Inverse Probability of Treatment Weighting and Confounder Missingness in Electronic Health Record-based Analyses: A Comparison of Approaches Using Plasmode Simulation.

BACKGROUND: Electronic health record (EHR) data represent a critical resource for comparative effectiveness research, allowing investigators to study intervention effects in real-world settings with large patient samples. However, high levels of missingness in confounder variables is common, challenging the perceived validity of EHR-based investigations. METHODS: We investigated performance of multiple imputation and propensity score (PS) calibration when conducting inverse probability of treatment weights (IPTW)-based comparative effectiveness research using EHR data with missingness in confounder variables and outcome misclassification. Our motivating example compared effectiveness of immunotherapy versus chemotherapy treatment of advanced bladder cancer with missingness in a key prognostic variable. We captured complexity in EHR data structures using a plasmode simulation approach to spike investigator-defined effects into resamples of a cohort of 4361 patients from a nationwide deidentified EHR-derived database. We characterized statistical properties of IPTW hazard ratio estimates when using multiple imputation or PS calibration missingness approaches. RESULTS: Multiple imputation and PS calibration performed similarly, maintaining ≤0.05 absolute bias in the marginal hazard ratio even when ≥50% of subjects had missing at random or missing not at random confounder data. Multiple imputation required greater computational resources, taking nearly 40 times as long as PS calibration to complete. Outcome misclassification minimally increased bias of both methods. CONCLUSION: Our results support multiple imputation and PS calibration approaches to missingness in missing completely at random or missing at random confounder variables in EHR-based IPTW comparative effectiveness analyses, even with missingness ≥50%. PS calibration represents a computationally efficient alternative to multiple imputation.

Telemedicine use among patients with metastatic breast cancer during the COVID-19 pandemic: Differences by race, age, and region.

PURPOSE: Our objective was to describe differences in telemedicine use among women with metastatic breast cancer (mBC) by race, age, and geographic region. METHODS: This was a retrospective cohort study of women with recurrent or de novo mBC treated in US community cancer practices that initiated a new line of therapy between March 2020 and February 2021. Multivariable modified Poisson regression models were used to calculate adjusted rate ratios (RR) and robust 95% confidence intervals (CI) associated with telemedicine visits within 90 days of therapy initiation. RESULTS: Overall, among 3412 women with mBC, 751 (22%) patients had telemedicine visits following therapy initiation, with lower risks observed among older women (<50 years: 24%; 50-64 years: 22%; 65-74 years: 21%; ≥75 years: 20%). Greater telemedicine use was observed among Asian women (35%) compared to White (21%), Black (18%), and Hispanic (21%) women. Fewer telemedicine visits occurred in Southern (12%) and Midwestern (17%) states versus Northeastern (37%) or Western (36%) states. In multivariable models, women ages ≥75 years had significantly lower risks of telemedicine visits (RR = 0.76, 95% CI 0.62-0.95) compared to ages <50 years. Compared to patients in Northeastern states, women in Midwestern (RR = 0.46, 95% CI 0.37-0.57) and Southern (RR = 0.31, 95% CI 0.26-0.37) states had significantly lower risks of telemedicine visits; but not women in Western states (RR = 0.96, 95% CI 0.83-1.12). No statistically significant differences in telemedicine use were found between racial groups in overall multivariable models. CONCLUSIONS: In this study of community cancer practices, older mBC patients and those living in Southern and Midwestern states were less likely to have telemedicine visits. Preferences for communication and delivery of care may have implications for measurement of exposures and endpoints in pharmacoepidemiologic studies of cancer patients.

Differential frequency in imaging-based outcome measurement: Bias in real-world oncology comparative-effectiveness studies.

BACKGROUND: Comparative-effectiveness studies using real-world data (RWD) can be susceptible to surveillance bias. In solid tumor oncology studies, analyses of endpoints such as progression-free survival (PFS) are based on progression events detected by imaging assessments. This study aimed to evaluate the potential bias introduced by differential imaging assessment frequency when using electronic health record (EHR)-derived data to investigate the comparative effectiveness of cancer therapies. METHODS: Using a nationwide de-identified EHR-derived database, we first analyzed imaging assessment frequency patterns in patients diagnosed with advanced non-small cell lung cancer (aNSCLC). We used those RWD inputs to develop a discrete event simulation model of two treatments where disease progression was the outcome and PFS was the endpoint. Using this model, we induced bias with differential imaging assessment timing and quantified its effect on observed versus true treatment effectiveness. We assessed percent bias in the estimated hazard ratio (HR). RESULTS: The frequency of assessments differed by cancer treatment types. In simulated comparative-effectiveness studies, PFS HRs estimated using real-world imaging assessment frequencies differed from the true HR by less than 10% in all scenarios (range: 0.4% to -9.6%). The greatest risk of biased effect estimates was found comparing treatments with widely different imaging frequencies, most exaggerated in disease settings where time to progression is very short. CONCLUSIONS: This study provided evidence that the frequency of imaging assessments to detect disease progression can differ by treatment type in real-world patients with cancer and may induce some bias in comparative-effectiveness studies in some situations.

An evaluation of the impact of missing deaths on overall survival analyses of advanced non-small cell lung cancer patients conducted in an electronic health records database.

PURPOSE: The aim of this study was to assess the impact of missing death data on survival analyses conducted in an oncology EHR-derived database. METHODS: The study was conducted using the Flatiron Health oncology database and the National Death Index (NDI) as a gold standard. Three analytic frameworks were evaluated in advanced non-small cell lung cancer (aNSCLC) patients: median overall survival [mOS]), relative risk estimates conducted within the EHR-derived database, and "external control arm" analyses comparing an experimental group augmented with mortality data from the gold standard to a control group from the EHR-derived database only. The hazard ratios (HRs) obtained within the EHR-derived database (91% sensitivity) and the external control arm analyses, were compared with results when both groups were augmented with mortality data from the gold standard. The above analyses were repeated using simulated lower mortality sensitivities to understand the impact of more extreme levels of missing deaths. RESULTS: Bias in mOS ranged from modest (0.6-0.9 mos.) in the EHR-derived cohort with (91% sensitivity) to substantial when lower sensitivities were generated through simulation (3.3-9.7 mos.). Overall, small differences were observed in the HRs for the EHR-derived cohort across comparative analyses when compared with HRs obtained using the gold standard data source. When only one treatment arm was subject to estimation bias, the bias was slightly more pronounced, but increased substantially when lower sensitivities were simulated. CONCLUSIONS: The impact on survival analysis is minimal with high mortality sensitivity with only modest impact associated within external control arm applications.

Prevalence and Predictors of Depression Among Patients With Epilepsy, Stroke, and Multiple Sclerosis Using the Cleveland Clinic Knowledge Program Within the Neurological Institute.

BACKGROUND: Neurological conditions carry a high risk of depression. Given this risk, the Neurological Institute (NI) at Cleveland Clinic has initiated systematic screening for depression using the 9-item Patient Health Questionnaire-9 (PHQ-9) embedded within its electronic medical record and its data capture system, the Knowledge Program (KP)1. OBJECTIVE: We sought to (1) estimate the prevalence of depression among patients with epilepsy, stroke, and multiple sclerosis (MS); (2) identify risk factors for depression within each disease; and (3) determine differential risks and predictors across neurological disorders. METHODS: The KP1 database provided information on approximately 23,000 visits involving 7946 outpatients with epilepsy, stroke, or MS seen in neurology specialty clinics. The primary outcome measure was depression as defined as a PHQ-9 ≥ 10. RESULTS: Overall, the point prevalence of depression was 29.0%. For stroke, epilepsy, and MS, prevalence of depression was 23% (95% CI: 21-25%), 33% (95% CI: 31-35%), and 29% (95% CI: 28-30%), respectively. For all 3 conditions, increasing disease severity and decreased health-related quality of life were independent predictors of depression. In multivariable models, there was a significant interaction between age and condition, and condition with disease severity. In stroke and MS, increasing age was associated with reduced odds for depression, whereas in epilepsy, increasing age was associated with an increased odds for depression. CONCLUSIONS: Although depression is common among patients with neurological disorders, our data suggest that predictors of depression such as age and disease severity varied by condition, supporting important possible phenomenological and pathophysiological differences of depression across these neurological conditions.

Proximity and gravity: modeling heaped self-reports.

Self-reported daily cigarette counts typically exhibit a preponderance of round numbers, a phenomenon known as heaping or digit preference. Heaping can be a substantial nuisance, as scientific interest lies in the distribution of the underlying true values rather than that of the heaped data. In principle, we can estimate parameters of the underlying distribution from heaped data if we know the conditional distribution of the heaped count given the true count, denoted the heaping mechanism (analogous to the missingness mechanism for missing data). In general, it is not possible to estimate the heaping mechanism robustly from heaped data only. A doubly-coded smoking cessation trial data set that includes daily cigarette count as both a conventional heaped retrospective recall measurement and a precise instantaneous measurement offers the rare opportunity to directly estimate the heaping mechanism. We propose a novel model that describes the conditional probability of the self-reported count as a function of its proximity to the truth and its intrinsic attractiveness, denoted its gravity. We apply variations of the model to the cigarette count data, illuminating the cognitive processes that influence self-reporting behaviors. The principal application of the model will be to enabling the correct analysis of heaped-only data sets. Copyright © 2017 John Wiley & Sons, Ltd.

Scale Linking to Enable Patient-Reported Outcome Performance Measures Assessed with Different Patient-Reported Outcome Measures.

BACKGROUND: Patient-reported outcome performance measures (PRO-PMs) incorporate outcomes from the patient's perspective into performance measures and may have great potential to impact health care. The various patient-reported outcome measures (PROMs) used to assess the same outcome challenge widespread use of PRO-PMs. A potential solution is to statistically link PROMs to provide equivalent PRO-PM conclusions to be drawn regardless of which PROM was used. OBJECTIVES: To determine the level of agreement in the performance of two depression-related PRO-PMs assessed using the nine-item Patient Health Questionnaire (PHQ-9) depression scale and the eight-item Patient-Reported Outcomes Measurement Information System (PROMIS) Depression short form and the PHQ-9 cocalibrated on the PROMIS metric. METHODS: We conducted a retrospective cohort study of patients who visited one of eight ambulatory neurological and psychiatric clinics at the Cleveland Clinic between January 23 and June 15, 2012, and who completed both the PHQ-9 and PROMIS Depression scales at the same visit. The level of agreement was measured between PRO-PM performance assessed with standard scoring of the PHQ-9, the PROMIS cocalibrated scoring of the PHQ-9, and the PROMIS score for two depression-related PRO-PMs. RESULTS: Of the 5736 enrolled patients, 701 had PROMs from two or more visits. Differences in performance of the depression remission PRO-PM ranged from 0.4% to 2.1%, and differences in the progress toward remission PRO-PM ranged from 0.9% to 5.1%, depending on which depression score was used. CONCLUSIONS: There was a high level of agreement in the PRO-PM for depression when incorporating different PROMs. These findings support the ability to use linkage of scale scores to assess performance of PRO-PMs with different PROMs.

A new Elixhauser-based comorbidity summary measure to predict in-hospital mortality.

BACKGROUND: Recently, van Walraven developed a weighted summary score (VW) based on the 30 comorbidities from the Elixhauser comorbidity system. One of the 30 comorbidities, cardiac arrhythmia, is currently excluded as a comorbidity indicator in administrative datasets such as the Nationwide Inpatient Sample (NIS), prompting us to examine the validity of the VW score and its use in the NIS. METHODS: Using data from the 2009 Maryland State Inpatient Database, we derived weighted summary scores to predict in-hospital mortality based on the full (30) and reduced (29) set of comorbidities and compared model performance of these and other comorbidity summaries in 2009 NIS data. RESULTS: Weights of our derived scores were not sensitive to the exclusion of cardiac arrhythmia. When applied to NIS data, models containing derived summary scores performed nearly identically (c statistics for 30 and 29 variable-derived summary scores: 0.804 and 0.802, respectively) to the model using all 29 comorbidity indicators (c=0.809), and slightly better than the VW score (c=0.793). Each of these models performed substantially better than those based on a simple count of Elixhauser comorbidities (c=0.745) or a categorized count (0, 1, 2, or ≥ 3 comorbidities; c=0.737). CONCLUSIONS: The VW score and our derived scores are valid in the NIS and are statistically superior to summaries using simple comorbidity counts. Researchers wishing to summarize the Elixhauser comorbidities with a single value should use the VW score or those derived in this study.

Comparison of Electronic Screening for Suicidal Risk With the Patient Health Questionnaire Item 9 and the Columbia Suicide Severity Rating Scale in an Outpatient Psychiatric Clinic.

BACKGROUND: Patient-reported data can improve clinical care, including identifying patients who are at risk for suicide. METHODS: In a tertiary care, psychiatric outpatient clinic, we compared computerized self-assessments of suicidal risk based on item 9 of the Patient Health Questionnaire-9 and an electronic version of the Columbia Suicide Severity Rating Scale (C-SSRS), using retrospective medical record review of clinical psychiatric assessments as the reference standard. We also surveyed patients׳ attitudes about participating in the process. We compared prevalence of suicidal risk rates by the 3 assessment methods as well as their sensitivity, specificity, and predictive value. RESULTS: Observed prevalence of positive suicidal risk screenings differed significantly, ranking (1) Patient Health Questionnaire-9 item 9, 24% (343/1416; 95% CI: 22%-26%) < (2) C-SSRS, 6.0% (85/1416; 95% CI: 5.0%-7.4%) < (3) clinical assessment, 1.4% (20/1416; 95% CI: 0.9%-2.2%). The sensitivity of Patient Health Questionnaire-9 item 9 was 92% (78/85; 95% CI: 86%-98%) and the specificity was 81% (1107/1376; 95% CI: 78%-82%). The sensitivity of the C-SSRS was 95.0% (19/20; 95% CI: 75%-100%) and the specificity was 95% (1330/1396; 95% CI: 94%-96%). Of 100 patients surveyed, the screening was well accepted, with some concerns about confidentiality and adequate clinical follow-up. CONCLUSIONS: As expected, Patient Health Questionnaire-9 item 9 generated much higher rates of apparently false-positive findings than the C-SSRS did, when compared with clinical assessment. C-SSRS backed with timely clinical assessment may be a useful and efficient method of screening for suicidal risk, provided that adequate, immediate clinical follow-up is available.

Thoughts of death and self-harm in patients with epilepsy or multiple sclerosis in a tertiary care center.

BACKGROUND: Patients with epilepsy or multiple sclerosis (MS) have high risks of depression and increased risks of suicide, but little is known about their risks of suicidal ideation. OBJECTIVE: We sought to (1) estimate the prevalence of thoughts of being better off dead or of self-harm among patients with epilepsy or MS, (2) identify risk factors for such thoughts, and (3) determine whether any risk factors interact with depression to predict such thoughts. METHODS: A Cleveland Clinic database provided information on 20,734 visits of 6586 outpatients with epilepsy or MS. Outcome measures were thoughts of death or self-harm (Patient Health Questionnaire [PHQ] item-9), and total score ≥10 for the 8 remaining PHQ items (probable major depression). Generalized estimating equations accounted for repeat visits in tests of associations of PHQ item-9 responses with depression, age, sex, race, household income, disease severity, and quality of life. RESULTS: Prevalence of thoughts of death or self-harm averaged 14.4% overall (epilepsy, 14.0% and MS, 14.7%). Factors associated with positive PHQ item-9 responses in epilepsy were depression and male sex, modified by poor quality of life. Factors associated with positive PHQ item-9 in MS were depression, male sex, medical comorbidity, and poor quality of life; the effect of depression was worse with greater MS severity and being unmarried. CONCLUSIONS: Among patients with common neurologic disorders (epilepsy or MS), 14%-15% reported thoughts of death or self-harm associated with illness severity, depression, quality of life, male sex, and being unmarried. Such patients require further evaluation of clinical outcomes and effects of treatment.

Incorporating patient-reported outcome measures into the electronic health record for research: application using the Patient Health Questionnaire (PHQ-9).

PURPOSE: Electronic health records (EHRs) present an opportunity to access large stores of data for research, but mapping raw EHR data to clinical phenotypes is complex. We propose adding patient-reported data to the EHR to improve phenotyping performance and describe a retrospective cohort study demonstrating a test case in depressive disorder. METHODS: We compared four EHR-phenotyping methods based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, medication records, and the Patient Health Questionnaire 9 (PHQ-9) regarding the ability to identify cases with depression and characteristics of patients identified with depression. Our sample included 168,884 patients seen (2007-2013) at our neurological institute. We assessed the diagnostic performance in a subset of 225 patients who had a reference standard measurement available. RESULTS: ICD-9-CM codes identified the fewest number of patients as depressed (4,658), followed by PHQ-9 (46,565), and medication data (50,505). The presence of at least one of these criteria identified the largest number (78,322). The PHQ-9 identified a higher proportion of elderly, disabled, Medicaid, and rural patients, as compared to ICD-9-CM codes. ICD-9-CM codes were least sensitive (6.7% sensitivity), whereas the method using at least one of the criteria identified the highest number of truly depressed patients (93.3% sensitivity); however, specificity dropped from 97.7 to 58.1%. CONCLUSIONS: The choice of phenotyping method may disproportionately exclude patient groups from research. Patient-reported data hold potential to improve sensitivity while maintaining an acceptable loss of specificity, depending on the context. Researchers should consider including patient-reported data in EHR-driven phenotyping methods.

Real-World Outcomes for Advanced Non-Small Cell Lung Cancer Patients Treated With a PD-L1 Inhibitor Beyond Progression.

BACKGROUND: Clinical trials of anti-programmed cell death ligand 1 (PD-L1) inhibitor to treat advanced non-small-cell lung cancer (aNSCLC) have permitted treatment beyond progression (TBP). However, the outcomes of patients receiving TBP in routine clinical care are unknown. MATERIALS AND METHODS: The present retrospective, observational, multicenter analysis evaluated de-identified electronic health record-derived data from community-based clinics in the United States. The patients had confirmed aNSCLC, had started anti-PD-L1 inhibitor therapy (nivolumab, pembrolizumab, or atezolizumab) before October 1, 2018, and had experienced a real-world progression (rwP) event. The study period ended March 31, 2019. The primary objective was to compare the overall survival (OS) of patients who had discontinued immunotherapy ≤ 30 days (non-TBP) compared with > 30 days after rwP (TBP). Descriptive analyses were performed. The Kaplan-Meier method and log-rank test were conducted for OS. An adjusted multivariable Cox proportional hazards regression model was also used. RESULTS: Overall, the data from 4223 patients were analyzed; 2555 (60.5%) and 1668 (39.5%) in the non-TBP and TBP cohorts, respectively. The median treatment duration for the non-TBP and TBP patients was 2.8 and 9.1 months (log-rank test, P < .001), respectively. The TBP group experienced longer unadjusted OS compared with the non-TBP group (11.5 vs. 5.1 months; log-rank test, P < .001). After adjusting for clinically relevant patient characteristics, the TBP OS benefit persisted (adjusted hazard ratio, 0.69; P < .001). CONCLUSIONS: TBP with PD-L1 inhibitor therapy is common in aNSCLC routine care and is potentially effective. These results support clinical trial observations likely to affect practice patterns.

A method comparison study of timeline followback and ecological momentary assessment of daily cigarette consumption.

INTRODUCTION: Uncertainty exists about how best to measure daily cigarette consumption. Two common measures are timeline followback (TLFB), which involves structured, prompted recall, and ecological momentary assessment (EMA), which involves recording consumption, as it occurs, on a handheld electronic device. METHODS: We evaluated the agreement between TLFB and EMA measures collected for 14 days prior to the target quit date from 236 smokers in a smoking cessation program. We performed a Bland-Altman analysis to assess agreement of TLFB and EMA using a regression-based model that allows for a nonuniform difference between methods and limits of agreement that can vary with the number of cigarettes smoked. RESULTS: For pairs of measurements taken on the same smoker, TLFB counts were on average 3.2 cigarettes higher than EMA counts; this difference increased for larger numbers of cigarettes. Using a model that allows for variable limits of agreement, the width of the 95% interval ranged from 8.7 to 61.8 cigarettes, with an average of 26.4 cigarettes. Variation between the methods increased substantially for larger cigarette counts, leading to wider limits and poorer agreement for heavy smokers. DISCUSSION: Throughout the measurement range, the estimated limits of agreement were far wider than the limits of clinical significance, defined a priori to be 20% of the number of cigarettes smoked. We conclude that TLFB and EMA cannot be considered equivalent for the assessment of daily cigarette consumption, especially for heavy smokers.

Generating Real-World Tumor Burden Endpoints from Electronic Health Record Data: Comparison of RECIST, Radiology-Anchored, and Clinician-Anchored Approaches for Abstracting Real-World Progression in Non-Small Cell Lung Cancer.

INTRODUCTION: Real-world evidence derived from electronic health records (EHRs) is increasingly recognized as a supplement to evidence generated from traditional clinical trials. In oncology, tumor-based Response Evaluation Criteria in Solid Tumors (RECIST) endpoints are standard clinical trial metrics. The best approach for collecting similar endpoints from EHRs remains unknown. We evaluated the feasibility of a RECIST-based methodology to assess EHR-derived real-world progression (rwP) and explored non-RECIST-based approaches. METHODS: In this retrospective study, cohorts were randomly selected from Flatiron Health's database of de-identified patient-level EHR data in advanced non-small cell lung cancer. A RECIST-based approach tested for feasibility (N = 26). Three non-RECIST approaches were tested for feasibility, reliability, and validity (N = 200): (1) radiology-anchored, (2) clinician-anchored, and (3) combined. Qualitative and quantitative methods were used. RESULTS: A RECIST-based approach was not feasible: cancer progression could be ascertained for 23% (6/26 patients). Radiology- and clinician-anchored approaches identified at least one rwP event for 87% (173/200 patients). rwP dates matched 90% of the time. In 72% of patients (124/173), the first clinician-anchored rwP event was accompanied by a downstream event (e.g., treatment change); the association was slightly lower for the radiology-anchored approach (67%; 121/180). Median overall survival (OS) was 17 months [95% confidence interval (CI) 14, 19]. Median real-world progression-free survival (rwPFS) was 5.5 months (95% CI 4.6, 6.3) and 4.9 months (95% CI 4.2, 5.6) for clinician-anchored and radiology-anchored approaches, respectively. Correlations between rwPFS and OS were similar across approaches (Spearman's rho 0.65-0.66). Abstractors preferred the clinician-anchored approach as it provided more comprehensive context. CONCLUSIONS: RECIST cannot adequately assess cancer progression in EHR-derived data because of missing data and lack of clarity in radiology reports. We found a clinician-anchored approach supported by radiology report data to be the optimal, and most practical, method for characterizing tumor-based endpoints from EHR-sourced data. FUNDING: Flatiron Health Inc., which is an independent subsidiary of the Roche group.

Characterizing the Feasibility and Performance of Real-World Tumor Progression End Points and Their Association With Overall Survival in a Large Advanced Non-Small-Cell Lung Cancer Data Set.

PURPOSE: Large, generalizable real-world data can enhance traditional clinical trial results. The current study evaluates reliability, clinical relevance, and large-scale feasibility for a previously documented method with which to characterize cancer progression outcomes in advanced non-small-cell lung cancer from electronic health record (EHR) data. METHODS: Patients who were diagnosed with advanced non-small-cell lung cancer between January 1, 2011, and February 28, 2018, with two or more EHR-documented visits and one or more systemic therapy line initiated were identified in Flatiron Health's longitudinal EHR-derived database. After institutional review board approval, we retrospectively characterized real-world progression (rwP) dates, with a random duplicate sample to ascertain interabstractor agreement. We calculated real-world progression-free survival, real-world time to progression, real-world time to next treatment, and overall survival (OS) using the Kaplan-Meier method (index date was the date of first-line therapy initiation), and correlations between OS and other end points were assessed at the patient level (Spearman's ρ). RESULTS: Of 30,276 eligible patients,16,606 (55%) had one or more rwP event. Of these patients, 11,366 (68%) had subsequent death, treatment discontinuation, or new treatment initiation. Correlation of real-world progression-free survival with OS was moderate to high (Spearman's ρ, 0.76; 95% CI, 0.75 to 0.77; evaluable patients, n = 20,020), and for real-world time to progression correlation with OS was lower (Spearman's ρ, 0.69; 95% CI, 0.68 to 0.70; evaluable patients, n = 11,902). Interabstractor agreement on rwP occurrence was 0.94 (duplicate sample, n = 1,065) and on rwP date 0.85 (95% CI, 0.81 to 0.89; evaluable patients n = 358 [patients with two independent event captures within 30 days]). Median rwP abstraction time from individual EHRs was 18.0 minutes (interquartile range, 9.7 to 34.4 minutes). CONCLUSION: We demonstrated that rwP-based end points correlate with OS, and that rwP curation from a large, contemporary EHR data set can be reliable, clinically relevant, and feasible on a large scale.

Development and Validation of a High-Quality Composite Real-World Mortality Endpoint.

OBJECTIVE: To create a high-quality electronic health record (EHR)-derived mortality dataset for retrospective and prospective real-world evidence generation. DATA SOURCES/STUDY SETTING: Oncology EHR data, supplemented with external commercial and US Social Security Death Index data, benchmarked to the National Death Index (NDI). STUDY DESIGN: We developed a recent, linkable, high-quality mortality variable amalgamated from multiple data sources to supplement EHR data, benchmarked against the highest completeness U.S. mortality data, the NDI. Data quality of the mortality variable version 2.0 is reported here. PRINCIPAL FINDINGS: For advanced non-small-cell lung cancer, sensitivity of mortality information improved from 66 percent in EHR structured data to 91 percent in the composite dataset, with high date agreement compared to the NDI. For advanced melanoma, metastatic colorectal cancer, and metastatic breast cancer, sensitivity of the final variable was 85 to 88 percent. Kaplan-Meier survival analyses showed that improving mortality data completeness minimized overestimation of survival relative to NDI-based estimates. CONCLUSIONS: For EHR-derived data to yield reliable real-world evidence, it needs to be of known and sufficiently high quality. Considering the impact of mortality data completeness on survival endpoints, we highlight the importance of data quality assessment and advocate benchmarking to the NDI.

The role of very high high-sensitivity C-reactive protein levels on mortality after stroke.

High-sensitivity C-reaction protein (hsCRP) levels are correlated with risk of vascular disease. The clinical implications of markedly elevated hsCRP levels in the setting of acute stroke are less understood. OBJECTIVE: To determine the association of very high admission hsCRP levels (≥10mg/L) on all-cause mortality in patients with acute stroke. METHODS: A retrospective cohort study of 1176 patients hospitalized with acute stroke between 2005 and 2012 who had a hsCRP assay drawn within seven days of admission. Mortality data was obtained using Ohio Death Index. Cox proportional analysis was used to determine hazard ratios of death among patients with hsCRP values >10mg/L after adjusting for age, sex, race and relevant co-morbid conditions. NIHSS was also adjusted in a subset of patients in whom it was available (n=378). RESULTS: Patients with hsCRP >10mg/L had 2.65 (95% CI: 1.99, 3.53) increased hazard of all-cause mortality compared to stroke patients with hsCRP ≤10mg/L. Association of hsCRP with mortality was greater for patients with low NIHSS and declined in patients with increasing NIHSS scores. A higher proportion of patients with hsCRP >10 died from nonvascular causes than patients with hsCRP ≤10mg/L. CONCLUSION: Very high levels of hsCRP may be a useful marker to identify stroke patients at increased risk for death, especially those with stroke of mild severity. More research is needed to determine how this information can be used to improve patient care, especially in patients with mild stroke.

The PROMIS physical function scale: A promising scale for use in patients with ischemic stroke.

OBJECTIVE: To evaluate the performance of the Patient Reported Outcomes Measurement Information System (PROMIS) physical function (PF) scale compared to the validated Stroke Impact Scale-16 (SIS-16) in ischemic stroke patients seen in an ambulatory cerebrovascular clinic. METHODS: This was a retrospective cohort study. PROMIS PF (computer adaptive testing version) and SIS-16 measures were electronically collected on 1,946 ischemic stroke patients seen in a cerebrovascular clinic using an electronic platform from September 12, 2012, to June 16, 2015. Distribution of scores was compared to assess ceiling and floor effects. Correlations with other commonly used functional status scales were performed to assess convergent validity. RESULTS: The SIS-16 and PROMIS PF had a 19.6% and <1% ceiling effect, respectively. Patients completed 16 SIS-16 items and a median of 4 (interquartile range 4-4) PROMIS PF items. Internal consistency of both SIS-16 and PROMIS PF was excellent. The SIS-16 had slightly but significantly higher correlations with the other functional scales than PROMIS PF. CONCLUSIONS: The use of PROMIS to obtain electronic patient-reported functional status in an ambulatory setting is feasible. PROMIS PF is an option for measurement of physical function in ischemic stroke patients. It had similar test characteristics as the SIS-16 but with lower patient burden and minimal ceiling effect.

Model-based imputation of latent cigarette counts using data from a calibration study.

In addition to dichotomous measures of abstinence, smoking studies may use daily cigarette consumption as an outcome variable. These counts hold the promise of more efficient and detailed analyses than dichotomous measures, but present serious quality issues - measurement error and heaping - if obtained by retrospective recall. A doubly-coded dataset with a retrospective recall measurement (timeline followback, TLFB) and a more precise instantaneous measurement (ecological momentary assessment, EMA) serves as a calibration dataset, allowing us to predict EMA given TLFB and baseline factors. We apply this model to multiply impute precise cigarette counts for a randomized, placebo-controlled trial of bupropion with only TLFB measurements available. To account for repeated measurements on a subject, we induce correlation in the imputed counts. Finally, we analyze the imputed data in a longitudinal model that accommodates random subject effects and zero inflation. Both raw and imputed data show a significant drug effect for reducing the odds of non-abstinence and the number of cigarettes smoked among non-abstainers, but the imputed data provide efficiency gains. This method permits the analysis of daily cigarette consumption data previously deemed suspect due to reporting error and is applicable to other self-reported count data sets for which calibration samples are available. Copyright © 2015 John Wiley & Sons, Ltd.

Impact of Sleep-Disordered Breathing Treatment on Patient Reported Outcomes in a Clinic-Based Cohort of Hypertensive Patients.

STUDY OBJECTIVES: We hypothesized that patient reported outcomes (PROs) improve with positive airway pressure (PAP) in patients with sleep-disordered breathing (SDB) and hypertension (HTN). METHODS: Questionnaire-based PROs (sleepiness [Epworth Sleepiness Scale, (ESS)], depression [Patient Health Questionnaire-9 (PHQ-9)], and fatigue [Fatigue Severity Scale (FSS)]) were retrospectively examined in patients with SDB and HTN at baseline and within a year following PAP initiation. PRO changes were estimated using multivariable linear mixed-effect models adjusted for baseline age, sex, race, body mass index, resistant hypertension (RHTN) status, cardiac and diabetes history, and correlation between repeated measurements. Age and race by PAP interaction terms (mean change, 95% CI) were examined. RESULTS: 894 patients with HTN and SDB were examined. 130 (15%) had baseline RHTN (age 58 ± 12 y, 52.9 % male, BMI 36.2 ± 9.1 kg/m2). In multivariable models, a significant improvement in sleepiness ESS (-2.09, 95% CI: -2.37, -1.82), PHQ-9 (-1.91, 95% CI: -2.25, -1.56), and FSS scores (-4.06 95% CI: -4.89, -3.22) was observed. A significant race by PAP effect interaction was observed (p < 0.0001 for all PROs); Caucasians had greater improvements than non-Caucasians. The interaction term of effect of PAP and age was significant for ESS (p = 0.04) and PHQ-9 (p = 0.0003), indicating greater improvement in younger patients. CONCLUSIONS: Consistent improvement of broad PRO domains in response to PAP in SDB was observed in this clinic-based hypertensive cohort; Caucasians and younger patients derived greater benefit.