Bio

Stylianos (Stelios) Serghiou, MD, MS, PhD, is an Assistant Professor of Epidemiology and Population Health and of Medicine (Computational Medicine) at Stanford University. His lab develops and deploys the methods and technologies needed to make high-quality care proactive, personalized, continuous, and accessible at scale.

Medicine remains organized around the episodic visit. Care often begins when symptoms appear, decisions rest on sparse and fragmented records, and the evidence base exceeds any clinician’s capacity to assimilate it. Meanwhile, patients’ physiology, behavior, and experience between visits remain largely unobserved. Continuous, multimodal data and AI systems capable of interpreting them at scale now make an alternative attainable. Stelios’s research aims to translate that possibility into rigorously validated practice: care embedded in daily life, tailored to the individual, and continually improved through evidence generated in practice.

Stelios trained in medicine at the University of Edinburgh, graduating with honors in the top 10% of his class, and practiced as an Academic Foundation Doctor with NHS Lothian before earning an MS in Statistics and a PhD in Epidemiology and Clinical Research at Stanford, where he was a Data Science Scholar. As one of four AI Residents at Google Health, he applied machine learning to public health at national and international scales. His work included exposure risk modeling for the Google-Apple COVID-19 Exposure Notification System and quantifying urban park use across the United States at the census-tract level.

He then joined Prolaio, a cardiology startup, as its first employee and Vice President of Clinical Data Science. He built the data science team from the ground up and led development of its flagship pipeline, integrating wearable signals and clinical records to support risk prediction, LLM-generated clinical summaries, and the company’s first clinical decision support product for heart failure.

Stelios is committed to carrying innovation into clinical practice. His lab pursues the full translational lifecycle, moving from data to evidence, from evidence to technology, and from technology to products validated, deployed, and evaluated in clinical workflows. He has released open-source R packages (rtransparent, clinicaltrialr, and metareadr), with applications extending as far as evolutionary biology, and collaborated with clinical groups across cardiology, neurosurgery, pediatrics, ophthalmology, and gastroenterology.

A longstanding commitment to transparent, rigorous, and reproducible research runs through this work. With John Ioannidis, Stelios developed computational methods to study the biomedical literature at scale, mapping indicators of transparency across 2.75 million open-access articles. He also coauthored a practitioner’s handbook on open, rigorous, and reproducible research and teaches these practices at Stanford. His lab now extends this agenda to clinical AI, investigating how agentic systems can embed rigor and reproducibility throughout clinical research workflows.

His work has been recognized with the Journal of Clinical Epidemiology’s David Sackett Young Investigator of the Year Award, a Stanford Data Science Scholarship, and selection for the Google Health AI Residency. He serves on the editorial board of the Journal of Clinical Epidemiology and has received awards for outstanding peer review from Annals of Internal Medicine, the Journal of Clinical Epidemiology, and the Machine Learning for Health (ML4H) conference.

The lab is actively recruiting postdoctoral fellows, students at every level, clinicians, and engineers who wish to shape its work from the start. Learn more at https://serghioulab.org/.

Academic Appointments

Honors & Awards

  • Top Reviewer Award, Annals of Internal Medicine (2023)
  • Best Reviewer of the Year, Journal of Clinical Epidemiology (2021)
  • Stanford Data Science Scholar, Stanford University (2020)
  • David Sackett Young Investigator of the Year, Journal of Clinical Epidemiology (2017)
  • Honorary Clinical Fellow, University of Edinburgh (2016)
  • Patric-Trevor Award, Royal College of Ophthalmology (2014)
  • Intercalated BSc Scholarship, Royal College of Physicians of UK and Ireland (2011)
  • AMGEN Scholar, University of Cambridge, AMGEN (2009)

Professional Education

  • PhD, Stanford University, Epidemiology & Clinical Research (2020)
  • MS, Stanford University, Statistics (2019)
  • Academic Foundation Doctor, NHS Lothian and University of Edinburgh, Medicine (2016)
  • MBChB (Honors), University of Edinburgh, Medicine (2015)
  • BSc (Hons) Neuroscience, University of Edinburgh, Neuroscience (2012)

Additional Info

Links

Current Research and Scholarly Interests

The Serghiou Lab develops and rigorously evaluates AI-native systems in real-world settings to make high-quality care proactive, personalized, continuous, and accessible at scale. We integrate longitudinal, multimodal data from clinical records, remote patient monitoring, and patient-reported outcomes with published evidence. Drawing on advances in computer science, statistics, and clinical epidemiology, we translate these sources into clinically reliable tools and actionable knowledge.

Our research spans three complementary directions:

Clinicians. Clinicians make high-stakes decisions from fragmented records, guided by an evidence base no individual can fully assimilate. We develop and evaluate decision support systems that make guideline-directed care the default. Our flagship project addresses statin underuse: millions of eligible U.S. adults are not receiving treatment. We combine longitudinal electronic health records with AI-derived clinical context to identify treatment gaps proactively and test whether this approach changes clinical decisions.

Patients. Patients’ daily lives contain information that medicine rarely captures, while their clinical records are often incomplete or inaccurate. In a large survey, one in five patients who read their notes reported an error; more than 40% of those respondents considered it serious. We are developing conversational, agentic systems that elicit and structure clinical histories and patient-reported outcomes. Using statins as our first test case, we investigate whether patient-generated records can identify eligible patients otherwise overlooked, creating a direct path from patient engagement to evidence-based prevention.

Researchers. Clinical research still relies on fragmented, labor-intensive workflows that are often difficult to reproduce. We investigate how agentic systems can serve as scientific companions, orchestrating end-to-end data analysis and evidence synthesis within transparent, auditable, and reproducible pipelines. Our aim is to accelerate the generation and updating of reliable evidence, extending our longstanding work on research rigor and reproducibility to the AI systems being developed to support scientific discovery.

Across all three directions, we pursue the full translational lifecycle: from data to evidence, from evidence to technology, and from technology to products validated, deployed, and evaluated in clinical workflows, with real-world use informing the next cycle of research.

We are actively recruiting doctoral students, as well as postdoctoral fellows, students at every level, clinicians, and engineers.

All Publications

  • CONTINUOUS MONITORING REVEALS DYNAMIC VITAL SIGN STABILIZATION FOLLOWING HEART FAILURE DISCHARGE AND MEDICATION TITRATION Martin, C. M., Sideris, K., Hanff, T., Schmalfuss, C., Lewis, N. P., Sallam, K., Carter, S., Hanson, H., Steinhubl, S. R., Edelberg, J. M., Tison, G. H., Kurio, G., Wegerich, S., Serghiou, S., Stehlik, J. ELSEVIER SCIENCE INC. 2026

    View details for Web of Science ID 001752251000053

  • Clinical Associations with Lenticulostriatal Vasculopathy (LSV) at Birth: A Case-Control Study. Children (Basel, Switzerland) Kyriakopoulou, A., Samikos, K., Kanavaki, A., Alexopoulou, E., Argyropoulou, M., Psaltopoulou, T., Kanaka-Gantenbein, C., Dinopoulos, A., Giorgi, M., Antoniadou, A., Filippa, I., Siafakas, N., Serghiou, S., Papaevangelou, V. 2025; 12 (2)

    Abstract

    To investigate the clinical characteristics associated with the presence of LSV at birth.Prospective 1:1 case-control study.Two tertiary neonatal units in Athens, Greece.Premature neonates (≤36 weeks gestational age) who underwent cerebral ultrasound within the first 3 weeks of life, where LSV was detected.Associations between LSV and clinical characteristics at birth. Both unmatched and matched analyses stratifying the study population by gestational week were conducted. Two-sided p-values were computed using the likelihood ratio test.This study included 166 participants (83 cases and 83 controls). Neonates with LSV exhibited more concurrent cerebral findings, notably periventricular echogenicity. LSV was correlated with higher z-scores for head circumference and body length. LSV was not associated with congenital CMV.This study indicated a relationship between LSV and increased head circumference and body length. Further research is warranted to explore LSV's pathophysiological mechanisms.

    View details for DOI 10.3390/children12020223

    View details for PubMedID 40003325

    View details for PubMedCentralID PMC11853899

  • Deep Learning for Epidemiologists: An Introduction to Neural Networks. American journal of epidemiology Serghiou, S., Rough, K. 2023; 192 (11): 1904-1916

    Abstract

    Deep learning methods are increasingly being applied to problems in medicine and health care. However, few epidemiologists have received formal training in these methods. To bridge this gap, this article introduces the fundamentals of deep learning from an epidemiologic perspective. Specifically, this article reviews core concepts in machine learning (e.g., overfitting, regularization, and hyperparameters); explains several fundamental deep learning architectures (convolutional neural networks, recurrent neural networks); and summarizes training, evaluation, and deployment of models. Conceptual understanding of supervised learning algorithms is the focus of the article; instructions on the training of deep learning models and applications of deep learning to causal learning are out of this article's scope. We aim to provide an accessible first step towards enabling the reader to read and assess research on the medical applications of deep learning and to familiarize readers with deep learning terminology and concepts to facilitate communication with computer scientists and machine learning engineers.

    View details for DOI 10.1093/aje/kwad107

    View details for PubMedID 37139570

    View details for PubMedCentralID PMC13368595

  • Lessons learnt from registration of biomedical research. Nature human behaviour Serghiou, S., Axfors, C., Ioannidis, J. P. 2023

    View details for DOI 10.1038/s41562-022-01499-0

    View details for PubMedID 36604496

  • Large Pediatric Randomized Clinical Trials in ClinicalTrials.gov. Pediatrics Cho, S. M., Serghiou, S., Ioannidis, J. P., Klassen, T. P., Contopoulos-Ioannidis, D. G. 2021; 148 (3)

    Abstract

    BACKGROUND: Large, randomized controlled trials (RCTs) are essential in answering pivotal questions in child health.METHODS: We created a bird's eye view of all large, noncluster, nonvaccine pediatric RCTs with ≥1000 participants registered in ClinicalTrials.gov (last search January 9, 2020). We analyzed the funding sources, countries, outcomes, publication status, and correlation with the pediatric global burden of disease (GBD) for eligible trials.RESULTS: We identified 247 large, nonvaccine, noncluster pediatric RCTs. Only 17 mega-trials with ≥5000 participants existed. Industry funding was involved in only 52 (21%) and exclusively funded 47 (19%) trials. Participants were from high-income countries (HICs) in 100 (40%) trials, from lower-middle-income countries (LMICs) in 122 (49%) trials, and from both HICs and LMICs in 19 (8%) trials; 6 trials did not report participants' country location. Of trials conducted in LMIC, 43% of investigators were from HICs. Of non-LMIC participants trials (HIC or HIC and LMIC), 39% were multicountry trials versus 11% of exclusively LMIC participants trials. Few trials (18%; 44 of 247) targeted mortality as an outcome. 35% (58 of 164) of the trials completed ≥12 months were unpublished at the time of our assessment. The number of trials per disease category correlated well with pediatric GBD overall (rho = 0.76) and in LMICs (rho = 0.69), but not in HICs (rho = 0.29).CONCLUSIONS: Incentivization of investigator collaborations across diverse country settings, timely publication of results of large pediatric RCTs, and alignment with the pediatric GBD are of pivotal importance to ultimately improve child health globally.

    View details for DOI 10.1542/peds.2020-049771

    View details for PubMedID 34465592

  • Media and social media attention to retracted articles according to Altmetric. PloS one Serghiou, S. n., Marton, R. M., Ioannidis, J. P. 2021; 16 (5): e0248625

    Abstract

    The number of retracted articles has grown fast. However, the extent to which researchers and the public are made adequately aware of these retractions and how the media and social media respond to them remains unknown. Here, we aimed to evaluate the media and social media attention received by retracted articles and assess also the attention they receive post-retraction versus pre-retraction. We downloaded all records of retracted literature maintained by the Retraction Watch Database and originally published between January 1, 2010 to December 31, 2015. For all 3,008 retracted articles with a separate DOI for the original and its retraction, we downloaded the respective Altmetric Attention Score (AAS) (from Altmetric) and citation count (from Crossref), for the original article and its retraction notice on June 6, 2018. We also compared the AAS of a random sample of 572 retracted full journal articles available on PubMed to that of unretracted full articles matched from the same issue and journal. 1,687 (56.1%) of retracted research articles received some amount of Altmetric attention, and 165 (5.5%) were even considered popular (AAS>20). 31 (1.0%) of 2,953 with a record on Crossref received >100 citations by June 6, 2018. Popular articles received substantially more attention than their retraction, even after adjusting for attention received post-retraction (Median difference, 29; 95% CI, 17-61). Unreliable results were the most frequent reason for retraction of popular articles (32; 19%), while fake peer review was the most common reason (421; 15%) for the retraction of other articles. In comparison to matched articles, retracted articles tended to receive more Altmetric attention (23/31 matched groups; P-value, 0.01), even after adjusting for attention received post-retraction. Our findings reveal that retracted articles may receive high attention from media and social media and that for popular articles, pre-retraction attention far outweighs post-retraction attention.

    View details for DOI 10.1371/journal.pone.0248625

    View details for PubMedID 33979339

  • Assessment of transparency indicators across the biomedical literature: How open is open? PLoS biology Serghiou, S. n., Contopoulos-Ioannidis, D. G., Boyack, K. W., Riedel, N. n., Wallach, J. D., Ioannidis, J. P. 2021; 19 (3): e3001107

    Abstract

    Recent concerns about the reproducibility of science have led to several calls for more open and transparent research practices and for the monitoring of potential improvements over time. However, with tens of thousands of new biomedical articles published per week, manually mapping and monitoring changes in transparency is unrealistic. We present an open-source, automated approach to identify 5 indicators of transparency (data sharing, code sharing, conflicts of interest disclosures, funding disclosures, and protocol registration) and apply it across the entire open access biomedical literature of 2.75 million articles on PubMed Central (PMC). Our results indicate remarkable improvements in some (e.g., conflict of interest [COI] disclosures and funding disclosures), but not other (e.g., protocol registration and code sharing) areas of transparency over time, and map transparency across fields of science, countries, journals, and publishers. This work has enabled the creation of a large, integrated, and openly available database to expedite further efforts to monitor, understand, and promote transparency and reproducibility in science.

    View details for DOI 10.1371/journal.pbio.3001107

    View details for PubMedID 33647013

  • Antenatal imaging and clinical outcome in congenital CMV infection: A field-wide systematic review and meta-analysis JOURNAL OF INFECTION Kyriakopoulou, A., Serghiou, S., Dimopoulou, D., Arista, I., Psaltopoulou, T., Dinopoulos, A., Papaevangelou, V. 2020; 80 (4): 407-418

    Abstract

    Postnatal outcome in fetuses with congenital cytomegalovirus infection (cCMV) varies from asymptomatic infection to severe neurodevelopmental impairment. Αntenatal biomarkers of long-term clinical outcome, have yet to be established. Α systematic review and meta-analysis was performed to examine whether prenatal cerebral ultrasonography (US) and magnetic resonance imaging (MRI) findings in cCMV fetuses may predict clinical outcome.PubMed and the Web of Science were systematically searched to identify studies reporting on any prenatal US and/or MRI imaging of fetuses with cCMV as well as their postnatal clinical outcome. All reported associations between imaging and postnatal clinical outcome were systematically extracted. Where appropriate, the reported associations were quantitatively synthesized within Bayesian random-effects meta-analyses.A total of 1336 studies were screened to identify 26 eligible observational studies. Overall, 4181 fetuses were studied, of which 1518 had been diagnosed with cCMV. All studies performed fetal US while in 14 (54%) MRI was also performed. Studies substantially varied in timing of fetal imaging, reporting of abnormalities, definition of poor outcome and statistical analysis. Among studies reporting on statistical significance, 6/6 for US and 3/4 for MRI identified significant associations between imaging findings and outcome. In our meta-analyses, within isolated abnormalities, only microcephaly had greater than 95% probability of being associated with poor outcome (OR 26.7; 95% CI, 1.44-1464.5; I2, 19%). Effect sizes for US were higher than those for MRI findings.Although studies displayed significant heterogeneity in both methodology and analytical decisions, it became evident that when both prenatal cerebral US and MRI are normal the negative predictive value of poor outcome is high. This is important for clinicians when consulting pregnant women. Need to standardize practices and definitions become evident.There was no source of funding.

    View details for DOI 10.1016/j.jinf.2020.02.012

    View details for Web of Science ID 000521239700006

    View details for PubMedID 32097687

  • Evaluation of confounding in epidemiologic studies assessing alcohol consumption on the risk of ischemic heart disease. BMC medical research methodology Wallach, J. D., Serghiou, S. n., Chu, L. n., Egilman, A. C., Vasiliou, V. n., Ross, J. S., Ioannidis, J. P. 2020; 20 (1): 64

    Abstract

    Among different investigators studying the same exposures and outcomes, there may be a lack of consensus about potential confounders that should be considered as matching, adjustment, or stratification variables in observational studies. Concerns have been raised that confounding factors may affect the results obtained for the alcohol-ischemic heart disease relationship, as well as their consistency and reproducibility across different studies. Therefore, we assessed how confounders are defined, operationalized, and discussed across individual studies evaluating the impact of alcohol on ischemic heart disease risk.For observational studies included in a recent alcohol-ischemic heart disease meta-analysis, we identified all variables adjusted, matched, or stratified for in the largest reported multivariate model (i.e. potential confounders). We recorded how the variables were measured and grouped them into higher-level confounder domains. Abstracts and Discussion sections were then assessed to determine whether authors considered confounding when interpreting their study findings.85 of 87 (97.7%) studies reported multivariate analyses for an alcohol-ischemic heart disease relationship. The most common higher-level confounder domains included were smoking (79, 92.9%), age (74, 87.1%), and BMI, height, and/or weight (57, 67.1%). However, no two models adjusted, matched, or stratified for the same higher-level confounder domains. Most (74/87, 85.1%) articles mentioned or alluded to "confounding" in their Abstract or Discussion sections, but only one stated that their main findings were likely to be affected by residual confounding. There were five (5/87, 5.7%) authors that explicitly asked for caution when interpreting results.There is large variation in the confounders considered across observational studies evaluating the impact of alcohol on ischemic heart disease risk and almost all studies spuriously ignore or eventually dismiss confounding in their conclusions. Given that study results and interpretations may be affected by the mix of potential confounders included within multivariate models, efforts are necessary to standardize approaches for selecting and accounting for confounders in observational studies.

    View details for DOI 10.1186/s12874-020-0914-6

    View details for PubMedID 32171256

  • Calibrating the Scientific Ecosystem Through Meta-Research Annual Review of Statistics and Its Application Tom, H. E., Stylianos, S., Janiaud, P., Danchev, V., Crüwell, S., Goodman, S. N., Ioannidis, J. P. 2020; 7

    View details for DOI 10.1146/annurev-statistics-031219-041104

  • Random-Effects Assumption in Meta-analyses-Reply. JAMA Goodman, S. N., Serghiou, S. 2019; 322 (1): 82

    View details for DOI 10.1001/jama.2019.5447

    View details for PubMedID 31265095

  • Use of machine learning for prediction of ocular conservation and visual outcomes after proton beam radiotherapy for choroidal melanoma Serghiou, S., Damato, B. E., Afshar, A. R. ASSOC RESEARCH VISION OPHTHALMOLOGY INC. 2019

    View details for Web of Science ID 000488628102135

  • New clinical trial designs in the era of precision medicine: An overview of definitions, strengths, weaknesses, and current use in oncology CANCER TREATMENT REVIEWS Janiaud, P., Serghiou, S., Ioannidis, J. P. A. 2019; 73: 20–30

    View details for DOI 10.1016/j.ctrv.2018.12.003

    View details for Web of Science ID 000458711900003

  • Random-EffectsMeta-analysis Summarizing Evidence With Caveats JAMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION Serghiou, S., Goodman, S. N. 2019; 321 (3): 301-302

    View details for DOI 10.1001/jama.2018.19684

    View details for Web of Science ID 000456347000021

  • Random-Effects Meta-analysis: Summarizing Evidence With Caveats. JAMA Serghiou, S., Goodman, S. N. 2018

    View details for PubMedID 30566189

  • New clinical trial designs in the era of precision medicine: An overview of definitions, strengths, weaknesses, and current use in oncology. Cancer treatment reviews Janiaud, P., Serghiou, S., Ioannidis, J. P. 2018; 73: 20–30

    Abstract

    With expanding knowledge in tumor biology and biomarkers, oncology therapies are increasingly moving away from the "one-size-fits-all" rationale onto biomarker-driven therapies tailored according to patient-specific characteristics, most commonly the tumor's molecular profile. The advent of precision medicine in oncology has been accompanied by the introduction of novel clinical trial designs that aim to identify biomarker-matched subgroups of patients that will benefit the most from targeted therapies. This innovation comes with the promise of answering more treatment questions, more efficiently and in less time. In this article, we give an overview of the different biomarker-based designs, comparing the features of enrichment, randomize-all, umbrella, and basket trials, and highlighting their advantages and disadvantages. We focus more on the novel designs known as master protocols, which include umbrella and basket trials. We have also conducted a search in ClinicalTrials.gov for registered oncology-related protocols of ongoing or completed trials labeled as umbrella or basket trials for solid tumors; we also included additional relevant trials retrieved from other reviews. We present and discuss the key features of the 30 eligible basket trials and 27 eligible umbrella trials. Only a minority of them are randomized (2 and 9, respectively), including three trials with adaptive randomization. Five of these trials have been completed as of July 2018. Precision medicine trial designs fuel new hopes for identifying best treatments, but there is also the potential for hype. The benefits and challenges associated with their use will need continued monitoring.

    View details for PubMedID 30572165

  • THE INTERLEUKIN-6 RECEPTOR AS A DRUG TARGET IN INFLAMMATORY BOWEL DISEASE; A MENDELIAN RANDOMISATION STUDY Parisinos, C., Serghiou, S., Patel, R., Hemingway, H., Hingorani, A. BMJ PUBLISHING GROUP. 2018: A55

    View details for DOI 10.1136/gutjnl-2018-BSGAbstracts.109

    View details for Web of Science ID 000439577600110

  • Altmetric Scores, Citations, and Publication of Studies Posted as Preprints JAMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION Serghiou, S., Ioannidis, J. P. A. 2018; 319 (4): 402–3

    View details for PubMedID 29362788

  • Variation in Interleukin 6 Receptor Gene Associates with Risk of Crohn's Disease and Ulcerative Colitis. Gastroenterology Parisinos, C. A., Serghiou, S. n., Katsoulis, M. n., George, M. J., Patel, R. S., Hemingway, H. n., Hingorani, A. D. 2018

    Abstract

    Interleukin 6 (IL6) is an inflammatory cytokine; signaling via its receptor (IL6R) is believed to contribute to development of inflammatory bowel diseases (IBD). The single nucleotide polymorphism rs2228145 in IL6R associates with increased levels of soluble IL6R (s-IL6R), as well as reduced IL6R signaling and risk of inflammatory disorders; its effects are similar to those of a therapeutic monoclonal antibody that blocks IL6R signaling. We used the effect of rs2228145 on s-IL6R level as an indirect marker to investigate whether reduced IL6R signaling associates with risk of ulcerative colitis (UC) or Crohn's disease (CD). In a genome-wide meta-analysis of 20,550 patients with CD, 17647 patients with UC, and more than 40,000 individuals without IBD (controls), we found that rs2228145 (scaled to a 2-fold increase in s-IL6R) was associated with reduced risk of CD (odds ratio, 0.876; 95% CI, 0.822-0.933; P=.00003) or UC (odds ratio, 0.932; 95% CI, 0.875-0.996; P=.036). These findings indicate that therapeutics designed to block IL6R signaling might be effective in treatment of IBD.

    View details for PubMedID 29775600

  • Curing World Blindness: The Lifeline Express Train Molera, R., Wren, S. M., Serghiou, S., Lee, S., Shields, R. A., Cho, H. ELSEVIER SCIENCE INC. 2017: S100

    View details for DOI 10.1016/j.jamcollsurg.2017.07.219

    View details for Web of Science ID 000413315300207

  • Long noncoding RNAs as novel predictors of survival in human cancer: a systematic review and meta-analysis MOLECULAR CANCER Serghiou, S., Kyriakopoulou, A., Ioannidis, J. P. 2016; 15

    Abstract

    Expression of various long noncoding RNAs (lncRNAs) may affect cancer prognosis. Here, we aim to gather and examine all evidence on the potential role of lncRNAs as novel predictors of survival in human cancer.We systematically searched through PubMed, to identify all published studies reporting on the association between any individual lncRNA or group of lncRNAs with prognosis in human cancer (death or other clinical outcomes). Where appropriate, we then performed quantitative synthesis of those results using meta-analytic methods to identify the true effect size of lncRNAs on cancer prognosis. The reliability of those results was then examined using measures of heterogeneity and testing for selective reporting biases.Three hundred ninety-two studies were screened to eventually identify 111 eligible studies on 127 datasets. In total, these represented 16,754 independent participants pertaining to 53 individual and 6 grouped lncRNAs within a total of 19 cancer sites. Overall, 83 % of the studies we identified addressed overall survival and 32 % of the studies addressed recurrence-free survival. For overall survival, 96 % (88/92) of studies identified a statistically significant association of lncRNA expression to prognosis. Meta-analysis of 6 out of 7 lncRNAs for which three or more studies were available, identified statistically significant associations with overall survival. The lncRNA HOTAIR was by far the most broadly studied lncRNA (n = 29; of 111 studies) and featured a summary hazard ratio (HR) of 2.22 (95 % confidence interval (CI), 1.86-2.65) with modest heterogeneity (I(2) = 49 %; 95 % CI, 14-79 %). Prominent excess significance was demonstrated across all meta-analyses (p-value = 0.0003), raising the possibility of substantial selective reporting biases.Multiple lncRNAs have been shown to be strongly associated with prognosis in diverse cancers, but substantial bias cannot be excluded in this field and larger studies are needed to understand whether these prognostic information may eventually be useful.

    View details for DOI 10.1186/s12943-016-0535-1

    View details for PubMedID 27352941

  • Comparative assessment of phototherapy protocols for reduction of oxidative stress in partially transected spinal cord slices undergoing secondary degeneration. BMC neuroscience Ashworth, B. E., Stephens, E., Bartlett, C. A., Serghiou, S., Giacci, M. K., Williams, A., Hart, N. S., Fitzgerald, M. 2016; 17 (1): 21

    Abstract

    Red/near-infrared light therapy (R/NIR-LT) has been developed as a treatment for a range of conditions, including injury to the central nervous system (CNS). However, clinical trials have reported variable or sub-optimal outcomes, possibly because there are few optimized treatment protocols for the different target tissues. Moreover, the low absolute, and wavelength dependent, transmission of light by tissues overlying the target site make accurate dosing problematic.In order to optimize light therapy treatment parameters, we adapted a mouse spinal cord organotypic culture model to the rat, and characterized myelination and oxidative stress following a partial transection injury. The ex vivo model allows a more accurate assessment of the relative effect of different illumination wavelengths (adjusted for equal quantal intensity) on the target tissue. Using this model, we assessed oxidative stress following treatment with four different wavelengths of light: 450 nm (blue); 510 nm (green); 660 nm (red) or 860 nm (infrared) at three different intensities: 1.93 × 10(16) (low); 3.85 × 10(16) (intermediate) and 7.70 × 10(16) (high) photons/cm(2)/s. We demonstrate that the most effective of the tested wavelengths to reduce immunoreactivity of the oxidative stress indicator 3-nitrotyrosine (3NT) was 660 nm. 860 nm also provided beneficial effects at all tested intensities, significantly reducing oxidative stress levels relative to control (p ≤ 0.05).Our results indicate that R/NIR-LT is an effective antioxidant therapy, and indicate that effective wavelengths and ranges of intensities of treatment can be adapted for a variety of CNS injuries and conditions, depending upon the transmission properties of the tissue to be treated.

    View details for DOI 10.1186/s12868-016-0259-6

    View details for PubMedID 27194427

    View details for PubMedCentralID PMC4872332

  • Field-wide meta-analyses of observational associations can map selective availability of risk factors and the impact of model specifications JOURNAL OF CLINICAL EPIDEMIOLOGY Serghiou, S., Patel, C. J., Tan, Y. Y., Koay, P., Ioannidis, J. P. 2016; 71: 58-67

    Abstract

    Instead of evaluating one risk factor at a time, we illustrate the utility of "field-wide meta-analyses" in considering all available data on all putative risk factors of a disease simultaneously.We identified studies on putative risk factors of pterygium (surfer's eye) in PubMed, EMBASE, and Web of Science. We mapped which factors were considered, reported, and adjusted for in each study. For each putative risk factor, four meta-analyses were done using univariate only, multivariate only, preferentially univariate, or preferentially multivariate estimates.A total of 2052 records were screened to identify 60 eligible studies reporting on 65 putative risk factors. Only 4 of 60 studies reported both multivariate and univariate regression analyses. None of the 32 studies using multivariate analysis adjusted for the same set of risk factors. Effect sizes from different types of regression analyses led to significantly different summary effect sizes (P-value < 0.001). Observed heterogeneity was very high for both multivariate (median I(2), 76.1%) and univariate (median I(2), 85.8%) estimates. No single study investigated all 11 risk factors that were statistically significant in at least one of our meta-analyses.Field-wide meta-analyses can map availability of risk factors and trends in modeling, adjustments and reporting, as well as the impact of differences in model specification.

    View details for DOI 10.1016/j.jclinepi.2015.09.004

    View details for PubMedID 26415577

  • Risk of Bias in Reports of In Vivo Research: A Focus for Improvement PLOS BIOLOGY Macleod, M. R., McLean, A. L., Kyriakopoulou, A., Serghiou, S., de Wilde, A., Sherratt, N., Hirst, T., Hemblade, R., Bahor, Z., Nunes-Fonseca, C., Potluru, A., Thomson, A., Baginskitae, J., Egan, K., Vesterinen, H., Currie, G. L., Churilov, L., Howells, D. W., Sena, E. S. 2015; 13 (10)

    Abstract

    The reliability of experimental findings depends on the rigour of experimental design. Here we show limited reporting of measures to reduce the risk of bias in a random sample of life sciences publications, significantly lower reporting of randomisation in work published in journals of high impact, and very limited reporting of measures to reduce the risk of bias in publications from leading United Kingdom institutions. Ascertainment of differences between institutions might serve both as a measure of research quality and as a tool for institutional efforts to improve research quality.

    View details for DOI 10.1371/journal.pbio.1002273

    View details for Web of Science ID 000364457500008

    View details for PubMedID 26460723

    View details for PubMedCentralID PMC4603955

https://orcid.org/0000-0002-2477-6060