Medical information call centers have an opportunity to transform the way they capture, code, and analyze adverse events (AEs) and product quality complaints (PQCs) with artificial intelligence (AI) and automation.
The use of such innovative technology improves data quality and consistency, compliance, and operational efficiency. It helps reduce the frequency of your pharmacovigilance (PV) operations resources going home, saying, “I have more to do at the end of the day than I did when I started."
Our one-hour, on-demand webinar shows you how you can use AI and automation to turbo-charge your end-to-end PV system. Use cases and demonstrations will include:
Analyzing safety data
Auto-coding verbatim terms to official medical dictionary terms
Auto-creating an AE case in your database
Converting speech to text
2. 2
AT-A-GLANCE
Perficient Profile
Alliance partnerships
with major tech
vendors
Multiple vendor &
industry tech and
growth awards
Global delivery centers
in China, India, Eastern
Europe and Latin
America
Dedicated
solution
practices
~35
Global Locations
$565M
2019 Revenue
1997
Founded in
~4,500
Employees
~90%
Repeat Business Rate
PRFT
Public, NASDAQ
3. 3
Pharmacovigilance Consulting Capabilities
• Medical Information (MI)
or Intake
• Regulatory Intelligence
• PV Workflow Optimization
• Regulatory Reporting
• PV Technology
• Metrics and Oversight
• PV Compliance
Documentation (SOPs,
Best Practices, etc.)
• PV PM
• Vendor Oversight
• Governance
• RMP
• Safety Surveillance
• Signal Detection
• RMP for New MDR
• Aggregate Report Data
and Compilation
• PV Training (Regulatory,
Database)
4. 4
Christine Livingston
Chief Strategist, AI
christine.livingston@perficient.com
Kari Blaho-Owens, Ph.D.
Director, Safety and Pharmacovigilance
kari.owens@perficient.com
Today’s Speakers
Prabha Ranganathan
Director, Clinical Data Warehousing and Analytics
prabha.ranganathan@perficient.com
5. 55
Agenda
• PV Landscape Challenges
• Integrating AI and Automation Into the PV Workflow
• Transforming the End-to-End Data Flow With AI and Automation
• Demo and Use Cases
• Q&A
6. 6
Challenges in Pharmacovigilance
• Large Case Volumes
• Higher Volume and Longer MI
Calls
• More Sources of
Information/Misinformation
• Regulatory Pressure/Dynamics
• Aggregate Reports
• Efficiency Gains With Processes
vs. Technology or Both
• Global Challenges: Data On-
Demand, RWD/RWE, Technology
and Language
• End-to-End Case Processing
Costs/Resources
• Pressure to Move From Reactive
to Proactive PV
• Payer/ Provider Considerations
• Signal Detection and Risk-Benefit
Analysis
7. 7
Adverse Events Case Processing Workflow
Call Center or
Other Intake
Method
Case Data Entry Case Processing Medical Review
Aggregate
Reporting
Submission to
RA/SDEA
QC of Case
10. 10
NLP Applied
Text Classification
Identify logical groupings
Entity Extraction
Transform unstructured data to structured
Question-Answer System
Interpret intent & surface relevant answers
Language Translation
Preserve syntactical information
Sentiment Analysis
Interpret tone and emotion
NLG
Synopsize structured data
11. 11
Breaking Down Complex Topics
Evening of vaccination developed left and neck pain that progressed to multiple muscle, joint, and skin pain. Alternating chills/sweats and
severe fatigue. Very similar to influenza without the respiratory symptoms. This lasted approximately 2 days, then developed nausea and
diarrhea with intermittent fatigue and chills/sweats. Day 5 I feel better, but still have sweats, diarrhea and intermittent nausea.
Primary Reaction Source=vaccination
Primary Source MedDRA Code=10046859
Symptom Covered Text=sweats
Normalized MedDRA Term=Hyperhidrosis
Modality=Positive
Symptom MedDRA Code=10042661
Reaction Start Date=20200910
Symptom Covered Text=respiratory symptoms
Normalized MedDRA Term=Respiratory symptom
Modality=Negative
Symptom MedDRA Code=10075535
Reaction Start Date=
13. 13
Automation and AI Outcomes in PV
Harmonized data intake
processes globally
Increase case volume will not
increase headcount
Reduced training for new
resources allocated for intake;
simplified process
Decreased compliance risks of
missing AEs
Reduced number of re-work from
investigator sites
Decreased “touch” time per case
for intake activities
Resource and time savings
Overcoming language barriers
14. 14
ROI of Automation in Intake
Estimate Cost and Time Savings for Intake Form Only
Task Serious (S)
Case
Non-Serious
Case (NS)
Hours Without
Automation
Hours Saved With AI
and Automation
AE Intake from MI Call
New Process/Automation:
15 minutes @ 10K
9 minutes @ 10K
10 minutes @ 10K
7 minutes @ 10K
2,500 for S
1,667 for NS
1,000 for S
500 for NS
Complete Data Entry
With Automation
30 minutes @10K
15 minutes @ 10K
20 minutes @ 10K
10 minutes @ 10K
5,000 for S
3,333 for NS
2,500 for S
1,667 for NS
MedDRA Coding
NLP MedDRA Coding
8 minutes @ 10K
2 minutes @10K
3.5 minutes @ 10K
2 minutes @ 10K
1,333 S
583
1,000 for S
250 for NS
13,083 6,917
Savings in Dollars
5K cases: 3,458 hours at $100/hour = $345,800/year
10K cases: 6,917 hours at $100/hour = $691,700/year
50K cases: 34,585 hours at $100/hour = $3.5M/year