AI and Formulary Decisions: Drafting P&T Committee Memos

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-creators-healthcare-AI-and-pharmacy-formulary-r13a6-adults

1. What is the primary time-saving benefit of using AI to draft a P&T committee memo for a new biologic agent?

   1. AI eliminates the need for the pharmacist to read the clinical trial publications
   2. AI drafts the complete 10-page memo in under an hour without any human input
   3. AI automatically negotiates with drug manufacturers on pricing
   4. AI synthesizes evidence and generates draft sections, allowing the pharmacist to focus on verification

2. A pharmacist is preparing a formulary addition for a new biologic. Which task would be MOST appropriate to delegate entirely to AI?

   1. Verifying that the wholesaler price matches the 340B contract price
   2. Determining whether the committee previously rejected a similar agent due to budget constraints
   3. Identifying undisclosed conflicts of interest in the trial authors' disclosures
   4. Synthesizing efficacy data from four randomized controlled trials into a comparison table

3. The lesson warns that AI models may 'confidently misstate' which elements of clinical trial data?

   1. Primary endpoints, p-values, and adverse event rates
   2. FDA approval dates and labeling information
   3. The drug's mechanism of action and molecular structure
   4. The names of all co-investigators listed in the author byline

4. Before presenting an AI-generated cost-per-responder calculation to the P&T committee, the pharmacist must independently verify:

   1. The specific 340B and wholesaler contract prices applicable to their health system
   2. The publication date of the clinical trials analyzed
   3. The number of citations included in the reference list
   4. The statistical software version used by the AI model

5. Why might an AI-generated P&T memo fail to capture relevant conflicts of interest?

   1. Journal-level conflicts of interest are often in separate disclosure statements not included in trial PDFs
   2. AI models automatically detect and remove all industry funding disclosures
   3. AI is required by regulation to reproduce all conflict of interest information verbatim
   4. Pharmacy & Therapeutics committees do not consider conflicts of interest relevant to formulary decisions

6. To generate an accurate cost analysis section, the pharmacist must provide which input to the AI?

   1. The specific drug prices from the institution's 340B and wholesaler contracts
   2. The complete list of other drugs the committee rejected last year
   3. The FDA package insert for the new biologic agent
   4. The names of all P&T committee members and their voting history

7. What prevents AI from serving as the final author of a formulary status recommendation?

   1. AI does not have access to the committee's voting history on similar agents and cannot weigh institutional priorities
   2. AI lacks the professional license required to make formulary recommendations
   3. AI cannot distinguish between brand-name and generic medications
   4. AI is prohibited from accessing PubMed and clinical trial databases

8. When reviewing an AI-generated safety summary for a biologic agent, the pharmacist should verify:

   1. The number of pages in the original clinical trial publications
   2. The publishing house that printed the medical journals
   3. The exact adverse event frequencies and severity classifications reported in the source trials
   4. The font formatting used in the draft document

9. How does AI's ability to synthesize multiple RCTs change the pharmacist's traditional workflow for P&T memos?

   1. It shifts the pharmacist's role from manual synthesis to verification and contextualization
   2. It eliminates the need for pharmacists to read clinical trial papers entirely
   3. It allows pharmacists to skip the P&T committee presentation entirely
   4. It enables pharmacists to delegate all final decision-making to the AI system

10. A health system operates under a 340B drug pricing program. Why would AI-generated cost data be unreliable without pharmacist intervention?

    1. The FDA requires all drug pricing to be listed on the medication package insert
    2. AI models are legally prohibited from accessing 340B pricing information
    3. 340B pricing is not published in peer-reviewed literature and varies by institution
    4. Pharmacy & Therapeutics committees are not permitted to consider 340B pricing in decisions

11. Which type of information in a P&T memo would require knowledge of the committee's past decisions that AI cannot provide?

    1. The mechanism of action for the new biologic agent
    2. Whether the committee previously rejected a similar agent due to cost-effectiveness concerns
    3. The FDA-approved indications for the new drug
    4. The standard dosing regimen from clinical trials

12. In formulary decision-making, 'place in therapy' refers to:

    1. The specific hospital unit or clinic where the drug may be administered
    2. The position of the drug's name alphabetically in the formulary guide
    3. How the new agent compares to existing formulary options for treating the same condition
    4. The physical location where the drug will be stored in the pharmacy inventory

13. Why might AI-generated p-values from clinical trials be misleading?

    1. Clinical trials do not report p-values in their published results
    2. AI models have access to the actual raw patient-level data from trials
    3. The lesson warns that AI confidently misstates primary endpoints, p-values, and adverse-event rates
    4. P-values are not considered important for formulary decisions

14. What is the pharmacist's primary responsibility when using AI to draft a P&T committee memo?

    1. To delegate the entire memo to AI and simply submit the generated document
    2. To serve as a witness that the AI system is functioning properly
    3. To verify all quantitative claims, pricing, and context-specific information before committee presentation
    4. To only review the introduction and conclusion sections, trusting the AI with data tables

15. A pharmacist inputs four clinical trial PDFs and current formulary documents into an AI system requesting a P&T memo. What output can they reasonably expect without additional verification?

    1. A draft requiring verification of all data, prices, and context-specific elements
    2. A document that cannot be edited or revised by the pharmacist
    3. An automatically approved formulary status recommendation
    4. A final, committee-ready document requiring no further review