Article
Oncology
Ashley M. Hopkins, Natansh D. Modi, Ahmad Y. Abuhelwa, Ganessan Kichenadasse, Nicole M. Kuderer, Gary H. Lyman, Michael D. Wiese, Ross A. Mckinnon, Frank W. Rockhold, Aaron Mann, Andrew Rowland, Michael J. Sorich
Summary: In a quality improvement study, it was found that there was significant variability in the completeness of key data variables and supporting documents within the provided IPD packages from industry-sponsored clinical trials. To enhance the data sharing ecosystem, key areas for improvement include ensuring eligibility for IPD sharing, transparent access to eligible IPD, and ensuring that IPD packages meet a standard of utility and completeness.
Article
Ophthalmology
Declan C. Murphy, Mo Al-Zubaidy, Noemi Lois, Neil Scott, David H. Steel
Summary: The effect of symptom duration on outcomes in people undergoing surgery for idiopathic full-thickness macular holes (iFTMHs) was investigated using an individual participant data (IPD) study. The study found that symptom duration was associated with surgical outcomes and postoperative visual acuity.
Article
Medicine, Research & Experimental
Lisa Irvine, Jennifer Kirsty Burton, Myzoon Ali, Terence J. Quinn, Claire Goodman
Summary: Creating a project-specific repository of individual participant data (IPD) from trials conducted in care homes can improve understanding of the under-researched population and inform the development of a national minimum dataset for care homes. By cleaning and standardizing the collected IPD, pooled data analysis can enhance future practice, research, and policy regarding care homes.
Review
Medicine, General & Internal
Christian Ohmann, David Moher, Maximilian Siebert, Edith Motschall, Florian Naudet
Summary: The study found that while there is a high willingness to share IPD data from clinical trials, the actual data-sharing rates are suboptimal, and journals have poor to moderate enforcement of data-sharing policies. When data is requested, it is more often for secondary analysis and meta-analysis, rather than re-analysis. Studies on the real impact of data-sharing are rare and often use surrogate metrics like citation metrics.
Letter
Medicine, General & Internal
Laura C. Esmail, Philipp Kapp, Rouba Assi, Julie Wood, Gabriela Regan, Philippe Ravaud, Isabelle Boutron
Summary: This study examines the consistency between the intent to share data reported in registries, publications, or preprints and the actual access to individual patient-level data from randomized clinical trials during the COVID-19 pandemic.
JAMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION
(2023)
Review
Urology & Nephrology
Sarah Burdett, D. J. Fisher, C. L. Vale, J. F. Tierney, N. W. Clarke, M. K. B. Parmar, C. N. Sternberg, M. Stoeckle, J. Lehmann, U. E. Studer, R. W. Sonntag, F. M. Torti, S. Groshen, A. V. Bono, P. J. Goebell, F. Cognetti, R. J. Cote, S. Groshen, L. Collette, C. N. Sternberg, A. I. Rolevich, A. G. Zhegalik
Summary: The systematic review and meta-analysis support the use of adjuvant cisplatin-based chemotherapy for muscle-invasive bladder cancer, showing significant improvement in overall and recurrence-free survival rates. The findings suggest that this treatment option is beneficial in improving outcomes for patients with this condition.
Article
Medicine, Research & Experimental
Stephanie R. Morain, Juli Bollinger, Kevin Weinfurt, Jeremy Sugarman
Summary: Sharing data in pragmatic clinical trials presents ethical challenges, such as waivers of informed consent, risks for patient-subjects and health systems, use of electronic health records, and limited control over sensitive data. These challenges raise questions about the suitability of traditional research ethics governance structures.
Article
Mathematical & Computational Biology
Francisco J. Diaz
Summary: Researchers propose a method called partial empirical Bayes to individualize treatment in N-of-1 trials, combining information from previous population crossover trials with individual patient data. They investigate statistical conditions, optimal treatment cycles, and the benefits of N-of-1 trials compared to traditional approaches. The study demonstrates the consistency and performance of estimators under common N-of-1 designs and shows the superiority of their approach over common clinical practices and individualization methods in some situations.
STATISTICS IN MEDICINE
(2021)
Article
Immunology
Anna L. Fournier, Laurent Hocqueloux, Dominique L. Braun, Karin J. Metzner, Roger D. Kouyos, Francois Raffi, Anais R. Briant, Esteban Martinez, Elisa De Lazzari, Eugenia Negredo, Bart Rijnders, Casper Rokx, Huldrych F. Guenthard, Jean-Jacques Parienti
Summary: This study aimed to identify independent factors associated with virological failure under dolutegravir monotherapy (DTG-m) and explored the impact of viral reservoir size on treatment failure. Among the 4 randomized controlled trials analyzed, HIV-1 DNA explained 80% of the effect size heterogeneity in virological failure.
OPEN FORUM INFECTIOUS DISEASES
(2022)
Article
Geriatrics & Gerontology
Michael J. DiStefano, G. Caleb Alexander, Daniel Polsky, Gerard F. Anderson
Summary: The survey found that approximately three-quarters of respondents were initially unfamiliar with aducanumab, but became less supportive of its approval once they were provided with information about its potential clinical and economic impact. The majority of respondents support restricting access to aducanumab to patients who are most likely to benefit from it, and are willing to enroll family members in clinical trials to further study the drug. The findings also indicate that if confirmatory trials fail, a significant majority of respondents believe that aducanumab should be withdrawn from the market. The median respondent is willing to pay $1-5 in higher Part B premiums to cover the cost of aducanumab.
JOURNAL OF THE AMERICAN GERIATRICS SOCIETY
(2022)
Article
Health Care Sciences & Services
Janharpreet Singh, Sandro Gsteiger, Lorna Wheaton, Richard D. Riley, Keith R. Abrams, Clare L. Gillies, Sylwia Bujkiewicz
Summary: This study develops a method for synthesizing data from single-arm trials and randomized controlled trials in network meta-analysis to estimate relative treatment effects. The study finds that incorporating single-arm trials in network meta-analysis may be useful in some situations where a treatment is not directly compared to other treatments.
BMC MEDICAL RESEARCH METHODOLOGY
(2022)
Article
Medicine, General & Internal
Jennie Louise, Amanda J. Poprzeczny, Andrea R. Deussen, Christina Vinter, Mette Tanvig, Dorte Moller Jensen, Annick Bogaerts, Roland Devlieger, Fionnuala M. McAuliffe, Kristina M. Renault, Emma Carlsen, Nina Geiker, Lucilla Poston, Annette Briley, Shakila Thangaratinam, Jodie M. Dodd
Summary: This study shows that dietary and/or lifestyle interventions in overweight and obese pregnant women do not alter the risk of early childhood obesity. Future research may need to focus on pre-conception and early childhood interventions.
Article
Behavioral Sciences
Eric Robinson, Ashleigh Haynes
Summary: In this study, it was found that reducing meal portion size generally led to a decrease in meal energy intake among most participants. There was little evidence to suggest that a subgroup of participants consistently resisted portion size reductions. The results suggest that portion size may be a universal driver of energy intake and downsizing food portions could be an effective intervention approach to reducing population-level energy intake.
Article
Psychiatry
Jo Annika Reins, Claudia Buntrock, Johannes Zimmermann, Simon Grund, Mathias Harrer, Dirk Lehr, Harald Baumeister, Kiona Weisel, Matthias Domhardt, Kotaro Imamura, Norito Kawakami, Viola Spek, Stephanie Nobis, Frank Snoek, Pim Cuijpers, Jan Philipp Klein, Steffen Moritz, David Daniel Ebert
Summary: The study showed that Internet-based interventions are effective in treating subthreshold depression and preventing major depression, especially for older individuals with a substantial symptom burden.
PSYCHOTHERAPY AND PSYCHOSOMATICS
(2021)
Article
Medicine, Research & Experimental
Naomi Attard, Nikki Totton, Katie Gillies, Beatriz Goulao
Summary: This study aimed to identify the methods, challenges, and suggestions in determining non-inferiority or equivalence margins. The survey found that most trials used evidence-based reviews or opinion-seeking methods to determine margins, and communication with doctors and patients posed a significant challenge. Clearer guidelines, increased stakeholder involvement, and improved communication tools were recommended to address these issues.
Article
Medicine, General & Internal
Richard D. Riley, Sofia Dias, Sarah Donegan, Jayne F. Tierney, Lesley A. Stewart, Orestis Efthimiou, David M. Phillippo
Summary: Network meta-analysis combines evidence from randomized trials to compare the efficacy of multiple treatments. Individual participant data (IPD) have potential advantages in network meta-analysis, providing more precise, reliable, and informative results, allowing treatment comparisons for individual patients and targeted populations based on their specific characteristics.
BMJ EVIDENCE-BASED MEDICINE
(2023)
Review
Health Care Sciences & Services
Mohammed T. Hudda, Lucinda Archer, Maarten van Smeden, Karel G. M. Moons, Gary S. Collins, Ewout W. Steyerberg, Charlotte Wahlich, Johannes B. Reitsma, Richard D. Riley, Ben Van Calster, Laure Wynants
Summary: The study aims to assess the improvement in reporting completeness of COVID-19 prediction models after peer review. The findings suggest that the reporting quality of preprints is poor and did not improve significantly after peer review, indicating that peer review had minimal effect on the completeness of reporting during the COVID-19 pandemic.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2023)
Review
Health Care Sciences & Services
Constanza L. Andaur Navarro, Johanna A. A. Damen, Maarten van Smeden, Toshihiko Takada, Steven W. J. Nijman, Paula Dhiman, Jie Ma, Gary S. Collins, Ram Bajpai, Richard D. Riley, Karel G. M. Moons, Lotty Hooft
Summary: This study aimed to summarize the research design, modeling strategies, and performance measures of clinical prediction models developed using machine learning techniques. A total of 152 studies were included, and it was found that most studies only reported the development of the models, without reporting sample size calculation, handling of missing values, and internal validation. Therefore, further improvement is needed in the methodological conduct and reporting standards of studies on machine learning-based prediction models.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2023)
Article
Health Care Sciences & Services
Alexander Pate, Richard D. Riley, Gary S. Collins, Maarten van Smeden, Ben Van Calster, Joie Ensor, Glen P. Martin
Summary: Multinomial logistic regression models are used to predict the risk of a categorical outcome with more than two categories. Researchers need to ensure that the number of participants is appropriate relative to the number of events and predictor variables for each category. This study proposes three criteria to determine the minimum required sample size, aiming to minimize overfitting, difference between observed and adjusted R-2 Nagelkerke, and ensure accurate estimation of overall risk. The criteria were evaluated through simulation study and applied to a worked example, with code provided for implementation in R and Stata.
STATISTICAL METHODS IN MEDICAL RESEARCH
(2023)
Article
Medicine, General & Internal
Tom Hughes, Richard Riley, Michael J. Callaghan, Jamie C. Sergeant
Summary: This study explored whether variables derived from periodic health examinations (PHE) are prognostic factors for indirect muscle injuries (IMIs) in elite football players. The results showed that, apart from age, most variables had limited prognostic value for injury risk prediction. The only variable that added prognostic value was a hamstring IMI occurring more than 12 months but less than 3 years prior to PHE.
Article
Medicine, General & Internal
Ella Flemyng, Theresa Helen Moore, Isabelle Boutron, Julian P. T. Higgins, Asbjorn Hrobjartsson, Camilla Hansen Nejstgaard, Kerry Dwan
Summary: A systematic review evaluates and combines all the empirical evidence from studies that meet specific criteria to answer a research question, assessing the risk of bias in the included studies to enhance confidence in the conclusions. Cochrane Reviews have used a risk of bias tool since 2008, and a new version, RoB 2, was introduced in 2019 to improve usability and reflect current understanding of bias. This paper discusses lessons learned from the phased implementation of RoB 2 and provides tips for systematic reviewers.
BMJ EVIDENCE-BASED MEDICINE
(2023)
Review
Health Care Sciences & Services
Constanza L. Andaur Navarro, Johanna A. A. Damen, Toshihiko Takada, Steven W. J. Nijman, Paula Dhiman, Jie Ma, Gary S. Collins, Ram Bajpai, Richard D. Riley, Karel G. M. Moons, Lotty Hooft
Summary: This study evaluated the presence and frequency of spin practices and poor reporting standards in studies that developed and/or validated clinical prediction models using supervised machine learning techniques. A total of 152 studies were included, and the results revealed the existence of spin practices and poor reporting standards in these studies, emphasizing the need for a tailored framework to enhance the reporting quality of prediction model studies.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2023)
Review
Health Care Sciences & Services
Paula Dhiman, Jie Ma, Constanza L. Andaur Navarro, Benjamin Speich, Garrett Bullock, Johanna A. A. Damen, Lotty Hooft, Shona Kirtley, Richard D. Riley, Ben Van Calster, Karel G. M. Moons, Gary S. Collins
Summary: This article conducted a systematic review on oncology-related studies that developed and validated prognostic models using machine learning. The findings revealed the presence of spin, i.e., overinterpretation of findings, in these studies. The inconsistent reporting and use of overly strong or leading words in the publications indicate the need for caution when reading and using prognostic models in oncology.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2023)
Article
Sport Sciences
Garrett S. Bullock, Patrick Ward, Franco M. Impellizzeri, Stefan Kluzek, Tom Hughes, Paula Dhiman, Richard D. Riley, Gary S. Collins
Summary: Regression or machine learning models used in sports medicine often suffer from poor methodology, incomplete reporting, and inadequate performance evaluation, leading to unreliable predictions and limited clinical usefulness. Thorough evaluation and open science practices are crucial for improving the validity and utility of these models, but they are currently lacking in the field.
Article
Mathematical & Computational Biology
Valentijn M. T. de Jong, Jeroen Hoogland, Karel G. M. Moons, Richard D. Riley, Tri-Long Nguyen, Thomas P. A. Debray
Summary: External validation of prediction models requires careful interpretation, as discrimination depends on both sample characteristics and generalizability of predictor coefficients. To resolve differences in discriminative ability across validation samples, we propose propensity-weighted measures of discrimination. Our methods account for case-mix differences and allow for fair comparisons of discriminative ability in the target population of interest, providing valuable insights for model updating strategies.
STATISTICS IN MEDICINE
(2023)
Article
Mathematical & Computational Biology
Alexander Pate, Matthew Sperrin, Richard D. Riley, Jamie C. Sergeant, Tjeerd Van Staa, Niels Peek, Mamas A. Mamas, Gregory Y. H. Lip, Martin O'Flaherty, Iain Buchan, Glen P. Martin
Summary: This study focuses on predicting the time until two survival outcomes have occurred and compares different analytical methods for multi-morbidity prognosis. The performance of these methods is evaluated through simulated data and a clinical example.
STATISTICS IN MEDICINE
(2023)
Article
Mathematical & Computational Biology
Richard D. Riley, Gary S. Collins
Summary: Clinical prediction models estimate an individual's risk of a particular health outcome. Many models are developed using small datasets, leading to instability in the model and its predictions. Researchers should examine instability at the model development stage and propose instability plots and measures to assess model reliability and inform critical appraisal, fairness, and validation requirements.
BIOMETRICAL JOURNAL
(2023)
Article
Mathematical & Computational Biology
J. Hoogland, T. P. A. Debray, M. J. Crowther, R. D. Riley, J. Inthout, J. B. Reitsma, A. H. Zwinderman
Summary: This study proposes a method that combines flexible parametric survival modeling and regularization to improve risk prediction models for time-to-event data. By introducing different penalty terms, the models can be regularized to enhance prediction accuracy and model performance.
BIOMETRICAL JOURNAL
(2023)
Article
Mathematical & Computational Biology
Sarah Booth, Sarwar I. Mozumder, Lucinda Archer, Joie Ensor, Richard D. Riley, Paul C. Lambert, Mark J. Rutherford
Summary: This article introduces a method called temporal recalibration to improve the calibration of prognostic models for new patients by accounting for trends in survival over time. The method involves estimating predictor effects using the full dataset and re-estimating the baseline using a subset of the most recent data. The authors demonstrate the application of temporal recalibration in the context of colon cancer survival and discuss considerations for applying this method.
STATISTICS IN MEDICINE
(2023)
Review
Medicine, General & Internal
Matthew J. Page, Jonathan A. C. Sterne, Isabelle Boutron, Asbjorn Hrobjartsson, Jamie J. Kirkham, Tianjing Li, Andreas Lundh, Evan Mayo-Wilson, Joanne E. McKenzie, Lesley A. Stewart, Alex J. Sutton, Lisa Bero, Adam G. Dunn, Kerry Dwan, Roy G. Elbers, Raju Kanukula, Joerg J. Meerpohl, Erick H. Turner, Julian P. T. Higgins
Summary: This paper describes a structured approach, the ROB-ME tool, for assessing bias risk in meta-analysis, which can help identify high-risk meta-analyses and interpret results appropriately.
BMJ-BRITISH MEDICAL JOURNAL
(2023)
