Storm-WMA-discussion-HHS

Confidentiality & ethics
1981 – Council of Europe Convention no 108
• Protection of individuals with regard to the automatic
processing of personal data
• Data is collected and treated lawful
• Filed for specified, explicit and legitimate purposes
• Data must be relevant and sufficient, and not excessive
• Data must be accurate and if necessary kept updated
• Data to be filed in a format that only allow identification of
persons needed to fulfil the purpose
Public health studies under threat
Explicit consent difficult to obtain
• Audit
– cancer screening programmes
– HpV vaccination programmes
– Survival & outcome research
• Occupational hazards and cancer
– Nuclear power and other radiation
– EMF – mobile phone’s etc.
• Environment
– Social inequality
– Risk factors identified in large groups
Confidentiality & ethics
Personal data & health research
• Clinical – selected individuals
• Informed consent
• Scientific ethical committee system operates
• Data inspection agencies or other authorities
• Epidemiological – populations
• Often involves very large if not all population
• Register based – data linkage – no contact to individuals
• Data inspection + patient right law + ethical committee
• Biological data
•A combination of clinical and epidemiological
• The codes of conduct in medical and epidemiological
research should be considered separately
Confidentiality & ethics
Why not informed consent?
• Representativity & generalization
• Avoid bias – selective loss (Low response rates)
• Solid conclusions
• Need for power and statistical strength
• Completeness
• Few cases may determine risk
• Loss in linkage may bias results
• All population
• Disproportionate effort
• The legitimacy of weighting practicability, or
feasibility, in deciding whether it is necessesary to
seek individual consent for retrospective (registry,
biobank or clinical note based) studies should also be
endorsed with the WMA declaration.
Confidentiality & ethics
What if we miss a link!
1
10
100
1000
40-49 50-59 60-69 70-79 80+
Age
SIR
Error 0%
Error 2%
Error 5%
Error 10%
Error 20%
Influence of missed
link to mortality – by
error proportion
(E. Pukkala)
Leukaemia risk in airline pilots – Denmark:
5 cases – significant increased risk
4 cases – no signifcant risk – but elevated SIR
Germany NRW cancer registry linkage study 150000 records
Pseudonyms: 1% linked wrongly 2% Not linked at all
Why not TTP encryption
• Uncertainty about linkages
• Errors in ID may radically change ”ID”
• Researchers lose track of data and responsability for
key variables ”ID”
• Third party increase time, costs and uncertainty on
the linked data
• At the outset Public Health researchers are
considered criminal or careless neglecting existing law.
• Anonymous or pseudo anonumous data are not
preferable to identifiable data and cannot be
satisfactory for the purpose of a health database.
• Evidently security measures to avoid unintended
disclosure apply.
Confidentiality & ethics
Derogations in the new EU regulation- Who decides
public interest?
• Articles 81.2 and 81.3 are still problematic:
• Exemption for informed consent for reasons of public
interest in the area of public health if it serves a high
public interest, if that research cannot possibly be
carried out otherwise.
• Pseudonymisation apply
Why not specify the need for use of the best possible method to
achieve results – it concerns yours and my life!
Freedom of Research and Ethics
• No research without ethics
• Without research no ethics
• Ethics is about the good life (Aristoteles)
• Research contributes by promoting and creating the basis for
the good life
(E.Tiedeman, previous Chair of the Natl. Ethics Committee of Denmark)
Confidentiality & ethics
• Research on health data can be done in an unethical way – which should
be avoided
• Research on data of low or uncertain quality should also be avoided
• Failure to do health research is also unethical – and devastating for public
health
Confidentiality & ethics
Does one size fit all?
https://plus.google.com/+MarMDog/posts/JFV3qCDXsy6