Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World...
Transcript of Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World...
![Page 1: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/1.jpg)
Challenges and opportunities of Real
World Datasets
Paolo Bruzzi - Genova
Thoracic Oncology Padova
March 29-30, 2019
![Page 2: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/2.jpg)
Contents (15’)
• Glossary
– Datasets vs Registries vs Studies vs Aims
– Trials vs Observational Studies
– Efficacy vs Effectiveness
– Internal Validity vs External Validity
– Scope of observational studies
• Challenges and opportunities of Real World
Datasets?
![Page 3: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/3.jpg)
Contents (15’)
• Glossary?
“There is great chaos under heaven …”
Mao Zedong, 1966, letter to his wife
![Page 4: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/4.jpg)
What are “real life” data?
“Data used for decision-making that are not
collected in conventional randomized clinical
trials”
International Society of Pharmacoeconomics and Outcomes
Research
![Page 5: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/5.jpg)
• Pragmatic clinical trials
• Prospective observational studies and patient registries
• Administrative claims data
• Patient surveys
• Electronic health records/medical chart reviews
• Government- or third-party-sponsored systematic surveys that
assess public health, resource consumption, practice patterns,
and clinical trends
Garrison LP, et al. Value Health. 2007;10(5):326-335.
Sources of Real-World Data
![Page 6: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/6.jpg)
OBSERVATIONAL
STUDIES
REGISTRIES
ADMINISTRATIVE
DATABASES
![Page 7: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/7.jpg)
a)Descriptions
b) Inferences
REGISTRIES
ADMINISTRATIVE
DATA
Dati raccolti
ad Hoc
Data
collected
ad hoc
![Page 8: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/8.jpg)
STUDIES
REGISTRIES
ADMINISTRATIVE
DATA
Dati raccolti
ad Hoc
Data
collected
ad hoc
![Page 9: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/9.jpg)
Contents (15’)
• Glossary
– Datasets vs Registries vs Studies vs Aims
![Page 10: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/10.jpg)
Definitions
• Dataset -> A collection of data
– No aims
– Specific/generic aims ≈ Registry
• Registry -> A planned & structured data collection
• Study -> An endeavor aimed at answering
questions
• Aim -> Aim
![Page 11: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/11.jpg)
Examples
• Dataset ->
– No aims
– Specific/generic aims
• Registry ->
• Study ->
• Aim -> Aim
• Any set of records (computer, paper, etc.)
– Clinical database, big data
– Birth Records, Hospital DF’s
• Cancer Registry
• Trial, Observational Study
• To evaluate..
– the incidence of...
– The efficacy of...
![Page 12: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/12.jpg)
Main Differences
Dataset
A collection of data from one or
multiple sources
Registry
• Prior definition of
- Target Population
- Data to be collected
- Sources
Existing databases
Ad hoc data collection
- Format of data (structure of DB)
• Active search of missing data
• Multiple aims (No specific aim)
![Page 13: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/13.jpg)
Main Differences
Registry
A planned & structured collection
of data, usually with multiple
but generic aims
Study
• Prior definition of
- STUDY AIM(S) = Study question
- Eligible Population
- Data to be collected
- Sources
Existing databases
Ad hoc data collection
- Format of data (structure of DB)
• Active search of missing data
![Page 14: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/14.jpg)
Main Differences
Registry
A structured collection of data,
usually with multiple but
generic aims
Study
• Study design
• Statistical Plan
• Timeline
• Intervention (if any)
![Page 15: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/15.jpg)
Exam in any basic course in statistics or
research methodology
• Question:
– Why careful design and skilled statistical planning are
required before doing a study?
• Answer:
– To prevent or assess the classical errors due to
selection bias, assessment bias, etc.
– To avoid the confusion between the generation of
hypotheses and their demonstration
– To accomodate the problems related to multiplicity
![Page 16: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/16.jpg)
Ad hoc collected Data
Registries
Data extraction from datasets
Data Mining
= Studies !
Same
Methodological
Problems
![Page 17: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/17.jpg)
1st Conclusion
• Datasets and Registries DO NOT provide
knowledge, but information that may prove useful
to create knowledge
• Building Knowledge is possible only by means of
(adequately planned and conducted) STUDIES
• The Analysis of a dataset without a prior adequate
study design is doing a study of poor quality
![Page 18: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/18.jpg)
Observational Studies vs Trials
• Observational Studies = Studies in which
– No change is introduced in the standard routine clinical
practice
– With the exception of tests/exams/questionnaires/etc
aimed at gathering information needed for the study
– That is not used in the management of the patient
• Trials = All clinical studies that cannot be classified
as observational
![Page 19: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/19.jpg)
STUDIES
Observational study Experimental Study (= Trial)
• Prior definition of
- STUDY AIM(S) = Study question
- Eligible Population
- Data to be collected
- Sources
Existing databases
Ad hoc data collection
- Format of data (structure of DB)
• Active search of missing data
• Prior definition of
- STUDY AIM(S) = Study question
- Eligible Population
- Data to be collected
- Sources
Existing databases
Ad hoc data collection
- Format of data (structure of DB)
• Active search of missing data
![Page 20: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/20.jpg)
STUDIES
Observational study Experimental Study (= Trial)
• Study design
• Statistical Plan
• Timeline
• Intervention
• Study design
• Statistical Plan
• Timeline
![Page 21: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/21.jpg)
TRIALS
Uncontrolled Trial Randomised Trial
• Study Aims
• Random assignment of
treatments
• Unbiased assessment of
Outcome
• Statistical Plan
• Study Aim(s)
• Selection of treated &
Controls
• Unbiased assessment of
Outcome
• Statistical Plan
![Page 22: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/22.jpg)
Note: The GRADE method
According to the Grade method (Grading of
Recommendations Assessment, Development and
Evaluation) all clinical studies that are not
randomised are referred to as observational
studies
Observational studies = Uncontrolled trials?
Observational Trials???
![Page 23: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/23.jpg)
Observational Trial? -> Oximoron!
Observational study
• Therapies already in use
• Unselected centers
• Unselected patients
• Routine assessments
Uncontrolled trial
• New therapies
• Specialised centers
• Selected patients
• Special assessments
![Page 24: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/24.jpg)
Dominant Thinking
The Randomised Clinical Trial is the only source of
reliable evidence on the efficacy of medical
interventions
- Uncontrolled trials, observational studies
Bias in the selection of treated (and controls, if any)
Bias in the assessment of outcome
![Page 25: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/25.jpg)
Contents
• Glossary
– Datasets vs Registries vs Studies vs Aims
– Trials vs Observational Studies
– Efficacy vs Effectiveness
![Page 26: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/26.jpg)
Efficacy vs Effectiveness
• Efficacy is the extent to which an intervention
does more good than harm
under ideal circumstances
• Effectiveness assesses whether an intervention
does more good than harm when provided
under usual circumstances of healthcare
practice Cochrane AL. Effectiveness and efficiency: random reflection on
health services. London: Nuffield Provincial Hospitals Trust; 1972.
![Page 27: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/27.jpg)
Argument
• Do (Randomised) Clinical Trials provide evidence
that is sufficient to evaluate the true impact of a
new therapy in routine clinical practice?
– “Fit” patients
– Young patients
– Specialised Centers
– Special procedures
– “Carefulness”
![Page 28: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/28.jpg)
Argument
• Do (Randomised) Clinical Trials provide evidence that
is sufficient to evaluate the true impact of a new
therapy in routine clinical practice?
NO!
Solution
– Explanatory trials -> Efficacy
– Pragmatic trials -> Effectiveness
(Unselected patients and centers, routine care +/-
experimental at random) -> Large &Simple Clinical trials
![Page 29: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/29.jpg)
10.000
1000
100 Phase II
Explanatory
Activity Efficacy Effectiveness
Pragmatic
LSCT
![Page 30: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/30.jpg)
Problem
IT IS NOT ACCEPTABLE, FROM AN UNETHICAL VIEWPOINT, TO LAUNCH A PRAGMATIC RANDOMISED TRIAL WHEN ONE TREATMENT HAS BEEN SHOWN IN AN EXPLANATORY RANDOMISED TRIAL TO BE BETTER THAN THE OTHER
Solution: Meta-analysis!?
![Page 31: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/31.jpg)
10.000
1000
100 Phase II
Explanatory
Activity Efficacy Effectiveness
META-ANALYSIS
Explanatory
ExplanatoryExplanatory
![Page 32: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/32.jpg)
The EBCTCG involves almost all trialists worldwide who
have done relevant randomised trials of the treatment of
women with breast cancer, …
Several hundred research groups have shared individual
patient data on more than 450,000 women in 400
randomised trials for meta-analyses
![Page 33: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/33.jpg)
Problem
Modern Oncology
1 drug -> 1 condition -> 1 trial-> Registration
What do we meta-analyse?
How do we treat patients excluded from trials?
![Page 34: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/34.jpg)
10.000
1000
100 Phase II
Activity Efficacy Effectiveness
“REAL LIFE” studies
(Observational)
Explanatory
![Page 35: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/35.jpg)
Contents
• Glossary
– Datasets vs Registries vs Studies vs Aims
– Trials vs Observational Studies
– Efficacy vs Effectiveness
– Internal Validity vs External Validity
![Page 36: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/36.jpg)
Interpretation the result of a study
1. Internal Validity
2. External Validity
Is it true?
So what?
![Page 37: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/37.jpg)
Internal Validity
• Research protocol
• Identification of the primary aim(s)
• Randomization Procedures
• Endpoint – Choice/Assessment/Blinding
• Statistical Plan
• Lost to follow-up/not evaluated
• Analysis ‘intention to treat’
![Page 38: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/38.jpg)
Interpretation
• Internal Validity:
• External Validity:
- Extrapolate
Possibility to - Generalise
- Apply
the results of
the study
![Page 39: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/39.jpg)
Determinants of External Validity
• INTERNAL VALIDITY!!!
• Study Design (Contrast)
• Selection Criteria/Patients Characteristics
• Participating Centers
• Treatment Protocol –Follow-up Protocol
• Endpoint
• Compliance – Contamination
• Precision of the estimates
• Analysis ‘intention to treat’
![Page 40: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/40.jpg)
Observational Studies
Due to any (or most) of the following drawbacks
• Lack of
– active selection of cases receiving the newtherapy
– concurrent, comparable controls recevingstandard treatment (Randomization)
– adequate follow-up (ITT) & unbiased endpointassessment (Masking)
The Internal validity of observational studies isALWAYS QUESTIONABLE
![Page 41: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/41.jpg)
Observational Studies
METHODOLOGIC PROBLEMS
Depend entirely on the study aim!
![Page 42: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/42.jpg)
Observational Studies: Possible aims
To gather (further) information on
• Efficacy
• Toxicity/Adverse effects
• Modes of use
• Patterns of Care
![Page 43: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/43.jpg)
Observational Studies: Possible aims
To gather (further) information on
• Efficacy in specific care settings or patients
subgroups
![Page 44: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/44.jpg)
Question
• To estimate the efficacy of a new drug we require
a randomized trial with strict methodological &
statistical standards ( masking, Intention to treat,
Planned statistical analyses, etc)
• Is it conceivable that we modify these estimates in
specific care settings or patients subgroups based
on the results of (observational) studies of
questionable validity?
![Page 45: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/45.jpg)
2nd Conclusion
Observational studies should not be used to modify
the estimates of the efficacy of treatments
obtained in randomised clinical trials, unless
dramatic effects (lack of) are seen in subgroups
previously not included in efficacy trials
![Page 46: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/46.jpg)
Observational Studies: Possible aims
To gather (further) information on
• Efficacy
• Toxicity/Adverse effects
– Rare events
– Rare Subgroups
– Patients not included in trials
• Pharmacoepidemiology - Data Mining (Big Data)?
![Page 47: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/47.jpg)
Observational Studies: Possible aims
To gather (further) information on
• Efficacy
• Toxicity/Adverse effects
• Modes of use
• Patterns of Care
Quality of care
Critical Research area
Few Studies usually of
poor quality
![Page 48: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/48.jpg)
Observational Studies of quality of care
(appropriateness)
Methodological Problems
1) Selection Bias
2) Assessment
3) CONFOUNDING
Association does not mean causation!
(e.g. lower % of histological confirmation in
geriatric oncologies)
![Page 49: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/49.jpg)
Contents
• Glossary
– Datasets vs Registries vs Studies vs Aims
– Trials vs Observational Studies
– Efficacy vs Effectiveness
– Internal Validity vs External Validity
– Scope of observational studies
![Page 50: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/50.jpg)
Summary
Scope of observational real life studies
Descriptive Epidemiology
• Epidemiological Determinants
• Outcomes
• Efficiency/Costs
•Appropriateness
• (Effects of care??Efficacy??)
(Historical controls for uncontrolled trials?)
Needs
Quality of Care
![Page 51: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/51.jpg)
Contents (15’)
• Glossary
– Datasets vs Registries vs Studies vs Aims
– Trials vs Observational Studies
– Efficacy vs Effectiveness
– Internal Validity vs External Validity
– Scope of observational studies
• Challenges and opportunities of Real World
Datasets?
![Page 52: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/52.jpg)
Real world data?
• Real world datasets will not (are not likely to) prove
useful sources of information on the efficacy of
treatments
• They may prove very important source of information
for observational studies, including those aimed at the
assessment of the health impact of new technologies
• However, so far, most Real Life studies have been
poorly planned, conducted and reported, while their
methodology is more challenging than that of
Randomised Clinical Trials
![Page 53: Challenges and opportunities of Real World DatasetsChallenges and opportunities of Real World Datasets Paolo Bruzzi -Genova Thoracic Oncology Padova March 29-30, 2019 Contents (15’)](https://reader034.fdocuments.net/reader034/viewer/2022050210/5f5c378f42cc7a0b2c0c2b8e/html5/thumbnails/53.jpg)
Contents (15’)
• Glossary?
“There is great chaos under heaven ….
….and the situation is excellent”
Mao Zedong, 1966, letter to his wife