Thiago
@csthiago.bsky.social
Methodologist/Epidemiologist
London - UK
London - UK
Would it be possible to get the underlying data that they have used? In the paper, they just say is public
November 24, 2025 at 12:26 PM
Would it be possible to get the underlying data that they have used? In the paper, they just say is public
Reposted by Thiago
The solutions of the past 3 decades have failed to change the incentives of #PublishOrPerish. As a result, researcher funding, time, control, and trust has been lost.
The ONE CONSTANT in the wake of the serial crisis, #PlanS and #OpenAccess reform has been publish profit margins.
2/n
The ONE CONSTANT in the wake of the serial crisis, #PlanS and #OpenAccess reform has been publish profit margins.
2/n
November 11, 2025 at 11:52 AM
The solutions of the past 3 decades have failed to change the incentives of #PublishOrPerish. As a result, researcher funding, time, control, and trust has been lost.
The ONE CONSTANT in the wake of the serial crisis, #PlanS and #OpenAccess reform has been publish profit margins.
2/n
The ONE CONSTANT in the wake of the serial crisis, #PlanS and #OpenAccess reform has been publish profit margins.
2/n
It can only change if the rewards change. Funding / promotions / hiring /etc. I can only see this working in a top down approach
November 12, 2025 at 1:24 PM
It can only change if the rewards change. Funding / promotions / hiring /etc. I can only see this working in a top down approach
I have drafted one response about the cognitive dissonance. However, I was convinced by my co-authors to just comply for a less stressful publication (already back and forth in multiple journals).
November 11, 2025 at 12:27 PM
I have drafted one response about the cognitive dissonance. However, I was convinced by my co-authors to just comply for a less stressful publication (already back and forth in multiple journals).
I had one reviewer forcing me to do that (same family of journal). "You can't use effect", because is observational..... Ok, bro. We are going to infect pregnant women in a rct now
November 11, 2025 at 11:39 AM
I had one reviewer forcing me to do that (same family of journal). "You can't use effect", because is observational..... Ok, bro. We are going to infect pregnant women in a rct now
Not reassuring this early separation already in 2 days
November 9, 2025 at 4:53 PM
Not reassuring this early separation already in 2 days
Ipw in a context of time varying exp, right?
Btw what is the mortality rate of these patients? because of this competing risk (death in proph)
I think ordinal models would fit in this scenario as well. Tagging @f2harrell.bsky.social to please check if is a sensible option
Btw what is the mortality rate of these patients? because of this competing risk (death in proph)
I think ordinal models would fit in this scenario as well. Tagging @f2harrell.bsky.social to please check if is a sensible option
November 8, 2025 at 12:46 PM
Ipw in a context of time varying exp, right?
Btw what is the mortality rate of these patients? because of this competing risk (death in proph)
I think ordinal models would fit in this scenario as well. Tagging @f2harrell.bsky.social to please check if is a sensible option
Btw what is the mortality rate of these patients? because of this competing risk (death in proph)
I think ordinal models would fit in this scenario as well. Tagging @f2harrell.bsky.social to please check if is a sensible option
Because ££££££ \$$$$$ \ €€€€ \¥¥¥¥ (w/e currency)
November 7, 2025 at 9:19 AM
Because ££££££ \$$$$$ \ €€€€ \¥¥¥¥ (w/e currency)
It's a pain. Even more the projects from the North to "fix" the publishing system. Is that hard just admit that already exist something better in the South?
The government money support direct the publisher/scielo. No profit, equal to how science is funded
The government money support direct the publisher/scielo. No profit, equal to how science is funded
November 7, 2025 at 8:11 AM
It's a pain. Even more the projects from the North to "fix" the publishing system. Is that hard just admit that already exist something better in the South?
The government money support direct the publisher/scielo. No profit, equal to how science is funded
The government money support direct the publisher/scielo. No profit, equal to how science is funded
You know Latin America has it (non profit, reputable, diamond) (www.science.org/content/arti...) but because it is global south there is no "prestige" loved by academia/funders
Latin American journals are open-access pioneers. Now, they need an audience
SciElo, other platforms work to help journals reach readers around the world
www.science.org
November 7, 2025 at 7:54 AM
You know Latin America has it (non profit, reputable, diamond) (www.science.org/content/arti...) but because it is global south there is no "prestige" loved by academia/funders
Or because dengue is famous for causing hematological problems (hemorrhage), a substantial misclassification towards dengue of the more "severe cases"
November 6, 2025 at 9:03 PM
Or because dengue is famous for causing hematological problems (hemorrhage), a substantial misclassification towards dengue of the more "severe cases"
Ehr, I have noticed that (and ignored for space constraints). One possibility is the deaths in that being more reported for the underlying disease (we didn't had the information of the cause of death apart from the disease notified). There are some reports showing chikv affecting blood paramaeters
November 6, 2025 at 8:53 PM
Ehr, I have noticed that (and ignored for space constraints). One possibility is the deaths in that being more reported for the underlying disease (we didn't had the information of the cause of death apart from the disease notified). There are some reports showing chikv affecting blood paramaeters
@danlewer.bsky.social @tdbrothers.bsky.social
Hi Dan and Tommy.
Remembered your paper today, using it to explain some concepts to a colleague. And also remembered I have used a "risk factors" recently, could you please roast my paper? hehe
www.sciencedirect.com/science/arti...
Hi Dan and Tommy.
Remembered your paper today, using it to explain some concepts to a colleague. And also remembered I have used a "risk factors" recently, could you please roast my paper? hehe
www.sciencedirect.com/science/arti...
Hospitalisation, mortality and years of life lost among chikungunya and dengue cases in Brazil: a nationwide cohort study, 2015–2024
The incidence of infections from arthropod-borne viruses, including chikungunya and dengue, is increasing globally. We used nationwide data collected …
www.sciencedirect.com
November 6, 2025 at 5:35 PM
@danlewer.bsky.social @tdbrothers.bsky.social
Hi Dan and Tommy.
Remembered your paper today, using it to explain some concepts to a colleague. And also remembered I have used a "risk factors" recently, could you please roast my paper? hehe
www.sciencedirect.com/science/arti...
Hi Dan and Tommy.
Remembered your paper today, using it to explain some concepts to a colleague. And also remembered I have used a "risk factors" recently, could you please roast my paper? hehe
www.sciencedirect.com/science/arti...
Probably from the UK Biobank studies limitation section.
November 3, 2025 at 6:21 PM
Probably from the UK Biobank studies limitation section.
Yeah. I mainly use it with very specific data wrangling things where I know for sure that something worked properly or not. But apart from that, it's very complicated to be certain about the LMS code
October 31, 2025 at 9:59 PM
Yeah. I mainly use it with very specific data wrangling things where I know for sure that something worked properly or not. But apart from that, it's very complicated to be certain about the LMS code
I use 42 as seed. I m always trying to answer the ultimate question! Haha.
Ehr, about llm. Even within staff I m seeing this. Now everyone is a R expert, creating (llm) code for very complicated models out of thin air
Ehr, about llm. Even within staff I m seeing this. Now everyone is a R expert, creating (llm) code for very complicated models out of thin air
October 31, 2025 at 3:55 AM
I use 42 as seed. I m always trying to answer the ultimate question! Haha.
Ehr, about llm. Even within staff I m seeing this. Now everyone is a R expert, creating (llm) code for very complicated models out of thin air
Ehr, about llm. Even within staff I m seeing this. Now everyone is a R expert, creating (llm) code for very complicated models out of thin air
Hmm. I think the umbrella of the black box also catches the FAW. There's a commentary from Greenland and colleagues that defines black box epi as an atheoretical search for associations (10.1097/01.ede.0000134867.12896.23)
October 29, 2025 at 6:58 PM
Hmm. I think the umbrella of the black box also catches the FAW. There's a commentary from Greenland and colleagues that defines black box epi as an atheoretical search for associations (10.1097/01.ede.0000134867.12896.23)
Hmm. I would consider those tables as "factors associated", as there are some stratifications/etc. Also, in 1944 probably wasn't straightforward to do regression. I'm not totally against these types of studies, especially in fields with no substantial prior information
October 29, 2025 at 10:34 AM
Hmm. I would consider those tables as "factors associated", as there are some stratifications/etc. Also, in 1944 probably wasn't straightforward to do regression. I'm not totally against these types of studies, especially in fields with no substantial prior information
This discussion (factors associated) is very old. But the name before was "black box Epidemiology". Weiss (10.1097/01.ede.0000135175.11460.27) comment about how sudden infant death vs prone position sleep was first seen in black box epi.
October 29, 2025 at 8:45 AM
This discussion (factors associated) is very old. But the name before was "black box Epidemiology". Weiss (10.1097/01.ede.0000135175.11460.27) comment about how sudden infant death vs prone position sleep was first seen in black box epi.
The worst part is that a lot of people think the same, but without owning an AI/ knowing how biased Grok is towards Musk's opinions (my assumption of AI/person). The grok even checked Musk's Twitter to reply to "sensible topics"
October 20, 2025 at 5:13 PM
The worst part is that a lot of people think the same, but without owning an AI/ knowing how biased Grok is towards Musk's opinions (my assumption of AI/person). The grok even checked Musk's Twitter to reply to "sensible topics"