Emulation of the KEYNOTE-189 Trial Using Electronic Health Records

This is a non-randomized observational study that is being conducted as a part of the Coalition to Advance Real-World Evidence through Randomized Controlled Trial Emulation (CARE) Initiative, a multi-stakeholder collaboration led by Aetion. The study is intended to emulate the trial listed below/above. Although many features of the trial cannot be directly replicated in real-world data sources, key elements of the study design, including outcomes, exposures, and inclusion/exclusion criteria, were selected to proxy those features from the trial as closely as possible. Investigators assume that the trial provides the reference standard treatment effect estimate and that failure to replicate trial findings is indicative of the inadequacy of the real-world data source for replication for a range of possible reasons and does not provide information on the validity of the original trial finding.

Inclusion criteria

• Has evidence in electronic health record indicating a 'lung' tumor site, non-squamous morphology (e.g., adenocarcinoma, large cell, etc), and a stage IV or metastatic disease in the form of diagnosis codes or derived variables from physician notes or linked data from a tumor registry. There is no evidence of other lung cancer types (i.e., mesothelioma, small cell, etc.).
• Does not have evidence of HGVS codes in health record indicating EGFR or ALK mutations.
• Has no evidence of use of guideline-recommended systemic cancer therapy* for NSCLC on or after first date associated with stage IV (or metastatic) disease and prior to index treatment.

Exclusion criteria

• Has evidence of squamous cell lung cancer morphology in the form of diagnosis codes or derived variables from physician notes or linked information from a tumor registry.
• Has evidence of systemic cancer therapy utilization on or after first date associated with stage IV (or metastatic) disease and prior to index treatment. Patient has no evidence of biologic therapy at any time prior to the first record indicating an exposure of interest.
• Has evidence in medical record or tumor registry of primary malignancies beyond the lung tissue, except basal cell carcinoma of the skin, superficial bladder cancer, squamous cell carcinoma of the skin, in situ cervical cancer, or other in situ cancers.
• Has evidence of central nervous system (CNS) metastases and/or carcinomatous meningitis in the 2 weeks prior to initiation.
• Has a record of use of one of the following agents in the 2 years prior to the index date: methotrexate, sulfasalazine, hydroxychloroquine, leflunomide, azathioprine, etanercept, adalimumab, infliximab, certolizumab, golimumab, anakinra, tocilizumab, sarilumumab, abatacept, rituximab, tofacitinib, barivitinib, upadacitinib, mycophenolate.
• Has at least two records 60 days apart indicating use of one of the following agents in the 3 months prior to the index date: prednisone, prednisolone, methylprednisolone, dexamethasone, hydrocortisone.
• Has evidence of use of one of the following agents prior to the index date: pembrolizumab, nivolumab, cemiplimab, atezolizumab, durvalumab.