Assessing Real-World Data From Electronic Health Records for Health Technology Assessment: The SUITABILITY Checklist: A Good Practices Report of an ISPOR Task Force.

This ISPOR Good Practices report provides a framework for assessing the suitability of electronic health records data for use in health technology assessments (HTAs). Although electronic health record (EHR) data can fill evidence gaps and improve decisions, several important limitations can affect its validity and relevance. The ISPOR framework includes 2 components: data delineation and data fitness for purpose. Data delineation provides a complete understanding of the data and an assessment of its trustworthiness by describing (1) data characteristics; (2) data provenance; and (3) data governance. Fitness for purpose comprises (1) data reliability items, ie, how accurate and complete the estimates are for answering the question at hand and (2) data relevance items, which assess how well the data are suited to answer the particular question from a decision-making perspective. The report includes a checklist specific to EHR data reporting: the ISPOR SUITABILITY Checklist. It also provides recommendations for HTA agencies and policy makers to improve the use of EHR-derived data over time. The report concludes with a discussion of limitations and future directions in the field, including the potential impact from the substantial and rapid advances in the diffusion and capabilities of large language models and generative artificial intelligence. The report's immediate audiences are HTA evidence developers and users. We anticipate that it will also be useful to other stakeholders, particularly regulators and manufacturers, in the future.

Prevalence of covid-19 and long covid in collegiate student athletes from spring 2020 to fall 2021: a retrospective survey.

BACKGROUND: Symptomatic COVID-19 and Long COVID, also referred to as post-acute sequelae of SARS-CoV-2 (PASC) or post-COVID conditions, have been widely reported in young, healthy people, but their prevalence has not yet been determined in student athletes. We sought to estimate the prevalence of reported COVID-19, symptomatic COVID-19, and Long COVID in college athletes in the United States attending 18 schools from spring 2020 to fall 2021. METHODS: We developed an online survey to measure the prevalence of student athletes who tested positive for COVID-19, developed Long COVID, and did not return to their sport during the relevant time period. We surveyed a convenience sample of 18 collegiate school administrators, representing about 7,000 student athletes. Of those schools surveyed, 16 responded regarding the spring 2020 semester, and 18 responded regarding the full academic year of fall 2020 to spring 2021 (both semesters). RESULTS: According to the survey responses, there were 9.8% of student athletes who tested positive for COVID-19 in spring 2020 and 25.4% who tested positive in the academic year of fall 2020 to spring 2021. About 4% of student athletes who tested positive from spring 2020 to spring 2021 developed Long COVID, defined as new, recurring, or ongoing physical or mental health consequences occurring 4 or more weeks after SARS-CoV-2 infection. CONCLUSIONS: This study highlights that Long COVID occurs among young, healthy athletes and is a real consequence of COVID-19. Understanding the prevalence of Long COVID in this population requires longer follow-up and further study.

Approaches for Enhanced Extrapolation of Long-Term Survival Outcomes Using Electronic Health Records of Patients With Cancer.

OBJECTIVES: This study aimed to demonstrate enhanced survival extrapolation methods using electronic health record-derived real-world data (RWD). METHODS: The study population included patients diagnosed of ER+/HER2- metastatic breast cancer who started first-line treatment with anastrozole or letrozole between November 18, 2014, and November 18, 2015. Two patient cohorts were constructed: a clinical trial cohort from digitized MONARCH-3 clinical trial results and a RWD cohort from a deidentified electronic health record-derived database. RWD patients were weighted to trial baseline covariate distributions. Standard parametric approaches were applied to trial data and a "best-fit" model was selected. We demonstrate traditional and enhanced hybrid (pooling with weighted RWD at start, 75%, or end of trial) extrapolation approaches. RESULTS: Observed and estimated 5-year progression-free survival (PFS) rates in extrapolating the trial control arm (n = 165) were comparable across all methods. Compared with the observed 5-year mean PFS in the RWD cohort (n = 118) of 20.4 months (95% confidence interval [CI] 16.9-23.8), there was some variation among studied methods. Best-fit standard parametric model (log-normal) had 5-year mean PFS of 21.3 months (95% CI 18.2-24.9), and for the hybrid methods in order of estimate conservativeness was start of trial (20.8 months; 95% CI 18.5-23.2), 75% of trial (21.3 months; 95% CI 18.1-24.5), and end of trial (21.8 months; 95% CI 18.8-25.2). CONCLUSIONS: Our study leverages RWD to enhance long-term survival extrapolation. Future use cases should include applying patient eligibility criteria, weighting on baseline characteristics, and choice of time window to add RWD to trial data.

Machine Learning Methods in Health Economics and Outcomes Research-The PALISADE Checklist: A Good Practices Report of an ISPOR Task Force.

Advances in machine learning (ML) and artificial intelligence offer tremendous potential benefits to patients. Predictive analytics using ML are already widely used in healthcare operations and care delivery, but how can ML be used for health economics and outcomes research (HEOR)? To answer this question, ISPOR established an emerging good practices task force for the application of ML in HEOR. The task force identified 5 methodological areas where ML could enhance HEOR: (1) cohort selection, identifying samples with greater specificity with respect to inclusion criteria; (2) identification of independent predictors and covariates of health outcomes; (3) predictive analytics of health outcomes, including those that are high cost or life threatening; (4) causal inference through methods, such as targeted maximum likelihood estimation or double-debiased estimation-helping to produce reliable evidence more quickly; and (5) application of ML to the development of economic models to reduce structural, parameter, and sampling uncertainty in cost-effectiveness analysis. Overall, ML facilitates HEOR through the meaningful and efficient analysis of big data. Nevertheless, a lack of transparency on how ML methods deliver solutions to feature selection and predictive analytics, especially in unsupervised circumstances, increases risk to providers and other decision makers in using ML results. To examine whether ML offers a useful and transparent solution to healthcare analytics, the task force developed the PALISADE Checklist. It is a guide for balancing the many potential applications of ML with the need for transparency in methods development and findings.

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.

First-line immune checkpoint inhibitor use in cisplatin-eligible patients with advanced urothelial carcinoma: a secular trend analysis.

Aim: Standard first-line treatment of advanced urothelial cell carcinoma involves cisplatin-based chemotherapy, with carboplatin or immune checkpoint inhibitor therapy (ICI) reserved for cisplatin-ineligible individuals. Methods: Using a large de-identified electronic health record-derived database of patients with advanced urothelial cell carcinoma in the USA, we examined trends in utilization of first-line systemic therapies in cisplatin-eligible patients from 1 January 2015 to 31 March 2018. Results: Among 1181 cisplatin-eligible patients, the quarterly proportion who received first-line ICI increased from 1 to 42% (ptrend <0.001), while the proportion who received cisplatin-based chemotherapy decreased from 53 to 33% (ptrend = 0.018). Patients receiving ICI were older than those receiving cisplatin (median age: 75 vs 68). Conclusion: Our analysis suggests rising off-label ICI use in cisplatin-eligible individuals, potentially because of ICI's favorable toxicity profile.

Trends in Checkpoint Inhibitor Therapy for Advanced Urothelial Cell Carcinoma at the End of Life: Insights from Real-World Practice.

Several immune checkpoint inhibitor therapies (CPIs) have been approved to treat metastatic urothelial cell carcinoma (mUC). Because of the favorable toxicity profile of CPI compared with chemotherapy, oncologists may have a low threshold to prescribe CPI to patients near the end of life. We evaluated trends in initiation of end-of-life systemic therapy in 1,637 individuals in the Flatiron Health Database who were diagnosed with mUC between 2015 and 2017 and who died. Rates of systemic therapy initiation in the last 30 and 60 days of life were 17.0% and 29.8%, respectively. The quarterly proportion of patients who initiated CPI within 60 days of death increased from 1.0% to 23% during the study period (p trend < .001). After CPI approval, end-of-life CPI initiation significantly increased among patients with poor performance status (p trend = .020) and did not significantly change among individuals with good performance status. The quarterly proportion of patients who initiated any systemic therapy at the end of life doubled (17.4% to 34.8%) during the study period, largely explained by increased CPI use. These findings suggest a dramatic rise in CPI use at the end of life in patients with mUC, a finding that may have important guideline and policy implications.

The Impact of ACA and Medicaid Expansion on Progress Toward UNAIDS 90-90-90 Goals.

PURPOSE OF REVIEW: Passage of the Affordable Care Act (ACA) in 2010 and subsequent Medicaid expansion has influenced access to HIV treatment and care in the USA. This review aims to evaluate whether the implementation of these policies has impacted progress toward UNAIDS 90-90-90 goals. RECENT FINDINGS: Preliminary evidence has emerged suggesting that the ACA and Medicaid expansion has increased the likelihood of HIV testing and diagnosis, reduced the number of people unaware of HIV infection, and increased the number of people on antiretroviral therapy (ART) who are virally suppressed. While the ACA is associated with some progress toward 90-90-90 goals, more years of data after policy implementation are needed for robust analysis. Methods including difference-in-differences, instrumental variables, and propensity scores are recommended to minimize bias from unmeasured confounders and make causal inference about non-random Medicaid expansion among states.

The Cost-Effectiveness of Financial Incentives for Viral Suppression: HPTN 065 Study.

OBJECTIVE: To evaluate the cost-effectiveness of financial incentives for human immunodeficiency virus (HIV) viral suppression compared to standard of care. STUDY DESIGN: Mathematical model of 2-year intervention offering financial incentives ($70 quarterly) for viral suppression (<400 copies/ml3) based on the HPTN 065 clinical trial with HIV patients in the Bronx, NY and Washington, D.C. METHODS: A disease progression model with HIV transmission risk equations was developed following guidelines from the Second Panel on Cost-Effectiveness in Health and Medicine. We used health care sector and societal perspectives, 3% discount rate, and lifetime horizon. Data sources included trial data (baseline N = 16,208 patients), CDC HIV Surveillance data, and published literature. Outcomes were costs (2017 USD), quality-adjusted life years (QALYs), HIV infections prevented, and incremental cost-effectiveness ratio (ICER). RESULTS: Financial incentives for viral suppression were estimated to be cost-saving from a societal perspective and cost-effective ($49,877/QALY) from a health care sector perspective. Compared to the standard of care, financial incentives gain 0.06 QALYs and lower discounted lifetime costs by $4210 per patient. The model estimates that incentivized patients transmit 9% fewer infections than the standard-of-care patients. In the sensitivity analysis, ICER 95% credible intervals ranged from cost-saving to $501,610/QALY with 72% of simulations being cost-effective using a $150,000/QALY threshold. Modeling results are limited by uncertainty in efficacy from the clinical trial. CONCLUSIONS: Financial incentives, as used in HTPN 065, are estimated to improve quality and length of life, reduce HIV transmissions, and save money from a societal perspective. Financial incentives offer a promising option for enhancing the benefits of medication in the United States.

Enhancing diversity in the public health research workforce: the research and mentorship program for future HIV vaccine scientists.

OBJECTIVES: We developed and evaluated a novel National Institutes of Health-sponsored Research and Mentorship Program for African American and Hispanic medical students embedded within the international, multisite HIV Vaccine Trials Network, and explored its impact on scientific knowledge, acquired skills, and future career plans. METHODS: Scholars conducted social, behavioral, clinical, or laboratory-based research projects with HIV Vaccine Trials Network investigators over 8 to 16 weeks (track 1) or 9 to 12 months (track 2). We conducted an in-depth, mixed-methods evaluation of the first 2 cohorts (2011-2013) to identify program strengths, areas for improvement, and influence on professional development. RESULTS: A pre-post program assessment demonstrated increases in self-reported knowledge, professional skills, and interest in future HIV vaccine research. During in-depth interviews, scholars reported that a supportive, centrally administered program; available funding; and highly involved mentors and staff were keys to the program's early success. CONCLUSIONS: A multicomponent, mentored research experience that engages medical students from underrepresented communities and is organized within a clinical trials network may expand the pool of diverse public health scientists. Efforts to sustain scholar interest over time and track career trajectories are warranted.

Transporting Comparative Effectiveness Evidence Between Countries: Considerations for Health Technology Assessments.

Internal validity is often the primary concern for health technology assessment agencies when assessing comparative effectiveness evidence. However, the increasing use of real-world data from countries other than a health technology assessment agency's target population in effectiveness research has increased concerns over the external validity, or "transportability", of this evidence, and has led to a preference for local data. Methods have been developed to enable a lack of transportability to be addressed, for example by accounting for cross-country differences in disease characteristics, but their consideration in health technology assessments is limited. This may be because of limited knowledge of the methods and/or uncertainties in how best to utilise them within existing health technology assessment frameworks. This article aims to provide an introduction to transportability, including a summary of its assumptions and the methods available for identifying and adjusting for a lack of transportability, before discussing important considerations relating to their use in health technology assessment settings, including guidance on the identification of effect modifiers, guidance on the choice of target population, estimand, study sample and methods, and how evaluations of transportability can be integrated into health technology assessment submission and decision processes.

Evaluation of mpox vaccine dose-sparing strategies.

The spring-summer 2022 mpox outbreak had over 50,000 cases globally, most of them in gay, bisexual, and other men who have sex with men (MSM). In response to vaccine shortages, several countries implemented dose-sparing vaccination strategies, stretching a full-dose vaccine vial into up to five fractional-dose vaccines. Recent studies have found mixed results regarding the effectiveness of the mpox vaccine, raising the question of the utility of dose-sparing strategies. We used an age- and risk-stratified mathematical model of an urban MSM population in the United States with ∼12% high-risk MSM to evaluate potential benefits from implementing dose-sparing vaccination strategies in which a full dose is divided into 3.5 fractional doses. We found that results strongly depend on the fractional-dose vaccine effectiveness (VE) and vaccine supply. With very limited vaccines available, enough to protect with a full dose approximately one-third of the high-risk population, dose-sparing strategies are more beneficial provided that fractional doses preserved at least 40% of full-dose effectiveness (34% absolute VE), projecting 13% (34% VE) to 70% (68% absolute VE) fewer infections than full-dose strategies. In contrast, if vaccine supply is enough to cover the majority of the high-risk population, dose-sparing strategies can be outperformed by full-dose strategies. Scenarios in which fractional dosing was 34% efficacious resulted in almost three times more infections than full dosing. Our analysis suggests that when mpox vaccine supply is limited and fractional-dose vaccination retains moderate effectiveness, there are meaningful health benefits from providing a smaller dose to a larger number of people in the high-risk population. These findings should inform the public-health response to future mpox outbreaks.

Evaluation of US oncology electronic health record real-world data to reduce uncertainty in health technology appraisals: a retrospective cohort study.

OBJECTIVES: Examine whether data from early access to medicines in the USA can be used to inform National Institute for Health and Care Excellence (NICE) health technology assessments (HTA) in oncology. DESIGN: Retrospective cohort study. SETTING: Oncology-based community and academic treatment centres in the USA. PARTICIPANTS: Patients present in a nationwide electronic health record (EHR)-derived deidentified database. INTERVENTIONS: Cancer drugs that underwent NICE technology appraisal (TA) between 2014 and 2019. PRIMARY AND SECONDARY OUTCOME MEASURES: The count and follow-up time of US patients, available in the EHR, who were exposed to cancer drugs of interest in the period between Food and Drug Administration (FDA) approval and dates relevant to the NICE appraisal process. RESULTS: In 59 of 60 TAs analysed, the cancer therapy was approved in the USA before the final appraisal by NICE. The median time from FDA approval to the publication of NICE recommendations was 18.5 months, at which time the US EHR-derived database had, on average, 269 patients (SD=356) exposed to the new therapy, with a median of 75.3 person-years (IQR: 13.1-173) in time-at-risk. A case study generated evidence on real-world overall survival and treatment duration. CONCLUSIONS: Across different cancer therapies, there was substantial variability in US real-world data accumulated between FDA approval and NICE decision milestones. The applicability of these data to generate evidence for HTA decision-making should be assessed on a case-by-case basis depending on the intended HTA use case.

Replication of Real-World Evidence in Oncology Using Electronic Health Record Data Extracted by Machine Learning.

Meaningful real-world evidence (RWE) generation requires unstructured data found in electronic health records (EHRs) which are often missing from administrative claims; however, obtaining relevant data from unstructured EHR sources is resource-intensive. In response, researchers are using natural language processing (NLP) with machine learning (ML) techniques (i.e., ML extraction) to extract real-world data (RWD) at scale. This study assessed the quality and fitness-for-use of EHR-derived oncology data curated using NLP with ML as compared to the reference standard of expert abstraction. Using a sample of 186,313 patients with lung cancer from a nationwide EHR-derived de-identified database, we performed a series of replication analyses demonstrating some common analyses conducted in retrospective observational research with complex EHR-derived data to generate evidence. Eligible patients were selected into biomarker- and treatment-defined cohorts, first with expert-abstracted then with ML-extracted data. We utilized the biomarker- and treatment-defined cohorts to perform analyses related to biomarker-associated survival and treatment comparative effectiveness, respectively. Across all analyses, the results differed by less than 8% between the data curation methods, and similar conclusions were reached. These results highlight that high-performance ML-extracted variables trained on expert-abstracted data can achieve similar results as when using abstracted data, unlocking the ability to perform oncology research at scale.

Approach to machine learning for extraction of real-world data variables from electronic health records.

Background: As artificial intelligence (AI) continues to advance with breakthroughs in natural language processing (NLP) and machine learning (ML), such as the development of models like OpenAI's ChatGPT, new opportunities are emerging for efficient curation of electronic health records (EHR) into real-world data (RWD) for evidence generation in oncology. Our objective is to describe the research and development of industry methods to promote transparency and explainability. Methods: We applied NLP with ML techniques to train, validate, and test the extraction of information from unstructured documents (e.g., clinician notes, radiology reports, lab reports, etc.) to output a set of structured variables required for RWD analysis. This research used a nationwide electronic health record (EHR)-derived database. Models were selected based on performance. Variables curated with an approach using ML extraction are those where the value is determined solely based on an ML model (i.e. not confirmed by abstraction), which identifies key information from visit notes and documents. These models do not predict future events or infer missing information. Results: We developed an approach using NLP and ML for extraction of clinically meaningful information from unstructured EHR documents and found high performance of output variables compared with variables curated by manually abstracted data. These extraction methods resulted in research-ready variables including initial cancer diagnosis with date, advanced/metastatic diagnosis with date, disease stage, histology, smoking status, surgery status with date, biomarker test results with dates, and oral treatments with dates. Conclusion: NLP and ML enable the extraction of retrospective clinical data in EHR with speed and scalability to help researchers learn from the experience of every person with cancer.

A machine learning framework supporting prospective clinical decisions applied to risk prediction in oncology.

We present a general framework for developing a machine learning (ML) tool that supports clinician assessment of patient risk using electronic health record-derived real-world data and apply the framework to a quality improvement use case in an oncology setting to identify patients at risk for a near-term (60 day) emergency department (ED) visit who could potentially be eligible for a home-based acute care program. Framework steps include defining clinical quality improvement goals, model development and validation, bias assessment, retrospective and prospective validation, and deployment in clinical workflow. In the retrospective analysis for the use case, 8% of patient encounters were associated with a high risk (pre-defined as predicted probability ≥20%) for a near-term ED visit by the patient. Positive predictive value (PPV) and negative predictive value (NPV) for future ED events was 26% and 91%, respectively. Odds ratio (OR) of ED visit (high- vs. low-risk) was 3.5 (95% CI: 3.4-3.5). The model appeared to be calibrated across racial, gender, and ethnic groups. In the prospective analysis, 10% of patients were classified as high risk, 76% of whom were confirmed by clinicians as eligible for home-based acute care. PPV and NPV for future ED events was 22% and 95%, respectively. OR of ED visit (high- vs. low-risk) was 5.4 (95% CI: 2.6-11.0). The proposed framework for an ML-based tool that supports clinician assessment of patient risk is a stepwise development approach; we successfully applied the framework to an ED visit risk prediction use case.

Finding and treating early-stage HIV infections: A cost-effectiveness analysis of the Sabes study in Lima, Peru.

Background: Sabes, a treatment-as-prevention intervention among men who have sex with men and transgender women in Lima, Peru, was developed to identify HIV during early primary infection (<3 months from acquisition) through monthly serologic assays and HIV RNA tests. Newly diagnosed individuals were rapidly linked to care and offered to initiate ART. In this study we sought to study the cost-effectiveness of Sabes compared to the standard of care (SOC) for HIV testing and initiation of treatment. Methods: We adapted a compartmental model of HIV transmission to evaluate the cost-effectiveness of the Sabes approach compared to the SOC using a government health care perspective, 20-year time horizon, and 3% annual discounting. We estimated the proportion of cases of HIV detected during early primary infection, reduction in HIV incidence and prevalence, incremental cost-effectiveness ratio (ICER), and net monetary benefit. We analyzed costs using data from the Sabes study, the Peruvian Ministry of Health, published literature, and expert consultation. Findings: The Sabes intervention is projected to identify 9294 early primary HIV infections in Lima, Peru over 20 years. The intervention costs $6,896 per early primary infection diagnosed and by 2038 is expected to decrease the fraction of early infections among prevalent infections by 62%. Sabes is expected to improve health, resulting in greater total discounted QALYs per person than the SOC (16·7 vs 16·4, respectively). Sabes had an ICER of $1431 (22% per capita GDP in Peru) per QALY compared to SOC. Interpretation: Our analysis suggests that in Lima, Peru the Sabes intervention could be a cost-effective approach to reduce the burden of HIV even under stringent cost-effectiveness criteria. This finding suggests that programs that use frequent HIV testing, rapid linkage to care and initiation of ART should be considered as part of a comprehensive HIV prevention strategy. Funding: National Institutes of Health.

Real World Characterization of Advanced Non-Small Cell Lung Cancer in Never Smokers by Actionable Mutation Status.

BACKGROUND: Non-small cell lung cancer (NSCLC) in never-smokers (NS) is vastly different from those with a smoking history in terms of etiology, driver mutations, and immunotherapy responsiveness. This study compares the real-world overall survival (OS) of NSCLC patients by smoking history and mutation status. METHODS: The study included 30,310 advanced or metastatic NSCLC patients in the Flatiron Health EHR-derived database who received biomarker testing results (EGFR, ALK, ROS1, and BRAF), and initiated therapy between 2011 and 2017, with follow up through June 2018. OS by smoking and driver mutation groups was summarized via Kaplan-Meier survival estimates, and compared in the context of a multivariate Cox proportional hazard model. RESULTS: OS differed by smoking and driver-mutation categories (adjusted and stratified P< .001). The median OS for wild-type (WT) smoking patients was 9.6 months, for mutated (MT) smokers was 19.4 months (adjusted and stratified hazard ratio [HR] relative to WT smokers 0.65; 95% CI 0.60-0.71), for WT NS was 15.1 months (HR 0.78; 95% CI 0.73-0.83 relative to WT smokers), and for MT NS was 25.5 months (HR 0.52; 95% CI 0.48-0.58 relative to WT smokers). CONCLUSION: NS with NSCLC survived longer than those with smoking history, in both groups of WT and mutation-positive patients. Findings highlight that in NSCLC patients, a history of never smoking may have similar effect on hazard of death as the presence of an actionable mutation. Taken together, differences in heredity, mutations, and biologic history suggest that NS lung cancer is a distinct clinical entity and must be managed accordingly.

Programmed Death-Ligand 1 Tumor Proportion Score and Overall Survival From First-Line Pembrolizumab in Patients With Nonsquamous Versus Squamous NSCLC.

INTRODUCTION: For patients with NSCLC receiving immune checkpoint inhibitors, programmed death-ligand 1 (PD-L1) tumor proportion score (TPS) has been validated as a predictive biomarker for improved overall survival (OS). Nevertheless, its histology-specific predictive value in patients with advanced squamous versus nonsquamous cancers remains unclear. To evaluate the differential value of PD-L1 TPS as a predictive biomarker for OS after first-line pembrolizumab in patients with squamous versus nonsquamous NSCLC. METHODS: Retrospective, observational study of patients diagnosed with having advanced NSCLC who were treated between October 2015 and April 2019 at community oncology clinics and academic medical centers in a deidentified electronic health record-derived database. Included patients were diagnosed with having advanced or metastatic NSCLC, received treatment with first-line, single-agent pembrolizumab, and had documentation of PD-L1 testing with a numeric result. Exclusion criteria included alterations in EGFR, ALK, and ROS1. The primary end point was OS from start of first-line pembrolizumab therapy by squamous or nonsquamous histology and PD-1 expression level measured by TPS (low, <50% or high, ≥50%). RESULTS: The cohort of 1460 patients with NSCLC who received pembrolizumab as a first-line therapy had a mean age of 72 years. Histology was 28% squamous and 72% nonsquamous. PD-L1 expression was low in 13% and high in 87%. No meaningful differences in age, sex, or smoking history were observed by PD-L1 TPS or histology type. A generalized gamma model adjusting for sex and stage at diagnosis found that for patients with nonsquamous histology, high PD-L1 TPS was significantly associated with improved OS by a median OS difference of 8.4 months (p < 0.001). In contrast, for patients with squamous histology, there was no evidence of association between PD-L1 expression level and OS (p = 0.283). PD-L1-related incremental differences in median OS between the patients with squamous and nonsquamous tumors were significantly different (p = 0.034). CONCLUSIONS: Among patients with NSCLC treated with first-line pembrolizumab, high PD-L1 TPS is associated with OS among patients with nonsquamous NSCLC, but not among patients with squamous NSCLC.