Transcript Philippe Lambin

The barriers for primary and secondary use of
EHR systems:
The clinical point of view
The Maastricht experience
Philippe Lambin
Contents
• The MAASTRO experience:
MAASTRO = An independent Radiotherapy
centre receiving cancer patients from 5
different hospitals (interoperability is sorted
out!)
• The barriers from a clinical point of view
EHR MAASTRO: 240 RT protocols + workflow, made by MD’s,
costs 5 minutes extra per patient
List of Treatment
protocol
In the coming years moren then 500
protocols (see PWC Pharma 2005)
Predictive model allowing treatment individualization:
An holistic approach
Prospective gathering of
pre-treatment data (+CI) Part of EHR
Biological
Data
Clinical Data
Image Data
Data-based &
Knowledge
based models:
Probability
of Survival &
Complications
(+ CI)
for treatment
x, y, z…
- Survival: National Database (GBA)
- Complications: Module EMF,
Questionnaire CTC like GP-PatientsLong specialist
Treatment
administered
Feed-back Loop
Real Outcome
(Complications,
Survival)
Quality of treatment is important! Register it
e.g.Two Dimensional Dose Guide radiotherapy with Portal Dose
Verification
predicted
portal dose
measured
portal dose
vs
Equivalent for drug =
Compliance, PK
*van
Elmpt, Nijsten et al., Med. Phys.
2005.
W. van32(9),
Elmpt et al., WIP
3
=
gamma
3
evaluation
Computer Assisted Theragnostic model
Prospective
gathering of per, post
treatment data (+CI)
Prospective
gathering of pretreatment data (+CI)
Biological Data
Clinical Data
Image Data
Data-based &
Knowledge
based models:
Probability
of Survival &
Complications
(+CI)
for treatment
x, y, z…
Biological Data
Treatment
administered
Clinical Data
Image Data
Treatment
Data
(Description,
Quality)
Feed-back Loop
Data-based &
Knowledge
based models:
Probability
of Survival &
Complications (+
CI)
for treatment
administered
Real Outcome
(Complications,
Survival)
Contents
• The Maastricht experience
• The barriers from a clinical point of view
Barrier? The MD’s
EHR = decrease of efficiency (less patient
seen in consultation, more work for the
MD’s, MD’s can not type...)
Solution: Use defaults, create a “win-win”
Train-educate MD’s, improve interaction with IT
MD’s are responsible of the individual care!
Solution: Involve them upfront in the R&D
Barrier? Lack of common ontology language
Especially for multicentric use
Solution: Use standard, invest in ontology = high priority
An ontology is the representation of the entities, ideas and
events, together with their properties and relations. These
are structured according to a system of categories. It is
more abstract and generic than a data model, which is often
grounded in the organisation and business processes of a
particular enterprise. The process of creating an ontology for
a specific domain is known as ‘ontology engineering’.
Barrier? Conventional clinical research
Three problems: a) less than 3% of the patient
population included in trials; b) standard clinical
trials often exhibit a strong bias in patient
selection; c) the costs of R&D and clinical research
are increasing.
We need a new complementary paradigm:
Machine learning clinical research based on
the
“No objection rules” (e.g. The Netherlands)
only when standard treatment (observational
study, long. cohort, saftey monitoring.
Barrier? Privacy aspects
Partial solution:
GRID
SOKU:
Data mining
without moving
the data
Software for
imaging,
Coded data,
not anonymous!
Barrier? Methodological, need of
large numbers + independent
validation dataset
Multicentric approach to have:
1.Large numbers of patients
2. Independant data set for validation
Barrier? Clinical aspects: Follow-up
For Survival: National database(e.g. GBA in The
Netherlands)
For complications, other diseases...:
Standardized scoring system (CTC NCI)
(e)Questionnaire to the patients
Database of the GP or minimum
European EHR
Barrier? Higher requirement for
clinical research
More data needed: QoL, unusual
imaging…
Higher quality: check inconsistencies
Stricter rules :e.g. GCP certification
Thank you for your attention
Barrier? IT
No really: we did it
HL7 too limited for Radiotherapy: we need a
broader standard
Barrier? Summary
MD’s
Semantics – Ontology
Legal aspects, informed consent
Need of multicentric data
Access to follow-up data (including national database)
Higher requirement for clinical research
New paradigm for clinical research
Barrier? Summary of potential
solutions: Merge HER for care and
MD’s
research
No objection rule, new concept of clinical research +
safety monitoring
Common ontology
National - European database
Minimum EHR
GRID- SOKU – improved HL7
Certification, standard for EHR