{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,2]],"date-time":"2026-08-02T23:51:26Z","timestamp":1785714686291,"version":"3.56.0"},"reference-count":41,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2025,2,22]],"date-time":"2025-02-22T00:00:00Z","timestamp":1740182400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,2,22]],"date-time":"2025-02-22T00:00:00Z","timestamp":1740182400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"NIHR Pre-doctoral Fellowship","award":["NIHR301098"],"award-info":[{"award-number":["NIHR301098"]}]},{"name":"Wellcome Trust - Royal Society Sir Henry Dale Fellowship","award":["218554\/Z\/19\/Z"],"award-info":[{"award-number":["218554\/Z\/19\/Z"]}]},{"name":"NIHR Advanced Fellowship","award":["NIHR300593"],"award-info":[{"award-number":["NIHR300593"]}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["BMC Med Res Methodol"],"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:sec>\n            <jats:title>Background<\/jats:title>\n            <jats:p>Classical approaches to subgroup analysis in randomised controlled trials (RCTs) to identify heterogeneous treatment effects (HTEs) involve testing the interaction between each pre-specified possible treatment effect modifier and the treatment effect. However, individual significant interactions may not always yield clinically actionable subgroups, particularly for continuous covariates. Non-parametric causal machine learning approaches are flexible alternatives for estimating HTEs across many possible treatment effect modifiers in a single analysis.<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Methods<\/jats:title>\n            <jats:p>We conducted a secondary analysis of the VANISH RCT, which compared the early use of vasopressin with norepinephrine on renal failure-free survival for patients with septic shock at 28 days. We used classical (separate tests for interaction with Bonferroni correction), data-adaptive (hierarchical lasso regression), and non-parametric causal machine learning (causal forest) methods to analyse HTEs for the primary outcome of being alive at 28 days. Causal forests comprise honest causal trees, which use sample splitting to determine tree splits and estimate treatment effects separately. The modal initial (root) splits of the causal forest were extracted, and the mean value was used as a threshold to partition the population into subgroups with different treatment effects.<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Results<\/jats:title>\n            <jats:p>All three models found evidence of HTE with serum potassium levels. Univariable logistic regression OR 0.435 (95%CI [0.270, 0.683]. <jats:italic>p<\/jats:italic>\u2009=\u20090.0004), hierarchical lasso logistic regression standardised OR: 0.604 (95% CI 0.259, 0.701), lambda\u2009=\u20090.0049. Hierarchical lasso kept the interaction between the treatment and serum potassium, sodium level, minimum temperature, platelet count and presence of ischemic heart disease. The causal forest approach found some evidence of HTE (<jats:italic>p<\/jats:italic>\u2009=\u20090.124). When extracting root splits, the modal split was on serum potassium (mean applied threshold of 4.68 mmol\/L). When dividing the patient population into subgroups based on the mean initial root threshold, risk differences in being alive at 28 days were 0.069 (95%CI [-0.032, 0.169]) and \u2212\u20090.257 (95%CI [-0.368, -0.146]) with serum potassium\u2009\u2264\u20094.68 and &gt;\u20094.68 respectively.<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Conclusions<\/jats:title>\n            <jats:p>The causal forest agreed with the data-adaptive and classical method of subgroup analysis in identifying HTE by serum potassium. Whilst classical and data-adaptive methods may identify sources of HTE, they do not immediately suggest subgroup splits which are clinically actionable. The extraction of root splits in causal forests is a novel approach to obtaining data-derived subgroups, to be further investigated.<\/jats:p>\n          <\/jats:sec>","DOI":"10.1186\/s12874-025-02489-2","type":"journal-article","created":{"date-parts":[[2025,2,22]],"date-time":"2025-02-22T04:39:51Z","timestamp":1740199191000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":8,"title":["Application of causal forests to randomised controlled trial data to identify heterogeneous treatment effects: a case study"],"prefix":"10.1186","volume":"25","author":[{"given":"Eleanor","family":"Van Vogt","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anthony C.","family":"Gordon","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Karla","family":"Diaz-Ordaz","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Suzie","family":"Cro","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,2,22]]},"reference":[{"key":"2489_CR1","doi-asserted-by":"crossref","unstructured":"Rubin DB. Causal Inference Using Potential Outcomes. https:\/\/doi.org\/101198\/016214504000001880. 2011;100:322\u201331.","DOI":"10.1198\/016214504000001880"},{"key":"2489_CR2","doi-asserted-by":"publisher","first-page":"354","DOI":"10.7326\/M16-1756","volume":"166","author":"S Basu","year":"2017","unstructured":"Basu S, Sussman JB, Hayward RA. Detecting heterogeneous treatment effects to guide personalized blood pressure treatment a modeling study of randomized clinical trials. Ann Intern Med. 2017;166:354\u201360.","journal-title":"Ann Intern Med"},{"key":"2489_CR3","doi-asserted-by":"publisher","first-page":"2049","DOI":"10.1056\/NEJMsb2005114","volume":"382","author":"EJ Emanuel","year":"2020","unstructured":"Emanuel EJ, Persad G, Upshur R, Thome B, Parker M, Glickman A, et al. Fair allocation of Scarce Medical resources in the time of Covid-19. N Engl J Med. 2020;382:2049\u201355.","journal-title":"N Engl J Med"},{"key":"2489_CR4","first-page":"190","volume":"130","author":"R Pirracchio","year":"2019","unstructured":"Pirracchio R, Gropper MA. Heterogeneity in Intensive CareLow. Severity Does Not Mean Low Risk! Anesthesiology. 2019;130:190\u20131.","journal-title":"Severity Does Not Mean Low Risk! Anesthesiology"},{"key":"2489_CR5","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s12913-020-05133-5","volume":"20","author":"K Kaier","year":"2020","unstructured":"Kaier K, Heister T, Wolff J, Wolkewitz M. Mechanical ventilation and the daily cost of ICU care. BMC Health Serv Res. 2020;20:1\u20135.","journal-title":"BMC Health Serv Res"},{"key":"2489_CR6","doi-asserted-by":"publisher","first-page":"437","DOI":"10.1093\/jamia\/ocz211","volume":"27","author":"ZM Ibrahim","year":"2020","unstructured":"Ibrahim ZM, Wu H, Hamoud A, Stappen L, Dobson RJB, Agarossi A. On classifying sepsis heterogeneity in the ICU: insight using machine learning. J Am Med Inf Assoc. 2020;27:437.","journal-title":"J Am Med Inf Assoc"},{"key":"2489_CR7","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s12879-018-3515-9","volume":"18","author":"C Flateau","year":"2018","unstructured":"Flateau C, Le Bel J, Tubiana S, Blanc FX, Choquet C, Rammaert B, et al. High heterogeneity in community-acquired pneumonia inclusion criteria: does this impact on the validity of the results of randomized controlled trials? BMC Infect Dis. 2018;18:1\u20139.","journal-title":"BMC Infect Dis"},{"key":"2489_CR8","doi-asserted-by":"publisher","first-page":"639","DOI":"10.1016\/j.ccm.2014.08.004","volume":"35","author":"CM Shaver","year":"2014","unstructured":"Shaver CM, Bastarache JA. Clinical and biological heterogeneity in ARDS: Direct versus Indirect Lung Injury. Clin Chest Med. 2014;35:639.","journal-title":"Clin Chest Med"},{"key":"2489_CR9","doi-asserted-by":"publisher","first-page":"51","DOI":"10.1186\/s13054-017-1910-z","volume":"21","author":"JL Vincent","year":"2017","unstructured":"Vincent JL. The coming era of precision medicine for intensive care. Crit Care. 2017;21:51\u20134.","journal-title":"Crit Care"},{"key":"2489_CR10","doi-asserted-by":"publisher","first-page":"659","DOI":"10.1016\/S2213-2600(18)30279-0","volume":"6","author":"JG Laffey","year":"2018","unstructured":"Laffey JG, Kavanagh BP. Negative trials in critical care: why most research is probably wrong. Lancet Respir Med. 2018;6:659\u201360.","journal-title":"Lancet Respir Med"},{"key":"2489_CR11","first-page":"221","volume":"132","author":"DI Sessler","year":"2020","unstructured":"Sessler DI. Negative trials, and what to do with Them?First. Stop Call Them Negat Anesthesiology. 2020;132:221\u20134.","journal-title":"Stop Call Them Negat Anesthesiology"},{"key":"2489_CR12","doi-asserted-by":"publisher","first-page":"136","DOI":"10.1002\/sim.7064","volume":"36","author":"I Lipkovich","year":"2017","unstructured":"Lipkovich I, Dmitrienko A, D\u2019Agostino RB. Tutorial in biostatistics: data-driven subgroup identification and analysis in clinical trials. Stat Med. 2017;36:136\u201396.","journal-title":"Stat Med"},{"key":"2489_CR13","doi-asserted-by":"publisher","first-page":"2189","DOI":"10.1056\/NEJMsr077003","volume":"357","author":"R Wang","year":"2007","unstructured":"Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in Medicine \u2014 Reporting of Subgroup analyses in clinical trials. N Engl J Med. 2007;357:2189\u201394.","journal-title":"N Engl J Med"},{"key":"2489_CR14","doi-asserted-by":"publisher","first-page":"2327","DOI":"10.1093\/annonc\/mdx410","volume":"28","author":"MR Spears","year":"2017","unstructured":"Spears MR, James ND, Sydes MR. Thursday\u2019s child has far to go\u2019-interpreting subgroups and the STAMPEDE trial. Ann Oncol. 2017;28:2327\u201330.","journal-title":"Ann Oncol"},{"key":"2489_CR15","doi-asserted-by":"publisher","first-page":"65","DOI":"10.1093\/aje\/kww165","volume":"185","author":"MS Schuler","year":"2017","unstructured":"Schuler MS, Rose S. Targeted maximum likelihood estimation for causal inference in Observational studies. Am J Epidemiol. 2017;185:65\u201373.","journal-title":"Am J Epidemiol"},{"key":"2489_CR16","doi-asserted-by":"publisher","first-page":"C1","DOI":"10.1111\/ectj.12097","volume":"21","author":"V Chernozhukov","year":"2018","unstructured":"Chernozhukov V, Chetverikov D, Demirer M, Duflo E, Hansen C, Newey W, et al. Double\/debiased machine learning for treatment and structural parameters. Econom J. 2018;21:C1\u201368.","journal-title":"Econom J"},{"key":"2489_CR17","doi-asserted-by":"publisher","first-page":"37","DOI":"10.1353\/obs.2019.0001","volume":"5","author":"S Athey","year":"2019","unstructured":"Athey S, Wager S. Estimating Treatment effects with causal forests: an application. Obs Stud. 2019;5:37\u201351.","journal-title":"Obs Stud"},{"key":"2489_CR18","doi-asserted-by":"publisher","first-page":"4156","DOI":"10.1073\/pnas.1804597116","volume":"116","author":"SR K\u00fcnzel","year":"2019","unstructured":"K\u00fcnzel SR, Sekhon JS, Bickel PJ, Yu B. Metalearners for estimating heterogeneous treatment effects using machine learning. Proc Natl Acad Sci U S A. 2019;116:4156\u201365.","journal-title":"Proc Natl Acad Sci U S A"},{"key":"2489_CR19","doi-asserted-by":"publisher","first-page":"1228","DOI":"10.1080\/01621459.2017.1319839","volume":"113","author":"S Wager","year":"2018","unstructured":"Wager S, Athey S. Estimation and inference of Heterogeneous Treatment effects using Random forests. J Am Stat Assoc. 2018;113:1228\u201342.","journal-title":"J Am Stat Assoc"},{"key":"2489_CR20","unstructured":"Sadique Z, Grieve R, Diaz-Ordaz K, Mouncey P, Lamontagne F, O\u2019Neill S. A Machine-Learning Approach for Estimating Subgroup- and Individual-Level Treatment Effects: An Illustration Using the 65 Trial: https:\/\/doi.org\/101177\/0272989X221100717. 2022;:0272989X2211007."},{"key":"2489_CR21","doi-asserted-by":"publisher","first-page":"75","DOI":"10.21037\/atm.2017.02.05","volume":"5","author":"Z Zhang","year":"2017","unstructured":"Zhang Z, Murtagh F, Poucke S, Van, Lin S, Lan P. Hierarchical cluster analysis in clinical research with heterogeneous study population: highlighting its visualization with R. Ann Transl Med. 2017;5:75\u201375.","journal-title":"Ann Transl Med"},{"key":"2489_CR22","doi-asserted-by":"publisher","first-page":"119","DOI":"10.21037\/atm.2018.01.24","volume":"6","author":"Z Zhang","year":"2018","unstructured":"Zhang Z, Abarda A, Contractor AA, Wang J, Dayton CM. Exploring heterogeneity in clinical trials with latent class analysis. Ann Transl Med. 2018;6:119\u2013119.","journal-title":"Ann Transl Med"},{"key":"2489_CR23","unstructured":"Chernozhukov V, Demirer M, Duflo E, Athey S, Buchinsky M, Chetverikov D, Generic machine learning inference on heterogenous treatment effects in randomized experiments, with an application to immunization in india. 2022."},{"key":"2489_CR24","doi-asserted-by":"publisher","first-page":"509","DOI":"10.1001\/jama.2016.10485","volume":"316","author":"AC Gordon","year":"2016","unstructured":"Gordon AC, Mason AJ, Thirunavukkarasu N, Perkins GD, Cecconi M, Cepkova M, et al. Effect of early vasopressin vs norepinephrine on kidney failure in patients with septic shock: the VANISH Randomized Clinical Trial. JAMA. 2016;316:509\u201318.","journal-title":"JAMA"},{"key":"2489_CR25","doi-asserted-by":"crossref","unstructured":"Lim M, Hastie T. Learning Interactions via Hierarchical Group-Lasso Regularization. http:\/\/dx.doi.org\/101080\/106186002014938812. 2015;24:627\u201354.","DOI":"10.1080\/10618600.2014.938812"},{"key":"2489_CR26","unstructured":"Liu K, Markovic J, Tibshirani R. More powerful post-selection inference, with application to the Lasso."},{"key":"2489_CR27","unstructured":"Causal forest \u2014 causal_forest \u2022 grf. https:\/\/grf-labs.github.io\/grf\/reference\/causal_forest.html. Accessed 12 Oct 2022."},{"key":"2489_CR28","unstructured":"Zhan R, Ren Z, Athey S, Zhou Z. Policy Learning with Adaptively Collected Data. 2021."},{"key":"2489_CR29","doi-asserted-by":"crossref","unstructured":"Athey S, Tibshirani J, Wager S, Generalized. random forests. https:\/\/doi.org\/101214\/18-AOS1709. 2019;47:1148\u201378.","DOI":"10.1214\/18-AOS1709"},{"key":"2489_CR30","doi-asserted-by":"publisher","first-page":"112","DOI":"10.1093\/bioinformatics\/btr597","volume":"28","author":"DJ Stekhoven","year":"2012","unstructured":"Stekhoven DJ, B\u00fchlmann P. MissForest\u2014non-parametric missing value imputation for mixed-type data. Bioinformatics. 2012;28:112\u20138.","journal-title":"Bioinformatics"},{"key":"2489_CR31","doi-asserted-by":"publisher","first-page":"2074","DOI":"10.1002\/sim.8086","volume":"38","author":"TP Morris","year":"2019","unstructured":"Morris TP, White IR, Crowther MJ. Using simulation studies to evaluate statistical methods. Stat Med. 2019;38:2074\u2013102.","journal-title":"Stat Med"},{"key":"2489_CR32","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13613-019-0509-8","volume":"9","author":"F D\u00e9pret","year":"2019","unstructured":"D\u00e9pret F, Peacock WF, Liu KD, Rafique Z, Rossignol P, Legrand M. Management of hyperkalemia in the acutely ill patient. Ann Intensive Care. 2019;9:1\u201316.","journal-title":"Ann Intensive Care"},{"key":"2489_CR33","doi-asserted-by":"publisher","first-page":"794","DOI":"10.21037\/atm.2020.04.24","volume":"8","author":"R Shi","year":"2020","unstructured":"Shi R, Hamzaoui O, Vita N, De, Monnet X, Teboul J-L. Vasopressors in septic shock: which, when, and how much? Ann Transl Med. 2020;8:794\u2013794.","journal-title":"Ann Transl Med"},{"key":"2489_CR34","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1016\/j.ahj.2022.07.003","volume":"253","author":"U Winsl\u00f8w","year":"2022","unstructured":"Winsl\u00f8w U, Sakthivel T, Zheng C, Bosselmann H, Haugan K, Bruun N, et al. Targeted potassium levels to decrease arrhythmia burden in high risk patients with cardiovascular diseases (POTCAST): study protocol for a randomized controlled trial. Am Heart J. 2022;253:59\u201366.","journal-title":"Am Heart J"},{"key":"2489_CR35","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13054-019-2446-1","volume":"23","author":"S Santhakumaran","year":"2019","unstructured":"Santhakumaran S, Gordon A, Prevost AT, O\u2019Kane C, McAuley DF, Shankar-Hari M. Heterogeneity of treatment effect by baseline risk of mortality in critically ill patients: re-analysis of three recent sepsis and ARDS randomised controlled trials. Crit Care. 2019;23:1\u20139.","journal-title":"Crit Care"},{"key":"2489_CR36","doi-asserted-by":"publisher","DOI":"10.1080\/00273171.2021.1994364\/SUPPL_FILE\/HMBR_A_1994364_SM3819.DOCX","author":"Y Suk","year":"2021","unstructured":"Suk Y, Kang H. Tuning Random forests for causal inference under cluster-level unmeasured confounding. Multivar Behav Res. 2021. https:\/\/doi.org\/10.1080\/00273171.2021.1994364\/SUPPL_FILE\/HMBR_A_1994364_SM3819.DOCX.","journal-title":"Multivar Behav Res"},{"key":"2489_CR37","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3310\/hta25140","volume":"25","author":"PR Mouncey","year":"2021","unstructured":"Mouncey PR, Richards-Belle A, Thomas K, Harrison DA, Zia Sadique M, Grieve RD, et al. Reduced exposure to vasopressors through permissive hypotension to reduce mortality in critically ill people aged 65 and over: the 65 RCT. Health Technol Assess (Rockv). 2021;25:1\u201390.","journal-title":"Health Technol Assess (Rockv)"},{"key":"2489_CR38","unstructured":"Medicines Agency E. Committee for Medicinal Products for Human Use (CHMP) Guideline on the investigation of subgroups in confirmatory clinical trials. 2019."},{"key":"2489_CR39","doi-asserted-by":"crossref","unstructured":"Begoli E, Bhattacharya T, Kusnezov D. The need for uncertainty quantification in machine-assisted medical decision making. Nat Mach Intell. 2019;1.","DOI":"10.1038\/s42256-018-0004-1"},{"key":"2489_CR40","first-page":"2121","volume":"18","author":"M Maadi","year":"2021","unstructured":"Maadi M, Khorshidi HA, Aickelin U. A review on Human\u2013AI Interaction in Machine Learning and insights for Medical Applications. Int J Environ Res Public Health 2021. 2021;18:18:2121.","journal-title":"Int J Environ Res Public Health 2021"},{"key":"2489_CR41","doi-asserted-by":"publisher","first-page":"1904","DOI":"10.1177\/09622802221104238","volume":"31","author":"T Carpenito","year":"2022","unstructured":"Carpenito T, Manjourides J. MISL: multiple imputation by super learning. Stat Methods Med Res. 2022;31:1904\u201315.","journal-title":"Stat Methods Med Res"}],"container-title":["BMC Medical Research Methodology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12874-025-02489-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12874-025-02489-2\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12874-025-02489-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,2,22]],"date-time":"2025-02-22T04:40:01Z","timestamp":1740199201000},"score":1,"resource":{"primary":{"URL":"https:\/\/bmcmedresmethodol.biomedcentral.com\/articles\/10.1186\/s12874-025-02489-2"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,2,22]]},"references-count":41,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["2489"],"URL":"https:\/\/doi.org\/10.1186\/s12874-025-02489-2","relation":{},"ISSN":["1471-2288"],"issn-type":[{"value":"1471-2288","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,2,22]]},"assertion":[{"value":"28 June 2023","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"3 February 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"22 February 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable for the statistical methods evaluation. The data for the case study comes from the VANISH trial which was conducted in full conformity with the Declaration of Helsinki (1996), and the principles of Good Clinical Practice (ICH GCP E6). All participants provided informed consent and ethical approval for the VANISH study, which was granted by the South Central - Oxford A REC on 30\/01\/2012 (ref: 12\/SC\/0014).","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"Not applicable.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent for publication"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"50"}}