How to consider multiple rows for a given Participant IDThis is a pretty elementary question, but how should one go about analysis of a large human clinical research data set with multiple rows for a given participant?
Lets say there are columns titled "ParticipantID", "Visit", and "HasBeard" each containing categorical data.
For each ParticipantID (i.e. 001; 002; 003) there are multiple study Visits (i.e. V1; V2; V3 etc) with each Visit containing unique data points on various measures such as "HasBeard".
Thus there are multiple rows for a single Participant ID (i.e. Participant 001, V1; Participant 001, V2; Participant 001, V3 etc.)
SEE ATTACHED.
If I were to do an analysis on a certain column measure such as "HasBeard" that contains categorical data (0 = has no beard, 1 = has beard), and want to know the distribution of Participants that have a beard at V3 but exclude those that already had a beard at V1, what is the most efficient and elegant method?
Given the provided image, I want to only return Participant 002, once.
Do I want to combine a Data Filter and a Local Data Filter? Do I want to exclude certain rows first using the Exclude function?
What is the best way to leverage JMP PRO for this sort of basic work?
AnalyticsRe: How to consider multiple rows for a given Participant IDHi @colez,
Looks like you are asking for a basic distributional type report. Are you looking to do number of patients with a beard or the number of times a patient has been seen to have a beard? In other words what kind of graph are you looking for? Visit in X axis with counts (unique or multiple) on Y axis? Is there only one event or multiple events? Maybe post a picture of what your are looking for.
I don't think JMP Pro is going to give you much of an advantage. If you have have structured data (and many), maybe JMP Clinical is the right tool for you since it is designed for clinical trials.
Clinical Data Analysis Software | JMP Clinical
Re: How to consider multiple rows for a given Participant IDHello @chris_kirchberg, thank you for the swift response!
I'm looking for the number of Participants who had a beard by their third visit (V3) but with the caveat that they did not already have one at their first visit (V1). (This premise humorously assumes that you can't lose a beard once you have one, so if they get a beard by V2, they will also have one at V3)
In other words, how many participants converted to having a beard over the course of the study? (The correct answer for my given example is 1.)
Notably in the image I attached, there are NINE rows total, but only THREE unique participants.
Out of those THREE participants, only Participant 002 actually meets the criteria I outlined.
Even though Participant 001 also had a beard at V3, (HasBeard = Yes), they do not meet the criteria because they already had a beard at V1.
So my query ought to return a value of 1, representing Participant 002 as a single participant despite their multiple visits/timepoints.
Re: How to consider multiple rows for a given Participant IDI would just use the Summary Platform and group the data based on Participant ID and HasBeard. Then all you have to do is to select all of the rows that are designated as HasBeard=No, and delete them. What is remaining is all of the unique ParticipantIDs that have Beards.
