Skip to content
DraconiticResearch
Search Sources API
194,161 studies31 sourcesUpdated 34 min ago
Peer-reviewedCC BYInternational journal of population data science

The need for fully-effective federated analytics of data sources for clinical trials.

  • Stella Maris Fabiane1
  • Sharon B Love2
  • David Fisher3
  • Ian R White4
  • Jayne F Tierney1
  • Catherine Dampney5
  • Folkert W Asselbergs
  • Philip R Quinlan6
  • Macey L Murray7
  • Matthew R Sydes8
8 affiliations
  1. 1UCL Innovative Clinical Trials Unit, Institute of Clinical Trials and Methodology, UCL, London, WC1V 6LJ, UK; ORCID: 0000-0001-5957-7275.
  2. 2UCL Innovative Clinical Trials Unit, Institute of Clinical Trials and Methodology, UCL, London, WC1V 6LJ, UK; ORCID: 0000-0002-6695-5390.
  3. 3UCL Innovative Clinical Trials Unit, Institute of Clinical Trials and Methodology, UCL, London, WC1V 6LJ, UK; ORCID: 0000-0002-2512-2296.
  4. 4UCL Innovative Clinical Trials Unit, Institute of Clinical Trials and Methodology, UCL, London, WC1V 6LJ, UK; ORCID: 0000-0002-6718-7661.
  5. 5Kent, Medway and Sussex Secure Data Environment, Kent, Medway and Sussex Secure Data Environment - NHS England Digital;ORCID: 0009-0004-1807-3758.
  6. 6School of Medicine, NIHR Nottingham Biomedical Research Centre, University of Nottingham, NG7 2UH, UK; ORCID: 0000-0002-3012-6646.
  7. 7UCL Innovative Clinical Trials Unit, Institute of Clinical Trials and Methodology, UCL, London, WC1V 6LJ, UK; Transforming Data for Trials programme, Health Data Research UK (HDR UK), London, NW1 2BE, UK; ORCID: 0000-0001-6418-0854.
  8. 8UCL Innovative Clinical Trials Unit, Institute of Clinical Trials and Methodology, UCL, London, WC1V 6LJ, UK; Data for R&D, "Data Access & Partnerships, Technology Data & Digital", NHS England, London, SE1 8UG, UK; ORCID: 0000-0002-9323-1371.
Read full textPDFDOICite

Abstract

Introduction

Clinical trials are usually analysed in a single environment allowing for flexible analysis including adjustment or stratification by subgroup: 'one-stage' analysis of individual-level data. Health systems datasets, often distributed across geography and providers, can streamline clinical trials. Data are increasingly accessible in secure data environments (SDEs). Future trial analyses may involve working across multiple SDEs. Row-level data and identifiable data often cannot leave, requiring a 'two-stage approach', where summary data from each SDE are meta-analysed.Objective

To quantify the potential loss of precision and concomitant increases in required sample sizes, and to make recommendations for trial design and conduct, if clinical trial data are split across silos (e.g. SDEs).Methods

Simulations used data from clinical trials in breast cancer, tuberculosis and prostate cancer with time-to-event, binary and continuous outcome measures. Silos were mimicked by 1000 random partitions into 2, 4, 10 and 25 equal silos and 4 unequal silos proportionate to the UK nations. Data were analysed as if the data could be pooled ignoring silo, pooled accounting for silo (one-stage) or not pooled (two-stage). Estimates and standard errors were presented graphically.Results

For all three outcome measure types, standard errors increased while point estimates spread out as more silos were introduced. Small biases occurred for binary and time-to-event outcomes. This did not always appreciably reduce efficiency. However, in one example with time-to-event data and the largest number of silos, a near-doubling of sample size would have been required to pre-emptively offset the loss of efficiency.Conclusion

Any need to use a two-stage analysis approach has a negative effect compared to doing a one-stage analysis. Technical and data governance solutions to support one-stage analyses are recommended.

Related

  1. Multivariate Analysis of Data From Crossover Clinical Trials
    Yi Jiang, Vernon M. Chinchilli
    Biometrical journal. Biometrische ZeitschriftPeer-reviewedCC BYFull textStatisticsEurope PMC
  2. Sample size determination for randomized decentralized clinical trials with longitudinal outcomes
    Chunyi Zhang, Ying Yuan, Ruitao Lin +1
    BiometricsPeer-reviewedPublisher termsFull textStatisticsPMC
  3. A Multi‐Task Learning Approach for Identifying Prognostic Variables Across Multiple Historical Clinical Trials
    Xueying Liu, Huijing Jiang, Jennie Zhang +1
    Statistics in medicinePeer-reviewedCC BY-NC-NDFull textStatisticsEurope PMC
  4. Automated, Efficient, and Model-Free Inference for Randomized Clinical Trials via Data-Driven Covariate Adjustment
    Kelly Van Lancker, Ivan Matos Diaz, Stijn Vansteelandt
    The American StatisticianPeer-reviewedLicence not statedFull text15 citedStatisticsOpenAlex
  5. Valid Data‐Driven Subgroup Selection With Uniform Guarantees in Clinical Trials
    Manuel M. Müller, Björn Bornkamp, Frank Bretz +7
    Statistics in medicinePeer-reviewedCC BYFull textStatisticsEurope PMC

Cite

BibTeXRIS
BibTeX
@article{fabiane2026need,
  title = {{The need for fully-effective federated analytics of data sources for clinical trials.}},
  author = {Stella Maris Fabiane and Sharon B Love and David Fisher and Ian R White and Jayne F Tierney and Catherine Dampney and Folkert W Asselbergs and Philip R Quinlan and Macey L Murray and Matthew R Sydes},
  journal = {International journal of population data science},
  year = {2026},
  month = aug,
  doi = {10.23889/ijpds.v11i1.3201},
  url = {https://europepmc.org/article/MED/42899104},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - The need for fully-effective federated analytics of data sources for clinical trials.
AU  - Stella Maris Fabiane
AU  - Sharon B Love
AU  - David Fisher
AU  - Ian R White
AU  - Jayne F Tierney
AU  - Catherine Dampney
AU  - Folkert W Asselbergs
AU  - Philip R Quinlan
AU  - Macey L Murray
AU  - Matthew R Sydes
PY  - 2026
DA  - 2026/08/27
JO  - International journal of population data science
DO  - 10.23889/ijpds.v11i1.3201
UR  - https://europepmc.org/article/MED/42899104
AB  - Introduction  Clinical trials are usually analysed in a single environment allowing for flexible analysis including adjustment or stratification by subgroup: 'one-stage' analysis of individual-level data. Health systems datasets, often distributed across geography and providers, can streamline clinical trials. Data are increasingly accessible in secure data environments (SDEs). Future trial analyses may involve working across multiple SDEs. Row-level data and identifiable data often cannot leave, requiring a 'two-stage approach', where summary data from each SDE are meta-analysed.Objective  To quantify the potential loss of precision and concomitant increases in required sample sizes, and to make recommendations for trial design and conduct, if clinical trial data are split across silos (e.g. SDEs).Methods  Simulations used data from clinical trials in breast cancer, tuberculosis and prostate cancer with time-to-event, binary and continuous outcome measures. Silos were mimicked by 1000 random partitions into 2, 4, 10 and 25 equal silos and 4 unequal silos proportionate to the UK nations. Data were analysed as if the data could be pooled ignoring silo, pooled accounting for silo (one-stage) or not pooled (two-stage). Estimates and standard errors were presented graphically.Results  For all three outcome measure types, standard errors increased while point estimates spread out as more silos were introduced. Small biases occurred for binary and time-to-event outcomes. This did not always appreciably reduce efficiency. However, in one example with time-to-event data and the largest number of silos, a near-doubling of sample size would have been required to pre-emptively offset the loss of efficiency.Conclusion  Any need to use a two-stage analysis approach has a negative effect compared to doing a one-stage analysis. Technical and data governance solutions to support one-stage analyses are recommended.
SN  - 2399-4908
ER  -