Evaluating the impact of database heterogeneity on observational study results

Prediction of medical admissions after psychiatric inpatient hospitalization in bipolar disorder: a retrospective cohort study

Objective

Bipolar Disorder (BD) is a severe mental illness associated with high rates of general medical comorbidity, reduced life expectancy, and premature mortality. Although BD has been associated with high medical hospitalization, the factors that contribute to this risk remain largely unexplored. We used baseline medical and psychiatric records to develop a supervised machine learning model to predict general medical admissions after discharge from psychiatric hospitalization.

Methods

In this retrospective three-year cohort study of 71 patients diagnosed with BD (mean age=52.19 years, females=56.33%), lasso regression models combining medical and psychiatric records, as well as those using them separately, were fitted and their predictive power was estimated using a leave-one-out cross-validation procedure.

Results

The proportion of medical admissions in patients with BD was higher compared with age- and sex-matched hospitalizations in the same region (25.4% vs. 8.48%). The lasso model fairly accurately predicted the outcome (area under the curve [AUC]=69.5%, 95%C.I.=55-84.1; sensitivity=61.1%, specificity=75.5%, balanced accuracy=68.3%). Notably, pre-existing cardiovascular, neurological, or osteomuscular diseases collectively accounted for more than 90% of the influence on the model. The accuracy of the model based on medical records was slightly inferior (AUC=68.7%, 95%C.I. = 54.6-82.9), while that of the model based on psychiatric records only was below chance (AUC=61.8%, 95%C.I.=46.2-77.4).

Conclusion

Our findings support the need to monitor medical comorbidities during clinical decision-making to tailor and implement effective preventive measures in people with BD. Further research with larger sample sizes and prospective cohorts is warranted to replicate these findings and validate the predictive model.

Using electronic health records for clinical pharmacology research: Challenges and considerations

Electronic health records (EHRs) contain a vast array of phenotypic data on large numbers of individuals, often collected over decades. Due to the wealth of information, EHR data have emerged as a powerful resource to make first discoveries and identify disparities in our healthcare system. While the number of EHR-based studies has exploded in recent years, most of these studies are directed at associations with disease rather than pharmacotherapeutic outcomes, such as drug response or adverse drug reactions. This is largely due to challenges specific to deriving drug-related phenotypes from the EHR. There is great potential for EHR-based discovery in clinical pharmacology research, and there is a critical need to address specific challenges related to accurate and reproducible derivation of drug-related phenotypes from the EHR. This review provides a detailed evaluation of challenges and considerations for deriving drug-related data from EHRs. We provide an examination of EHR-based computable phenotypes and discuss cutting-edge approaches to map medication information for clinical pharmacology research, including medication-based computable phenotypes and natural language processing. We also discuss additional considerations such as data structure, heterogeneity and missing data, rare phenotypes, and diversity within the EHR. By further understanding the complexities associated with conducting clinical pharmacology research using EHR-based data, investigators will be better equipped to design thoughtful studies with more reproducible results. Progress in utilizing EHRs for clinical pharmacology research should lead to significant advances in our ability to understand differential drug response and predict adverse drug reactions.

Applicability of an unsupervised cluster model developed on first wave COVID-19 patients in second/third wave critically ill patients

OBJECTIVE

To validate the unsupervised cluster model (USCM) developed during the first pandemic wave in a cohort of critically ill patients from the second and third pandemic waves.

DESIGN

Observational, retrospective, multicentre study.

SETTING

Intensive Care Unit (ICU).

PATIENTS

Adult patients admitted with COVID-19 and respiratory failure during the second and third pandemic waves.

INTERVENTIONS

None.

MAIN VARIABLES OF INTEREST

Collected data included demographic and clinical characteristics, comorbidities, laboratory tests and ICU outcomes. To validate our original USCM, we assigned a phenotype to each patient of the validation cohort. The performance of the classification was determined by Silhouette coefficient (SC) and general linear modelling. In a post-hoc analysis we developed and validated a USCM specific to the validation set. The model's performance was measured using accuracy test and area under curve (AUC) ROC.

RESULTS

A total of 2330 patients (mean age 63 [53-82] years, 1643 (70.5%) male, median APACHE II score (12 [9-16]) and SOFA score (4 [3-6]) were included. The ICU mortality was 27.2%. The USCM classified patients into 3 clinical phenotypes: A (n = 1206 patients, 51.8%); B (n = 618 patients, 26.5%), and C (n = 506 patients, 21.7%). The characteristics of patients within each phenotype were significantly different from the original population. The SC was -0.007 and the inclusion of phenotype classification in a regression model did not improve the model performance (0.79 and 0.78 ROC for original and validation model). The post-hoc model performed better than the validation model (SC -0.08).

CONCLUSION

Models developed using machine learning techniques during the first pandemic wave cannot be applied with adequate performance to patients admitted in subsequent waves without prior validation.

Describing diversity of real world data sources in pharmacoepidemiologic studies: The DIVERSE scoping review

PURPOSE

Real-world evidence (RWE) is increasingly used for medical regulatory decisions, yet concerns persist regarding its reproducibility and hence validity. This study addresses reproducibility challenges associated with diversity across real-world data sources (RWDS) repurposed for secondary use in pharmacoepidemiologic studies. Our aims were to identify, describe and characterize practices, recommendations and tools for collecting and reporting diversity across RWDSs, and explore how leveraging diversity could improve the quality of evidence.

METHODS

In a preliminary phase, keywords for a literature search and selection tool were designed using a set of documents considered to be key by the coauthors. Next, a systematic search was conducted up to December 2021. The resulting documents were screened based on titles and abstracts, then based on full texts using the selection tool. Selected documents were reviewed to extract information on topics related to collecting and reporting RWDS diversity. A content analysis of the topics identified explicit and latent themes.

RESULTS

Across the 91 selected documents, 12 topics were identified: 9 dimensions used to describe RWDS (organization accessing the data source, data originator, prompt, inclusion of population, content, data dictionary, time span, healthcare system and culture, and data quality), tools to summarize such dimensions, challenges, and opportunities arising from diversity. Thirty-six themes were identified within the dimensions. Opportunities arising from data diversity included multiple imputation and standardization.

CONCLUSIONS

The dimensions identified across a large number of publications lay the foundation for formal guidance on reporting diversity of data sources to facilitate interpretation and enhance replicability and validity of RWE.

Pooling of primary care electronic health record (EHR) data on Huntington's disease (HD) and cancer: establishing comparability of two large UK databases

OBJECTIVES

To explore whether UK primary care databases arising from two different software systems can be feasibly combined, by comparing rates of Huntington's disease (HD, which is rare) and 14 common cancers in the two databases, as well as characteristics of people with these conditions.

DESIGN

Descriptive study.

SETTING

Primary care electronic health records from Clinical Practice Research Datalink (CPRD) GOLD and CPRD Aurum databases, with linked hospital admission and death registration data.

PARTICIPANTS

4986 patients with HD and 1 294 819 with an incident cancer between 1990 and 2019.

PRIMARY AND SECONDARY OUTCOME MEASURES

Incidence and prevalence of HD by calendar period, age group and region, and annual age-standardised incidence of 14 common cancers in each database, and in a subset of 'overlapping' practices which contributed to both databases. Characteristics of patients with HD or incident cancer: medical history, recent prescribing, healthcare contacts and database follow-up.

RESULTS

Incidence and prevalence of HD were slightly higher in CPRD GOLD than CPRD Aurum, but with similar trends over time. Cancer incidence in the two databases differed between 1990 and 2000, but converged and was very similar thereafter. Participants in each database were most similar in terms of medical history (median standardised difference, MSD 0.03 (IQR 0.01-0.03)), recent prescribing (MSD 0.06 (0.03-0.10)) and demographics and general health variables (MSD 0.05 (0.01-0.09)). Larger differences were seen for healthcare contacts (MSD 0.27 (0.10-0.41)), and database follow-up (MSD 0.39 (0.19-0.56)).

CONCLUSIONS

Differences in cancer incidence trends between 1990 and 2000 may relate to use of a practice-level data quality filter (the 'up-to-standard' date) in CPRD GOLD only. As well as the impact of data curation methods, differences in underlying data models can make it more challenging to define exactly equivalent clinical concepts in each database. Researchers should be aware of these potential sources of variability when planning combined database studies and interpreting results.

Risk of Arterial and Venous Thrombotic Events Among Patients with COVID-19: A Multi-National Collaboration of Regulatory Agencies from Canada, Europe, and United States

Purpose

Few studies have examined how the absolute risk of thromboembolism with COVID-19 has evolved over time across different countries. Researchers from the European Medicines Agency, Health Canada, and the United States (US) Food and Drug Administration established a collaboration to evaluate the absolute risk of arterial (ATE) and venous thromboembolism (VTE) in the 90 days after diagnosis of COVID-19 in the ambulatory (eg, outpatient, emergency department, nursing facility) setting from seven countries across North America (Canada, US) and Europe (England, Germany, Italy, Netherlands, and Spain) within periods before and during COVID-19 vaccine availability.

Patients and Methods

We conducted cohort studies of patients initially diagnosed with COVID-19 in the ambulatory setting from the seven specified countries. Patients were followed for 90 days after COVID-19 diagnosis. The primary outcomes were ATE and VTE over 90 days from diagnosis date. We measured country-level estimates of 90-day absolute risk (with 95% confidence intervals) of ATE and VTE.

Results

The seven cohorts included 1,061,565 patients initially diagnosed with COVID-19 in the ambulatory setting before COVID-19 vaccines were available (through November 2020). The 90-day absolute risk of ATE during this period ranged from 0.11% (0.09-0.13%) in Canada to 1.01% (0.97-1.05%) in the US, and the 90-day absolute risk of VTE ranged from 0.23% (0.21-0.26%) in Canada to 0.84% (0.80-0.89%) in England. The seven cohorts included 3,544,062 patients with COVID-19 during vaccine availability (beginning December 2020). The 90-day absolute risk of ATE during this period ranged from 0.06% (0.06-0.07%) in England to 1.04% (1.01-1.06%) in the US, and the 90-day absolute risk of VTE ranged from 0.25% (0.24-0.26%) in England to 1.02% (0.99-1.04%) in the US.

Conclusion

There was heterogeneity by country in 90-day absolute risk of ATE and VTE after ambulatory COVID-19 diagnosis both before and during COVID-19 vaccine availability.

The State of Use and Utility of Negative Controls in Pharmacoepidemiologic Studies

Uses of real-world data in drug safety and effectiveness studies are often challenged by various sources of bias. We undertook a systematic search of the published literature through September 2020 to evaluate the state of use and utility of negative controls to address bias in pharmacoepidemiologic studies. Two reviewers independently evaluated study eligibility and abstracted data. Our search identified 184 eligible studies for inclusion. Cohort studies (115, 63%) and administrative data (114, 62%) were, respectively, the most common study design and data type used. Most studies used negative control outcomes (91, 50%), and for most studies the target source of bias was unmeasured confounding (93, 51%). We identified 4 utility domains of negative controls: 1) bias detection (149, 81%), 2) bias correction (16, 9%), 3) P-value calibration (8, 4%), and 4) performance assessment of different methods used in drug safety studies (31, 17%). The most popular methodologies used were the 95% confidence interval and P-value calibration. In addition, we identified 2 reference sets with structured steps to check the causality assumption of the negative control. While negative controls are powerful tools in bias detection, we found many studies lacked checking the underlying assumptions. This article is part of a Special Collection on Pharmacoepidemiology.

Integrating Multi-Omics Data With EHR for Precision Medicine Using Advanced Artificial Intelligence

With the recent advancement of novel biomedical technologies such as high-throughput sequencing and wearable devices, multi-modal biomedical data ranging from multi-omics molecular data to real-time continuous bio-signals are generated at an unprecedented speed and scale every day. For the first time, these multi-modal biomedical data are able to make precision medicine close to a reality. However, due to data volume and the complexity, making good use of these multi-modal biomedical data requires major effort. Researchers and clinicians are actively developing artificial intelligence (AI) approaches for data-driven knowledge discovery and causal inference using a variety of biomedical data modalities. These AI-based approaches have demonstrated promising results in various biomedical and healthcare applications. In this review paper, we summarize the state-of-the-art AI models for integrating multi-omics data and electronic health records (EHRs) for precision medicine. We discuss the challenges and opportunities in integrating multi-omics data with EHRs and future directions. We hope this review can inspire future research and developing in integrating multi-omics data with EHRs for precision medicine.

Using A Standardized Nomenclature to Semantically Map Oncology-Related Concepts from Common Data Models to a Pediatric Cancer Data Model

The Pediatric Cancer Data Commons (PCDC) comprises an international community whose ironclad commitment to data sharing is combatting pediatric cancer in an unprecedented way. The byproduct of their data sharing efforts is a gold-standard consensus data model covering many types of pediatric cancer. This article describes an effort to utilize SSSOM, an emerging specification for semantically-rich data mappings, to provide a "hub and spoke" model of mappings from several common data models (CDMs) to the PCDC data model. This provides important contributions to the research community, including: 1) a clear view of the current coverage of these CDMs in the domain of pediatric oncology, and 2) a demonstration of creating standardized mappings. These mappings can allow downstream crosswalk for data transformation and enhance data sharing. This can guide those who currently create and maintain brittle ad hoc data mappings in order to utilize the growing volume of viable research data.

Analysis of postoperative complications related to cannabis and tobacco usage in patients undergoing mandible facial fracture surgeries

BACKGROUND

Cannabis is the third most used controlled substance in the world. Despite its widespread use, minimal research investigates cannabis usage in patients undergoing facial fracture surgeries. This study aimed to evaluate patterns of postoperative complications related to cannabis and tobacco usage after mandible fracture surgeries.

MATERIALS AND METHODS

PearlDiver™, a commercially available healthcare database, was used to identify patients endorsing the use of cannabis, tobacco, or both who underwent mandible fracture surgeries for cross-sectional analysis. The study population was categorized into groups using the Classification of Diseases, 9th revision (ICD-9), ICD-10, and Current Procedural Terminology (CPT) codes. A chi-square analysis was performed to assess the influence of cannabis and tobacco use on postoperative complications.

RESULTS

A total of 8288 patients met the inclusion criteria, with 72 patients with cannabis-only usage, 914 patients with cannabis and tobacco usage, 3236 patients with tobacco-only usage, and 4066 in the control group. For patients using only cannabis, there was not an increased risk of developing postoperative complications compared with the control population. Patients with concurrent cannabis and tobacco usage and those with tobacco-only usage had an increased risk of surgical site infection, facial nonunion, facial abscess, debridement, and malocclusion after surgical repair of mandibular facial fracture.

CONCLUSION

Patients with tobacco-only as well as cannabis and tobacco usage had an increased risk of all postoperative complications, except malocclusion, compared with cannabis-only. Based on the results of this study, it is recommended that healthcare providers consider a patient's history of tobacco use when planning and performing surgical treatment for traumatic mandible fractures.

Modelling of intensive care unit (ICU) length of stay as a quality measure: a problematic exercise

BACKGROUND

Intensive care unit (ICU) length of stay (LOS) and the risk adjusted equivalent (RALOS) have been used as quality metrics. The latter measures entail either ratio or difference formulations or ICU random effects (RE), which have not been previously compared.

METHODS

From calendar year 2016 data of an adult ICU registry-database (Australia & New Zealand Intensive Care Society (ANZICS) CORE), LOS predictive models were established using linear (LMM) and generalised linear (GLMM) mixed models. Model fixed effects quality-metric formulations were estimated as RALOSR for LMM (geometric mean derived from log(ICU LOS)) and GLMM (day) and observed minus expected ICU LOS (OMELOS from GLMM). Metric confidence intervals (95%CI) were estimated by bootstrapping; random effects (RE) were predicted for LMM and GLMM. Forest-plot displays of ranked quality-metric point-estimates (95%CI) were generated for ICU hospital classifications (metropolitan, private, rural/regional, and tertiary). Robust rank confidence sets (point estimate and 95%CI), both marginal (pertaining to a singular ICU) and simultaneous (pertaining to all ICU differences), were established.

RESULTS

The ICU cohort was of 94,361 patients from 125 ICUs (metropolitan 16.9%, private 32.8%, rural/regional 6.4%, tertiary 43.8%). Age (mean, SD) was 61.7 (17.5) years; 58.3% were male; APACHE III severity-of-illness score 54.6 (25.7); ICU annual patient volume 1192 (702) and ICU LOS 3.2 (4.9). There was no concordance of ICU ranked model predictions, GLMM versus LMM, nor for the quality metrics used, RALOSR, OMELOS and site-specific RE for each of the ICU hospital classifications. Furthermore, there was no concordance between ICU ranking confidence sets, marginal and simultaneous for models or quality metrics.

CONCLUSIONS

Inference regarding adjusted ICU LOS was dependent upon the statistical estimator and the quality index used to quantify any LOS differences across ICUs. That is, there was no "one best model"; thus, ICU "performance" is determined by model choice and any rankings thereupon should be circumspect.

Assessing heterogeneity of electronic health-care databases: A case study of background incidence rates of venous thromboembolism

PURPOSE

Heterogeneous results from multi-database studies have been observed, for example, in the context of generating background incidence rates (IRs) for adverse events of special interest for SARS-CoV-2 vaccines. In this study, we aimed to explore different between-database sources of heterogeneity influencing the estimated background IR of venous thromboembolism (VTE).

METHODS

Through forest plots and random-effects models, we performed a qualitative and quantitative assessment of heterogeneity of VTE background IR derived from 11 databases from 6 European countries, using age and gender stratified background IR for the years 2017-2019 estimated in two studies. Sensitivity analyses were performed to assess the impact of selection criteria on the variability of the reported IR.

RESULTS

A total of 54 257 284 subjects were included in this study. Age-gender pooled VTE IR varied from 5 to 421/100 000 person-years and IR increased with increasing age for both genders. Wide confidence intervals (CIs) demonstrated considerable within-data-source heterogeneity. Selecting databases with similar characteristics had only a minor impact on the variability as shown in forest plots and the magnitude of the I2 statistic, which remained large. Solely including databases with primary care and hospital data resulted in a noticeable decrease in heterogeneity.

CONCLUSIONS

Large variability in IR between data sources and within age group and gender strata warrants the need for stratification and limits the feasibility of a meaningful pooled estimate. A more detailed knowledge of the data characteristics, operationalisation of case definitions and cohort population might support an informed choice of the adequate databases to calculate reliable estimates.

Padé approximant meets federated learning: A nearly lossless, one-shot algorithm for evidence synthesis in distributed research networks with rare outcomes

OBJECTIVE

We developed and evaluated a novel one-shot distributed algorithm for evidence synthesis in distributed research networks with rare outcomes.

MATERIALS AND METHODS

Fed-Padé, motivated by a classic mathematical tool, Padé approximants, reconstructs the multi-site data likelihood via Padé approximant whose key parameters can be computed distributively. Thanks to the simplicity of [2,2] Padé approximant, Fed-Padé requests an extremely simple task and low communication cost for data partners. Specifically, each data partner only needs to compute and share the log-likelihood and its first 4 gradients evaluated at an initial estimator. We evaluated the performance of our algorithm with extensive simulation studies and four observational healthcare databases.

RESULTS

Our simulation studies revealed that a [2,2]-Padé approximant can well reconstruct the multi-site likelihood so that Fed-Padé produces nearly identical estimates to the pooled analysis. Across all simulation scenarios considered, the median of relative bias and rate of instability of our Fed-Padé are both <0.1%, whereas meta-analysis estimates have bias up to 50% and instability up to 75%. Furthermore, the confidence intervals derived from the Fed-Padé algorithm showed better coverage of the truth than confidence intervals based on the meta-analysis. In real data analysis, the Fed-Padé has a relative bias of <1% for all three comparisons for risks of acute liver injury and decreased libido, whereas the meta-analysis estimates have a substantially higher bias (around 10%).

CONCLUSION

The Fed-Padé algorithm is nearly lossless, stable, communication-efficient, and easy to implement for models with rare outcomes. It provides an extremely suitable and convenient approach for synthesizing evidence in distributed research networks with rare outcomes.

The IMPACT framework and implementation for accessible in silico clinical phenotyping in the digital era

Clinical phenotyping is often a foundational requirement for obtaining datasets necessary for the development of digital health applications. Traditionally done via manual abstraction, this task is often a bottleneck in development due to time and cost requirements, therefore raising significant interest in accomplishing this task via in-silico means. Nevertheless, current in-silico phenotyping development tends to be focused on a single phenotyping task resulting in a dearth of reusable tools supporting cross-task generalizable in-silico phenotyping. In addition, in-silico phenotyping remains largely inaccessible for a substantial portion of potentially interested users. Here, we highlight the barriers to the usage of in-silico phenotyping and potential solutions in the form of a framework of several desiderata as observed during our implementation of such tasks. In addition, we introduce an example implementation of said framework as a software application, with a focus on ease of adoption, cross-task reusability, and facilitating the clinical phenotyping algorithm development process.

Should lethal arrhythmias in hypertrophic cardiomyopathy be predicted using non-electrophysiological methods?

While sudden cardiac death (SCD) in hypertrophic cardiomyopathy (HCM) is due to arrhythmias, the guidelines for prediction of SCD are based solely on non-electrophysiological methods. This study aims to stimulate thinking about whether the interests of patients with HCM are better served by using current, 'risk factor', methods of prediction or by further development of electrophysiological methods to determine arrhythmic risk. Five published predictive studies of SCD in HCM, which contain sufficient data to permit analysis, were analysed to compute receiver operating characteristics together with their confidence bounds to compare their formal prediction either by bootstrapping or Monte Carlo analysis. Four are based on clinical risk factors, one with additional MRI analysis, and were regarded as exemplars of the risk factor approach. The other used an electrophysiological method and directly compared this method to risk factors in the same patients. Prediction methods that use conventional clinical risk factors and MRI have low predictive capacities that will only detect 50-60% of patients at risk with a 15-30% false positive rate [area under the curve (AUC) = ∼0.7], while the electrophysiological method detects 90% of events with a 20% false positive rate (AUC = ∼0.89). Given improved understanding of complex arrhythmogenesis, arrhythmic SCD is likely to be more accurately predictable using electrophysiologically based approaches as opposed to current guidelines and should drive further development of electrophysiologically based methods.

How to Create an Orthopaedic Arthroplasty Administrative Database Project: A Step-by-Step Guide Part I: Study Design

BACKGROUND

Use of clinical and administrative databases in orthopaedic surgery research has grown substantially in recent years. It is estimated that approximately 10% of all published lower extremity arthroplasty research have been database studies. The aim of this review is to serve as a guide on how to (1) design, (2) execute, and (3) publish an orthopaedic administrative database arthroplasty project.

METHODS

In part I, we discuss how to develop a research question and choose a database (when databases should/should not be used), detailing advantages/disadvantages of those most commonly used. To date, the most commonly published databases in orthopaedic research have been the National Inpatient Sample, Medicare, National Surgical Quality Improvement Program, and those provided by PearlDiver. General advantages of most database studies include accessibility, affordability compared to prospective research studies, ease of use, large sample sizes, and the ability to identify trends and aggregate outcomes of multiple health care systems/providers.

RESULTS

Disadvantages of most databases include their retrospective observational nature, limitations of procedural/billing coding, relatively short follow-up, limited ability to control for confounding variables, and lack of functional/patient-reported outcomes.

CONCLUSION

Although this study is not all-encompassing, we hope it will serve as a starting point for those interested in conducting and critically reviewing lower extremity arthroplasty database studies.

Can Observational Analyses of Routinely Collected Data Emulate Randomized Trials? Design and Feasibility of the Observational Patient Evidence for Regulatory Approval Science and Understanding Disease Project

OBJECTIVES

The Observational Patient Evidence for Regulatory Approval Science and Understanding Disease (OPERAND) project examines whether real-world data (RWD) can be used to inform regulatory decision making.

METHODS

OPERAND evaluates whether observational analyses using RWD to emulate index trials can produce effect estimates similar to those of the trials and examines the impact of relaxing the eligibility criteria of the observational analyses to obtain samples that more closely match the real-world populations receiving the treatments. In OPERAND, 2 research teams independently attempt to emulate the ROCKET Atrial Fibrillation and LEAD-2 trials using OptumLabs data. This article describes the design of the project, summarizes the approaches of the 2 research teams, and presents feasibility results for 2 emulations using new-user designs.

RESULTS

There were differences in the teams' conceptualizations of the emulation, design decisions for cohort identification, and resulting RWD cohorts. These differences occurred even though both teams were guided by the same index trials and had access to the same source of RWD.

CONCLUSIONS

Reasonable alternative design and analysis approaches may be taken to answer the same research question, even when attempting to emulate the same index trial. Researcher decision making is an understudied and potentially important source of variability across RWD analyses.

An overview of methodological flaws of real-world studies investigating drug safety in the post-marketing setting

INTRODUCTION

The evaluation of the post-marketing safety profile of drugs is a continuous monitoring process for approved and marketed medicines and it is crucial for detecting new adverse drug reactions. As such, real-world studies are essential to complement pre-marketing evidence with information concerning drug risk-benefit profile and use in wider patient populations and they have a great potential to support post-marketing drug safety evaluations.

AREAS COVERED

A detailed description of the main limitations of real-world data sources (i.e. claims databases, electronic healthcare records, drug/disease registers and spontaneous reporting system databases) and of the main methodological challenges of real-world studies in generating real-world evidence is provided.

EXPERT OPINION

Real-world evidence biases can be ascribed to both the methodological approach and the specific limitations of the different real-world data sources used to carry out the study. As such, it is crucial to characterize the quality of real-world data, by establishing guidelines and best practices for the assessment of data fitness for purpose. On the other hand, it is important that real-world studies are conducted using a rigorous methodology, aimed at minimizing the risk of bias.

Survivorship after reverse total shoulder arthroplasty and predictors of 1-year and overall mortality

BACKGROUND

Patient survivorship and risk factors of mortality after reverse total shoulder arthroplasty (RTSA) are seldom and inadequately studied. The purpose of this study was to evaluate the mortality rates and predictors of 1-year and overall mortality after RTSA.

METHODS

We retrospectively reviewed 1518 consecutive adult patients who underwent RTSA at our institution. The Social Security Death Index and institutional electronic medical records were queried to verify patient living status. Patients were censored at date of death if deceased, the date that living status was verified if alive, or latest follow-up if living status could not be verified. Mortality rates and risk factors of 1-year and overall mortality after RTSA were identified on univariate and multivariate analysis.

RESULTS

Mean follow-up was 5.1 ± 3.8 years. Thirty-day (0.1%), 90-day (0.7%), and 1-year (1.8%) mortality rates were low but increased to 11% at 5 years. Increased odds of 1-year mortality were independently associated with heart disease (odds ratio [OR] 2.64, 95% confidence interval [CI] 1.07-6.50, P = .035) and use of a cemented stem (OR 2.64, 95% CI 1.04-6.69, P = .041). Independent risk factors of overall mortality included older age at surgery (hazard ratio [HR] 1.05, 95% CI 1.03-1.07, P < .001), minority ethnicity (protective risk factor, HR 0.37, 95% CI 0.15-0.91, P = .031), heart disease (HR 1.42, 95% CI 1.00-2.02, P = .048), diabetes mellitus (HR 1.47, 95% CI 1.04-2.08, P = .028), tobacco use (HR 1.79, 95% CI 1.08-2.98, P = .025), post renal transplant (HR 12.69, 95% CI 3.92-41.05, P < .001), chronic liver failure (HR 4.40, 95% CI 1.38-14.09, P = .013), and receiving a cemented stem (HR 1.60, 95% CI 1.13-2.26, P = .008).

CONCLUSIONS

RTSA carries a low risk of short-term mortality postoperatively. When counseling patients preoperatively, surgeons should consider the predictors of mortality after RTSA reported herein to ensure appropriate patient selection and counseling.

Trends in Surgical Approach for Single-Level Lumbar Fusion Over the Past Decade

STUDY DESIGN

Retrospective Comparative Study.

OBJECTIVE

The purpose of this study was to characterize trends in surgical approach for single-level lumbar fusion over the past decade.

SUMMARY OF BACKGROUND DATA

The number of elective lumbar fusion cases performed is increasing annually. Several different surgical approaches exist for lumbar spinal fusion including novel anterior approaches developed in recent years. With ongoing innovation, trends in the utilization of common surgical approaches in recent years are unclear.

MATERIALS AND METHODS

A retrospective cohort study was conducted using the PearlDiver database (Fort Wayne, IN). Patients undergoing single-level lumbar fusion between 2010 and 2019 were identified using Current Procedural Technology codes and divided into 4 mutually exclusive cohorts based on surgical approach: (1) anterior-only, (2) anterior approach with posterior instrumentation, (3) posterolateral, and (4) posterior-only interbody. Trend analyses of surgical approach utilization over the last decade were performed with the Cochran-Armitage test to evaluate the 2-tailed null hypothesis that utilization of each surgical approach for single-level lumbar fusion remained constant.

RESULTS

A total of 53,234 patients met inclusion criteria and were stratified into 4 cohorts: anterior-only (n=5104), anterior with posterior instrumentation (n=23,515), posterolateral (n=5525), and posterior-only interbody (n=19,090). Trend analysis revealed the utilization of a posterior-only interbody approach significantly decreased from 36.7% to 29.2% (P<0.001), whereas the utilization of a combined anterior and posterior approach significantly increased from 45.8% to 50.4% (P<0.001). The utilization of an anterior-only approach also significantly increased from 7.9% to 10.5% (P<0.001).

CONCLUSIONS

Utilization of anterior-only and anterior with posterior instrumentation approaches for single-level lumbar fusion have been significantly increasing over the past decade while use of posterior-only interbody approach trended significantly downward. These data may be particularly useful for trainees and spine surgeons as new techniques and technology become available.

LEVEL OF EVIDENCE

Level III-retrospective cohort study.

Demographic Predictors of Treatment and Complications for Adult Spinal Deformity: An Analysis of the National Inpatient Sample

PURPOSE

To examine the role of demographics on surgical management and inpatient complications in patients with spinal deformity between 2010 and 2014 via retrospective analysis.

METHODS

Data were obtained from the National Inpatient Sample (NIS). International Classification of Diseases 9th revision codes were used to identify patients with a primary diagnosis of adult spinal deformity (ASD). Multivariable Poisson regression analyses were used to determine whether any individual demographic variables were predictive of surgical management, surgical complexity, postoperative complications and revision operations.

RESULTS

17,433 patients were identified for analysis. Surgical intervention was performed for 94.5% of patients with a primary diagnosis of ASD. Patients at urban teaching hospitals were the most likely to receive surgery (OR= 2.13; 95% CI 1.51-2.95; p < 0.001) relative to rural patients. Female patients were the majority undergoing surgery and were more likely to receive a complication or require a revision when controlling for surgical complexity. Medicare patients were the least likely to undergo surgery and the most likely to receive complex fusion when undergoing an operation. Medicare patients were the least likely to experience complications (OR=0.89; 95% CI 0.80-0.98; p = 0.022) after adjusting for surgical complexity. With regards to race and ethnicity, Hispanics had a decreased likelihood of receiving a revision surgery.

CONCLUSION

There were substantial differences in rates of surgical management, postoperative complications, and revisions among individuals of different demographics including sex, insurance status, ethnicity and hospital teaching status. Further research evaluating the effect of demographics in spine surgery is warranted to fully understand their influence on patient outcomes.

Evaluating the performance of temporal pattern discovery: new application using statins and rhabdomyolysis in OMOP databases

Background

Temporal pattern discovery (TPD) is a method of signal detection using electronic healthcare databases, serving as an alternative to spontaneous reporting of adverse drug events. Here, we aimed to replicate and optimise a TPD approach previously used to assess temporal signals of statins with rhabdomyolysis (in The Health Improvement Network (THIN) database) by using the OHDSI tools designed for OMOP data sources.

Methods

We used data from the Truven MarketScan US Commercial Claims and the Commercial Claims and Encounters (CCAE). Using an extension of the OHDSI ICTemporalPatternDiscovery package, we ran positive and negative controls through four analytical settings and calculated sensitivity, specificity, bias and AUC to assess performance.

Results

Similar to previous findings, we noted an increase in the Information Component (IC) for simvastatin and rhabdomyolysis following initial exposure and throughout the surveillance window. For example, the change in IC was 0.266 for the surveillance period of 1–30 days as compared to the control period of − 180 to − 1 days. Our modification of the existing OHDSI software allowed for faster queries and more efficient generation of chronographs.

Conclusion

Our OMOP replication matched the we can account forwe can account for of the original THIN study, only simvastatin had a signal. The TPD method is a useful signal detection tool that provides a single statistic on temporal association and a graphical depiction of the temporal pattern of the drug outcome combination. It remains unclear if the method works well for rare adverse events, but it has been shown to be a useful risk identification tool for longitudinal observational databases. Future work should compare the performance of TPD with other pharmacoepidemiology methods and mining techniques of signal detection. In addition, it would be worth investigating the relative TPD performance characteristics using a variety of observational data sources.

Secondary Prevention Using Cholesterol-Lowering Medications in Patients with Prior Atherosclerotic Cardiovascular Disease Events: A Retrospective Cohort Analysis

Background: Secondary prevention with lipid-lowering medications in patients with atherosclerotic cardiovascular disease (ASCVD) is known to reduce the risk of clinical events and death. Current guidelines codify recommendations for implementing secondary prevention in appropriate patients. However, in real-world practice, secondary prevention is frequently initiated only after the patient experiences a cardiovascular-related hospitalization. The impact of these delays is not well known. Objectives: To estimate the effects of delaying treatment on the risk of cardiovascular-related hospitalization and on costs for patients who meet the criteria for secondary prevention as specified in the 2013 American College of Cardiology/American Heart Association (ACC/AHA) Guidelines for Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults. Methods: This is a retrospective cohort analysis using Humana data. Eligible patients were categorized by treatment group: (1) patients who initiated treatment before an ASCVD-related hospitalization and (2) patients who either did not initiate treatment until after an ASCVD hospitalization or never initiated treatment. The associations between the timely initiation of cholesterol-lowering medications for secondary prevention and (1) the risk for an ASCVD hospitalization and (2) health-care costs over one year, were estimated using multivariate regressions. Results: A total of 272 899 secondary prevention patients were identified who met study selection criteria. Early treatment was associated with significant reductions in the risk of an ASCVD hospitalization at any time following the identification of the patient's eligibility for secondary prevention (by 33% compared to those treated late or never, P<.0001), but was significantly associated with higher total cost over the first post-index year (by US $509, P<.001). Patients whose low-density lipoprotein cholesterol (LDL-C) levels were >130 mg/dL experienced higher ASCVD hospitalization risks, and also larger risk reductions if treated before an ASCVD hospitalization compared to patients with lower LDL-C levels who were treated late or never treated. Conclusions: More widespread implementation of the treatment policies specified in the 2013 ACC/AHA Guidelines for secondary prevention should significantly reduce cardiovascular disease hospitalizations and reduce costs.

Transformation and Evaluation of the MIMIC Database in the OMOP Common Data Model: Development and Usability Study

Background

In the era of big data, the intensive care unit (ICU) is likely to benefit from real-time computer analysis and modeling based on close patient monitoring and electronic health record data. The Medical Information Mart for Intensive Care (MIMIC) is the first open access database in the ICU domain. Many studies have shown that common data models (CDMs) improve database searching by allowing code, tools, and experience to be shared. The Observational Medical Outcomes Partnership (OMOP) CDM is spreading all over the world.

Objective

The objective was to transform MIMIC into an OMOP database and to evaluate the benefits of this transformation for analysts.

Methods

We transformed MIMIC (version 1.4.21) into OMOP format (version 5.3.3.1) through semantic and structural mapping. The structural mapping aimed at moving the MIMIC data into the right place in OMOP, with some data transformations. The mapping was divided into 3 phases: conception, implementation, and evaluation. The conceptual mapping aimed at aligning the MIMIC local terminologies to OMOP's standard ones. It consisted of 3 phases: integration, alignment, and evaluation. A documented, tested, versioned, exemplified, and open repository was set up to support the transformation and improvement of the MIMIC community's source code. The resulting data set was evaluated over a 48-hour datathon.

Results

With an investment of 2 people for 500 hours, 64% of the data items of the 26 MIMIC tables were standardized into the OMOP CDM and 78% of the source concepts mapped to reference terminologies. The model proved its ability to support community contributions and was well received during the datathon, with 160 participants and 15,000 requests executed with a maximum duration of 1 minute.

Conclusions

The resulting MIMIC-OMOP data set is the first MIMIC-OMOP data set available free of charge with real disidentified data ready for replicable intensive care research. This approach can be generalized to any medical field.

Comparative effectiveness and safety of pharmaceuticals assessed in observational studies compared with randomized controlled trials

BACKGROUND

There have been ongoing efforts to understand when and how data from observational studies can be applied to clinical and regulatory decision making. The objective of this review was to assess the comparability of relative treatment effects of pharmaceuticals from observational studies and randomized controlled trials (RCTs).

METHODS

We searched PubMed and Embase for systematic literature reviews published between January 1, 1990, and January 31, 2020, that reported relative treatment effects of pharmaceuticals from both observational studies and RCTs. We extracted pooled relative effect estimates from observational studies and RCTs for each outcome, intervention-comparator, or indication assessed in the reviews. We calculated the ratio of the relative effect estimate from observational studies over that from RCTs, along with the corresponding 95% confidence interval (CI) for each pair of pooled RCT and observational study estimates, and we evaluated the consistency in relative treatment effects.

RESULTS

Thirty systematic reviews across 7 therapeutic areas were identified from the literature. We analyzed 74 pairs of pooled relative effect estimates from RCTs and observational studies from 29 reviews. There was no statistically significant difference (based on the 95% CI) in relative effect estimates between RCTs and observational studies in 79.7% of pairs. There was an extreme difference (ratio < 0.7 or > 1.43) in 43.2% of pairs, and, in 17.6% of pairs, there was a significant difference and the estimates pointed in opposite directions.

CONCLUSIONS

Overall, our review shows that while there is no significant difference in the relative risk ratios between the majority of RCTs and observational studies compared, there is significant variation in about 20% of comparisons. The source of this variation should be the subject of further inquiry to elucidate how much of the variation is due to differences in patient populations versus biased estimates arising from issues with study design or analytical/statistical methods.

A narrative review on the validity of electronic health record-based research in epidemiology

Electronic health records (EHRs) are widely used in epidemiological research, but the validity of the results is dependent upon the assumptions made about the healthcare system, the patient, and the provider. In this review, we identify four overarching challenges in using EHR-based data for epidemiological analysis, with a particular emphasis on threats to validity. These challenges include representativeness of the EHR to a target population, the availability and interpretability of clinical and non-clinical data, and missing data at both the variable and observation levels. Each challenge reveals layers of assumptions that the epidemiologist is required to make, from the point of patient entry into the healthcare system, to the provider documenting the results of the clinical exam and follow-up of the patient longitudinally; all with the potential to bias the results of analysis of these data. Understanding the extent of as well as remediating potential biases requires a variety of methodological approaches, from traditional sensitivity analyses and validation studies, to newer techniques such as natural language processing. Beyond methods to address these challenges, it will remain crucial for epidemiologists to engage with clinicians and informaticians at their institutions to ensure data quality and accessibility by forming multidisciplinary teams around specific research projects.

Association Between Testosterone Treatment and Risk of Incident Cardiovascular Events Among US Male Veterans With Low Testosterone Levels and Multiple Medical Comorbidities

Background

Testosterone treatment is common in men, although risks for major cardiovascular events are unclear.

Methods and Results

A study was conducted in US male veterans, aged ≥40 years, with low serum testosterone and multiple medical comorbidities and without history of myocardial infarction, stroke, venous thromboembolism, prostate cancer, or testosterone treatment in the prior year. For the primary outcome, we examined if testosterone treatment was associated with a composite cardiovascular outcome (incident myocardial infarction, ischemic stroke, or venous thromboembolism). Testosterone use was modeled as intramuscular or transdermal and as current use, former use, and no use. Current testosterone users were compared with former users to reduce confounding by indication. The cohort consisted of 204 857 men with a mean (SD) age of 60.9 (9.9) years and 4.7 (3.5) chronic medical conditions. During follow‐up of 4.3 (2.8) years, 12 645 composite cardiovascular events occurred. In adjusted Cox regression analyses, current use of transdermal testosterone was not associated with risk for the composite cardiovascular outcome (hazard ratio [HR], 0.89; 95% CI, 0.76–1.05) in those without prevalent cardiovascular disease, and in those with prevalent cardiovascular disease was associated with lower risk (HR, 0.80; 95% CI, 0.70–0.91). In similar analyses, current use of intramuscular testosterone was not associated with risk for the composite cardiovascular outcome in men without or with prevalent cardiovascular disease (HR, 0.91; 95% CI, 0.80–1.04; HR, 0.98; 95% CI, 0.89–1.09, respectively).

Conclusions

In a large cohort of men without a history of myocardial infarction, stroke, or venous thromboembolism, testosterone treatment was not associated with increased risk for incident composite cardiovascular events.

Stroke Risk Among Non-Elderly Users of Haloperidol or First-Generation Antipsychotics vs Second-Generation Antipsychotics: A Cohort Study from a US Health Insurance Claims Database

Background

Previous studies have reported an increased risk of stroke in patients taking antipsychotics. However, most of these studies have been conducted in the elderly population.

Objective

We estimated stroke risk in new users of any first-generation antipsychotic or haloperidol, vs second-generation antipsychotics among patients aged 18–64 years without a recent dementia diagnosis and, separately, regardless of a recent dementia diagnosis.

Methods

Data were obtained from IBM MarketScan^® Commercial Database (1 January, 2001–31 December, 2017). Among new users without a recent dementia diagnosis, stroke risk for first-generation antipsychotics (FGAw/oD cohort) or haloperidol (HALw/oD cohort) was compared with second-generation antipsychotics (SGAw/oD cohort). A similar comparison was conducted among new users regardless of dementia diagnosis: first-generation antipsychotics (FGA cohort) or haloperidol (HAL cohort) vs second-generation antipsychotics (SGA cohort). Crude incident stroke rates within each cohort were determined. For hazard ratios, three propensity score matching strategies were used: unadjusted (crude), Sentinel propensity score strategy, and large-scale regularized regression model (adapted propensity score strategy).

Results

Each cohort included ≥12,000 patients. The incident rates for stroke per 1000 person-years were 3.10 (FGAw/oD), 5.99 (HALw/oD), 0.85 (SGAw/oD), 3.14 (FGA), 6.12 (HAL), and 0.90 (SGA). Pre-planned analysis with adapted propensity score strategy matching yielded calibrated hazard ratios for stroke: FGAw/oD vs SGAw/oD: 2.05 (calibrated confidence interval 1.13–3.89); HALw/oD vs SGAw/oD: 2.47 (1.14–5.48), FGA vs SGA: 1.64 (0.94–2.97), and HAL vs SGA: 1.98 (0.99–4.00). A post-hoc sensitivity analysis to address potential bias introduced by the 2015 change from the International Classification of Diseases, Ninth Revision to the International Classification of Diseases, Tenth Revision yielded calibrated hazard ratios for FGAw/oD vs SGAw/oD: 1.59 (0.87–3.01), HALw/oD vs SGAw/oD: 2.79 (1.24–6.42), FGA vs SGA: 1.41 (0.79–2.62), and HAL vs SGA: 3.47 (1.63–7.92).

Conclusions

Among adults aged ≤64 years, without a recent dementia diagnosis, stroke risk is higher among those exposed to haloperidol compared with those exposed to second-generation antipsychotics.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40801-021-00267-2.

Risk of cardiovascular disease following gonadotropin-releasing hormone agonists vs antagonists in prostate cancer: Real-world evidence from five databases

Observational studies in prostate cancer (PCa) have shown an increased risk of cardiovascular disease (CVD) following gonadotropin-releasing hormone (GnRH) agonists, whereas randomised-controlled trials have shown no associations. Compared to GnRH agonists, GnRH antagonists have shown less atherosclerotic effects in preclinical models. We used real-world data from five countries to investigate CVD risk following GnRH agonists and antagonists in PCa men. Data sources included cancer registries, primary and secondary healthcare databases. CVD event was defined as an incident or fatal CVD. Multivariable Cox proportional hazard models estimated hazard ratios (HRs) and 95% confidence intervals (CIs), which were pooled using random-effects meta-analysis. Stratified analyses were conducted by history of CVD and age (75 years). A total of 48 757 men were on GnRH agonists and 2144 on GnRH antagonists. There was no difference in risk of any CVD for men on GnRH antagonists and agonists (HR: 1.25; 95% CI: 0.96-1.61; I2 : 64%). Men on GnRH antagonists showed increased risk of acute myocardial infarction (HR: 1.62; 95% CI: 1.11-2.35; I2 : 0%) and arrhythmia (HR: 1.55; 95% CI: 1.11-2.15, I2 : 17%) compared to GnRH agonists. Having a history of CVD was found to be an effect modifier for the associations with some CVD subtypes. Overall, we did not observe a difference in risk of overall CVD when comparing GnRH antagonists with agonists-though for some subtypes of CVD we noted an increased risk with antagonists. Further studies are required to address potential confounding caused by unadjusted variables such as severity of CVD history and PCa stage.

Spine Surgery in the Octogenarian Population: A Comparison of Demographics, Surgical Approach, and Healthcare Utilization With the PearlDiver Database

Background

With the recent advances in technology and healthcare, increasing numbers of individuals over the age of 80 will require surgical intervention for spinal pathology. Given the risk of increased complications in the elderly, a limited number of spinal surgeries are performed on octogenarians every year. This makes it difficult to generalize the trends and outcomes of these surgeries to a greater population. This study attempts to understand the trends in the safety profile and healthcare utilization across the United States for octogenarians undergoing spinal fusion and/or decompression surgery for spinal stenosis and/or degenerative disease using the PearlDiver database.

Methodology

Patients who underwent fusion and/or decompression for stenosis and/or degenerative diseases were extracted using International Classification of Disease ninth and tenth revisions (ICD-9 prior to October 2015, ICD-10 after) from 2007 to 2016 in the PearlDiver database. Three comparative groups were considered: (1) primary fusion without concurrent decompression, (2) primary decompression with concurrent fusion, and (3) fusion with concurrent decompression. Outcomes of interest were patient characteristics, demographics, length of stay, surgery hospitalization payments, and discharge disposition. These outcomes were compared to patients over the age of 20 who also underwent spinal surgery.

Results

A total of 9,715 patients who underwent spinal surgery were identified in the search. Of the 9,139 patients, 503 were octogenarians and 73 were nonagenarians. Octogenarians and nonagenarians diagnosed with spinal stenosis were more likely to undergo decompression alone rather than fusion or both fusion and decompression (21 for both fusion and decompression; p < 0.0001). Patients diagnosed with both spinal stenosis and degeneration were more likely to undergo both fusion and decompression than fusion or decompression alone (239 for both, 208 for decompression alone, and 23 for fusion alone; p < 0.0001). No statistically significant difference was found in the percentage of patients discharged home following either fusion or decompression or both surgeries (p = 0.0737). The mean length of stay for patients in the 20-79-year age group was 2.79 days, whereas for the octogenarian and nonagenarian cohort it was 3.85 days. The index hospitalization pay for patients in the 20-79-year age group was $19,220, whereas for the octogenarians and nonagenarians cohort it was $15,091.

Conclusions

Patients over the age of 80 were more likely to undergo either a fusion procedure or a decompression procedure alone rather than both unless they were diagnosed with spinal degeneration. The PearlDiver database analysis indicates that the length of stay for octogenarians and nonagenarians is longer than that for patients in the 20-79-year age group, and that younger patients are more likely to be discharged earlier than patients over the age of 80. Moreover, we observed that the index hospitalization pay was higher for patients over the age of 20 than for octogenarians and nonagenarians in all cases except for a fusion procedure.

Swiss Validation of the Enhanced Recovery After Surgery (ERAS) Database

Background

Enhanced recovery after surgery (ERAS) pathways have considerably improved postoperative outcomes and are in use for various types of surgery. The prospective audit system (EIAS) could be a powerful tool for large-scale outcome research but its database has not been validated yet.

Methods

Swiss ERAS centers were invited to contribute to the validation of the Swiss chapter for colorectal surgery. A monitoring team performed on-site visits by the use of a standardized checklist. Validation criteria were (I) coverage (No. of operated patients within ERAS protocol; target threshold for validation: ≥ 80%), (II) missing data (8 predefined variables; target ≤ 10%), and (III) accuracy (2 predefined variables, target ≥ 80%). These criteria were assessed by comparing EIAS entries with the medical charts of a random sample of patients per center (range 15–20).

Results

Out of 18 Swiss ERAS centers, 15 agreed to have onsite monitoring but 13 granted access to the final dataset. ERAS coverage was available in only 7 centers and varied between 76 and 100%. Overall missing data rate was 5.7% and concerned mainly the variables “urinary catheter removal” (16.4%) and “mobilization on day 1” (16%). Accuracy for the length of hospital stay and complications was overall 84.6%. Overall, 5 over 13 centers failed in the validation process for one or several criteria.

Conclusion

EIAS was validated in most Swiss ERAS centers. Potential patient selection and missing data remain sources of bias in non-validated centers. Therefore, simplified validation of other centers appears to be mandatory before large-scale use of the EIAS dataset.

Supplementary Information

The online version contains supplementary material available at (10.1007/s00268-020-05926-z).

Reoperation Rates Following Meniscus Transplantation Using the Truven Database

PURPOSE

The purpose of this study was to determine the (1) reoperation rate and (2) 30-day complication rate in a large insurance database.

METHODS

The Truven Database was queried for subjects that underwent meniscus allograft transplantation (Current Procedural Terminology code 29868) in the outpatient setting with minimal 2-year follow-up. Patients without confirmed laterality and patients that underwent concomitant ligament reconstruction were excluded. Reoperation was defined by ipsilateral knee procedure after the index surgery. The 30-day postoperative complication rates were assessed using International Classification of Diseases, 9th Revision, Clinical Modification codes.

RESULTS

A total of 284 patients (mean age of 26.2 ± 10.4 years; 49.6% females) were included in this study with mean follow up of 43.2 ± 19.2 months. One hundred and sixty-seven subjects (58.8%) undergoing meniscus allograft transplantation underwent reoperation at an average of 11.9 ± 12.2 months postoperatively. There was a low number of subjects that required ipsilateral unicompartmental knee arthroplasty and total knee arthroplasty postoperatively (0.7% and 1.1%, respectively). The overall 30-day complication rate following meniscus allograft transplantation was 1.4%.

CONCLUSIONS

Patients undergoing meniscus allograft transplantation have a 58.8% reoperation rate at final follow up with low (1.4%) 30-day complication rates in a large insurance database.

LEVEL OF EVIDENCE

Level IV, case series.

The necessity and risk factors of subsequent fusion after decompression alone for lumbar spinal stenosis with lumbar spondylolisthesis: 5 years follow-up in two different large populations

BACKGROUND/CONTEXT

Although decompression without fusion is a reasonable surgical treatment option for some patients with lumbar spinal stenosis (LSS) secondary to spondylolisthesis, some of these patients will require secondary surgery for subsequent fusion. Long-term outcome and need for subsequent fusion in patients treated with decompression alone in the setting of lumbar spondylolisthesis remains controversial.

PURPOSE

The aim of this study was to examine the rate, timing, and risk factors of subsequent fusion for patients after decompression alone for LSS with spondylolisthesis.

STUDY DESIGN/SETTING

A retrospective cohort study.

PATIENT SAMPLE

Patients who had LSS with spondylolisthesis and underwent decompression alone at 1 or 2 levels as a primary lumbar surgery with more than 5 year follow-up.

OUTCOME MEASURES

The rate, timing, and risk factors for subsequent fusion.

METHODS

Subjects were extracted from both public and private insurance resources in a nationwide insurer database. Risk factors for subsequent fusion were evaluated by multivariate cox proportion-hazard regression controlling for age, gender, comorbidities and the presence or absence of claudication.

RESULTS

Five thousand eight hundred and seventy-five patients in the public insurance population (PI population) and 1,456 patients in the private insurance population (PrI population) were included in this study. The rates of patients who needed subsequent fusion were 1.9% at 1 year, 3.5% at 2 years, and 6.7% at 5 years in the PI population, whereas they were 4.3% at 1 year, 8.9% at 2 years, 14.6% at 5 years in the PrI population. The time to subsequent fusion was 730 (365-1234) days in the PI population and 588 (300-998) days in the PrI population. Age less than 70 years, presence of neurogenic claudication and rheumatoid arthritis (RA)/collagen vascular diseases (CVD) were independent risk factors for subsequent fusion in both populations.

CONCLUSIONS

Decompression surgery alone can demonstrate good outcomes in some patients with LSS with spondylolisthesis. It is important for surgeons to recognize, however, that patient age less than 70 years, symptomatic neurogenic claudication, and presence of RA and/or CVD are significant independent factors associated with greater likelihood of needing secondary fusion surgery.

The Weight of Evidence From Electrophysiology, Observational, and Cardiovascular End Point Studies Demonstrates the Safety of Azithromycin

Increased use of azithromycin (AZ) in treating infections associated with coronavirus disease 2019 (COVID-19) and reports of increased incidence of prolonged corrected QT (QTc) interval associated with AZ used with hydroxychloroquine prompted us to review the latest evidence in the literature, present additional analyses of human cardiovascular (CV) electrophysiology studies, and to describe sequential steps in research and development that were undertaken to characterize the benefit-risk profile of AZ. Combined QTc findings from electrocardiograms taken during oral and i.v. pharmacokinetic-pharmacodynamic studies of AZ suggest that clinically meaningful QTc prolongation is unlikely. Findings from several observational studies were heterogeneous and not as consistent as results from at least two large randomized controlled trials (RCTs). The QTc findings presented and observational data from studies with large numbers of events are not consistent with either a proarrhythmic action of AZ or an increase in frequency of CV deaths. Well-powered RCTs do not suggest a presence of increased risk of CV or sudden cardiac death after short-term or protracted periods of AZ usage, even in patients at higher risk from pre-existing coronary disease.

Comparative safety and effectiveness of alendronate versus raloxifene in women with osteoporosis

Alendronate and raloxifene are among the most popular anti-osteoporosis medications. However, there is a lack of head-to-head comparative effectiveness studies comparing the two treatments. We conducted a retrospective large-scale multicenter study encompassing over 300 million patients across nine databases encoded in the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM). The primary outcome was the incidence of osteoporotic hip fracture, while secondary outcomes were vertebral fracture, atypical femoral fracture (AFF), osteonecrosis of the jaw (ONJ), and esophageal cancer. We used propensity score trimming and stratification based on an expansive propensity score model with all pre-treatment patient characteritistcs. We accounted for unmeasured confounding using negative control outcomes to estimate and adjust for residual systematic bias in each data source. We identified 283,586 alendronate patients and 40,463 raloxifene patients. There were 7.48 hip fracture, 8.18 vertebral fracture, 1.14 AFF, 0.21 esophageal cancer and 0.09 ONJ events per 1,000 person-years in the alendronate cohort and 6.62, 7.36, 0.69, 0.22 and 0.06 events per 1,000 person-years, respectively, in the raloxifene cohort. Alendronate and raloxifene have a similar hip fracture risk (hazard ratio [HR] 1.03, 95% confidence interval [CI] 0.94–1.13), but alendronate users are more likely to have vertebral fractures (HR 1.07, 95% CI 1.01–1.14). Alendronate has higher risk for AFF (HR 1.51, 95% CI 1.23–1.84) but similar risk for esophageal cancer (HR 0.95, 95% CI 0.53–1.70), and ONJ (HR 1.62, 95% CI 0.78–3.34). We demonstrated substantial control of measured confounding by propensity score adjustment, and minimal residual systematic bias through negative control experiments, lending credibility to our effect estimates. Raloxifene is as effective as alendronate and may remain an option in the prevention of osteoporotic fracture.

Risk of sudden cardiac arrest and ventricular arrhythmia with sulfonylureas: An experience with conceptual replication in two independent populations

Sulfonylureas are commonly used to treat type 2 diabetes mellitus. Despite awareness of their effects on cardiac physiology, a knowledge gap exists regarding their effects on cardiovascular events in real-world populations. Prior studies reported sulfonylurea-associated cardiovascular death but not serious arrhythmogenic endpoints like sudden cardiac arrest (SCA) or ventricular arrhythmia (VA). We assessed the comparative real-world risk of SCA/VA among users of second-generation sulfonylureas: glimepiride, glyburide, and glipizide. We conducted two incident user cohort studies using five-state Medicaid claims (1999–2012) and Optum Clinformatics commercial claims (2000–2016). Outcomes were SCA/VA events precipitating hospital presentation. We used Cox proportional hazards models, adjusted for high-dimensional propensity scores, to generate adjusted hazard ratios (aHR). We identified 624,406 and 491,940 sulfonylurea users, and 714 and 385 SCA/VA events, in Medicaid and Optum, respectively. Dataset-specific associations with SCA/VA for both glimepiride and glyburide (vs. glipizide) were on opposite sides of and could not exclude the null (glimepiride: aHR_Medicaid 1.17, 95% CI 0.96–1.42; aHR_Optum 0.84, 0.65–1.08; glyburide: aHR_Medicaid 0.87, 0.74–1.03; aHR_Optum 1.11, 0.86–1.42). Database differences in data availability, populations, and documentation completeness may have contributed to the incongruous results. Emphasis should be placed on assessing potential causes of discrepancies between conflicting studies evaluating the same research question.

Understanding the challenge of comparative effectiveness research in focal epilepsy: A review of network meta-analyses and real-world evidence on antiepileptic drugs

OBJECTIVE

Head-to-head randomized controlled trials (RCTs) are the gold standard for assessing comparative treatment effects. In the absence of direct comparisons between all possible antiepileptic drugs (AEDs), however, clinical decision-making in focal (partial onset) epilepsy relies on alternative evidence borne from indirect comparisons including network meta-analyses (NMAs) and from real-world evidence (RWE) studies. We review NMAs and observational RWE studies comparing AEDs in the adjunctive setting to compare the robustness of these methods and to formulate recommendations for future evidence development.

METHODS

A literature review identified NMAs and RWE studies comparing AEDs for the adjunctive treatment of focal seizures published between January 2008 and October 2018. NMAs were evaluated for robustness using a framework based on guidelines from the National Institute for Health and Care Excellence Decision Support Unit and the International Society for Pharmacoeconomics and Outcomes Research. RWE studies were evaluated using the GRACE checklist.

RESULTS

From a total of 1993 records, 11 NMAs and six RWE studies were eligible. Key limitations identified in the NMAs include nonsystematic selection of RCTs, unexplored heterogeneity between included RCTs in terms of study and patient characteristics, and selection of AEDs and AED doses or dosing strategies that are not reflective of clinical practice. The main limitations of RWE studies concern sample size, design, and analysis methods. Approximately 90% of comparisons between individual AEDs were nonsignificant in the NMAs. None of the RWE studies adjusted for baseline differences between comparator groups; therefore, they lack the validity to make comparative conclusions.

SIGNIFICANCE

Current NMAs and RWE studies provide only nominal comparative evidence for AED treatments in focal epilepsy, and should be used with caution for decision-making due to their methodological limitations. To overcome these hurdles, adherence to methodological guidelines and concerted efforts to collect relevant outcome data in the real world are needed.

On the Convergence of Epidemiology, Biostatistics, and Data Science

Epidemiology, biostatistics, and data science are broad disciplines that incorporate a variety of substantive areas. Common among them is a focus on quantitative approaches for solving intricate problems. When the substantive area is health and health care, the overlap is further cemented. Researchers in these disciplines are fluent in statistics, data management and analysis, and health and medicine, to name but a few competencies. Yet there are important and perhaps mutually exclusive attributes of these fields that warrant a tighter integration. For example, epidemiologists receive substantial training in the science of study design, measurement, and the art of causal inference. Biostatisticians are well versed in the theory and application of methodological techniques, as well as the design and conduct of public health research. Data scientists receive equivalently rigorous training in computational and visualization approaches for high-dimensional data. Compared to data scientists, epidemiologists and biostatisticians may have less expertise in computer science and informatics, while data scientists may benefit from a working knowledge of study design and causal inference. Collaboration and cross-training offer the opportunity to share and learn of the constructs, frameworks, theories, and methods of these fields with the goal of offering fresh and innovate perspectives for tackling challenging problems in health and health care. In this article, we first describe the evolution of these fields focusing on their convergence in the era of electronic health data, notably electronic medical records (EMRs). Next we present how a collaborative team may design, analyze, and implement an EMR-based study. Finally, we review the curricula at leading epidemiology, biostatistics, and data science training programs, identifying gaps and offering suggestions for the fields moving forward.

Assessment of the impact of EHR heterogeneity for clinical research through a case study of silent brain infarction

Background

The rapid adoption of electronic health records (EHRs) holds great promise for advancing medicine through practice-based knowledge discovery. However, the validity of EHR-based clinical research is questionable due to poor research reproducibility caused by the heterogeneity and complexity of healthcare institutions and EHR systems, the cross-disciplinary nature of the research team, and the lack of standard processes and best practices for conducting EHR-based clinical research.

Method

We developed a data abstraction framework to standardize the process for multi-site EHR-based clinical studies aiming to enhance research reproducibility. The framework was implemented for a multi-site EHR-based research project, the ESPRESSO project, with the goal to identify individuals with silent brain infarctions (SBI) at Tufts Medical Center (TMC) and Mayo Clinic. The heterogeneity of healthcare institutions, EHR systems, documentation, and process variation in case identification was assessed quantitatively and qualitatively.

Result

We discovered a significant variation in the patient populations, neuroimaging reporting, EHR systems, and abstraction processes across the two sites. The prevalence of SBI for patients over age 50 for TMC and Mayo is 7.4 and 12.5% respectively. There is a variation regarding neuroimaging reporting where TMC are lengthy, standardized and descriptive while Mayo’s reports are short and definitive with more textual variations. Furthermore, differences in the EHR system, technology infrastructure, and data collection process were identified.

Conclusion

The implementation of the framework identified the institutional and process variations and the heterogeneity of EHRs across the sites participating in the case study. The experiment demonstrates the necessity to have a standardized process for data abstraction when conducting EHR-based clinical studies.

Predicting Adverse Drug Reactions on Distributed Health Data using Federated Learning

Using electronic health data to predict adverse drug reaction (ADR) incurs practical challenges, such as lack of adequate data from any single site for rare ADR detection, resource constraints on integrating data from multiple sources, and privacy concerns with creating a centralized database from person-specific, sensitive data. We introduce a federated learning framework that can learn a global ADR prediction model from distributed health data held locally at different sites. We propose two novel methods of local model aggregation to improve the predictive capability of the global model. Through comprehensive experimental evaluation using real-world health data from 1 million patients, we demonstrate the effectiveness of our proposed approach in achieving comparable performance to centralized learning and outperforming localized learning models for two types of ADRs. We also demonstrate that, for varying data distributions, our aggregation methods outperform state-of-the-art techniques, in terms of precision, recall, and accuracy.

Comparative studies on ion-pair energetic, distribution among three domains of life: Archaea, eubacteria, and eukarya

Salt-bridges play a unique role in the structural and functional stability of proteins, especially under harsh environments. How these salt-bridges contribute to the overall thermodynamic stability of protein structure and function across different domains of life is elusive still date. To address the issue, statistical analyses on the energies of salt-bridges, involved in proteins' structure and function, are performed across three domains of life, that is, archaea, eubacteria, and eukarya. Results show that although the majority of salt-bridges are stable and conserved, yet the stability of archaeal proteins (∆∆G_net = -5.06 ± 3.8) is much more than that of eubacteria (∆∆G_net = -3.7 ± 2.9) and eukarya (∆∆G_net = -3.54 ± 3.1). Unlike earlier study with archaea, in eukarya and eubacteria, not all buried salt-bridge in our dataset are stable. Buried salt-bridges play surprising role in protein stability, whose variations are clearly observed among these domains. Greater desolvation penalty of buried salt-bridges is compensated by stable network of salt-bridges apart from equal contribution of bridge and background energy terms. On the basis proteins' secondary structure, topology, and evolution, our observation shows that salt-bridges when present closer to each other in sequence tend to form a greater number. Overall, our comparative study provides insight into the role of specific electrostatic interactions in proteins from different domains of life, which we hope, would be useful for protein engineering and bioinformatics study.

Analytic and Data Sharing Options in Real-World Multidatabase Studies of Comparative Effectiveness and Safety of Medical Products

A wide range of analytic and data sharing options are available in nonexperimental multidatabase studies designed to assess the real-world benefits and risks of medical products. Researchers often consider six scientific domains when choosing among these options-study design, exposure type, outcome type, covariate summarization technique, covariate adjustment method, and data sharing approach. This article reviews available analytic and data sharing options and discusses key scientific and practical considerations when choosing among these options in multidatabase studies of comparative effectiveness and safety of medical products. The scientific considerations must be balanced against what the data-contributing sites are able or willing to share. While pooling of person-level data sets remains the most familiar and analytically flexible approach, newer analytic and data sharing approaches that share less granular summary-level information may be equally valid and preferred in some multidatabase studies, especially when sharing of person-level data is challenging or infeasible.

Impact of pre-exposure time bias in self-controlled case series when the event conditions the exposure: Hip/femur fracture and use of benzodiazepines as a case study

BACKGROUND

In self-controlled case series (SCCS), the event should not condition the probability of subsequent exposure. If this assumption is not met, an important bias could take place. The association of hip/femur fracture (HFF) and use of benzodiazepines (BDZ) has a bidirectional causal relationship and can serve as case study to investigate the impact of this methodological issue.

OBJECTIVES

To assess the magnitude of bias introduced in a SCCS when HFF conditions the posterior exposure to BDZ and explore ways to correct it.

METHODS

Four thousand four hundred fifty cases of HFF who had at least one BZD prescription were selected from the primary care health record database BIFAP. Exposure to BZD was divided into non-use, current, recent, and past use. Conditional Poisson regression was used to estimate incidence rate ratios (IRRs) of HFF among current vs non-use/past, adjusted for age. To investigate possible event-exposure dependence, a pre-exposure time of different lengths (15, 30, and 60 days) was excluded from the reference category to evaluate the IRR.

RESULTS

IRR of HHF for current use was 0.79 (0.72-0.86); removing 30 days, IRR was 1.43 (1.31-1.57). Removing 15 days, IRR was 1.29 (1.18-1.41), and removing 60 days, IRR was 1.56 (1.42-1.72). A pre-exposure period up to 182 days was necessary to remove such effect giving an IRR of 1.64 (1.48-1.81).

CONCLUSIONS

HFF remarkably conditioned the use of BDZs resulting in seriously biased IRRs when this association was studied through a SCCS design. The use of pre-exposure periods of different lengths helped to correct this error.

Comprehensive Real-World Assessment of Marketed Medications to Guide Parkinson's Drug Discovery

BACKGROUND

Parkinson's disease is a disorder growing in prevalence, disability, and deaths. Healthcare databases provide a 'real-world' perspective for millions of individuals. We envisioned helping accelerate drug discovery by using these databases.

OBJECTIVES

The objectives of this study were to assess the association of marketed medications with the risk of parkinsonism in four US claims databases and to evaluate the consistency of the association of β-adrenoreceptor modulation with parkinsonism.

METHODS

The study was conducted using a self-controlled cohort design in which subjects served as their own control. The time from treatment initiation until discontinuation or end of observation was the exposed period and a similar time preceding medication was the unexposed period. Medications were studied at ingredient and class level. The incidence rate ratio (IRR) and combined IRR were calculated.

RESULTS

We assessed 2181 drugs and 117,015,066 people. Diphenhydramine, isradipine, methylphenidate, armodafinil, and modafinil were associated with reduced risk for parkinsonism in at least two databases. Armodafinil, modafinil, methylphenidate, and the β-agonist albuterol were associated with a 56%, 54%, 39%, and 17% reduction in the risk of having parkinsonism, respectively. Isradipine results were heterogeneous and no significant association was found. Propranolol was associated with a 32% increased risk, the only β-adrenoceptor antagonist (β-blocker) associated with an increased risk.

CONCLUSIONS

Armodafinil, modafinil, and methylphenidate were associated with a decreased risk of parkinsonism, as were β-agonists. Of the β-blockers, only propranolol was associated with increased risk. Healthcare database analyses that incorporate scientific rigor provide insight and direction for drug discovery efforts. These findings show association not causality; however, they offer considerable support to the association between β-adrenergic receptor modulation and risk of Parkinson's disease.

Does health informatics have a replication crisis?

Objective

Many research fields, including psychology and basic medical sciences, struggle with poor reproducibility of reported studies. Biomedical and health informatics is unlikely to be immune to these challenges. This paper explores replication in informatics and the unique challenges the discipline faces.

Methods

Narrative review of recent literature on research replication challenges.

Results

While there is growing interest in re-analysis of existing data, experimental replication studies appear uncommon in informatics. Context effects are a particular challenge as they make ensuring replication fidelity difficult, and the same intervention will never quite reproduce the same result in different settings. Replication studies take many forms, trading-off testing validity of past findings against testing generalizability. Exact and partial replication designs emphasize testing validity while quasi and conceptual studies test generalizability of an underlying model or hypothesis with different methods or in a different setting.

Conclusions

The cost of poor replication is a weakening in the quality of published research and the evidence-based foundation of health informatics. The benefits of replication include increased rigor in research, and the development of evaluation methods that distinguish the impact of context and the nonreproducibility of research. Taking replication seriously is essential if biomedical and health informatics is to be an evidence-based discipline.

High-Risk Phenotypes of Early Psychiatric Readmission in Bipolar Disorder With Comorbid Medical Illness

BACKGROUND

Individuals with co-existing serious mental illness and non-psychiatric medical illness are at high risk of acute care utilization. Mining of electronic health record data can help identify and categorize predictors of psychiatric hospital readmission in this population.

OBJECTIVE

This study aimed to identify modifiable predictors of psychiatric readmission among individuals with comorbid bipolar disorder and medical illness. This goal was accomplished by applying objective variable selection via machine learning techniques.

METHOD

This was a retrospective analysis of electronic health record data derived from 77,296 episodes of care from 2006 to 2016 within the University of California Health Care System. Data included 1,250 episodes of care involving patients with bipolar disorder and serious comorbid medical illnesses (defined by transfer between medicine and psychiatry services or concomitant primary medical and psychiatric admission diagnoses). Machine learning (classification trees) was used to identify potential predictors of 30-day psychiatric readmission across hospital encounters. Predictors included demographics, medical and psychiatric diagnoses, medication regimen, and disposition. The algorithm was internally validated using 10-fold cross-validation.

RESULTS

The model predicted 30-day readmission with high accuracy (98% unbalanced model, 88% balanced model). Modifiable predictors of readmission were length of stay, transfers between medical and psychiatric services, discharge disposition to home, and all-cause acute health service utilization in the year before the index hospitalization.

CONCLUSION

Among bipolar disorder patients with comorbid medical conditions, characteristics of the index hospitalization (e.g., duration, transfer, and disposition) emerged as more predictive than static properties of the patient (e.g., sociodemographic factors and psychiatric comorbidity burden). Findings identified phenotypes of patients at high risk for rehospitalization and suggest potential ways of modifying the risk of early readmission.