Diagnostic Algorithms for Cardiovascular Death in Administrative Claims Databases: A Systematic Review

Comparative Effectiveness of Second-Line Antihyperglycemic Agents for Cardiovascular Outcomes: A Multinational, Federated Analysis of LEGEND-T2DM

BACKGROUND

Sodium-glucose cotransporter 2 inhibitors (SGLT2is) and glucagon-like peptide-1 receptor agonists (GLP-1 RAs) reduce the risk of major adverse cardiovascular events (MACE) in patients with type 2 diabetes mellitus (T2DM). However, their effectiveness relative to each other and other second-line antihyperglycemic agents is unknown, without any major ongoing head-to-head clinical trials.

OBJECTIVES

The aim of this study was to compare the cardiovascular effectiveness of SGLT2is, GLP-1 RAs, dipeptidyl peptidase-4 inhibitors (DPP4is), and clinical sulfonylureas (SUs) as second-line antihyperglycemic agents in T2DM.

METHODS

Across the LEGEND-T2DM (Large-Scale Evidence Generation and Evaluation Across a Network of Databases for Type 2 Diabetes Mellitus) network, 10 federated international data sources were included, spanning 1992 to 2021. In total, 1,492,855 patients with T2DM and cardiovascular disease (CVD) on metformin monotherapy were identified who initiated 1 of 4 second-line agents (SGLT2is, GLP-1 RAs, DPP4is, or SUs). Large-scale propensity score models were used to conduct an active-comparator target trial emulation for pairwise comparisons. After evaluating empirical equipoise and population generalizability, on-treatment Cox proportional hazards models were fit for 3-point MACE (myocardial infarction, stroke, and death) and 4-point MACE (3-point MACE plus heart failure hospitalization) risk and HR estimates were combined using random-effects meta-analysis.

RESULTS

Over 5.2 million patient-years of follow-up and 489 million patient-days of time at risk, patients experienced 25,982 3-point MACE and 41,447 4-point MACE. SGLT2is and GLP-1 RAs were associated with lower 3-point MACE risk than DPP4is (HR: 0.89 [95% CI: 0.79-1.00] and 0.83 [95% CI: 0.70-0.98]) and SUs (HR: 0.76 [95% CI: 0.65-0.89] and 0.72 [95% CI: 0.58-0.88]). DPP4is were associated with lower 3-point MACE risk than SUs (HR: 0.87; 95% CI: 0.79-0.95). The pattern for 3-point MACE was also observed for the 4-point MACE outcome. There were no significant differences between SGLT2is and GLP-1 RAs for 3-point or 4-point MACE (HR: 1.06 [95% CI: 0.96-1.17] and 1.05 [95% CI: 0.97-1.13]).

CONCLUSIONS

In patients with T2DM and CVD, comparable cardiovascular risk reduction was found with SGLT2is and GLP-1 RAs, with both agents more effective than DPP4is, which in turn were more effective than SUs. These findings suggest that the use of SGLT2is and GLP-1 RAs should be prioritized as second-line agents in those with established CVD.

Comparative safety of tramadol and other opioids following total hip and knee arthroplasty

BACKGROUND

Tramadol is increasingly used to treat acute postoperative pain among older adults following total hip and knee arthroplasty (THA/TKA). However, tramadol has a complex pharmacology and may be no safer than full opioid agonists. We compared the safety of tramadol, oxycodone, and hydrocodone among opioid-naïve older adults following elective THA/TKA.

METHODS

This retrospective cohort included Medicare Fee-for-Service beneficiaries ≥ 65 years with elective THA/TKA between January 1, 2010 and September 30, 2015, 12 months of continuous Parts A and B enrollment, 6 months of continuous Part D enrollment, and no opioid use in the 6 months prior to THA/TKA. Participants initiated single-opioid therapy with tramadol, oxycodone, or hydrocodone within 7 days of discharge from THA/TKA hospitalization, regardless of concurrently administered nonopioid analgesics. Outcomes of interest included all-cause hospitalizations or emergency department visits (serious adverse events (SAEs)) and a composite of 10 surgical- and opioid-related SAEs within 90-days of THA/TKA. The intention-to-treat (ITT) and per-protocol (PP) hazard ratios (HRs) for tramadol versus other opioids were estimated using inverse-probability-of-treatment-weighted pooled logistic regression models.

RESULTS

The study population included 2,697 tramadol, 11,407 oxycodone, and 14,665 hydrocodone initiators. Compared to oxycodone, tramadol increased the rate of all-cause SAEs in ITT analyses only (ITT HR 1.19, 95%CLs, 1.02, 1.41; PP HR 1.05, 95%CLs, 0.86, 1.29). Rates of composite SAEs were not significant across comparisons. Compared to hydrocodone, tramadol increased the rate of all-cause SAEs in the ITT and PP analyses (ITT HR 1.40, 95%CLs, 1.10, 1.76; PP HR 1.34, 95%CLs, 1.03, 1.75), but rates of composite SAEs were not significant across comparisons.

CONCLUSIONS

Postoperative tramadol was associated with increased rates of all-cause SAEs, but not composite SAEs, compared to oxycodone and hydrocodone. Tramadol does not appear to have a superior safety profile and should not be preferentially prescribed to opioid-naïve older adults following THA/TKA.

Comparative Effectiveness of Second-line Antihyperglycemic Agents for Cardiovascular Outcomes: A Large-scale, Multinational, Federated Analysis of the LEGEND-T2DM Study

Background

SGLT2 inhibitors (SGLT2is) and GLP-1 receptor agonists (GLP1-RAs) reduce major adverse cardiovascular events (MACE) in patients with type 2 diabetes mellitus (T2DM). However, their effectiveness relative to each other and other second-line antihyperglycemic agents is unknown, without any major ongoing head-to-head trials.

Methods

Across the LEGEND-T2DM network, we included ten federated international data sources, spanning 1992-2021. We identified 1,492,855 patients with T2DM and established cardiovascular disease (CVD) on metformin monotherapy who initiated one of four second-line agents (SGLT2is, GLP1-RAs, dipeptidyl peptidase 4 inhibitor [DPP4is], sulfonylureas [SUs]). We used large-scale propensity score models to conduct an active comparator, target trial emulation for pairwise comparisons. After evaluating empirical equipoise and population generalizability, we fit on-treatment Cox proportional hazard models for 3-point MACE (myocardial infarction, stroke, death) and 4-point MACE (3-point MACE + heart failure hospitalization) risk, and combined hazard ratio (HR) estimates in a random-effects meta-analysis.

Findings

Across cohorts, 16·4%, 8·3%, 27·7%, and 47·6% of individuals with T2DM initiated SGLT2is, GLP1-RAs, DPP4is, and SUs, respectively. Over 5·2 million patient-years of follow-up and 489 million patient-days of time at-risk, there were 25,982 3-point MACE and 41,447 4-point MACE events. SGLT2is and GLP1-RAs were associated with a lower risk for 3-point MACE compared with DPP4is (HR 0·89 [95% CI, 0·79-1·00] and 0·83 [0·70-0·98]), and SUs (HR 0·76 [0·65-0·89] and 0·71 [0·59-0·86]). DPP4is were associated with a lower 3-point MACE risk versus SUs (HR 0·87 [0·79-0·95]). The pattern was consistent for 4-point MACE for the comparisons above. There were no significant differences between SGLT2is and GLP1-RAs for 3-point or 4-point MACE (HR 1·06 [0·96-1·17] and 1·05 [0·97-1·13]).

Interpretation

In patients with T2DM and established CVD, we found comparable cardiovascular risk reduction with SGLT2is and GLP1-RAs, with both agents more effective than DPP4is, which in turn were more effective than SUs. These findings suggest that the use of GLP1-RAs and SGLT2is should be prioritized as second-line agents in those with established CVD.

Funding

National Institutes of Health, United States Department of Veterans Affairs.

Electrocardiographic deep learning for predicting post-procedural mortality: a model development and validation study

BACKGROUND

Preoperative risk assessments used in clinical practice are insufficient in their ability to identify risk for postoperative mortality. Deep-learning analysis of electrocardiography can identify hidden risk markers that can help to prognosticate postoperative mortality. We aimed to develop a prognostic model that accurately predicts postoperative mortality in patients undergoing medical procedures and who had received preoperative electrocardiographic diagnostic testing.

METHODS

In a derivation cohort of preoperative patients with available electrocardiograms (ECGs) from Cedars-Sinai Medical Center (Los Angeles, CA, USA) between Jan 1, 2015 and Dec 31, 2019, a deep-learning algorithm was developed to leverage waveform signals to discriminate postoperative mortality. We randomly split patients (8:1:1) into subsets for training, internal validation, and final algorithm test analyses. Model performance was assessed using area under the receiver operating characteristic curve (AUC) values in the hold-out test dataset and in two external hospital cohorts and compared with the established Revised Cardiac Risk Index (RCRI) score. The primary outcome was post-procedural mortality across three health-care systems.

FINDINGS

45 969 patients had a complete ECG waveform image available for at least one 12-lead ECG performed within the 30 days before the procedure date (59 975 inpatient procedures and 112 794 ECGs): 36 839 patients in the training dataset, 4549 in the internal validation dataset, and 4581 in the internal test dataset. In the held-out internal test cohort, the algorithm discriminates mortality with an AUC value of 0·83 (95% CI 0·79-0·87), surpassing the discrimination of the RCRI score with an AUC of 0·67 (0·61-0·72). The algorithm similarly discriminated risk for mortality in two independent US health-care systems, with AUCs of 0·79 (0·75-0·83) and 0·75 (0·74-0·76), respectively. Patients determined to be high risk by the deep-learning model had an unadjusted odds ratio (OR) of 8·83 (5·57-13·20) for postoperative mortality compared with an unadjusted OR of 2·08 (0·77-3·50) for postoperative mortality for RCRI scores of more than 2. The deep-learning algorithm performed similarly for patients undergoing cardiac surgery (AUC 0·85 [0·77-0·92]), non-cardiac surgery (AUC 0·83 [0·79-0·88]), and catheterisation or endoscopy suite procedures (AUC 0·76 [0·72-0·81]).

INTERPRETATION

A deep-learning algorithm interpreting preoperative ECGs can improve discrimination of postoperative mortality. The deep-learning algorithm worked equally well for risk stratification of cardiac surgeries, non-cardiac surgeries, and catheterisation laboratory procedures, and was validated in three independent health-care systems. This algorithm can provide additional information to clinicians making the decision to perform medical procedures and stratify the risk of future complications.

FUNDING

National Heart, Lung, and Blood Institute.

Validation of Algorithms to Identify Acute Myocardial Infarction, Stroke, and Cardiovascular Death in German Health Insurance Data

Purpose

Validation of outcomes allows measurement of and correction for potential misclassification and targeted adjustment of algorithms for case definition. The purpose of our study was to validate algorithms for identifying cases of acute myocardial infarction (AMI), stroke, and cardiovascular (CV) death using patient profiles, ie, chronological tabular summaries of relevant available information on a patient, extracted from pseudonymized German claims data.

Patients and Methods

Based on the German Pharmacoepidemiological Research Database (GePaRD), 250 cases were randomly selected (50% males) for each outcome between 2016 and 2017 based on the inclusion criteria age ≥50 years and continuous insurance ≥1 year and applying the following algorithms: hospitalization with a main diagnosis of AMI (ICD-10-GM codes I21.- and I22.-) or stroke (I63, I61, I64) or death with a hospitalization in the 60 days before with a main diagnosis of CV disease. Patient profiles were built including (i) age and sex, (ii) hospitalizations incl. diagnoses, procedures, discharge reasons, (iii) outpatient diagnoses incl. diagnostic certainty, physician specialty, (iv) outpatient encounters, and (v) outpatient dispensings. Using adjudication criteria based on clinical guidelines and risk factors, two trained physicians independently classified cases as “certain”, “probable”, “unlikely” or “not assessable”. Positive predictive values (PPVs) were calculated as percentage of confirmed cases among all assessable cases.

Results

For AMI, the overall PPV was 97.6% [95% confidence interval 94.8–99.1]. The PPV for any stroke was 94.8% [91.3–97.2] and higher for ischemic (98.3% [95.0–99.6]) than for hemorrhagic stroke (86.5% [76.5–93.3]). The PPV for CV death was 79.9% [74.4–84.4]. It increased to 91.7% [87.2–95.0] after excluding 32 cases with data insufficient for a decision.

Conclusion

Algorithms based on hospital diagnoses can identify AMI, stroke, and CV death from German claims data with high PPV. This was the first study to show that German claims data contain information suitable for outcome validation.

Association of Glucagon-Like Peptide-1 Receptor Agonist vs Dipeptidyl Peptidase-4 Inhibitor Use With Mortality Among Patients With Type 2 Diabetes and Advanced Chronic Kidney Disease

IMPORTANCE

Glucagon-like peptide-1 (GLP-1) receptor agonist use is associated with reduced mortality and improved cardiovascular outcomes in the general population with diabetes. Dipeptidyl peptidase-4 (DPP-4) inhibitors are commonly used antidiabetic agents for patients with advanced-stage chronic kidney disease (CKD). The association of these 2 drug classes with outcomes among patients with diabetes and advanced-stage CKD or end-stage kidney disease (ESKD) is not well understood.

OBJECTIVE

To assess whether use of GLP-1 receptor agonists in a population with diabetes and advanced-stage CKD or ESKD is associated with better outcomes compared with use of DPP-4 inhibitors.

DESIGN, SETTING, AND PARTICIPANTS

This retrospective cohort study used data on patients with type 2 diabetes and stage 5 CKD or ESKD obtained from the National Health Insurance Research Database of Taiwan. The study was conducted between January 1, 2012, and December 31, 2018. Data were analyzed from June 2020 to July 2021.

EXPOSURES

Treatment with GLP-1 receptor agonists compared with treatment with DPP-4 inhibitors.

MAIN OUTCOMES AND MEASURES

All-cause mortality, sepsis- and infection-related mortality, and mortality related to major adverse cardiovascular and cerebrovascular events were compared between patients treated with GLP-1 receptor agonists and patients treated with DPP-4 inhibitors. Propensity score weighting was used to mitigate the imbalance among covariates between the groups.

RESULTS

Of 27 279 patients included in the study, 26 578 were in the DPP-4 inhibitor group (14 443 [54.34%] male; mean [SD] age, 65 [13] years) and 701 in the GLP-1 receptor agonist group (346 [49.36%] male; mean [SD] age, 59 [13] years). After weighting, the use of GLP-1 receptor agonists was associated with lower all-cause mortality (hazard ratio [HR], 0.79; 95% CI, 0.63-0.98) and lower sepsis- and infection-related mortality (HR, 0.61; 95% CI, 0.40-0.91). Subgroup analysis demonstrated a lower risk of mortality associated with use of GLP-1 receptor agonists compared with DDP-4 inhibitors among patients with cerebrovascular disease (HR, 0.33; 95% CI, 0.12-0.86) than among those without cerebrovascular disease (HR, 0.89; 95% CI, 0.71-1.12) (P = .04 for interaction).

CONCLUSIONS AND RELEVANCE

Treatment with GLP-1 receptor agonists was associated with lower all-cause mortality among patients with type 2 diabetes, advanced-stage CKD, and ESKD than was treatment with DPP-4 inhibitors. Additional well-designed, prospective studies are needed to confirm the potential benefit of GLP-1 receptor agonist treatment for patients with advanced CKD or ESKD.

Data Consult Service: Can we use observational data to address immediate clinical needs?

OBJECTIVE

A number of clinical decision support tools aim to use observational data to address immediate clinical needs, but few of them address challenges and biases inherent in such data. The goal of this article is to describe the experience of running a data consult service that generates clinical evidence in real time and characterize the challenges related to its use of observational data.

MATERIALS AND METHODS

In 2019, we launched the Data Consult Service pilot with clinicians affiliated with Columbia University Irving Medical Center. We created and implemented a pipeline (question gathering, data exploration, iterative patient phenotyping, study execution, and assessing validity of results) for generating new evidence in real time. We collected user feedback and assessed issues related to producing reliable evidence.

RESULTS

We collected 29 questions from 22 clinicians through clinical rounds, emails, and in-person communication. We used validated practices to ensure reliability of evidence and answered 24 of them. Questions differed depending on the collection method, with clinical rounds supporting proactive team involvement and gathering more patient characterization questions and questions related to a current patient. The main challenges we encountered included missing and incomplete data, underreported conditions, and nonspecific coding and accurate identification of drug regimens.

CONCLUSIONS

While the Data Consult Service has the potential to generate evidence and facilitate decision making, only a portion of questions can be answered in real time. Recognizing challenges in patient phenotyping and designing studies along with using validated practices for observational research are mandatory to produce reliable evidence.

Machine-learning model to predict the cause of death using a stacking ensemble method for observational data

OBJECTIVE

Cause of death is used as an important outcome of clinical research; however, access to cause-of-death data is limited. This study aimed to develop and validate a machine-learning model that predicts the cause of death from the patient's last medical checkup.

MATERIALS AND METHODS

To classify the mortality status and each individual cause of death, we used a stacking ensemble method. The prediction outcomes were all-cause mortality, 8 leading causes of death in South Korea, and other causes. The clinical data of study populations were extracted from the national claims (n = 174 747) and electronic health records (n = 729 065) and were used for model development and external validation. Moreover, we imputed the cause of death from the data of 3 US claims databases (n = 994 518, 995 372, and 407 604, respectively). All databases were formatted to the Observational Medical Outcomes Partnership Common Data Model.

RESULTS

The generalized area under the receiver operating characteristic curve (AUROC) of the model predicting the cause of death within 60 days was 0.9511. Moreover, the AUROC of the external validation was 0.8887. Among the causes of death imputed in the Medicare Supplemental database, 11.32% of deaths were due to malignant neoplastic disease.

DISCUSSION

This study showed the potential of machine-learning models as a new alternative to address the lack of access to cause-of-death data. All processes were disclosed to maintain transparency, and the model was easily applicable to other institutions.

CONCLUSION

A machine-learning model with competent performance was developed to predict cause of death.

Idiosyncratic Drug-Induced Liver Injury (DILI) and Herb-Induced Liver Injury (HILI): Diagnostic Algorithm Based on the Quantitative Roussel Uclaf Causality Assessment Method (RUCAM)

Causality assessment in liver injury induced by drugs and herbs remains a debated issue, requiring innovation and thorough understanding based on detailed information. Artificial intelligence (AI) principles recommend the use of algorithms for solving complex processes and are included in the diagnostic algorithm of Roussel Uclaf Causality Assessment Method (RUCAM) to help assess causality in suspected cases of idiosyncratic drug-induced liver injury (DILI) and herb-induced liver injury (HILI). From 1993 until the middle of 2020, a total of 95,865 DILI and HILI cases were assessed by RUCAM, outperforming by case numbers any other causality assessment method. The success of RUCAM can be traced back to its quantitative features with specific data elements that are individually scored leading to a final causality grading. RUCAM is objective, user friendly, transparent, and liver injury specific, with an updated version that should be used in future DILI and HILI cases. Support of RUCAM was also provided by scientists from China, not affiliated to any network, in the results of a scientometric evaluation of the global knowledge base of DILI. They highlighted the original RUCAM of 1993 and their authors as a publication quoted the greatest number of times and ranked first in the category of the top 10 references related to DILI. In conclusion, for stakeholders involved in DILI and HILI, RUCAM seems to be an effective diagnostic algorithm in line with AI principles.

Bias Implications of Outcome Misclassification in Observational Studies Evaluating Association Between Treatments and All-Cause or Cardiovascular Mortality Using Administrative Claims

Background

The bias implications of outcome misclassification arising from imperfect capture of mortality in claims‐based studies are not well understood.

Methods and Results

We identified 2 cohorts of patients: (1) type 2 diabetes mellitus (n=8.6 million), and (2) heart failure (n=3.1 million), from Medicare claims (2012–2016). Within the 2 cohorts, mortality was identified from claims using the following approaches: (1) all‐place all‐cause mortality, (2) in‐hospital all‐cause mortality, (3) all‐place cardiovascular mortality (based on diagnosis codes for a major cardiovascular event within 30 days of death date), or (4) in‐hospital cardiovascular mortality, and compared against National Death Index identified mortality. Empirically identified sensitivity and specificity based on observed values in the 2 cohorts were used to conduct Monte Carlo simulations for treatment effect estimation under differential and nondifferential misclassification scenarios. From National Death Index, 1 544 805 deaths (549 996 [35.6%] cardiovascular deaths) in the type 2 diabetes mellitus cohort and 1 175 202 deaths (523 430 [44.5%] cardiovascular deaths) in the heart failure cohort were included. Sensitivity was 99.997% and 99.207% for the all‐place all‐cause mortality approach, whereas it was 27.71% and 33.71% for the in‐hospital all‐cause mortality approach in the type 2 diabetes mellitus and heart failure cohorts, respectively, with perfect positive predicted values. For all‐place cardiovascular mortality, sensitivity was 52.01% in the type 2 diabetes mellitus cohort and 53.83% in the heart failure cohort with positive predicted values of 49.98% and 54.45%, respectively. Simulations suggested a possibility for substantial bias in treatment effects.

Conclusions

Approaches to identify mortality from claims had variable performance compared with the National Death Index. Investigators should anticipate the potential for bias from outcome misclassification when using administrative claims to capture mortality.

A comparison of two algorithms to identify sudden cardiac deaths in computerized databases

PURPOSE

Two previously validated algorithms to identify sudden cardiac death using administrative data showed high positive predictive value. We evaluated the agreement between the algorithms using data from a common source population.

METHODS

We conducted a cross-sectional study to assess the percent agreement between deaths identified by two sudden cardiac death algorithms using Tennessee Medicaid and death certificate data from 2007 through 2014. The source population included all deceased patients aged 18 to 64 years with Medicaid enrollment in the 6 months prior to death. To identify sudden cardiac deaths, algorithm 1 used only hospital/emergency department (ED) claims from encounters at the time of death, and algorithm 2 required death certificates and used claims data for specific exclusion criteria.

RESULTS

We identified 34 107 deaths in the source population over the study period. The two algorithms identified 4372 potential sudden cardiac deaths: Algorithm 1 identified 3117 (71.3%) and algorithm 2 identified 1715 (39.2%), with 460 (10.5%) deaths identified by both algorithms. Of the deaths identified by algorithm 1, 1943 (62.3%) had an underlying cause of death not specified in algorithm 2. Of the deaths identified by algorithm 2, 1053 (61.4%) had no record of a hospital or ED encounter at the time of death, and 202 (11.8%) had a discharge diagnosis code not specified in algorithm 1.

CONCLUSIONS

We found low agreement between the two algorithms for identification of sudden cardiac deaths because of differences in sudden cardiac death definitions and data sources.

The risk of new-onset epilepsy and refractory epilepsy in older adult stroke survivors

Objective

Our study objectives were to identify factors associated with new-onset epilepsy and refractory epilepsy among older adult stroke survivors and to evaluate the receipt of diagnostic care and mortality for participants who developed epilepsy.

Methods

We conducted a population-based, retrospective cohort study using linked, administrative health care databases. The Ontario Stroke Registry was used to identify patients 67 years and older who were hospitalized for a stroke at a designated stroke center in Ontario, Canada, between April 1, 2003, and March 31, 2009, and were previously free of epilepsy. Multivariable Fine–Gray hazard models were used to examine risk factors of epilepsy and refractory epilepsy, accounting for the competing risk of death.

Results

Among 19,138 older adults hospitalized for a stroke, 210 (1.1%) developed epilepsy and 27 (12.9%) became refractory to antiepileptic drugs. Within 1 year of epilepsy diagnosis, 24 (11.4%) patients were assessed with EEG and 19 (9.0%) with MRI. In multivariable analysis, younger age and thrombolysis receipt significantly increased epilepsy risk. Lesser stroke severity and anticoagulant medication receipt also significantly increased epilepsy risk; however, these effects decreased over time. Younger age and female sex were the only risk factors of refractory epilepsy. In the 5 years following epilepsy diagnosis, 97 (46.2%) participants died of any cause.

Conclusions

Older adult stroke survivors are less likely to develop epilepsy and pharmacologically refractory epilepsy. An estimated 86.6% of deaths among older adult stroke survivors with new-onset epilepsy are attributed to causes other than stroke or epilepsy.