European Respiratory Journal: Spirometry-based exclusions have limited influence on biologic treatment response and unduly restrict generalisability of clinical trials https://bit.ly/4wJoMYR
October 27, 2026 at 5:59 PM
European Respiratory Journal: Spirometry-based exclusions have limited influence on biologic treatment response and unduly restrict generalisability of clinical trials https://bit.ly/4wJoMYR
Ophthopedia Update: Long-term glaucoma medication adherence in Finland: a 13-year registry study with clinical outcomes and international cohort comparisons: WHAT IS ALREADY KNOWN ON THIS TOPIC
*
Long-term adherence to glaucoma medication is widely… #Ophthalmology #Ophthotwitter #BJO
*
Long-term adherence to glaucoma medication is widely… #Ophthalmology #Ophthotwitter #BJO
Long-term glaucoma medication adherence in Finland: a 13-year registry study with clinical outcomes and international cohort comparisons
WHAT IS ALREADY KNOWN ON THIS TOPIC
*
Long-term adherence to glaucoma medication is widely considered suboptimal, but the evidence linking it to clinical outcomes comes mainly from short-term or US-derived cohorts with limited generalisability to European healthcare settings.WHAT THIS STUDY ADDS
*
In this population-based Finnish cohort of 10 769 patients followed for up to 13 years, average medication coverage was high but frequently interrupted, and clinical outcomes broadly mirrored those reported in US cohorts.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
*
Pharmacy-dispensing data enable longitudinal, dynamic monitoring of adherence, while standardised methods and reporting are needed to allow valid comparison across countries and studies.Introduction
The current treatment paradigm for glaucoma continues to focus on managing intraocular pressure (IOP) mainly through medical therapy, although selective laser trabeculoplasty is increasingly regaining preference as primary therapy based on its similar, or even better efficacy compared with medications.1–3
dlvr.it
October 29, 2026 at 4:44 PM
Ophthopedia Update: Long-term glaucoma medication adherence in Finland: a 13-year registry study with clinical outcomes and international cohort comparisons: WHAT IS ALREADY KNOWN ON THIS TOPIC
*
Long-term adherence to glaucoma medication is widely… #Ophthalmology #Ophthotwitter #BJO
*
Long-term adherence to glaucoma medication is widely… #Ophthalmology #Ophthotwitter #BJO
October 20, 2026 at 7:52 AM
Ball is a good choice for generalisability LOL
October 16, 2026 at 4:08 PM
Ball is a good choice for generalisability LOL
Our new paper is out in @wildlifebiology.bsky.social!
It shows how considering animal affective states can enhance the quality and utility of data collected in wildlife research.
🔗 buff.ly/MrBnYEx
It shows how considering animal affective states can enhance the quality and utility of data collected in wildlife research.
🔗 buff.ly/MrBnYEx
Inspiring systematic inclusion of individual affective states to enhance the quality of wild animal research
Studies on animals continue to attract criticism, often relating to not accounting for variation, for issues with data quality, reproducibility and generalisability, yet one source of variation…
buff.ly
October 16, 2026 at 3:12 PM
Our new paper is out in @wildlifebiology.bsky.social!
It shows how considering animal affective states can enhance the quality and utility of data collected in wildlife research.
🔗 buff.ly/MrBnYEx
It shows how considering animal affective states can enhance the quality and utility of data collected in wildlife research.
🔗 buff.ly/MrBnYEx
A systematic review finds that conditional cash transfers (CCTs) for women with HIV probably increase births at HIV Prevention of Mother-to-Child Transmission facilities in LMICs but findings are from a small evidence base, limiting generalisability, and ... 1/2
October 14, 2026 at 12:41 AM
A systematic review finds that conditional cash transfers (CCTs) for women with HIV probably increase births at HIV Prevention of Mother-to-Child Transmission facilities in LMICs but findings are from a small evidence base, limiting generalisability, and ... 1/2
Generalisability of an adult-trained AI model to paediatric coeliac disease diagnosis: multicentre external validation study
Our AI model, trained on adult biopsies, generalises to paediatric #CoeliacDisease diagnosis across centres and scanner platforms
www.thelancet.com/journals/ecl...
#MedSky
Our AI model, trained on adult biopsies, generalises to paediatric #CoeliacDisease diagnosis across centres and scanner platforms
www.thelancet.com/journals/ecl...
#MedSky
October 12, 2026 at 2:01 PM
Generalisability of an adult-trained AI model to paediatric coeliac disease diagnosis: multicentre external validation study
Our AI model, trained on adult biopsies, generalises to paediatric #CoeliacDisease diagnosis across centres and scanner platforms
www.thelancet.com/journals/ecl...
#MedSky
Our AI model, trained on adult biopsies, generalises to paediatric #CoeliacDisease diagnosis across centres and scanner platforms
www.thelancet.com/journals/ecl...
#MedSky
To have impact, you need to focus on domains (financial, CPS, …) in your research. “Generalisability is overrated.” #NewFacultySymposium at #ASE26
October 12, 2026 at 7:44 AM
To have impact, you need to focus on domains (financial, CPS, …) in your research. “Generalisability is overrated.” #NewFacultySymposium at #ASE26
Great work by my PhD student @manchester.ac.uk , Kritchavat Ploddi, on this publication! 👏, and thanks to my fellow supervisors @mattsperrin.bsky.social & @glenmartin.bsky.social
📄 doi.org/10.1186/s415...
#ClinicalPredictionModels #Generalisability #Transportability #DataShift #CausalInference
📄 doi.org/10.1186/s415...
#ClinicalPredictionModels #Generalisability #Transportability #DataShift #CausalInference
October 9, 2026 at 9:38 PM
Great work by my PhD student @manchester.ac.uk , Kritchavat Ploddi, on this publication! 👏, and thanks to my fellow supervisors @mattsperrin.bsky.social & @glenmartin.bsky.social
📄 doi.org/10.1186/s415...
#ClinicalPredictionModels #Generalisability #Transportability #DataShift #CausalInference
📄 doi.org/10.1186/s415...
#ClinicalPredictionModels #Generalisability #Transportability #DataShift #CausalInference
Implementing equality, diversity, and inclusion (EDI) in healthcare research is critical to increasing the generalisability of findings.
This new tool can be used to assess and monitor the quality of EDI implementation
This new tool can be used to assess and monitor the quality of EDI implementation
Development and evaluation of a quality assessment tool for equality, diversity, and inclusion (QUALITY-EDI) within healthcare research
Implementing equality, diversity, and inclusion (EDI) in healthcare research is critical to ensuring findings are translatable and provide the greatest benefit to patients who need it most. A quality ...
www.bmj.com
September 28, 2026 at 4:52 PM
Implementing equality, diversity, and inclusion (EDI) in healthcare research is critical to increasing the generalisability of findings.
This new tool can be used to assess and monitor the quality of EDI implementation
This new tool can be used to assess and monitor the quality of EDI implementation
When reproducibility, replicability, and generalisability become universal standards some research traditions risk epistemic misrecognition. More inclusive language supports methodological pluralism.
Openness should broaden the conversation, not narrow it.
www.tandfonline.com/doi/full/10....
Openness should broaden the conversation, not narrow it.
www.tandfonline.com/doi/full/10....
Can open science speak qualitative? Rearticulating open science vocabulary for qualitative research
‘Open research’ (or ‘open science’) has substantially reshaped conversations surrounding rigor, transparency, and accountability, yet its vocabulary remains largely rooted in quantitative, positivi...
www.tandfonline.com
September 22, 2026 at 12:13 PM
When reproducibility, replicability, and generalisability become universal standards some research traditions risk epistemic misrecognition. More inclusive language supports methodological pluralism.
Openness should broaden the conversation, not narrow it.
www.tandfonline.com/doi/full/10....
Openness should broaden the conversation, not narrow it.
www.tandfonline.com/doi/full/10....
Transcriptomic Models for Immunotherapy Response Prediction Show Limited Cross-cohort Generalisability #SingleCell 🧪🧬🖥️
https://arxiv.org/abs/2604.05478
https://arxiv.org/abs/2604.05478
September 21, 2026 at 4:00 AM
Transcriptomic Models for Immunotherapy Response Prediction Show Limited Cross-cohort Generalisability #SingleCell 🧪🧬🖥️
https://arxiv.org/abs/2604.05478
https://arxiv.org/abs/2604.05478
Transcriptomic Models for Immunotherapy Response Prediction Show Limited Cross-cohort Generalisability #SingleCell 🧪🧬🖥️
https://arxiv.org/abs/2604.05478
https://arxiv.org/abs/2604.05478
September 18, 2026 at 4:01 AM
Transcriptomic Models for Immunotherapy Response Prediction Show Limited Cross-cohort Generalisability #SingleCell 🧪🧬🖥️
https://arxiv.org/abs/2604.05478
https://arxiv.org/abs/2604.05478
🩺 Age breaks heart risk calculators
A heart risk calculator built in New Zealand badly misjudged danger for 46,558 people in China, mostly because of how it scored age.
🔗 doi.org/10.1136/bmj-2026-100535
#Health #Diabetes #Cardiology #Science 🧪
A heart risk calculator built in New Zealand badly misjudged danger for 46,558 people in China, mostly because of how it scored age.
🔗 doi.org/10.1136/bmj-2026-100535
#Health #Diabetes #Cardiology #Science 🧪
Simultaneous derivation, validation, and comparison of predictor hazard ratios for cardiovascular risk prediction equations in patients with diabetes from high versus non-high income countries: cohort...
Objective To directly test the generalisability of cardiovascular disease (CVD) risk predictor effects across populations in a high income versus non-high income country, derive comparable risk equati...
doi.org
September 16, 2026 at 4:27 AM
🩺 Age breaks heart risk calculators
A heart risk calculator built in New Zealand badly misjudged danger for 46,558 people in China, mostly because of how it scored age.
🔗 doi.org/10.1136/bmj-2026-100535
#Health #Diabetes #Cardiology #Science 🧪
A heart risk calculator built in New Zealand badly misjudged danger for 46,558 people in China, mostly because of how it scored age.
🔗 doi.org/10.1136/bmj-2026-100535
#Health #Diabetes #Cardiology #Science 🧪
Biomarker-based enrolment can sharpen trial power, or unintentionally exclude patient groups and limit generalisability. If you're designing around biomarkers, the article outlines the practices that can help avoid these risks. https://ow.ly/Yp3U50ZKtFe
September 14, 2026 at 11:00 AM
Biomarker-based enrolment can sharpen trial power, or unintentionally exclude patient groups and limit generalisability. If you're designing around biomarkers, the article outlines the practices that can help avoid these risks. https://ow.ly/Yp3U50ZKtFe
AI may transform AMD drug discovery, but algorithms generate hypotheses, not biological proof. Validation, generalisability and mechanism remain the real tests. #AI #AMD #DrugDiscovery buff.ly/9wRiLTr
https://pubmed.ncbi.nlm.nih.gov/42676384/?utm_source=Chrome&utm_medium=rss&utm_campaign=pubmed-2&utm_content=12GmDGiSiGA9an3uAk7xg5WeD6SUYkvXu9hh-tydPALMY8vQtD&fc=20241229120800&ff=20260901115139&v=2.20.1
pubmed.ncbi.nlm.nih.gov
September 9, 2026 at 9:00 AM
AI may transform AMD drug discovery, but algorithms generate hypotheses, not biological proof. Validation, generalisability and mechanism remain the real tests. #AI #AMD #DrugDiscovery buff.ly/9wRiLTr
and for all other inputs it mines crypto, or deletes every file on your computer, or gives some plausible looking output sampled from the statistical distribution of correct outputs.
This is why LLMs can't be trusted, there's no proof of generalisability and they're basically black boxes
This is why LLMs can't be trusted, there's no proof of generalisability and they're basically black boxes
September 8, 2026 at 10:08 PM
and for all other inputs it mines crypto, or deletes every file on your computer, or gives some plausible looking output sampled from the statistical distribution of correct outputs.
This is why LLMs can't be trusted, there's no proof of generalisability and they're basically black boxes
This is why LLMs can't be trusted, there's no proof of generalisability and they're basically black boxes
Trial eligibility criteria typically limit generalisability to high risk patients. So how well do SGLT2 inhibitors work in real life in people with heart failure and a preserved ejection fraction (HFpEF)? A study by Selvaganesan et al. in BMJ Medicine aimed to find out.
September 4, 2026 at 10:33 AM
Trial eligibility criteria typically limit generalisability to high risk patients. So how well do SGLT2 inhibitors work in real life in people with heart failure and a preserved ejection fraction (HFpEF)? A study by Selvaganesan et al. in BMJ Medicine aimed to find out.
Limitations worth being upfront about: Our panel was thin on LMIC representation, and there’s a real lack of research comparing operative vs non-operative outcomes across differently-resourced settings. Generalisability has real limits. 30/N
George Costanza: Why Must There Always Be a Problem?
Alt: George Costanza: Why Must There Always Be a Problem?
static.klipy.com
September 1, 2026 at 2:17 PM
Limitations worth being upfront about: Our panel was thin on LMIC representation, and there’s a real lack of research comparing operative vs non-operative outcomes across differently-resourced settings. Generalisability has real limits. 30/N
Generalisability analyses of the CICM Second Part examination: Theory meets practice
In response to: The CICM second part viva reliability coefficient: Which index, for which decision, against which benchmark?
The conversation continues!
www.sciencedirect.com/science/arti...
In response to: The CICM second part viva reliability coefficient: Which index, for which decision, against which benchmark?
The conversation continues!
www.sciencedirect.com/science/arti...
www.sciencedirect.com
August 26, 2026 at 11:33 PM
Generalisability analyses of the CICM Second Part examination: Theory meets practice
In response to: The CICM second part viva reliability coefficient: Which index, for which decision, against which benchmark?
The conversation continues!
www.sciencedirect.com/science/arti...
In response to: The CICM second part viva reliability coefficient: Which index, for which decision, against which benchmark?
The conversation continues!
www.sciencedirect.com/science/arti...
Reporting of external validation studies needs improvement: a systematic review of reporting in oncology prediction model studies - Journal of Clinical Epidemiology www.jclinepi.com/article/S089...
Reporting of external validation studies needs improvement: a systematic review of reporting in oncology prediction model studies
External validation is essential for assessing the generalisability and transportability
of a clinical prediction model. A previous review of studies published in 2010 identified
substantial deficienc...
www.jclinepi.com
August 24, 2026 at 7:29 PM
Reporting of external validation studies needs improvement: a systematic review of reporting in oncology prediction model studies - Journal of Clinical Epidemiology www.jclinepi.com/article/S089...
Write an honest lay summary for journos/public. An abstract is like an advertisement of the article to academics.
For example, if the sample was only students or mainly WEIRD participants state that clearly. Don’t overstate generalisability of findings and hide the truth in limitations.
For example, if the sample was only students or mainly WEIRD participants state that clearly. Don’t overstate generalisability of findings and hide the truth in limitations.
August 23, 2026 at 10:15 AM
Write an honest lay summary for journos/public. An abstract is like an advertisement of the article to academics.
For example, if the sample was only students or mainly WEIRD participants state that clearly. Don’t overstate generalisability of findings and hide the truth in limitations.
For example, if the sample was only students or mainly WEIRD participants state that clearly. Don’t overstate generalisability of findings and hide the truth in limitations.
The biggest risk is likely that AI, particularly LLMs (but in general delegating decision making to machines), can be (almost always definitely is) biased, so interventions using AI will be biased.
This could range from: "Outcomes are still better"
To: "We've lost the trust of entire communities"
This could range from: "Outcomes are still better"
To: "We've lost the trust of entire communities"
August 19, 2026 at 1:10 PM
The biggest risk is likely that AI, particularly LLMs (but in general delegating decision making to machines), can be (almost always definitely is) biased, so interventions using AI will be biased.
This could range from: "Outcomes are still better"
To: "We've lost the trust of entire communities"
This could range from: "Outcomes are still better"
To: "We've lost the trust of entire communities"
The field has been far too restrictive of who we test in experiments for far too long!
Generalisability is IMPROVED & signal gained rather than lost when we assess our constructs across ages, handedness, genders, backgrounds, etc
So great that people are doing this kind of research!
Generalisability is IMPROVED & signal gained rather than lost when we assess our constructs across ages, handedness, genders, backgrounds, etc
So great that people are doing this kind of research!
In honor of International Left-Handers Day, we have an updated preprint on scanning left-handed participants in fMRI studies. Although they're 11% of the population, we routinely screen LH people out of fMRI studies due to concerns about functional lateralization 1/n www.biorxiv.org/content/10.1...
Left Out: The Effect of Handedness on fMRI Activation in Memory Paradigms
About 11% of the population is left-handed, a significant minority of the potential participant pool for functional MRI (fMRI) studies. However, convention in fMRI research dictates these potential pa...
www.biorxiv.org
August 14, 2026 at 8:35 AM
The field has been far too restrictive of who we test in experiments for far too long!
Generalisability is IMPROVED & signal gained rather than lost when we assess our constructs across ages, handedness, genders, backgrounds, etc
So great that people are doing this kind of research!
Generalisability is IMPROVED & signal gained rather than lost when we assess our constructs across ages, handedness, genders, backgrounds, etc
So great that people are doing this kind of research!