Kelsey
Menu

Features

Ask the way you would ask a colleague: the drug, the side effect, and what you need to know. Kelsey answers from the public record, and every answer links to source documents. Below, each feature with an example question and what comes back.

Ask in plain English

Kelsey
Product illustration. Question: What did published studies report about eye risk with redacted before the label warned of it? Answer: A six-patient case series published in 2018 reported pigmentary maculopathy in patients with long-term redacted exposure and called for further investigation of causality [1]. FDA’s review describes the manufacturer’s June 2019 labeling submission [2]. FDA approved the revised warning in June 2020 [3]. Sources: [1] Pearce et al., Ophthalmology, 2018, published; [2] FDA review describing the manufacturer’s June 2019 labeling submission, 24 Jun 2019, submitted; [3] redacted label, Warnings, 16 Jun 2020, approved. Source 3 is open: redacted label · Warnings · approved 16 June 2020. Highlighted passage: Summary of the June 2020 label: retinal pigmentary changes have been reported with long-term use of the drug, redacted.

Kelsey has masked names for presentation; the masking is not part of the cited originals. Animation timing does not represent measured product speed.

Every answer shows the cited passage behind it, what kind of date each date is (published, submitted to FDA, or approved by FDA), and where each calculated figure comes from. When nothing is found, it says which sources were searched and what period they cover. Under every answer: "Written by AI. Check the figures against the exhibits and every quotation against its source."

What did published studies report about eye risk with before the label warned of it?

What comes back: a six-patient case series published in 2018 that called for further investigation of causality, FDA's review describing the manufacturer's June 2019 labeling submission, and the label after FDA's approval of 16 June 2020, opened at the warning sentence.

Adverse-event reports

Signal

Reported more often than other drugs?

· gastroparesis

All reports, including those filed by lawyers.

Bar lengths are illustrative and do not represent results for a particular drug. Adverse-event reports do not establish incidence or causation.

Product illustration, schematic with no real figures: a bar for one drug’s reports mentioning gastroparesis against all other drugs, with and without lawyer-filed reports.

Kelsey has masked names for presentation; the masking is not part of the cited originals. Animation timing does not represent measured product speed.

Kelsey searches reports of suspected side effects sent to FDA, from 2004. It checks whether a drug is reported with a side effect more often than other drugs are, and shows the count with and without reports filed by lawyers. Reports that look like repeats are flagged, and the result is shown with and without them. A higher-than-expected count is a reason to look closer. It does not show that the drug caused the event.[1]

Is reported with gastroparesis more often than other drugs? Show how it developed over time.

What comes back: using the current dataset and the documented calculation, the reconstructed series first crosses Kelsey's threshold in the quarter ending 30 September 2018, the count by quarter since, and the date of the first lawyer-filed report (February 2024). Later follow-ups may affect that result; it does not show what was detectable in 2018.

Label history

Label history
Question

Our client started in 2016. Which label versions were available before then?

Among the versions reviewed, the latest published before then did not name the reaction1. The warning was approved on 16 June 20202.

  • 8 earlier versions did not mention it
  • 6 versions carry the warning
  1. 1Label versions before the warningApr 2010 to Sep 2018published
  2. 2 label, Warnings16 Jun 2020approved
Product illustration. Label history: 14 label versions published from April 2010 to September 2024. The 8 earlier versions, published April 2010 to September 2018, did not mention the reaction; the warning appears from June 2020, approved 16 June 2020. The mock shows availability, not the labeling supplied with a particular dispensed product.

Kelsey has masked names for presentation; the masking is not part of the cited originals. Animation timing does not represent measured product speed.

Kelsey holds label versions and FDA's approval records. You see what the available label versions said before a given day, when a warning first appeared, and who held the application. Brand and generic makers are listed apart, because a generic label has to match the brand's.[2]

Show the label history of .

What comes back: "FDA approved the warning on 16 June 2020. The identified SPL version is dated 19 June 2020; the published label lists its publication date as 23 June 2020. The 8 earlier versions (published 1 April 2010 to 6 September 2018) did not mention it." Document, approval and publication dates are different.

FDA reviews and decisions

Kelsey holds FDA's approval reviews and quotes them by page. It also holds FDA's safety communications, safety-related label changes, advisory committee material, citizen petitions, warning letters and guidance. When FDA has taken a page down, Kelsey links the archived copy and says so.

What did FDA's reviewers say about this risk at approval, and did the company later propose a warning FDA declined?

What comes back: the reviewers' passages with page numbers, the company's submissions and FDA's response, each dated.

Studies and trials

Timeline

: from study to warning

  1. 2018 published

    A study in Ophthalmology describes eye damage in long-term users

    Pearce et al.

  2. 24 Jun 2019 submitted

    The manufacturer asks FDA to add a warning

    Request to FDA

  3. 16 Jun 2020 approved

    FDA approves the warning in the label

    label, Warnings

Kelsey has masked names for presentation; the masking is not part of the cited originals. Animation timing does not represent measured product speed.

Kelsey holds studies and trial results. It also holds the actions of regulators abroad, each dated. Trial numbers are not pooled across trials.

List the published studies on and tendon rupture.

What comes back: the publications found by the documented search, with the first publication date and the counts by study design, then each study with its date, journal, type and quoted finding.

Devices

Kelsey holds FDA's device reports, recalls, clearances and approvals, and shows whether FDA cleared a device or approved it through premarket approval. Reports do not show how many devices were in use, and a report does not prove cause.

What was reported to FDA about this device before the recall?

What comes back: the reports and the recall in date order, with FDA's recall class and a link to FDA's record for each.

Law and case law

Kelsey
Product illustration. Question: What did the generic maker argue on preemption in the redacted case? Answer: The preemption motion in the 2019 redacted decision came from the brand maker, not a generic maker [1]. For generic makers, a 2011 Supreme Court decision holds that federal law requires their label to match the brand’s [2]. Sources: [1] A 2019 federal district court decision, decided 2019; [2] A 2011 Supreme Court decision on generic labels, decided 2011. Source 2 is open: A 2011 Supreme Court decision on generic labels. Highlighted holding, paraphrased: federal law requires a generic drug's label to be the same as the brand-name label, so a generic maker cannot change its label on its own to meet a state-law duty to warn; state claims that would require it are preempted.

Kelsey has masked names for presentation; the masking is not part of the cited originals. Animation timing does not represent measured product speed.

Kelsey holds US court decisions, FDA guidance, the current Federal Food, Drug, and Cosmetic Act, federal regulations as they read on a past date where Kelsey holds that year, and four states' product-liability statutes. They are cited in the same answer as the labels and reports. Counsel determines the applicable law and its application.

Which Rule 702 rulings accepted or rejected statistics drawn from FDA's adverse-event reports?

What comes back: the rulings, with court, date and the passage on the method, for both outcomes.

Watch lists

Alert

Watch list

Added since the previous check

  • New studyJournalpublished
  • Foreign regulator warningRegulator websitedecided
  • New label versionFDA labelapproved

Each check compares with the one before and links every item to its source.

Product illustration, schematic: a watch list lists three items added since the previous check, a new study, a foreign regulator warning and a new label version, each linked to its source.

Kelsey has masked names for presentation; the masking is not part of the cited originals. Animation timing does not represent measured product speed.

A watch list is a saved question about a drug or a condition. Each check compares the result with the previous one. When the dated evidence has changed, Kelsey adds one alert listing what was added and what dropped out, each linked to its source.

Which GLP-1 drugs gained new evidence of pancreatitis in the last 90 days, and did any foreign regulator act first?

What comes back: each drug's new dated items, with foreign regulators' actions and their dates.

Exports and reports

Every answer can leave Kelsey as a PDF or an editable Word file. Eligible stored-data calculations can be rerun by an expert with access to the specified code, software environment and dataset versions. Document-derived findings are excluded, and the pack identifies those exclusions. Tables can be downloaded as spreadsheets.

Which terms make up "eye damage", and where does each number come from?

What comes back: the list of terms and the data behind each calculated figure.

Kelsey is a working product. Ask for a test link.

Sources (2)
  1. FDA, "FDA Adverse Event Monitoring System (AEMS) Public Dashboard," http://www.fda.gov/drugs/fda-adverse-event-monitoring-system-aems/fda-adverse-event-monitoring-system-aems-public-dashboard
  2. PLIVA, Inc. v. Mensing, 564 U.S. 604 (2011), decided June 23, 2011. https://www.law.cornell.edu/supct/html/09-993.ZO.html