How will ClinSpark and Nabios MUSE interact?
An AWS (Amazon) S3 bucket will be created and used to enable the communication between Nabios and ClinSpark.
Nabios will drop XML and PDF files for each performed ECG. ClinSpark will pick up the dropped files, parse them and record the results in a master JSON file.
ClinSpark Dashboard component will read the results from the JSON file and load them in ClinSpark for users to be able to match the results with the correct Volunteer/Day/Timepoint und upload the results into the forms.
What is in the S3 bucket?
There will be several folders in the S3:
Study-specific folder - this folder will be named using the Protocol Name from ClinSpark Study > Configure > Basics section. This folder will be used by Nabios to drop the XML and PDF files that need processing. Within the Study-specific folder, there will be 2 other folders.
ClinSpark folder will hold a master JSON Cache file - this file will hold the checksums of all parsed ECG and communicate with the Dashboard component in ClinSpark. This means, the Dashboard component does not need to go and read the actual XML and PDF files and parse them each time, which should help with the performance.
“processed” folder will hold the files that have been parsed previously and no longer need “attention” from ClinSpark.
What does ClinSpark check in the Nabios XML file?
Subject screening number defined in: RestingECG/PatientDemographics/PatientID
Very critical is the Timepoint identifier in: RestingECG/TestDemographics/TestReason; defines in minutes the timepoint
Consists of 3 tokens separated by underscores; token 2 is the day (can be negative) and token 3 is the relative day, hours, minutes (can also be negative)
Example MOX-P2_D1_00D03H00M - Token 1: MOX-P2 refers to the study, Token 2: D1 means that it is an ECG for Day 1 of that study, Token 3: 00D03H00M means that the timepoint is 3 hours postdose.
ClinSpark will parse Tokens 2 and 3 and “marry” that information with the information in ClinSpark.
ECG interval values are parsed from RestingECG/RestingECGMeasurements
ECG diagnostic statement is parsed by inspecting each RestingECG/Diagnosis/DiagnosisStatement/StmtText and joining the values with a pipe
Acquisition time is parsed out of RestingECG/TestDemographics/AcquisitionDate and AcquisitionTime
It is assumed that the timezone in the XML file and the timezone of the Site in ClinSpark are the same.
What needs to be configured in ClinSpark?
Device Profile
Device profile must be added to the environment, and configured by the ClinSpark team.
New Device profile available: Nabios MUSE ECG
Variety of Parameters available, not all will be used
Needs the Settings Key and Settings value input by superadmin to indicate which Amazon S3 should be used. This will require the Access Key / Secret Key for the AWS S3 connectivity.
Object ID Prefix should not be used in Prod! This is only useful, while working on different testing environments in order to properly name the folders on the Amazon S3 and avoid confusion which environment should read from which folder
User Role
We strongly advise that only trained designated users should interact with the Dashboard component and the MUSE ECG Forms. The customer should create a dedicated User Role to control this access.
Dashboard Component and async job
Dashboard component and async job must be added to the environment, and configured by the ClinSpark team. The Dashboard component can be added to the “Home page” Dashboard or to the “Study > Subjects” screen.
Study Setup
Customers will use the new Device Integrated Form for Nabios MUSE ECG in the CRF Design > Study Library > Forms and select the relevant Parameters to be used in the Study.
This form will be included in the Activity Plans, at specific Timepoints, as required.
Nabios will also use the value in the “Protocol” field from Study > Configure to name the folder in the S3 bucket, so that ClinSpark “knows” which study the results belong to.
Collecting the ECG
Data collector connects the holter to the subject.
Data collector indicates when a 12-lead ECG needs to be taken. The 10s + 3 min before that particular indication by the data collector are selected.
Selected data is sent over to the MUSE server for conversion.
MUSE server posts the ECGs in XML and PDF pairs to Amazon S3 bucket and will do so with the protocol name prefix (folder). The .pdf and .xml files will have the same name minus the file suffix.
As ECGs are processed, the XML and PDF pairs are moved to a ‘processed’ directory in the same protocol directory and parsed to the master JSON file.
The ECG results are ready to be imported into ClinSpark using the Dashboard component.
Using the Dashboard Component to import the ECG results
The Dashboard component will be added by the ClinSpark team.
The dashboard component is visible when:
Nabios MUSE Integrated Form is used on a Study.
The Study is in Active State.
The user has appropriate Role to access the component.
To use the Dashboard, the user will need to
apply at least one filter (Subject or Study Event),
match at least one ClinSpark Form with Nabios ECG results (this can be done automatically or manually)
select the “matched” ECG in the list by ticking the appropriate tickbox
click the green “Save” button to import the matched values into the Form.
When filters are applied, the system loads the unsaved forms in a table below the filters.
Subject/Timepoint/Status columns refer to the Activity Plan in ClinSpark.
ECG column is used to select and display the results from Nabios.
The component will check the JSON file in the S3 bucket for “unmatched” ECG results from Nabios.
Automatic Match
If ClinSpark can find the exact match based on the Screening Number, Study Day and Timepoint for the Subject, it will automatically match the results and display them against the correct form in the list.
An exact match is only possible when the Study Day (Event) and Timepoint match perfectly between ClinSpark planned Timepoint and Nabios planned Timepoint, ie the delta between those is 0. The actual ECG acquisition time is not taken into consideration.
The user can override the automatic selection by using the dropdown to select a different ECG.
When there is an exact match, the system also automatically preselects the ECGs to import the values on the next “Save”.
Manual Matching
If there is no exact match, the user can use the dropdown in the ECG column to find the appropriate ECG. To facilitate the selection, the deviation between the “planned” and “actual” time will be displayed. If the Deviation is big, the user should double-check their selection.
They will need to manually select the ECG for import, by ticking the appropriate tickbox.
Once all the desired ECG have been matched and selected in the list, the user will need to click the green “Save” button to “import” the values into the form(s).
Importing Results
When the user clicks “Save”, ClinSpark will do a few validation checks:
is at least one tickbox ticked in the list?
is the selected ECG Form “married up” with results from Nabios?
did the user select the same ECG result for 2 different forms in the list?
A confirmation message will pop up to confirm when the import is done.
The values and the PDF file will be imported to the form and the form will show as “Completed” on the Study > Data page.
For “mock” Testing
Device profile, dashboard component and async job need to be configured in the ClinSpark instance
A simple mock study must be built using the relevant device profile (client)
XML / PDF ECG pairs must be created with the sample protocol (ClinSpark or Nabios team)
XML / PDF ECGs must be uploaded for testing to the S3 destination
Any of the 3 respective teams can do this with an S3 utility; some customers use FileZilla Pro: https://filezillapro.com
Nabios MUSE in Headless Mode
The Nabios MUSE interface can also run in a fully headless mode where ClinSpark will automatically assign all ‘deviation of 0’ ECGs in the same way that the algorithm in the dashboard component executes with the user actions.
This automated option eliminates the need for user interactions. Assuming there is proper alignment on the timepoint calculations between the two systems, the interface should work perfectly.