A participant timeline can make a complicated clinical history easier to examine. Its value comes from organizing context, not from automatically proving a causal relationship. A sequence of dosing, symptoms, laboratory values, and concomitant treatments is evidence to review, with uncertainty and provenance preserved.
The FDA’s premarketing safety guidance provides a broader evaluation reference. This article proposes design and review practices for patient profiles used alongside aggregate analyses.
1. Use a common time reference
A profile should make clear whether its axis uses calendar time, study day, or time relative to an intervention. Explain the reference point and preserve relevant dates. Misaligned clocks can make an event appear to precede or follow a treatment incorrectly.
For a hypothetical laboratory excursion, show the measurement date, dose history, and relevant treatment interruptions on a consistent axis. If only a partial date is known, communicate that uncertainty rather than placing the event at an apparently exact point without explanation.
2. Separate observation from interpretation
An elevated value is an observation. A judgment that it is clinically significant or treatment-related is an interpretation requiring appropriate expertise. Use labels that distinguish collected assessments, derived flags, and reviewer annotations.
Keep the derivation rules for plotted values and flags accessible. If units are standardized, preserve the original units and conversion method in the supporting record. A visually smooth trajectory should not hide incomparable assays or changes in reference ranges.
3. Show missingness honestly
Connecting points across a long gap can imply continuous observation. A profile should reveal assessment gaps and distinguish planned assessments from those actually collected. Avoid substituting a value for an absent measurement merely to make the chart easier to read.
ICH E9(R1) is relevant when patient events affect the treatment-effect question. Profiles can inform that discussion, but an exploratory visual review should not quietly redefine the prespecified analysis.
4. Preserve the route back to source
Every plotted item should have a recoverable source reference or derivation path. Review comments should attach to the profile’s data version, so later refreshes do not make an earlier comment ambiguous. Record whether the profile is exploratory or part of a controlled deliverable.
Astraea’s workspace includes patient-level review alongside study exploration. Our proposed acceptance exercise is to select one clinically important event and reconstruct its sequence using the profile and linked records, then compare that account with the source material.
A good profile helps clinicians ask better questions: what changed, what was observed, what is uncertain, and what evidence is still needed? It should support that judgment without turning a compelling visual sequence into an unsupported conclusion.
