Every step, and who it lands on.
This is the run between a patient mentioning a symptom and a monitor closing a query on it. Eight steps, four different people, and four durations nobody has measured. If yours looks different, we’d like to see it.
The run as it happens today.
Where a step carries a number, that number has a source and we’ll give it to you. Where it says unmeasured, nobody has measured it and we’re not going to fill the gap with an estimate that looks better.
~30 min
The research nurse interviews the patient
In a room, on the day of the visit, working through what has happened since the last one. The room itself is a constrained resource at most centres.
unmeasured
A coordinator reads the note and finds the events
Everything reported, plus everything implied by a lab value or a medication change. Nobody has measured this step for us and we are not going to guess at it.
15 to 20 min each
Each event is graded against CTCAE
For the routine case, when everything goes well. A grade 3 or 4 event costs hours of documentation instead.
- Two CTCAE versions run at once where a site has studies open on both
- 837 criteria changed between v5 and v6
unmeasured
Each event is attributed to a drug
Scored against every agent in the regimen. Published inter-rater agreement on attribution sits at kappa 0.59 to 0.68, and attribution changes in 31 to 36% of cases between investigator and central review.
~30 min per patient
The investigator reviews and signs
Sorting through everything that has accumulated since the last visit before signing. This is the step the design partner measured for us.
about 30 days
The record reaches the study database
The average time to enter safety data on an oncology trial, as described by a pharmaceutical safety physician on our advisory board. Same-day entry is rare.
unmeasured
A monitor visits and raises queries
Months after the encounter, against a record written by someone who has since moved on. Coordinator turnover has run 35 to 61% a year since 2020.
unmeasured
Someone reconstructs the reasoning and answers
Going back to the source note to work out why the grade was what it was. The answer already existed; it was never written down beside the grade.
The same run, with the drafting done.
Burna reads the note, finds the events, grades each one against the version the protocol names, scores attribution per drug, and attaches the citation that produced each answer. Your coordinator reviews a page instead of assembling one, and your investigator signs.
The interview still happens. The review still happens. The signature still happens, and it’s still the investigator’s. What changes is that the query at the end arrives against a record that already carries its own reasoning.
“If I'm going to go to our CEO, I would have to justify and say: look, I see this allowing us to open up this many more trials, or have this many more patients on study, because my sub-I is now only taking two minutes to review this AE rather than a half hour trying to sort through all of the AEs that occurred since the last visit.”
What we would want to measure at your site.
Four of the eight steps above have no number, and the two that do come from one site and one advisory board. Before claiming anything about your workflow we’d rather measure it: time in the chart per encounter, taken from your medical record’s own telemetry rather than from a log anyone has to keep.