Author: okorie

  • Did Treatment Expectations Predict Response in a Psilocybin–Escitalopram Trial?

    Participants in a randomized comparison of investigational psilocybin and escitalopram entered the trial with higher expectations for psilocybin. Yet higher expectations for psilocybin did not predict therapeutic response in that arm. Higher expectations for escitalopram were associated with better outcomes among participants assigned to escitalopram.

    Expectancy and suggestibility measured different things

    The analysis used data from 55 participants with major depressive disorder. The researchers used mixed linear models to examine how pretreatment efficacy expectations and baseline traits related to outcomes in each treatment arm. Trait suggestibility was associated with therapeutic response in the psilocybin arm. The analysis did not find that association in the escitalopram arm.

    What randomization can tell us

    The trial randomized treatment assignment. Participants’ expectations and suggestibility were measured characteristics. Their associations with outcomes are observational, even within a randomized trial. The sample size also limits how much weight to give an association that appears in one arm or a result that fails to reach statistical significance in the other.

    Pretreatment expectancy answers a specific question about beliefs before exposure. Whether participants inferred their assignment after experiencing treatment effects is a separate question about masking. A report assessing bias should specify which of those questions the available evidence addresses.

    Source Assessing expectancy and suggestibility in a trial of escitalopram v. psilocybin for depression, Psychological Medicine.

  • Why a Panic fMRI Study Included Generally Aversive Pictures

    Three picture cards sit beside a stylized brain scanner

    Engel and colleagues compared 19 participants with panic disorder and agoraphobia with 21 healthy matched controls. They used functional MRI while participants viewed pictures. For participants in the clinical group, the study selected panic-related scenes based on individual ratings. The experiment also included generally aversive and neutral pictures.

    The two comparison sets served different purposes. Neutral images gave the researchers a reference for examining the response to panic-related scenes. Generally aversive images let them examine how the groups responded to unpleasant material outside those panic-related scenes. The researchers reported greater activation in several regions in the clinical group for the panic-related versus neutral contrast. The generally aversive images activated similar regions in both groups.

    For someone evaluating a claim based on this experiment, the contrast belongs alongside the finding. A response to a particular set of rated pictures is conditional on how those pictures were selected and what they were compared with. The study reported group patterns under laboratory conditions and did not establish a diagnostic rule for individual scans.

    Engel et al., European Archives of Psychiatry and Clinical Neuroscience (2016)

  • What the Wide Confidence Interval Means in an AKT1 and Cannabis Study

    A point on a horizontal line with end caps illustrates an estimate and its confidence interval without a data scale

    Di Forti and colleagues studied 489 people with first-episode psychosis and 278 controls. They examined whether the association between cannabis use and psychosis varied by genotype at a site in the AKT1 gene. The headline result for daily users compared people with the C/C genotype with people with the T/T genotype within that daily-use group.

    The estimated odds ratio was 7.23, with a 95 percent confidence interval from 1.37 to 38.12. The interval conveys how imprecisely the study estimated the size of the association. A summary that gives only the point estimate leaves out how widely the values compatible with the data and model assumptions span.

    The comparison was between two genotype groups among daily users. It did not compare daily users with people who had never used cannabis. The study selected participants by case or control status, so the estimate also does not give the probability that a particular person will develop psychosis.

    When citing the result, keep the comparison group and interval in the same sentence as the estimate. Those details make clear which question the study addressed and how much uncertainty accompanied its answer.

    Di Forti et al., Biological Psychiatry

  • Why a Stop-Signal Study Used Orthopedic Patients as Controls

    A response arrow interrupted by a stop signal beside two comparison folders

    A stop-signal task estimates how long it takes a person to cancel a response already underway. Successful cancellation leaves no button press to time, so researchers estimate the stopping time from performance across trials. In a 2003 study, Rieger and colleagues used that estimate to examine whether the location of a brain lesion was associated with stopping performance.

    The study enrolled 17 patients with frontal lesions, 20 with nonfrontal lesions, eight with basal ganglia lesions, and 20 orthopedic patients as controls. The researchers selected the nonfrontal and orthopedic groups to match the frontal and basal ganglia groups as closely as possible on age, sex and years of education. The orthopedic group supplied a clinical comparison outside the brain-lesion groups.

    The frontal and basal ganglia groups had longer estimated stopping times than the orthopedic group. That result is a group comparison under the study’s task conditions. The study did not establish a stopping-time cutoff that identifies the location of an individual’s lesion. The basal ganglia group included just eight patients, an important detail when judging how much weight to give the group result.

    When a report cites this study, check which comparison it relies on. Comparing a lesion group with orthopedic patients addresses a different question from comparing two groups with lesions in different locations. The comparator belongs alongside the finding when describing what the study showed.

    Rieger et al., Neuropsychology (2003)

  • Who Made It Into the CBD and THC Expectations Survey?

    Two questionnaires separated by a funnel representing the familiarity screen

    A survey of CBD and THC expectations started with 520 adults in Canada and analyzed responses from 345. The difference reflects a decision about whose beliefs the researchers wanted to measure. Spinella and colleagues included respondents who reported at least some knowledge of both compounds. The 175 excluded respondents reported no knowledge of one or both.

    Who was represented in the analysis?

    The screen was about familiarity, which the respondents reported themselves. The analyzed group included 287 people who had previously used cannabis and 58 who had never used it. So the comparison between those groups was a comparison among people who said they knew at least something about both CBD and THC. Someone could qualify without ever having used cannabis.

    The remaining respondents rated statements about CBD, THC and the combination. Ratings generally favored CBD-containing options for perceived therapeutic effects. Those results describe expectations under the survey design. The researchers recruited through an online panel, measured beliefs at one point in time and did not administer either compound.

    What to look for when the survey is cited

    Check which denominator a summary uses. The 520 respondents describe the larger survey intake, while 345 describes the group selected for this analysis. The paper also reports removing extreme outliers in individual models, so the number contributing to a particular result could be smaller still. Keeping those stages distinct makes it easier to see how far a statement about the results can reach.

    Spinella et al. (2023), Addictive Behaviors

  • Who Was Represented on a Historical Forensic Science Review Board

    A meeting table with seats arranged around it

    During a discussion of oversight, Barry Scheck pointed to a New York forensic science review board whose membership included judges and defense lawyers. The example came up as workshop participants considered ways that institutions outside individual laboratories might contribute to oversight.

    Membership identifies who can bring a perspective to the table. Evaluating how oversight worked in practice would also require information about what the board could review, how issues reached it and what happened after review. The summary reports the example without measuring the board’s effect on laboratory performance.

    Institute of Medicine, Medicolegal Death Investigation System (2003), p. 36

  • Why a Prosecutor Valued a Second Signature on Autopsy Reports

    A report bearing two stylized signatures

    At the 2003 workshop, prosecutor Jeff Dusek described San Diego County’s practice of having a second medical examiner sign the autopsy report. He emphasized continuity if the original examiner became unavailable to testify. The account gives a specific reason for the practice from the prosecutor’s perspective.

    A signature identifies another person associated with the report. To understand the scope of that involvement in a case, the record would need to show what the second examiner reviewed and whether that examiner documented any separate observations. The workshop summary describes the practice without specifying that level of detail.

    Institute of Medicine, Medicolegal Death Investigation System (2003), p. 31

  • Why Corrective Action Records Matter in Quality Assurance

    Two panels labeled Review finding and Corrective action

    The Institute of Medicine’s 2003 workshop discussed quality assurance programs that used scheduled review and documentation of corrective actions. A written policy describes how a review is intended to work. The resulting records show how an identified problem was handled.

    When reviewing a particular discrepancy, trace the finding to any documented response. The records can show whether the issue affected the case and what steps the office recorded in response.

    Institute of Medicine, Medicolegal Death Investigation System (2003), p. 21

  • Why Access to Forensic Specialists Was a Focus of Reform

    Two panels labeled Local caseload and Specialist access

    The Institute of Medicine’s 2003 workshop recorded a proposal for referral arrangements where local offices lacked access to qualified forensic pathologists. The discussion linked the proposal to differences in population and resources.

    The historical proposal illustrates why the location of the investigating office and the location of the specialist examination can differ. A case review can identify who performed the examination and which office held the underlying records.

    Institute of Medicine, Medicolegal Death Investigation System (2003), p. 11

  • How to Attribute a Statement From a Forensic Workshop Summary

    Two panels labeled Speaker viewpoint and Workshop record

    The 2003 Institute of Medicine workshop summary records presentations and discussions about medicolegal death investigation. The introduction explains that its committee organized the workshop and took responsibility for the accuracy of the summary. The committee was convened to oversee the workshop.

    When using the report in an evidence review, identify the presenter or discussant whose view is being cited. A statement presented at the workshop has a different status from a conclusion adopted by a committee. Precise attribution makes that distinction clear.

    Institute of Medicine, Medicolegal Death Investigation System (2003), p. 2