Machine learning and data analytics can help improve pharmacovigilance in several ways: 1) Machine learning algorithms can automatically extract adverse drug reactions from biomedical literature and FDA drug labels, helping pharmacovigilance teams more efficiently identify all potential ADRs. 2) Large healthcare datasets and sophisticated algorithms can help pharmaceutical companies with drug discovery, clinical trials, personalized treatment, and epidemic outbreak prediction. 3) Advances in machine learning are reshaping healthcare and have the potential to cut clinical trial costs, improve quality, speed up trials, and facilitate tasks like reviewing literature, recruiting patients, and making diagnoses.

1. Machine Learning In Health
Data Analytics And
Pharmacovigilance
1

2. Introduction
• Intelligence of machines and the branch of computer
science which aims to create it.
• “Machines will be capable, within 20 years, of doing any work a
man can do.” –Herbert Simon, 1965(AI innovator)
2
Three Steps
Three elements of machine
learning
Computers and
programs
Massive amount ofdata
Sophisticated algorithms The Turing test
High performanceparallel
processors
The Darmont
Conference

3. Pharmacovigilance
• The word “Pharmacovigilance” was derived from the Greek literature
“pharmakon” (means drug) and the word “vigilare” (means keep watch) in Latin.
In 1961, the World Health Organization (WHO) has established the
pharmacovigilance (PV) program in response to the thalidomide disaster, for
global drug monitoring.
• Many rare adverse effects remain undetected due to a limited number of
sampled individuals in a clinical trial; hence, it is necessary to monitor the drugs
even after their release into the market.
• In this context, “pharmacovigilance” helps to collect, analyze, and disseminate
adverse drug reaction reports collected during the post-marketing phase.
3

4. Machine learning and pharmacovigilance
• Monitoring the scientific literature for adverse drug reactions
(ADRs) is critical to maintaining drug safety, and there is no
room for error.
• As regulations tighten, pharmacovigilance teams are seeking
better strategies and methods for ensuring that all ADRs are
identified in the most effective and efficient way possible.
• Machine learning has been doing great work on the automated
extraction of ADRs from biomedical literature and FDA drug
labels.
• As a part of outreach to the global pharmacovigilance
community
4

5. Health data analytics and
pharmacovigilance
5

6. Challenges for machine learning
Reasoning, Problem Solving
Knowledge representation
Planning
Learning
Natural language processing
Perception
Motion manipulation
Social Intelligence
Creativity
General Intelligence
6
Approaches
Cybernetics
Symbolic
Statistical
Integrating the approaches
Applications
Healthcare and Medicines
Automotive
Finance and economic
Video Games
Heavy Industries
Robotics

7. Machine learning in Healthcare
Managing Medical Records and other data
Doing repetitive jobs
Treatment Design
Digital Consultation
Virtual Nurses
Medication Management
Drug Discovery
Precision Medicine
Healthcare Monitoring
Healthcare SystemAnalysis
7

8. CT Participant Identifier
Connected Machines
Dosage error Detection
Fraud detection
Adm. workflowAssistance
Virtual Nurshing…
Robot-assisted Surgery
Cybersecurity Advance Image Diagnosis and Preliminary Diagnosis by
using data analytics and machine learning.
8
estimated potential
annual benefit for each
application by 2026(in
billon USD)
0 10 20 30 40 50
Estimated potential annual benefit for each application
by 2026(in billon USD)
Source: AccentureAnalysis
Total= $150Billions

9. Many big Pharmaceutical companies began investing
in machine learning in order to develop better
diagnostics or biomarkers, to identify drug targets
and to design new drugs and products.
Merck partnership with Numerate in March 2012
focusing on generating novel small molecule drug
leads for unnamed cardiovascular disease target.
In december, 2016 Pfizer and IBM announced
partnership to accelerate drug discovery in immuno-
oncology.
Current Scenario
9

10. Disease Identification
10
2015- Report by Pharmaceutical Research and
Manufacturers of America- more than 800 drugs and
vaccines are in trial phase to treat cancer.
Google’s DeepMind Health, announced multiple
partnerships including some eye hospitals in which
they are developing technology to address macular
degeneration in agingeyes.
Oxford’s Pivital® Predicting Response to Depression
Treatment (PReDicT) project is aiming to produce
commercially-available emotional test battery for use
in clinical setting.

11. Personalized Treatment
11
Micro biosensors and devices, mobile apps with more
sophisticated health-measurement and remote
monitoring capabilities; these data can further be used
for R&D.
DermCheck; app available in Google play store in
which images are sent to dermatologists(human not
machines)

12. Drug Discovery/Manufacturing
12
From initial screening of drug compounds to predicted
success rate based on biological factors.
R&D discovery technology; next-generation
sequencing.
Previous experiments are used to train the model
Optimization softwares (example:FormRules)
Designing of the processes

13. Clinical Trial Research
13
Machine learning- to shape, direct clinical trials
Advanced predictive analysis in identifying candidates
for clinical trials
Remote monitoring and real time data access for
increased safety; biological and other signals for any
sign of harm or death to participants.
Finding best sample sizes for increased efficiency;
addressing and adapting to differences in sites for
patient recruitments; using electronic medical records
to reduce data errors.

14. Epidemic Outbreak Prediction
14
To predict malaria outbreaks, from data like
temperature, average monthly rainfall, total number of
positive cases,etc.
ProMED-mail is a internet based reporting program
for monitoring emerging diseases and providing
outbreak reports.

15. Radiology and Radiotherapy
15
Google’s DeepMind Health is working with University
College London Hospital (UCLH) to develop machine
learning algorithms capable of detecting differences in
healthy and cancerous tissues.

16. Smart Electronic Health Records
16
Artificial intelligence to help diagnosis, clinical
decisions, and personalized treatment
suggestions.
Handwriting recognition and transforming cursive or
other sketched handwriting into digitized characters.

17. Regulating Use in Digital Health
Products
17
Incomplete insight from US FDA for products utilizing AI.
Medical devices provisions of Federal Food, Drug and
Cosmetic Act-1970s
FDA created Digital Health Program tasked with
developing and implementing a new regulatory model for
digital health technology.
Over the last five years different guidelines like Mobile
Medical Applications Guidelines.

18. In Clinical Research
18
Cutting costs
Improving trial quality
Improving trial time by almost half
Finding biomarkers and gene signatures that cause
diseases
Recruiting trial patients in minutes
Reading volumes of text and data in seconds
On verse of discovering involving new diagnostic tools
and treatments for Alzimer’s disease, cancer, and other
chronic and terminalillness.

19. Conclusion
• Machine learning, data analytics are reshaping human relations,
promoting the global economy and triggering societal and political
reforms on a world scale.
• Those machine learning methods may perform differently, if non-
medical or non-healthcare datasets are processed or tested. However,
learning the machine learning methodologies is more important in
general big data science and cloud computing applications in
healthcare and pharmacovigilance data management.
• These technologies should be included in the academic curriculum for
healthcare course, which are going to play crucial role in healthcare
system in coming future.
• Students should be encouraged about the advances in various
technologies at graduation level.
19

20. References
20
BAksu, AParadkar; Quality by design approach: Application of
Artificial Intellegence Techniques of Tablets Manufactured by Direct
Compression; PharmsciTech; 2012;13(4); 1138-1146
JADimasi, RW Hansen; The price of innovation: new estimates of drug
development costs. JHealth Econ;2003;22(2);151-185
S Behjati and PS Tarpey; What is next generation sequencing?; Arch
Dis Child Pract Ed;2013; 98(6);236-238
https://doi.org/10.1080/23808993.2017.1380516
http://artint.info
http://www.fda.gov
http://www.clinicalinformaticsnews.com