1. Visualizing Subject Level Data in Clinical Trial Data
PHUSE Data Visualization Working Group,
Subject Level Data Visualization Sub Group
October 16, 2018

2. Agenda Ranking Graphical Patient Profiles Tabular Patient Profiles
Presentation of Considered Elements
Discussion/Ranking
Tabular Patient Profiles
Patient Narratives

3. Ranking Features
Critical, e.g. would not consider using a tool without this feature
Very useful/important
Useful
Nice to have
Not Needed

4. Graphical Patient Profiles
Time oriented, visual presentations of subject level study data requiring little or no external data preparation or programming. These profiles may focus on a specific area such as liver function or depict a comprehensive view of the subject’s time in the trial and include exposure, adverse event, lab, disposition or other data.
Feature Areas
Display/Use
Data
Export/Printing
Licensing
Other

5. Graphical Patient Profiles Some examples

6. Graphical Patient Profiles, Display/Use
Scrollable display - horizontally – through available days axis range, and vertically – through different types of data displays (labs, AEs, conmeds, dosage, etc.) Ranking:
Ability to click a point or duration bar of interest to drill into the details of other events, interventions and/or findings that occurred during that same time point or time duration for that patient. Includes both the surrounding events, interventions and findings that occurred during the time period as well as the details of each event, intervention or finding that occurred (e.g. dose of the conmed or severity of the event). Ranking:

7. Graphical Patient Profiles, Display/Use
Ability to include any data item from study data to be presented graphically on time scale that is in a Standard CDISC Domain. Ranking:
Ability to include any data item from study data to be presented graphically on time scale that is in a custom CDISC Domain or not in a standard format, even if some configuration is required. Ranking:
The ability to display reference ranges for
Vital Signs
Labs
ECG Data
Other ___________
Ranking:

8. Graphical Patient Profiles, Display/Use
The ability to display duration bars for
AEs
ConMeds
Dosing
Other _______________ Ranking:
The ability to calculate time (days since any chosen reference point (e.g. reference start date, reference exposure start date, start of a specific adverse event) on the fly – without external data setup or processing. Ranking:
The ability to easily drill down from an aggregated view to a patient profile. Ranking:

9. Graphical Patient Profiles, Display/Use
Settings that allow for bookmarking and commenting on interesting finds when reviewing patient profiles, preferably the reviewer should be able to save what they see and share it with others (this includes any marking and filter settings) Ranking:
Ability to call other software (such as R and SAS) to run on underlying data and display the results within the current tool. Ranking:
Ability to define/save multiple graph patient profiles – with different content – for different targeted use. Ranking:

10. Graphical Patient Profiles, Data
The ability to work with CDISC SDTM. Ranking:
The ability to work with CDISC ADaM. Ranking:
The ability to work with both CDISC SDTM and ADaM and merge the 2 datasets even if some configuration is required for ADaM (not SDTM), including the ability to use population and other subgroup flags defined in ADaM to filter subjects in any dataset. Ranking:
The ability to work with any clinical data source (e.g. a legacy data format) even if some configuration required. Ranking:
The ability handle large data sets (data sets 2 gigabytes or larger, even up to 30 gigabytes). Ranking:

11. Graphical Patient Profiles, Export/Printing
The ability to save and print patient profiles both individually and in groups (batches). Ranking:
Ability to export graphic display for inclusion in
Word
PowerPoint
PDF
Other__________ Ranking:

12. Graphical Patient Profiles, Licensing
Determine metrics to assess the cost of the development license and the viewing license.  For example, Spitfire and Tableau have a license that allows a person to develop visuals and then a reviewer would have a license that allows them to view the visuals online without the need for the full software.  We’d also want to know how much a sever costs and how many users and how big of files it can handle.  Ranking:

13. Graphical Patient Profiles, Discussion
What type(s) of interactive display features for graphical patient profile visualization not mentioned would you like to have?

14. Tabular Patient Profiles
Also called case summaries, display information for a single patient in tabular format, allowing the user to view summarized data about a single patient’s clinical experience. For example, counts and SOCs of adverse events during exposure and/or other information relevant to the trial can be displayed.

15. Tabular Patient Profiles Some Examples

16. Ranking Features (Reminder)
Critical, e.g. would not consider using a tool without this feature
Very useful/important
Useful
Nice to have
Not Needed

17. Tabular Patient Profiles, Display/Use
Display of nicely formatted tabular patient profile for a selected patient. Ranking:
Sufficient content, display, formatting for use both internally as well as submitting to regulatory agencies as case summary. Ranking:
Scrollable display – vertically and horizontally. Ranking:
Export tabular profile as a PDF or print. Ranking:
Batch processing (generation of tabular patient profiles – resulting in one PDF per patient) to reduce time of generation. Ranking:
Ability to specify criteria (e.g. based on country or treatment group) for patients to be included in generation of pdfs showing the tabular patient profiles Ranking:

18. Tabular Patient Profiles, Display/Use
The ability to easily drill down from an aggregated view to a patient profile. Ranking:
Settings that allow for bookmarking and commenting on interesting finds when reviewing patient profiles. Ranking:
Highlight new or updated data since last patient review. Ranking:

19. Tabular Patient Profiles, Data
 The ability to work with CDISC SDTM. Ranking:
The ability to work with CDISC ADaM. Ranking:
The ability to work with both CDISC SDTM and ADaM. Ranking:
The ability to work with any clinical data source (e.g. a legacy data format). Ranking:
No external set up time, e.g. no need for external data transformation or data programming/preparation – all handled by the application. Ranking:
The ability handle large studies with large amount of data per patient (should we define a size here- is this related to number of patients or amount of data for each patient or both?) Ranking:

20. Tabular Patient Profiles, Definition within Product
Ability to include any data item from study to be presented in tabular format – with multiple tables/domains presented for one patient at a time. Ranking:
Ability to filter at the row level – to decide which rows should be displayed for specific tables/domains, for example, to display only ConMeds from a drug class of interest. Ranking:
Ability to sort data in any of the included domains on selected items as appropriate for the data of that domain, for example, ordering AEs based on onset date rather than alphabetic sort on AE verbatim text or visit number collected. Ranking:
Ability to control formatting, font, color highlighting. Ranking:
Ability to define/save multiple tabular patient profiles – with different content – for different targeted use. Ranking:
Ability to calculate time (days since reference date) on the fly – without external data setup or processing. Ranking:

21. Tabular Patient Profiles, Export/Printing
The ability to save and print patient profiles

22. Discussion of Tabular Patient Profiles Features
What type(s) of interactive display features for graphical patient profile visualization not mentioned would you like to have?

23. Patient Narratives
Summaries that describe a subject’s clinical experience resulting from participation in a clinical trial. Narratives contain elements of a patient profile but are primarily generated to communicate reasons for the event(s) the patient experienced during their participation in the trial.

24. Patient Narratives Some Examples

25. Patient Narratives Execution/Display/Use
Interactive display of generated ‘automated text’ – with the ability to insert ‘free text’ additional content where desired – after review of the ‘automated text’ for any of the patients. Ranking:
Generate the completed narrative including any embedded reports or graphs (only including data for the current patient). Ranking:
Save the generated patient narrative as a PDF or RTF or Other(___). Ranking:
Batch schedule/batch processing of patient narratives – resulting in separate RTF files per patient. Ranking:

26. Patient Narratives, Definition within product/environment
Ability to define a patient narrative template – defining where/when to insert data items from the study (similar to MailMerge in Word). Ranking:
Ability to define ‘conditional’ inclusion of template text based on a condition definition, i.e., if patient reported a serious AE, then include a paragraph describing the serious AE. Ranking:
Ability to include ‘aggregate function versions (min, max, mean, etc.) of data item – with definition of filtering of which rows should be included, for example, max systolic BP post baseline. Ranking:
Ability to insert reports or graphs to further illustrate information within the narrative. Ranking:

27. Patient Narratives, Definition within product/environment
Ability to define/save multiple patient narrative definitions – with different content – for different targeted use. Ranking:
Ability to define the criteria for selection of patients who need to have narratives generated. Ranking:
Works with any clinical data source – CDISC, or legacy data format. Ranking:
No need for external data transformation or data programming/preparation – all handled by the application. Ranking:

28. Patient Narratives, Data
The ability to work with CDISC SDTM
The ability to work with CDISC ADaM
The ability to work with both CDISC SDTM and ADaM
The ability to work with any clinical data source (e.g. a legacy data format)
The ability handle big data (should we define a size here- is this related to number of patients or amount of data for each patient or both?)

29. Patient Narratives, Export/Printing
The ability to save and print patient narratives. Ranking:

30. Discussion of Patient Narratives
What type(s) of features for patient narrative generation not mentioned would you like to have?

31. Thank you!