as a safer alternative to tenofovir disoproxil fumarate (TDF) for human immunodeficiency virus oral preexposure prophylaxis (PrEP). It is unknown if switching from TDF to TAF translates to improved renal function. We used electronic health record (EHR) data to assess changes in creatinine-estimated glomerular filtration rate (eGFR) after switching from TDF to TAF.

Methods

We conducted a retrospective cohort study using EHR data from Kaiser Permanente Southern California. We identified individuals who switched from TDF to TAF between October 2019 and May 2022 and used time-varying propensity score matching to identify controls who were on TDF ("nonswitchers"). We then used Bayesian longitudinal modeling to compare differences in eGFR between switching and nonswitching scenarios.

Results

Among 5246 eligible individuals, we included 118 TDF to TAF switchers and 114 nonswitchers. Compared to nonswitchers, switchers had older age of starting TDF but similar body weights at index date. A higher proportion of switchers were White, on Medicare or Medicaid, and had dyslipidemia at index date. Switching to TAF was associated with a higher eGFR compared to staying on TDF in 3-15 months post-switch, but the differences were not statistically significant (eg, month 9 difference: 1.27 [95% credible interval, -1.35 to 3.89]). While most of the estimated changes showed eGFR increase associated with switching, most were <2 eGFR units. Sensitivity analyses to address missingness or nonadherence showed similar results.

Conclusions

Switching from TDF to TAF for PrEP was associated with a nonsignificant increase in eGFR. Findings need to be confirmed using larger cohorts.

Mathematical Modeling of COVID-19 Cases and Deaths and the Impact of Vaccinations during Three Years of the Pandemic in Peru

The COVID-19 pandemic has caused widespread infections, deaths, and substantial economic losses. Vaccine development efforts have led to authorized candidates reducing hospitalizations and mortality, although variant emergence remains a concern. Peru faced a significant impact due to healthcare deficiencies. This study employed logistic regression to mathematically model COVID-19's dynamics in Peru over three years and assessed the correlations between cases, deaths, and people vaccinated. We estimated the critical time (tc) for cases (627 days), deaths (389 days), and people vaccinated (268 days), which led to the maximum speed values on those days. Negative correlations were identified between people vaccinated and cases (-0.40) and between people vaccinated and deaths (-0.75), suggesting reciprocal relationships between those pairs of variables. In addition, Granger causality tests determined that the vaccinated population dynamics can be used to forecast the behavior of deaths (p-value < 0.05), evidencing the impact of vaccinations against COVID-19. Also, the coefficient of determination (R2) indicated a robust representation of the real data. Using the Peruvian context as an example case, the logistic model's projections of cases, deaths, and vaccinations provide crucial insights into the pandemic, guiding public health tactics and reaffirming the essential role of vaccinations and resource distribution for an effective fight against COVID-19.

Antidepressant continuation and adherence in pregnancy, and risk of antenatal hospitalization for unipolar major depressive and/or anxiety disorders

BACKGROUND

Knowledge about the effectiveness of antidepressants in pregnancy is limited. We aimed to evaluate the association of antidepressant continuation in pregnancy and adherence with the risk of antenatal hospitalization for depression/anxiety.

METHODS

In a population-based study based on the healthcare databases of the Lombardy region, Italy (2010-2020), we included 17,033 live-birth pregnancies within 16,091 women with antidepressant use before pregnancy. Antidepressant exposure was classified as continued in pregnancy versus discontinued proximal to pregnancy. Outcome measure was antenatal hospitalization for depression/anxiety. Propensity score matching analysis was performed to control for measured confounding. Stratification by pre-pregnancy antidepressant adherence based on the proportion of days covered (PDC) with antidepressants served to address confounding by disease severity. We applied 60 days lag-time for antidepressant exposure to minimize the risk of protopathic bias.

RESULTS

There were 362 (2.1 %) antenatal hospitalizations for depression/anxiety. Among the matched pairs, the cumulative incidence was 3.5 (continued antidepressant) versus 2.1 (discontinued antidepressant) per 1000 person-months, yielding a hazard ratio (HR) of 1.76 (95 % confidence interval (CI): 1.34-2.33)). The HR declined to the null (1.02, 95 % CI: 0.62-1.69) in the stratified analysis of pregnancies with moderate-high adherence pre-pregnancy. Moderate-high adherence in pregnancy was associated with 85 % greater risk of the antenatal outcome, but the HR decreased with the 60 days lag-time (HR: 1.40, 95 % CI: 0.79-2.50).

LIMITATIONS

Lack of information regarding antidepressant dosage.

CONCLUSION

We found no difference in risk for antenatal hospitalization for depression/anxiety with antidepressant continuation or higher adherence in pregnancy, relative to discontinuation or lower adherence.

High Intensity Focused Ultrasound Ablation for Patients With Locally Advanced Pancreatic Adenocarcinoma: A Propensity Score-Matching Analysis

OBJECTIVES

This retrospective study was conducted to assess the efficacy and safety of high intensity focused ultrasound (HIFU) in combination with chemotherapy compared with chemotherapy alone in treating patients with unresectable locally advanced pancreatic cancer (LAPC).

METHODS

The data of unresectable LAPC patients who received chemotherapy with or without HIFU ablation were retrieved retrospectively. The overall survival (OS), objective response rate (ORR), cancer antigen 19-9 response rate, and safety were compared between these two groups before and after propensity score matching (PSM).

RESULTS

Overall, 254 patients with LAPC were included, of whom 92 underwent HIFU ablation. After PSM to control for potential biases, HIFU was associated with improved OS (12.8 versus 12.2 months, log-rank P = .046), as compared to patients without HIFU ablation. Patients with numeric rating scale (NRS) less than 4, and receiving HIFU ablation were significantly associated with improved OS (adjusted hazard ratio [aHR] = 0.365 [95% confidence interval (CI) = 0.148-0.655], P = .002; aHR = 0.490 [95% CI = 0.250-0.961], P = .038; respectively) by multivariate analyses with the adjustment of age, NRS, and tumor size. ORR was also observed to be higher in HIFU group of 30.0% than in the chemotherapy group of 13.3% (P = .039). No severe adverse events of special interest or HIFU-caused deaths were observed.

CONCLUSIONS

Patients with unresectable LAPC who received gemcitabine-based chemotherapy might benefit from additional HIFU ablation.

Smartphone-Based Remote Monitoring in Heart Failure With Reduced Ejection Fraction: Retrospective Cohort Study of Secondary Care Use and Costs

BACKGROUND

Despite effective therapies, the economic burden of heart failure with reduced ejection fraction (HFrEF) is driven by frequent hospitalizations. Treatment optimization and admission avoidance rely on frequent symptom reviews and monitoring of vital signs. Remote monitoring (RM) aims to prevent admissions by facilitating early intervention, but the impact of noninvasive, smartphone-based RM of vital signs on secondary health care use and costs in the months after a new diagnosis of HFrEF is unknown.

OBJECTIVE

The purpose of this study is to conduct a secondary care health use and health-economic evaluation for patients with HFrEF using smartphone-based noninvasive RM and compare it with matched controls receiving usual care without RM.

METHODS

We conducted a retrospective study of 2 cohorts of newly diagnosed HFrEF patients, matched 1:1 for demographics, socioeconomic status, comorbidities, and HFrEF severity. They are (1) the RM group, with patients using the RM platform for >3 months and (2) the control group, with patients referred before RM was available who received usual heart failure care without RM. Emergency department (ED) attendance, hospital admissions, outpatient use, and the associated costs of this secondary care activity were extracted from the Discover data set for a 3-month period after diagnosis. Platform costs were added for the RM group. Secondary health care use and costs were analyzed using Kaplan-Meier event analysis and Cox proportional hazards modeling.

RESULTS

A total of 146 patients (mean age 63 years; 42/146, 29% female) were included (73 in each group). The groups were well-matched for all baseline characteristics except hypertension (P=.03). RM was associated with a lower hazard of ED attendance (hazard ratio [HR] 0.43; P=.02) and unplanned admissions (HR 0.26; P=.02). There were no differences in elective admissions (HR 1.03, P=.96) or outpatient use (HR 1.40; P=.18) between the 2 groups. These differences were sustained by a univariate model controlling for hypertension. Over a 3-month period, secondary health care costs were approximately 4-fold lower in the RM group than the control group, despite the additional cost of RM itself (mean cost per patient GBP £465, US $581 vs GBP £1850, US $2313, respectively; P=.04).

CONCLUSIONS

This retrospective cohort study shows that smartphone-based RM of vital signs is feasible for HFrEF. This type of RM was associated with an approximately 2-fold reduction in ED attendance and a 4-fold reduction in emergency admissions over just 3 months after a new diagnosis with HFrEF. Costs were significantly lower in the RM group without increasing outpatient demand. This type of RM could be adjunctive to standard care to reduce admissions, enabling other resources to help patients unable to use RM.

Metformin and Cancer: Solutions to a Real-World Evidence Failure

The quest to repurpose metformin, an antidiabetes drug, as an agent for cancer prevention and treatment, which began in 2005 with an observational study that reported a reduction in cancer incidence among metformin users, generated extensive experimental, observational, and clinical research. Experimental studies revealed that metformin has anticancer effects via various pathways, potentially inhibiting cancer cell proliferation. Concurrently, multiple nonrandomized observational studies reported remarkable reductions in cancer incidence and outcomes with metformin use. However, these studies were shown, in 2012, to be affected by time-related biases, such as immortal time bias, which tend to greatly exaggerate the benefit of a drug. The observational studies that avoided these biases did not find an association. Subsequently, the randomized trials of metformin for the treatment of type 2 diabetes and as adjuvant therapy for the treatment of various cancers, advanced or metastatic, did not find reductions in cancer incidence or outcomes. Most recently, the largest phase 3 randomized trial of metformin as adjuvant therapy for breast cancer, which enrolled 3,649 women with a 5-year follow-up, found no benefit for disease-free survival or overall survival with metformin. This major failure of observational real-world evidence studies in correctly assessing the effects of metformin on cancer incidence and outcomes was caused by preventable biases which, surprisingly, are still prominent in 2022. Rigorous approaches for observational studies that emulate randomized trials, such as the incident and prevalent new-user designs along with propensity scores, avoid these biases and can provide more accurate real-world evidence for the repurposing of drugs such as metformin.

Extracorporeal cardiopulmonary resuscitation for adult out-of-hospital cardiac arrest patients: time-dependent propensity score-sequential matching analysis from a nationwide population-based registry

BACKGROUND

There is inconclusive evidence regarding the effectiveness of extracorporeal cardiopulmonary resuscitation (ECPR) for out-of-hospital cardiac arrest (OHCA) patients. We aimed to evaluate the association between ECPR and neurologic recovery in OHCA patients using time-dependent propensity score matching analysis.

METHODS

Using a nationwide OHCA registry, adult medical OHCA patients who underwent CPR at the emergency department between 2013 and 2020 were included. The primary outcome was a good neurological recovery at discharge. Time-dependent propensity score matching was used to match patients who received ECPR to those at risk for ECPR within the same time interval. Risk ratios (RRs) and 95% confidence intervals (CIs) were estimated, and stratified analysis by the timing of ECPR was also performed.

RESULTS

Among 118,391 eligible patients, 484 received ECPR. After 1:4 time-dependent propensity score matching, 458 patients in the ECPR group and 1832 patients in the no ECPR group were included in the matched cohort. In the matched cohort, ECPR was not associated with good neurological recovery (10.3% in ECPR and 6.9% in no ECPR; RR [95% CI] 1.28 [0.85-1.93]). In the stratified analyses according to the timing of matching, ECPR with a pump-on within 45 min after emergency department arrival was associated with favourable neurological outcomes (RR [95% CI] 2.51 [1.33-4.75] in 1-30 min, 1.81 [1.11-2.93] in 31-45 min, 1.07 (0.56-2.04) in 46-60 min, and 0.45 (0.11-1.91) in over 60 min).

CONCLUSIONS

ECPR itself was not associated with good neurological recovery, but early ECPR was positively associated with good neurological recovery. Research on how to perform ECPR at an early stage and clinical trials to evaluate the effect of ECPR is warranted.

Community–Clinical Linkages: The Effects of the Healthy Here Wellness Referral Center on Chronic Disease Indicators Among Underserved Populations in New Mexico

The majority of U.S. adults are living with at least one chronic condition, and people of color bear a disproportionate burden of chronic disease. Prior research identifies community–clinical linkages (CCLs) as a strategy for improving health. CCLs traditionally use health care providers to connect patients to community-based self-management programs. The purpose of this study was to examine the effectiveness of a centralized CCL system on health indicators and health disparities. Administrative health data were merged with referral system data to conduct a quasi-experimental comparative time series study with a comparison group of nonreferred patients. Interrupted time-series comparisons within referred patients were also conducted. Of the 2,920 patients meeting inclusion criteria, 972 (33.3%) received a referral during the study period (January 2019—September 2021). Hemoglobin A1c levels, used to diagnose diabetes, declined significantly among referred patients, as did disparities among Hispanic/Latinx participants compared with non-Hispanic White participants. No changes were observed in body mass index (BMI). Blood pressure increased among both referred and nonreferred patients. CCLs with a centralized referral system can effectively reduce markers of diabetes and may contribute to the maintenance of BMI. The observed increase in blood pressure may have been affected by the COVID-19 pandemic and warrants further study. Practitioners can work with community partners to implement a centralized CCL model, either on its own or to enhance existing clinician or community health worker-based models.

The Efficacy and Safety of Sodium Bicarbonate Ringer's Solution in Critically Ill Patients: A Retrospective Cohort Study

Background: Sodium bicarbonate Ringer’s solution has been widely used in clinical practice in recent years. There are few clinical studies on the efficacy and safety of this fluid among critically ill patients until now.

Method: This retrospective cohort study included critically ill adult patients in the intensive care unit (ICU) of Tongji Hospital from 1 January 2019 to 31 December 2020. By reviewing exclusively the use of either sodium bicarbonate Ringer’s solution or saline for resuscitation or maintenance, the patients were included into two groups, respectively. The primary outcome was the major adverse kidney event within 30 days (MAKE30), including death, new receipt of renal replacement therapy, or persistent renal dysfunction. Safety outcomes were focused on arterial blood gas and plasma biochemical alterations, which might potentially be induced by the administration of bicarbonate Ringer’s solution.

Result: A total of 662 patients were included in the cohort. Compared to the saline group, the bicarbonate Ringer’s group had a significantly lower rate of the new receipt of renal replacement therapy [adjusted odds ratio (OR) = 0.591, 95% confidence interval (CI), 0.406 to 0.861; p = 0.006]. There was no significant difference between the two groups in 30-day mortality, final creatinine level ≥200% of baseline, and major adverse kidney event within 30 days. In subgroup analysis, the incidence of MAKE30 was higher in the bicarbonate Ringer’s group than that of the saline group among patients with cardiovascular disease. The patients in the bicarbonate Ringer’s group had a longer length of intensive care unit stay than patients in the saline group, but their new renal replacement therapy days were shorter. No major alterations were found in arterial blood gas and plasma biochemical during the follow-up period.

Conclusion: Compared to saline, sodium bicarbonate Ringer’s solution exhibited a potential renal function protective effect while causing no major alterations in arterial blood gas and plasma biochemistry. However, the application in patients with cardiovascular disease diagnosis at ICU admission should be cautious.

Confounding adjustment methods in longitudinal observational data with a time-varying treatment: a mapping review

OBJECTIVES

To adjust for confounding in observational data, researchers use propensity score matching (PSM), but more advanced methods might be required when dealing with longitudinal data and time-varying treatments as PSM might not include possible changes that occurred over time. This study aims to explore which confounding adjustment methods have been used in longitudinal observational data to estimate a treatment effect and identify potential inappropriate use of PSM.

DESIGN

Mapping review.

DATA SOURCES

We searched PubMed, from inception up to January 2021, for studies in which a treatment was evaluated using longitudinal observational data.

ELIGIBILITY CRITERIA

Methodological, non-medical and cost-effectiveness papers were excluded, as were non-English studies and studies that did not study a treatment effect.

DATA EXTRACTION AND SYNTHESIS

Studies were categorised based on time of treatment: at baseline (interventions performed at start of follow-up) or time-varying (interventions received asynchronously during follow-up) and sorted based on publication year, time of treatment and confounding adjustment method. Cumulative time series plots were used to investigate the use of different methods over time. No risk-of-bias assessment was performed as it was not applicable.

RESULTS

In total, 764 studies were included that met the eligibility criteria. PSM (165/201, 82%) and inverse probability weighting (IPW; 154/502, 31%) were most common for studies with a treatment at baseline (n=201) and time-varying treatment (n=502), respectively. Of the 502 studies with a time-varying treatment, 123 (25%) used PSM with baseline covariates, which might be inappropriate. In the past 5 years, the proportion of studies with a time-varying treatment that used PSM over IPW increased.

CONCLUSIONS

PSM is the most frequently used method to correct for confounding in longitudinal observational data. In studies with a time-varying treatment, PSM was potentially inappropriately used in 25% of studies. Confounding adjustment methods designed to deal with a time-varying treatment and time-varying confounding are available, but were only used in 45% of the studies with a time-varying treatment.

Clinical practice of Best Practice Nursing Care Standards for Older Adults with Fragility Hip Fracture: A propensity score matched analysis

BACKGROUND

Fragility hip fracture (FHF) is a significant cause of morbidity and mortality in older adults. In 2018, Best Practice Nursing Care Standards for Older Adults with Fragility Hip Fracture (NSOF) were released by The International Collaboration of Orthopaedic Nursing (ICON). However, there are only limited clinical data about the application of this standard in clinical practice in China.

AIMS

To determine the clinical practice effect of the NSOF.

METHODS

A retrospective single-centre cohort study was performed from January 2016 to June 2020. Patients were divided into the standardized nursing care group (SN group) and the conventional nursing care group (CN group) depending on whether they were cared for according to the NSOF criteria. The propensity score matched (PSM) analysis was conducted in this study. The perioperative and follow-up outcomes between the two groups were analyzed.

RESULTS

A total of 204 patients diagnosed with FHF were included in the study. After a 1:1 matching, 56 cases were identified in the SN group as well as the CN group. Patients in the SN group had significantly shorter preoperative wait times for surgery (17.4 ± 4.6 vs. 24.4 ± 7.6 h, p < 0.05) and a higher proportion of individuals performing exercise within 24 h after surgery (94.6% vs. 66.1%, p < 0.05). Notably, patients in the SN group also had a significantly shorter length of stay than those in the CN group (9.4 ± 3.1 vs. 14.2 ± 5.1 days, p < 0.05). At the 6-month follow-up, the incidence of refracture was significantly lower (3.6% vs. 14.3%, p < 0.05), and the timed up and go mobility index was improved in the SN group compared to the CN group (20.3 ± 1.7 vs. 24.6 ± 2.2 s, p < 0.05).

CONCLUSION

This study showed that application of the NSOF resulted in a significant improvement in the treatment of older adults patients with FHF.

Comparison of statistical methodologies used to estimate the treatment effect on time-to-event outcomes in observational studies

The use of real-world data became more and more popular in the pharmaceutical industry. The impact of real-world evidence is now well emphasized by the regulatory authorities. Indeed, the analysis of this type of data can play a key role for treatment efficacy and safety. The aim of this work is to assess various methods and give guidance on the comparisons of drugs, mostly with respect to time-to-event data, in non-randomized studies with potentially confounding variables. For that purpose, several statistical methodologies are compared based on simulation studies. These methodologies belong to family classes of methods that are widely used for this type of problem: regression, matching, weighting and subclassification methods. The evaluation criteria used to compare methods performances are the relative bias, the mean square error, the coverage probability and the width of the confidence interval. In this paper, we consider different scenarios of dataset features in order to study the effect of the sample size, the number of covariates and the magnitude of the treatment effect on the statistical methodologies performances. These statistical analyses are conducted within a proportional hazard model framework. Furthermore, we highlight the advantage of using techniques to identify relevant covariates for time-to-event outcomes by comparing two variable selection methods under a frequentist and a Bayesian inference. Based on simulation results, recommendations on each of the family of methods are provided to guide decision making.

Propensity score matching with R: conventional methods and new features

It is increasingly important to accurately and comprehensively estimate the effects of particular clinical treatments. Although randomization is the current gold standard, randomized controlled trials (RCTs) are often limited in practice due to ethical and cost issues. Observational studies have also attracted a great deal of attention as, quite often, large historical datasets are available for these kinds of studies. However, observational studies also have their drawbacks, mainly including the systematic differences in baseline covariates, which relate to outcomes between treatment and control groups that can potentially bias results. Propensity score methods, which are a series of balancing methods in these studies, have become increasingly popular by virtue of the two major advantages of dimension reduction and design separation. Within this approach, propensity score matching (PSM) has been empirically proven, with outstanding performances across observational datasets. While PSM tutorials are available in the literature, there is still room for improvement. Some PSM tutorials provide step-by-step guidance, but only one or two packages have been covered, thereby limiting their scope and practicality. Several articles and books have expounded upon propensity scores in detail, exploring statistical principles and theories; however, the lack of explanations on function usage in programming language has made it difficult for researchers to understand and follow these materials. To this end, this tutorial was developed with a six-step PSM framework, in which we summarize the recent updates and provide step-by-step guidance to the R programming language. This tutorial offers researchers with a broad survey of PSM, ranging from data preprocessing to estimations of propensity scores, and from matching to analyses. We also explain generalized propensity scoring for multiple or continuous treatments, as well as time-dependent PSM. Lastly, we discuss the advantages and disadvantages of propensity score methods.

The protective role of statins in COVID-19 patients: a retrospective observational study

BACKGROUND

To evaluate and determine the protective role of statins in COVID-19 patients.

METHODS

This is a retrospective cohort study conducted across five hospitals in India. Patients diagnosed with COVID-19 and hospitalized with existing and valid medical documentation were included.

RESULTS

This study comprised 3252 COVID-19 patients, of whom 1048 (32.2%) were on statins, with 52.4% being males. The comorbidity prevalence of hypertension was 75%, followed by diabetes 62.51% and coronary artery disease being 47.5%. At the time of hospitalization, statin users had a higher incidence of dyspnea, cough, and fatigue (95.8, 93.3, and 92.7%). The laboratory results revealed a lower mean of WBC count (7.8 × 10³/μL), D-dimer (2.4 μg/mL), and C-reactive protein (103 mg/L) among statin users. They also had lower mortality rates (17.1%), a lesser requirement for mechanical ventilation (20%), and hemodialysis (5.4%).

CONCLUSION

This observation study elaborates on the beneficial effects of statins in COVID-19 patients. However, the inferences from this study should be viewed with caution due to the impending effect of confounding factors on its statistical results.