MétaCan
Menu
Back to cohort
Record W2312871991 · doi:10.1097/ede.0000000000000015

Estimating Causal Effect with RCTs

2013· letter· en· W2312871991 on OpenAlexaffabout
Ian Shrier

Bibliographic record

VenueEpidemiology · 2013
Typeletter
Languageen
FieldMathematics
TopicAdvanced Causal Inference Techniques
Canadian institutionsJewish General Hospital
Fundersnot available
KeywordsRandomized controlled trialContext (archaeology)MedicineCovertDeceptionPhysical therapyPsychologyInternal medicineSocial psychology

Abstract

fetched live from OpenAlex

The author responds: Dr. Wolfson1 highlights important nuances to my discussion2 of using randomized clinical trials (RCTs) to estimate causal effects in the clinical setting. We agree on the following points: A deception RCT (where patients are deceived into believing a treatment is effective) matches the clinical context when a patient is offered a choice between two “active” treatments, with only one actually being effective. The unblinded RCT matches the clinical context where patients are offered active treatment or no treatment. The unblinded RCT cannot distinguish between effects mediated through physiological processes versus psychological processes (which themselves, may be mediated through physiological or behavioral processes). A “covert” RCT measures a biological effect corresponding to the context where subjects believe they will receive inactive treatment. A covert RCT is one type of deception RCT. In a deception RCT, subjects believe they will receive a treatment (reference); the reference treatment may be active (eg, caffeinated coffee) or inactive (eg, decaffeinated coffee). Furthermore, subjects may not even know they are in a study. We differ in some minor areas. Dr. Wolfson’s definition of biological effect1 is restricted to causal effects when subjects do not know they are receiving active treatment. However, one could envision a medication that lowers sympathetic activity (heart rate) when sympathetic activity is elevated (eg, subjects believe they will receive caffeine) but not when sympathetic activity is “normal” (eg, subject believe they will not receive caffeine). The “biological” effect would be different for each level of the variable “sympathetic activity,” with each effect measured by a different deception RCT. Considering one variable level as more important than another variable level is a value judgment. Although deception RCTs may “generally be regarded as unethical”1 when evaluating treatments with unknown side-effect profiles, the method is applicable where the risk of adverse effects is nil or the minimal effects considered minor. Although clinicians may sometimes choose between offering versus not offering treatment, I believe they usually choose between several possible treatments that are potentially active. For example, there are many exercise programs promoted as effective treatment for ankle sprains, and effective treatment options for an anterior cruciate ligament tear include exercise, as well as several surgical procedures. In these contexts, the deception RCT where subjects are told a treatment is effective mimics the context of interest. One should evaluate on a case-by-case basis whether traditional RCTs or observational studies provide the less-biased estimate for the causal effect of interest. Ian Shrier Centre for Clinical Epidemiology, Jewish General Hospital, Montreal, QC, Canada, [email protected]

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.002
metaresearch head score (Gemma)0.019
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMetaresearch, Meta-epidemiology (narrow), Research integrity
Consensus categoriesResearch integrity
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Not applicable · Consensus signal: none
GenreCandidate signal: Methods · Consensus signal: Methods
Teacher disagreement score0.836
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0020.019
Meta-epidemiology (narrow)0.0010.000
Meta-epidemiology (broad)0.0020.000
Bibliometrics0.0000.000
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0010.000
Research integrity0.0020.004
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.198
GPT teacher head0.445
Teacher spread0.247 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; both teacher heads agree on what is shown here.

Study designNot applicable
Domainnot available
GenreMethods

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations1
Published2013
Admission routes2
Has abstractyes

Explore more

Same venueEpidemiologySame topicAdvanced Causal Inference TechniquesFrench-language works237,207