{"paper":{"title":"Digital Twins as Synthetic Controls in Single-Arm Trials","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Digital twins from machine learning models can serve as synthetic controls in single-arm clinical trials","cross_cats":["cs.LG","stat.ML"],"primary_cat":"stat.AP","authors_text":"Aaron M. Smith, Daniele Bertolini, Franklin Fuller, Jonathan R. Walsh, Run Zhuang","submitted_at":"2026-05-12T23:58:48Z","abstract_excerpt":"Single-arm trials are an important study design for evaluating drug efficacy and safety without enrolling patients into a control arm. Although they do not provide the gold-standard evidence of randomized controlled trials, they are increasingly used in clinical development as they offer an efficient, ethical, and practical alternative. A wide variety of approaches can be used to construct control comparators and estimate treatment effects, from fixed comparators informed by clinical knowledge to data-based and model-based patient-level comparators, also known as synthetic controls. Powerful a"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"outcome-model-based synthetic control arms are an important tool for single-arm trials... we focus on digital twins, personalized predictions of disease progression generated from machine learning models trained on historical datasets, which naturally leverage these flexible approaches.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That machine learning models trained on historical datasets will produce accurate and unbiased predictions of disease progression for patients in the current single-arm trial, even when populations differ in unmeasured ways.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Digital twins from outcome models trained on historical data can function as robust synthetic controls in single-arm trials, supported by doubly robust estimators, power formulas, and reanalyses in ALS and Huntington's disease.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Digital twins from machine learning models can serve as synthetic controls in single-arm clinical trials","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"0061f64b8e9b3c52b5934c475fbcf32772e07790e19d73a32e048f02c0ca6172"},"source":{"id":"2605.12832","kind":"arxiv","version":1},"verdict":{"id":"b899e881-3609-4e36-bacc-296c3a10309f","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:16:27.084601Z","strongest_claim":"outcome-model-based synthetic control arms are an important tool for single-arm trials... we focus on digital twins, personalized predictions of disease progression generated from machine learning models trained on historical datasets, which naturally leverage these flexible approaches.","one_line_summary":"Digital twins from outcome models trained on historical data can function as robust synthetic controls in single-arm trials, supported by doubly robust estimators, power formulas, and reanalyses in ALS and Huntington's disease.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That machine learning models trained on historical datasets will produce accurate and unbiased predictions of disease progression for patients in the current single-arm trial, even when populations differ in unmeasured ways.","pith_extraction_headline":"Digital twins from machine learning models can serve as synthetic controls in single-arm clinical trials"},"references":{"count":29,"sample":[{"doi":"","year":2020,"title":"Hern´ an and James M","work_id":"003f0819-8194-4781-95da-24313d961467","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2015,"title":"Imbens and Donald B","work_id":"4d38d767-43fe-48ab-920e-1c5c80c4c9bc","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1974,"title":"Donald B. Rubin. Estimating causal effects of treatments in randomized and nonrandomized studies.Journal of Educational Psychology, 66(5):688–701, 1974","work_id":"934e68ca-7b52-4460-bfab-dbb3a7096d33","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Long story short: Omitted variable bias in causal machine learning, 2024","work_id":"afe797be-259b-433a-9c7a-45778da3c953","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2014,"title":"Placebo effects: from the neurobiological paradigm to translational implica- tions.Neuron, 84(3):623–637, November 2014","work_id":"27d63d1f-8cf0-4538-8924-dca4c0dd9715","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":29,"snapshot_sha256":"04b897e852f462ee3884d7e4619b448788241cdc6596735fb03deaf68a60a42f","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"2959dfa24d3a8426b4ed0b80771c5534305296bdb7d64fe99a46292b99fe9293"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}