Date 01.12.2008
Drug regulatory process, the
supporting information
systems, and the Escher-
project
Tommi Tervonen
Faculty of Economics and Business
University of Groningen
t.p.tervonen@rug.nl
Date 01.12.2008
What is drug discovery and drug
development?
Ambit biosciences, commercial Example drug development costs of
compound DBs (ambitbio.com) Trimeris Inc. (thebody.com)
Date 01.12.2008
>What are Drug Information Systems
(DISs)?
>What are they being used for?
>Who uses them?
>What could they be used for?
Development
Launch
Discovery Pre- Clinical trials Marketing
clinical
trials Phase I Phase II Phase III Phase IV clinical trials
Drug lifecycle
Date 01.12.2008
> DIS classes
Compound DBs
Pre/clinical trial DBs
Summary of Product Characteristics (SmPC) DBs
Adverse Drug Reaction (ADR) DBs
CPOEs
Development
Launch
Discovery Pre- Clinical trials Marketing
clinical
trials Phase I Phase II Phase III Phase IV clinical trials
Compound
DBs
Pre/clinical trial DBs
SmPC DBs
Time
ADR DBs
CPOEs
Date 01.12.2008
DIS Data Organization
Quanti
Type tative
(Structural/Trial/Reg (Yes/N Aggregat Non-profit
Class Name ulative/Patient) Details o) ed/Raw Name (Yes/No)
Cambridge structural
Compound database S Molecule crystal structures Y R CCDC N
Compound NCI 3D database S Molecular 3d structures Y R NCI Y
CT clinicaltrials.gov T Only CT settings and objectives N A NIH / FDA Y
CT Janus T, R Not functional yet Y A, R FDA Y
CT EudraCT T No public functionality* N A EMEA Y
CT Cochrane T CT Meta-analyses Y A Cochrane Y
CT clinicalstudyresults.org T Detailed clinical study results N A PhRMA N
NCI Drug SmPC and clinical trials for cancer
CT / SmPC Dictionary/thesaurus T, R drugs N A NCI Y
Micromedex,
Cerner Multum,
SmPC, drug interactions, condition Wolter Kluwer,
SmPC Drug Information Online R medication, pill labeling N A and others N
SmPC, drug interactions, condition
SmPC RxList R medication, pill labeling N A RxList, Inc. N
Lung Association of
Saskatchewan lung SmPCs for drugs with indication in Lung Association
SmPC disease drug repository R lung diseases N A of Saskatchewan Y
SmPC, drug interactions, pill
SmPC Drug Digest R labeling N N
SmPC DailyMed R Detailed SmPC N A NLM Y
SmPC EMEA EPAR R EPAR, incl. detailed SmPC Y A EMEA Y
ADR MedEffect R Superficial ADR N A Health Canada Y
ADR FDA ADR R ADR quarterly reports N A FDA Y
ADR EMEA ADR R No public functionality* N A EMEA Y
CPOE Various R, P SmPC through internal SmPC DB N A Various N
Date 01.12.2008
> The drug regulatory process
Aims to make sure, that the drugs entering market are both safe and
efficient
Is laborious and slow
Has relatively poor dissemination of results
Doesn’t have transparent decision making
Has recently all participating parties (drug industry, academia, and
regulatory authorities) concerned about reforming the process
> The main reason for
reforming the regulatory
process is to limit
the linear growth
of costs,
but…
Regulatory Benefit-risk
Logic assessment
Data and
evidence
Date 01.12.2008
>The current DISs don’t store regulatory
information of sufficient precision; only
aggregated information is available
>Systematic, quantitative analysis is not
possible without suitable quantitative
information. Current Benefit-Risk analysis is
qualitative!
Development
Launch
Discovery Pre- Clinical trials Marketing
clinical
trials Phase I Phase II Phase III Phase IV clinical trials
Compound
DBs
Pre/clinical trial DBs
SmPC DBs
Time
ADR DBs
CPOEs
Date 01.12.2008
>Dose-response curve-fitting with various A-II
antagonists
Dose (mg/d) Dose (mg/d)
0 100 200 300 0 8 16 24 32
0 0
-2 -2
Trough DBP (mm Hg)
Trough DBP (mm Hg)
-4 -4
-6 -6
-8 -8
Dose (mg/d) Dose (mg/d)
-10 -10 0 50 100 150 200 0 80 160 240 320
0 0
-12 -12
-2
-2
Trough DBP (mm Hg)
Trough DBP (mm Hg)
-4 -4
Similar compounds, -6 -6
partially different -8 -8
indications, totally different -10 -10
clinical data! -12 -12
Date 01.12.2008
Escher-project
Workpackages 3.1 and 3.2
Date 01.12.2008
>ESCHER
Is a TI-Pharma project with an objective to
“demonstrate, that another way is possible”
Incorporates 3 universities (+medical centers), 4
PostDocs, 17 PhD students, 4 drug development
companies, and x external personnel
WP 3.1: develop a new framework for drug
benefit-risk assessment
WP 3.2: build a drug information system that
allows quantitative comparisons
>Benefit-risk analysis of WP 3.1 requires data
from the DIS of WP 3.2
Date 01.12.2008
>Escher 3.1
How can we measure benefits and risks?
- Rank drugs and placebo for the same indication
- Multiple criteria
Inherent value judgements
- But what about clashing / missing preference
information?
Quantitative data available
- But data is uncertain!
- Should it be used “as is”?
Multi-Criteria Decision Analysis (MCDA)
Date 01.12.2008
> Stochastic Multicriteria
Acceptability Analysis
(SMAA)
Allows MCDA with
imprecise criteria
measurements and
missing/incomplete
preference information
Criteria measurements
can be defined through
joint probability
distributions -> RCT
data can be used +-
directly
Date 01.12.2008
> SMAA central weights
Central weights are “typical” preferences that
favour different alternatives
Although drug A might not have “better” benefit-
risk ratio than drug B with all preferences, some
preferences usually support A as well
Elevator planning
with SMAA
Date 01.12.2008
>Rank acceptability indices describe stability of
ranking, and can be used in risk management
Rank acceptability indices
80
70
Choosing a location 60
BEN
CAS
for a new cargo 50 DAK
TAN
40 Acceptability
airport in Morocco 30
AGA
RAB
with SMAA b9
20 OUJ
10 MAR
b7
0 FEZ
b5
BEN
CAS
Rank
DAK
b3
TAN
AGA
RAB
b1
OUJ
MAR
Alternative
FEZ
Date 01.12.2008
>Escher 3.2
Supports various other workpackages by
building an information system that allows
quantitative analyses
Web-based drug repository (Java, Spring)
Agile development
Enables various new
research topics
Date 01.12.2008
Why Agile?
Date 01.12.2008
>How to model relevant data (SmPC)?
5.1 Pharmacodynamic properties 5.1 Pharmacodynamic properties
Pharmacotherapeutic group: {group [lowest available level]}, ATC Pharmacotherapeutic group: Drugs used in erectile
code: {code}
[For products approved under “conditional approval”, include the
dysfunction. ATC Code: G04B E03
following statement:] …
selectivity over PDE6 which is involved in the
[For products approved under “exceptional circumstances”, include
the following statement:]
phototransduction pathway in the retina. At maximum
it has not been possible to obtain
complete information on this medicinal product. 4,000-fold selectivity for PDE5 over PDE3,
The European Medicines Agency (EMEA) will review any new the cAMP-specific phosphodiesterase isoform involved in
information which may become available every year and this SPC will the control of cardiac contractility.
be updated as necessary.>
…
Template Viagra SmPC
Date 01.12.2008
>US Food and Drug
Administration
(FDA) is working to
build a DIS (Janus)
incorporating “raw”
data
>The European
Medicines Agency
(EMEA) doesn’t see
aggregated data as
a problem
>Cause? FDA is multi-
disciplinary, EMEA
consists of medical
doctors
Date 01.12.2008
>Conclusions
Drug regulatory process is in need of reform
Current drug information systems cannot
support the future needs, because they don’t
store the data in an appropriate format
Escher-project tries to show, that a different
“way of doing things” is possible
Department of B&IS participates in the
project through Bert, Tommi, Vahid, Douwe
(starting 1d/w@Jan), and 1 more PhD-
student (starting@Apr)
Date 01.12.2008
>Thank you!
>Q?
>Future publications:
T. Tervonen, V. Oskuee, E.O. de Brock, P.A. de
Graef, H.L. Hillege (2008). Current status and
future perspectives in Drug Information Systems
(manuscript)
T. Tervonen, D. Postmus, H.L. Hillege (2009)
Multi-criteria decision analysis in drug benefit-risk
analysis. Invited presentation, 23rd European
Conference on Operational Research, Bonn,
Germany. July 5-8, 2009
>/dev/null