How to prepare a REMS assessment report

How do you write a REMS assessment report that supports a defensible conclusion?

Use the operative REMS versions, assessment plan, reporting interval and actual data cutoff. Report each metric with its population, definitions, method, result and limitations, then assess each goal separately. Explain missing evidence and proposed next actions instead of converting incomplete data into success. The 2019 FDA assessment guidance remains draft; the approved program and applicable agency instructions establish the actual reporting baseline.

Before you begin

Assessment reporting for an actual REMS. Use the approved REMS, assessment timetable, agreed plan and agency correspondence as the operative baseline. FDA’s 2019 assessment guidance remains draft.

What you will prepare: An assessment report linking each required or agreed question to its method, result, limitation and proposed action.

Identify the period and program actually assessed

Collect the operative REMS versions, timetable, assessment plan, instruments, prior assessments and correspondence. Specify reporting period and data cutoff. If the program changed during the interval, identify which requirements and materials were in effect when. The 2019 draft suggests a report organization; it does not replace the product’s approved requirements or establish a universal deadline.

Report every assessment question transparently

Organize results by the actual plan. For each metric, identify the source, population, numerator/denominator where relevant, method, missingness and result. Distinguish process completion from meaningful risk mitigation. If a metric cannot be calculated, explain why, the impact on interpretation and the proposed remedy. Do not substitute a convenient available measure and report the original question as answered.

Explain what the evidence can and cannot establish

Compare findings with the relevant objective and prior period using compatible definitions. Identify operational changes, survey selection, incomplete follow-up or other limitations. A high training-completion rate does not automatically prove correct clinical behavior. Proposals to change or eliminate an element require a reasoned case and the appropriate separate process; a favorable report does not itself modify the REMS.

Fictional example: the denominator changed

A dashboard shows improvement, but the prior report counted all eligible participants while this report counts only respondents. Recalculate on a justified comparable basis or explain why a trend cannot be inferred. If the protocol version cannot be found, recover it before claiming the prespecified analysis was followed. Keep an unresolved result visible rather than turning missing data into a passing assessment.

Make each conclusion traceable to a defined result

Build a result record for every question in the actual assessment plan before writing the executive summary. Preserve measures that were not calculable or not collected; deleting them makes a partial assessment look complete.

Fictional missing-data exercise: 1,000 participants were eligible for a knowledge survey, 200 responded and 180 respondents answered a key question correctly. The observed result is 90% among respondents, with a 20% response rate. It does not establish 90% knowledge among all eligible participants. Conversely, 180 divided by 1,000 is not a measured 18% knowledge rate, because the other answers are unknown. Assess nonresponse and the planned analysis, and report the limitation explicitly.

Reconcile period-specific and cumulative numbers using their definitions. Do not add cumulative counts across reports and count the same participants repeatedly. When the program, eligibility rule or instrument changes, show the break in comparability and provide an appropriate stratified or otherwise justified analysis. Explain why a trend is interpretable before describing it as improvement.

The January 2019 draft recommends discussing results for individual goals and the program overall. A program can have evidence that one objective is being met while another is unresolved or not met. Avoid a single favorable headline that hides the distinction. Where evidence is insufficient, identify the additional data or analysis that could resolve the question.

For REMS with elements to assure safe use, the draft also discusses patient access and burden on the health-care system. Consider evidence about delays, abandonment, workflow difficulty and participant experience where relevant to the actual plan. Low use alone does not prove the REMS caused an access problem; investigate alternative explanations and the limitations of each source.

Use the methodology guide to describe what was planned and departures from it. Use the change-history guide for proposed modifications. An assessment recommendation to remove an element is not itself authority to stop performing an approved requirement.

Frequently asked questions

Sources and revisions

Requirements, source recommendations and editorial preparation advice have different roles. Review the scope and revision of the source you use.

Related guides