Nanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Is Big Data Always Good Data?
1. Is Big DataAlwaysGood Data?
MEDTEC EUROPE 2018
STUTTGART
17 APRIL 12:00 – 12:30
DR. ANDREW RUT
CHIEF EXECUTIVE
OFFICER
2. “
”
Each year about a million biomedical
studies are published in the scientific
literature. And many of them are
simply wrong.
Rigor Mortis: how sloppy science creates worthless
cures, crushes hope and wastes billions*
*Richard Harris
** Six red flags for suspect work C. Glenn Begley
3. What is big data?
“Data of a very large size, typically to the extent that its
manipulation and management present significant and logistical
challenges.” – OxfordEnglishDictionary
“The ability of society to harness information in novel ways to
produce useful insights or goods and services of significant
value [and] things one can do at a large scale that cannot be
done at a smaller one, to extract new insights or create new
forms of value.” – ViktorMayerSchönberger and Kenneth Cukier;BigData
4. Medical big data
“The totality of data related to patient
healthcare and wellbeing.”
– Raghupathi 2014
5. ‘The amount of data you slough off everyday—in lab tests, medical
images, genetic profiles, liquid biopsies, electrocardiograms……is
overwhelming…...
Throw in … medical claims, clinical trials, prescriptions,
academic research, and more, and the yield is ~ 750 quadrillion
(1015) bytes every day—or some 30% of the world’s data
production. ’
Fortune Magazine. April 2018
6. Sources of medical data
Medical research
Clinical trials
Cohort studies
Clinical information systems
Electronic health records
Patient examinations
Medical imaging
Mobile devices
Web searches
Social media
Insurance claims data
Public records
Routine population statistics and
major disease surveillance
Data reported by patients
7. AORTA (Always On Real Time Access): Continuous
Monitoring of Health Status
• New technologies and real-time, remote monitoring of health
status and treatment compliance – wearables, sensors, social
media
• New patterns of consumer/patient interaction with the healthcare
system and healthcare professionals (“expanded touch points”)
• Progressive evolution of a seamless blend of online and
physical services for clinical care and individual health risk Mx
8. Mobile devices, Wearables, Sensors and
Continuous Monitoring plus Social Media ……Your
Behaviour in now measurable
• Our actions can now be tracked and
measured with (GPS) precision
• Big data sets identify predictable behaviours
and individual risk patterns
• Ethical and legal issues – Security, Privacy,
ownership……… GDPR
10. Big data characteristics: 3Vs
Volume
• Scale of data
Velocity
• Rate of increase
Variety
• Different forms of data
11. What makes big data good data
Validity
‒ Information must be factually sound and reliable
Value
‒ Data must be meaningful, i.e. will it help people make decisions about
healthcare in the real world?
12.
13. Plots show updated GFT model estimates compared with weighted CDC ILI-Net data for
(B) Mid-Atlantic HHS-2 Region States (New Jersey, New York), and local ILI surveillance
from emergency department ILI visit data for (C) New York City. (Olson et al; 2013)
Scatter plots of weekly excess influenza-like illness (ILI) visit proportions against
updated Google Flu Trends (GFT) model
14. Challenges of prospective use of internet
search algorithms:
Substantial flaws in GFT models identified
‒ missing and overestimating the intensity of the
epidemic
Reasons
‒ changes in internet search behavior
‒ differences in seasonality or geography
‒ Changes in age-distribution of the epidemics
16. An adverse event (AE) is any untoward medical occurrence in a patient or clinical
investigation subject administered a pharmaceutical product and which does not
necessarily have a causal relationship
An adverse drug reaction (ADR) is suspected to be causally related to the drug.
Safety of Medicines
19. “
”
Adverse drug events (ADEs) are an
increasingly relevant issue for healthcare
systems as they are associated with
poorer health outcomes and avoidable
misuse of resources
The Economic Burden of Inappropriate Drug Prescribing, Lack of Adherence and Compliance,Adverse Drug
Events in Older People;A Systematic Review; Carlos Chiatti,1 Silvia Bustacchini,1 Gianluca Furneri,1 Lorenzo
Mantovani,2 Marco Cristiani,3Clementina Misuraca4 and Fabrizia Lattanzio
20. H e a l t h c a r e b u r d e n o f A D R s
3.7% 39% 40%
ADRs account for
3.7% of hospital
admissions in the
Developed world.
~39% of ADRs in
pediatric patients
can be life
threatening or fatal.
Non-
adherence
Howard RL, Avery AJ, Slavenburg S, Royal S, Pipe G, Lucassen P, et al. Which drugs cause preventable
admissions to hospital? A systematic review. Br J Clin Pharmacol. 2007;63:136–47.
21. U SA AE R epor ts 2016; D ata G a p s
Rx
4.5B
AERs
1.7M
Serious AERs
>50% basic data missing
37% Age missing
0.57M
Institute for Safe Medicines; Quarter Watch Annual Report: July 2017
22. High Value Safety Data informs understanding of
products & their effect on patients
Value of Safety reports
directly proportional to the
amount
of clinically relevant
information.
Overall, only 1 in 8 of
reports provide the
desired level of
information
Physicians & consumers
using e-reporting tools
generate greater proportions
of well-documented reports
23. Dismissed importance
“Doctor said would have to live with side effects and did not seem to
care. Ignored complaints about symptoms”
“Acted as if it was in imagination. Don’t think
doctor believed me.”
Dismissed existence
Source: Beatrice A. Golomb et al, Physician Response to Patient Reports of Adverse Drug
Effects Implications For Patient-Targeted Adverse Effect Surveillance, Drug Safety 2007
24. Getting Purer data, faster from the
source
Get the right data first time on medicine
safety
26. Configurable, intuitive design – Simple safety data collection
in real time
• Web, mobile APP (on/offline)
• Multi-language
• High availability on AWS GxP cloud
• Active workflow triggers questions based upon Event & Product
• Attachments and images supported (pathology data etc)
• Fully coded data (XML )
• Feeds AI and ML engines (learning and further processing)
27. User Interface flexibility is key
• Simple screens for rapid collection
• Body map selection mapped to a medical coding
library (MedDRA)
• Local languages available with coding libraries
32. HISTORICA
L
WEB
CALL
CENTR
E
DIRECT WITH
PSP TEAM
EMAIL LETTE
R
LITERATUR
E
FREE TEXT
WORKFLOW
FOLLOW UP
PASSIVE
HUMAN REPORTING
REPORTER
TEXT CONTAINING ‘REPORT’
PV FIRST
PV FIRST
PQC ICSR
EXCEPTION
REPORTING
AND QC
SAFETY EXPERT
REVIEW
CASE REVIEW
AND CODING
FULLY CODED
REPORTS
REAL TIME
AI ENGINE
AI ENGINE
33. Big & Good Data: Focus on the Individual
Real time
Monitoring &
Data integration
Data
Analytics
Individual risk
assessment &
Management
Genomics,
Diagnostics, Lifestyle
plus efficacy-safety
profile of medicines
AI/ML
Health
management plans
and optimized RX
A Drive to Personal Prediction, Prevention &
Management
Understanding the safety of medicines
What is the strength of evidence
GiGo
Alternative BMJ definition: “High-volume, high-velocity and/or high variety information assets that demand cost-effective innovative forms of information processing that enable enhanced insight, decision making and process automation.” - BMJ 2018; 360: j5910
Strava and US troops
Social media predictive of disease patterns???
GFT model problematic:
changes in internet search behavior
differences in seasonality or geography
Changes in age-distribution of the epidemics
What is driving the increase in volumes?
What is the contribution of intrinsic characteristics of patients?
What is the contribution of manufacturing issues; diverted products in the supply chain
How do MAHs currently determine whether the problem relates to the authentic medicine itself vs some other factor?
How do Tracelink-Reportum connect this process most rapidly to identify Batch issues or diverted product? Patient; Authentic manufacturing issue; Diverted products…….
Identify core issue
Reconciliation of the core issue rapidly………how many days is saved in the process in Pharma
Our belief is that the initial point of capture is key to quality processes downstream
Ask relevant questions of the reporter based upon intelligent triggers to maximise data quality to fuel the safety process
Do this by providing simple tools appropriate for the audience
Seamless integration of complex, diverse and dynamic data for real-time monitoring of health status and risk management ! shift from reactive episodic care encounters to increasingly proactive risk mitigation ! progressive shift from management of overt disease to sustained wellness and continuity in care