Indirect treatment comparison

Every placebo controlled trial found, matched on timepoint and estimand, and compared through placebo by Bucher and network meta analysis. Every input traced to its source

Customize this Agent

DOCXXLSX15 steps

The problem

Enlicitide cuts LDL cholesterol 56% against placebo at week 24, or 63% on treatment

CORALreef Lipids cut LDL cholesterol 55.8 percentage points more than placebo at week 24 counting every patient randomized, and 62.9 counting only patients still taking the drug. Inclisiran's pivotal trials report at day 510. Set the oral's on treatment figure against an injectable's treatment policy figure and the oral gains seven percentage points before any comparison is run

The request

Lipfendra was approved July 16 with no head to head trial against the injectables. For our P&T committee, compare enlicitide with evolocumab, alirocumab and inclisiran on percent change in LDL cholesterol, anchored on placebo. Find every placebo controlled trial of the four. Compare at week 24 and week 52, or each trial's nearest visit within 12 weeks, on the treatment policy estimand throughout. Run heterozygous familial hypercholesterolemia as its own network. Run Bucher comparisons and fixed and random effects network meta analysis, with heterogeneity, consistency and sensitivity analyses. Flag every trial whose timepoint, estimand, background statin or LDL assay differs. Reconcile to the oral PCSK9 network meta analysis in Endocrine Practice. I need the evidence workbook and technical report.

How Axe works

ApprovedChanges asked

Your reviewApprove, or say what to change

Approved or changes

Technical report and workbookReport, workbook and check log

Final reviewApprove the revised files

Revised technical report and workbookRevised files and check log

Step

Match timepoints and estimands

Prompt

Sent to the model at this step

Choose the one result from each trial that enters each analysis, by the protocol's rules, and show why. A trial's headline figure is often the wrong one. CORALreef Lipids reports 55.8 percentage points against placebo at week 24 on treatment policy and 62.9 on treatment: the treatment policy figure enters, and the on treatment figure goes to the sensitivity analysis. ORION-10 and ORION-11 report at day 510 and as a time adjusted average after day 90 up to day 540, neither of them week 24 or week 52: find the visit inside each window in the supplements and FDA review; where there is none, take the nearest visit no more than 12 weeks from the target, flagged with its distance. LAPLACE-2 and RUTHERFORD-2 report at week 12 and the mean of weeks 10 and 12, outside both windows: enter them at week 24, flagged at 12 weeks from target, and keep them out of week 52. Record where each visit falls in the dosing interval of a monthly or six monthly injection, at peak or at trough, since the daily oral has neither. Never convert an on treatment result into a treatment policy one; a trial that reports only on treatment enters the sensitivity analysis, flagged. Compare placebo arms before trusting any contrast: placebo rose 1.0 percent at day 510 in ORION-10, 4.0 in ORION-11 and 3.0 at week 24 in CORALreef Lipids, so every comparison runs on placebo adjusted differences, never on active arms alone. Build the analysis set: one row per trial, population, timepoint and contrast, with the estimand, the visit used and its distance from target, the dosing interval position, and a flag for each difference in timepoint, estimand, background statin or LDL cholesterol method.

Uses

Protocol
Trial register
Extraction

Produces

Analysis set
Selection notes

What you get back

DOCXTechnical reportResults first, then methods, similarity, network diagrams, forest plots, league tables and limitations

XLSXEvidence workbookSearch, screening, trials, risk of bias, arm data, analysis set, network, results, sensitivity, reconciliation

Check logOpen items first, then each value and result marked checked or corrected

Make it yours

Change the drugs, populations, timepoints or estimand, or start from your committee's own PICO. Every step and every prompt can be edited

Customize this Agent

Opens a private copy of this Agent in your workspace, ready to adapt to your own sources

More agent examples

Life sciences analystsClinical trial landscapeXLSXDOCX14 steps

Pharma forecastersPatient based forecastXLSXDOCX16 steps

Product strategy teamsTarget product profileDOCXXLSX15 steps

Capital markets analystsBroker opinion of valuePPTXXLSX16 steps

Financial planning analystsBudget variance analysisXLSXDOCX13 steps