“FROM OBSCURITY TO ANDREW DENTON IN 12 MONTHS: USING THE AUSTRALIAN OF THE YEAR AWARD TO IMPROVE CHILD HEALTH & WELLBEING”
Professor Fiona Stanley AC
Thank you very much for inviting me to this wonderful conference in such a lovely island; both Geoff and I are delighted to be here.
Whilst John Phillips was keen that I talked about the experience of using the Australian of the Year award to advocate for children and youth, I wanted also to include issues around population data and privacy issues. This invitation to speak has provided me a wonderful opportunity as a researcher in population health to present to you my arguments as to why we need to 1. collect population data and 2. link it together (without consent) and 3. use it for public good.
Professor Lawrence Gostin (Department of Law, Georgetown University Washington USA) wrote
“Law and ethics in population health are having a renaissance. Once fashionable during the Industrial and Progressive eras, the ideals of population health began to wither with the rise of liberalism in the late 20th century. In its place came a sharpened focus on personal and economic freedom. Political attention shifted from population health to individual health and from public health to private medicine.”
Gostin suggests that there is a conflict between individual rights (eg. to privacy) and activities that help the majority (eg. linking health data on individuals without their consent for public good).
What is record linkage? It is the bringing together of records from different sources, on the one individual, to provide a composite picture. The history of the population data base which we developed in Western Australia to monitor and study child health is the best Australian example of how powerful such data can be for public good research, evaluation of health services and improvement of health.
In the 1970s there were major debates about birth defects data following the thalidomide scare and how we could set up early warning systems to alert us to such problems. As well, many in the community (both paediatric and generally) were voicing concerns about the long term effects of more intensive perinatal care, which had resulted in the dramatic increase in survival of preterm and low birth weight babies. These encouraged paediatric epidemiologists like myself to think about the data and methods we needed to have to address these concerns.
In WA, we already had the beginnings of excellent population data and record linkage. Throughout the 1980s we set about developing the most complete record linked collections of data to monitor and study all pregnancies and their child and (eventually) youth outcomes. Birth and death information on all births was brought together for the first time in Australia. Neither death nor birth certificates recorded race, birth weight or gestational age. Bringing them together with midwives notification which had these items recorded enabled us to calculate still birth and infant death rates by birth weight and to provide the first data comparing Aboriginal and non-Aboriginal mothers and babies (their birth weight, prematurity rates, death rates by cause etc). At the same time as linking these existing records, we set about collecting new data (by establishing the first registers) on birth defects, cerebral palsies and other problems to answer questions about trends, and whether there were any increases in problems noted following antenatal and perinatal care practices.
These data enabled us to show that, in spite of the promise that electronic foetal monitoring and caesarean sections would reduce cerebral palsy, cerebral palsy rates actually rose following more widespread use of these practices. This helped in the research around the huge increase in obstetric litigation and encouraged research into the antenatal period, where we now know most of the cases of cerebral palsy and all birth defects arise.
The data also were used to sample cases and controls for our important studies which contributed to the discovery that folate (a vitamin found in leafy green vegetables and Mediterranean food!) around the time of conception could reduce the risk of spina bifida and related defects.
Additional linkages throughout the 1980s and 1990s included all hospitalisations for all causes in children and youth in all (public and private) hospitals in WA, the mental health register, the suicide data base, and we also linked address of each birth into the collection district of each census so that population census descriptions (eg socio-economic information) were available as well.
CURRENT SITUATION – THE DATA LINKAGE UNIT
With the move of all linkage to the WA Data Linkage Unit, we now have access to prescription and Medicare data which is very exciting, enabling us to detect the harmful effects of drugs in pregnancy or in paediatric usage.
All population health data linkage in WA is now carried out completely independently of any user (researcher) to further enhance privacy and to ensure that the data on any individual is kept completely confidential. This means that no-one linking the data has access to the sensitive clinical or exposure information and no-one using the data for analysis and eventual publication, has any access to any of the identifying information, thus ensuring privacy. This has actually meant that research requesting name identified data has decreased in WA as the availability of data linkage has increased.
ADVANTAGES OF RECORD LINKAGE:
Maternal and Child Health Research Database (MCHRDB)
Data on all birth cohorts from 1980 enables:
- Description of total burden of problems in children and youth
- Description of the risk and protective factors for these problems
- Sample unbiased groups for epidemiological studies (cohorts, cases and controls)
- Ascertainment of potential participation bias in epidemiological studies (and generalisability)
- Evaluation of the impact of interventions (public health or clinical programs) on all groups
- Complete ascertainment and thus not biased, no one excluded
- Results apply and are useful for the whole population and complete subsets (eg. rural, Indigenous, teenage mothers).
- Cheap compared with any studies where contact and consent from participants needs to be sought
- Valid and reliable data on issues often difficult to obtain in direct ways (eg. psychiatric illness, abortion, drug use etc)
- Reduce burden on population from surveys (many of which are unreliable)
- Technology for linkage now very advanced
- Privacy issues regarding record linkage now better handled eg: Data Linkage Unit
- Better data for policy and planning
- Improve data quality in administrative data sets
LIMITATIONS OF POPULATION DATA & RECORD LINKAGE
- Information only available on items and outcomes recorded in data bases (breadth > depth)
- Privacy issues still need to be addressed eg ethics committees, understanding of public good by the community
- Need better, complete denominators
- Changes in diagnostic classifications present challenges for temporal analyses
- (In)accuracy of recorded information
- Incomplete ascertainment
- Sample size for rare disorders (APSU)
The examples of public good research which are described below are from my area of expertise; there are even more examples from other health services researchers in WA such as Michael Hobbs (cardiovascular disease and aged care) and Darcy Holman (psychiatric, surgical and cancer services).
EXAMPLES OF PUBLIC GOOD RESEARCH
These data have been used to evaluate the impact of the increased survival following neonatal intensive care, on the rates of cerebral palsy in very preterm infants. Throughout the 1980s we were one of the only groups in the world to be able to monitor the increasing rates of neurological problems as more and more of these tiny neonates survived. We also analysed rates of cerebral palsy amongst twins and triplets as many more of such babies were born following IVF and similar interventions. This led to legislation to restrict the number of embryos transferred in infertility treatments to reduce multiple birth rates. Cerebral palsy rates amongst these births are now falling.
We used these data to evaluate the SIDS (cot death) prevention campaign in WA – showing a dramatic fall in these deaths in non-Aboriginal children but no significant fall in Aboriginal children. This led to a special commitment to reducing SIDS in this population. WA, in a program led by Professor Carol Bower, initiated the world’s first preventive program to increase periconceptional folate and we used the database to evaluate it. There was a 40-50% reduction in spina bifida and related defects but again in Aboriginal babies there was no impact. This information has spearheaded the approach to fortify flour with folate (mandatory) internationally.
Other examples include:
- Antenatal factors in childhood diseases
- Births to psychiatric patients
- Preventable illnesses and deaths in psychiatric patients
- Cannabis, drugs in motor vehicle accidents, suicides
- Evaluation of vaccine programs
- Cancer and birth defect occurrence – environmental monitoring
- Predicting future obstetric, neonatal services
- Adverse effects of prescribed drugs (Vioxx, Roaccutane etc)
WHY NOT SEEK CONSENT?
Seeking consent is neither practical nor feasible: it is actually impossible to contact 100% of the population. People move, cannot be traced, die or do not want to participate. It is not feasible because the costs of tracing people, informing them of the reasons the data are to be linked and obtaining consent would preclude this ever being funded. When asked face to face with full explanations as to why the data are important, how they will be used and how confidentiality can be assured, nearly 100% of people give consent.
SEEKING CONSENT RESULTS IN POOR SCIENCE (BIAS)
Seeking consent results in variable proportions of records being linked/available. The profiles of people who participate are very different from those who don’t. Non-participants are usually sicker, disabled or dead, have higher risk factors, are socially and geographically more mobile, are younger and come from deprived backgrounds. They are just the very people on whom we need the best information.
Seeking information on exposures from such studies results in even more bias from two major sources: 1. participation (as previously mentioned) and 2. recall bias. This has been shown to vary enormously on even major and important exposures such as alcohol, drugs, smoking, abortions, number of pregnancies/children, previous employment etc. Thus, if such exposure data are available from other complete and accurate sources which can be linked to complete and accurate outcome information, we then have accurate estimates of risks to enable the best information for prevention, evaluation of harmful treatments and better health care.
EXAMPLES OF BIAS
Bias results in misinterpretation and error, due to the sample in a study not being representative of the true risks.
Some examples resulting from studies which did not use complete population data are:
- Data on HIV status in women attending antenatal clinics in New York was sought to help plan appropriate services. Individuals who gave their consent had <1% HIV positive results whereas when all the bloods were tested (anonymously) nearly 15% were HIV positive. This resulted in a very different provision of services!
- A recent overview of studies was published to answer the question “Does termination of pregnancy (therapeutic abortion) increase the risk of breast cancer?” Those case control and cohort studies which obtained termination of pregnancy (TOP) data by contacting women and asking them about this history showed a significant increase in risk of TOP for subsequent cancer. Those studies using total population data and linking registers of TOP to registers of cancer showed the opposite – that TOP was actually protective! A very important finding for women, those counselling them and providing services. It is important to get the most accurate information.
- The Canadian Stroke register aimed to obtain representative cases for describing trends and aetiology, to evaluate treatments and to provide cases for randomised controlled trials. Their ethics committee insisted on them obtaining individual consent. This cost them $0.5M and even so resulted in a useless register, with only 34% of cases registered. Those cases were much younger, less sick, died less and were very unrepresentative of all cases.
- The Australian Women’s Longitudinal Study has a participation rate of less than 30% for women aged 18-24 years.
THE ROLE OF NHMRC AND NATIONAL PRIVACY GUIDELINES: WHY ARE WE WORRIED?
Both the National Health and Medical Research Council (NHMRC), the National Privacy Legislation and some (but not all) states (eg. not WA) have legislation which protects record linkage and access to data without consent.
These provisions state that seeking consent is best practice and not seeking it is tacked on to the end as an afterthought to be allowed if best practice is not achievable. I hope I have demonstrated that not seeking consent is actually best scientific practice. However, there is, as Gostin highlights in his paper, an increasing focus on individual rights over public/community good. Civil liberty and privacy groups, whilst actually very small in numbers, campaign vigorously (and effectively in some jurisdictions) against access and use of data. They influence the media, the public, consumer groups, lawyers and ethics committees. This means that unless properly informed (the way I hope you now are!), ethics committees can refuse any activities without consent in spite of the huge importance of the information so gained.
WHAT SHOULD HAPPEN?
What do “we” need to do? “We” are population researchers committed to doing the best research and obtaining the best evidence to enable the best planning and practice for health care in Australia. We need to communicate these ideas to those making decisions about privacy legislation, databases and their uses, to ethics committees and to the general public. We need to give examples of public good which such data provide and the worrying biases and wrong interpretations which come from studies which seek consent. We need advocates like yourselves and Lawrence Gostin who can influence those on committees and who make legislation,
I would like to end this section with another quote from Gostin:
“ In the late 20th century, scholars and politicians posed a key question ‘What desires and needs do you have as an autonomous rights bearing person to privacy, liberty and free enterprise?’ Now it is important to ask another kind of question ‘What kind of community do you want and deserve to live in, and what personal interests are you willing to forgo to achieve a good and healthy society?’ “
USING THE AUSTRALIAN OF THE YEAR AWARD TO ADVOCATE FOR CHILDREN
The Australian of the Year award was a fantastic opportunity to advocate for issues affecting children and young people in Australia. In 2003 there was much happening nationally and internationally. For example:
- Iraq war and its aftermath
- Australia – US relations
- Australia in Asia
- Children in detention, asylum seekers, Human Rights
- Aboriginal reconciliation, history wars
- Aboriginal family violence, child abuse, Governor General
- National research priorities
- University reform
- Criticism of the ABC
We planned the year to be strategic and focus only on the following key messages:
- Concerns regarding children and youth – future of Australia. How healthy child development links to National development
- Unacceptable increases in inequalities – Aboriginal AND non-Aboriginal
- How research and evidence should influence policy and practice
We wanted to have these messages always credible, supported by the best data and evidence available. Hence the importance of our population linked databases! Rather than accepting over 500 invitations to speak, the support team around me in the Institute and in the Australian Research Alliance for Children and Youth (ARACY) decided to accept only opportunities that (a) had National impact, (b) were appropriate for our key messages and (c) enabled us to target government, non-government, private sector and communities. Our Institute journalist (Elizabeth Chester) coordinated the media (print, radio, TV) to enable the most effective targeting for major impact.
WHAT ARE THE ISSUES FOR AUSTRALIA’S CHILDREN AND YOUTH?
In spite of our increasing prosperity we read much about children, youth and families in today’s media and hear about lots of reports and research on radio and TV. It seems we are all concerned for our children but we’re not quite sure what is really happening, why much appears to be so worrying and what we as ordinary people can do about it. So many aspects of life seem to have changed since we were young, in our families, in communities and in the workplace. Surely not all changes are impacting negatively on our children but which are and why? What does it mean for the future? If Australia is so well off, why is the picture for our children not hopeful, just and optimistic? Are similar things happening in other countries? What IS happening?
It has been taken for granted that as countries improved their economic prosperity and the GDP increased, that this would translate into improved health, wellbeing and happiness across the whole of society. That all a country had to do was to encourage wealth creation and the result would be more opportunities for people to live good lives, that the new economy would deliver a better world for the poor as well as the rich. Time saving new technologies would enable women to take on paid work, help all of us to work less and have more leisure time, and open up opportunities for improving our lives like no other era in our history. If we look back over a 100 years, this is broadly true. But it appears in recent decades that we have not done as well. The economists say “a rising tide lifts all boats” but it seems that the little boats have either got stuck in the mud or been swamped with the rising tide!
Canadian researchers have looked at the trends in their society and called it “Modernity’s Paradox”. They say that in spite of increasing economic prosperity and “globalisation” enabling greater access to opportunities, many key indicators of the health, development and well being of their children and youth are not improving and many are worsening. And the social gradients (we call them inequalities) – the differences in outcomes between the advantaged and disadvantaged groups in the population – are actually growing larger, not smaller as promised.
I would like to finish with a quote from Mamphela Ramphele, Managing Director of the World Bank:
“It is never too early to become involved but it can easily be too late.”
Copyright 2005. Greek/Australian International Legal and Medical Conference.
For more information contact Jenny Crofts at firstname.lastname@example.org