Professor Mika Gissler
Nordic registers for research – opportunities and challenges
All the five Nordic countries have comprehensive population-based health information systems based on individual-focused health registers. Unique identification numbers makes data linkages technically very easy. Registers on cancers have operated since the 1940s, registers on infectious diseases since the 1950s, hospital discharge registers, cause-of-death registers and birth and malformation registers since the 1960s, and health care quality registers and prescription registers since the 1990s. All countries have a register-based census since the 1990s: information on education, income, socioeconomic position, family background and relatives can be linked to the various registers. The current national and European legislations allow the collection of sensitive health and social welfare data without informed consent, and enable the use of such data in scientific research and in for statistical purposes. The main obstacles for Nordic science are differences in register contents, classifications, definitions and data coverages – the harmonisation process may take long time. Also the process to get permissions to combine data from multiple Nordic countries may be complex and time-consuming. The collection of new data sources, such as primary health care data and electronic patient records opens new possibilities for the Nordic countries.
Dr Siri Haberg
Multipurpose use of national health databases – surveillance, natural experiments and linkage with cohort studies
Health registries and national health data bases can serve several purposes, including surveillance, evaluations of policy changes and “natural experiments” like pandemics and other outbreaks. Another useful approach is to combine analyses based on national databases with analyses from sub populations participating in cohort studies with questionnaire data and biological samples. Norway has the longest running birth registry in the world, and the possibility to link data from several national sources on an individual level. This allows for long term follow up of perinatal outcomes and birth characteristics, and the availability of IDs from parents enables use of family designs. Several examples will be presented.
Professor Rodney Jackson
Using big data to improve vascular risk prediction and better targeted treatment
Readily available treatments can halve the risk of premature vascular disease but under - and over-treatment is common and there are substantial ethnicity- and deprivation-related inequities in vascular disease burden. The effectiveness of most treatments depends on patients' risks of developing vascular disease but estimating risk is difficult without risk prediction algorithms and few valid algorithms have been developed.
We have established three overlapping 'big-data' cohort studies, a primary care cohort, a hospital cohort and a national cohort. These cohorts are electronically linked to the same routine national health data-sets of laboratory investigations, drug treatment, hospitalisations and deaths. Using these linked data we are: i. developing new risk algorithms to assist clinicians estimate vascular risk in multiple high-risk populations; ii. investigating in whom, where and why, under- and over treatment and inequities in vascular risk and risk management occur; iii. developing and implementing a multi-algorithm risk prediction engine and a 'big-data' vascular health information platform to support initiatives to increase appropriate treatment, reduce inequities in vascular disease outcomes and improve vascular health.
Professor Louisa Jorm
Big data meet epidemiology
Health and medical big data come from a variety of sources, including administrative databases, clinical trials, electronic health records (EHRs), patient registries, genomic, and other ‘omic measurements and medical imaging. More recently, data are being integrated from social media, wearable and implantable devices, mobile applications, occupational and retail information and environmental monitoring. These data are ‘big’ in volume because they include large numbers of records (e.g. administrative data), large numbers of variables (e.g. ‘omics data), or both (e.g. EHRs). They are also characterised by great variety (including both structured data and unstructured data such as free text and images) and high velocity (generated in or near real-time). Health and medical big data present vast potential for the discovery of relationships among pieces of information that would not previously have been possible. This requires integrating data-driven and hypothesis-driven approaches, and deductive and inductive reasoning, and applying new and up-scaled analytic methods that draw on both statistics and computer science. It is time for epidemiologists to embrace this challenge!
Professor David Preen
The future of epidemiology
Scientific enquiry in the areas of health and medicine continue to evolve as the understanding of our world and access to new technologies expand. Since John Snow's sentinel work using statistical mapping to investigate cholera outbreaks, the discipline of epidemiology has also undergone considerable transformation. The expansion of focus from infectious to non-communicable disease, and the incorporation of socio-ecological paradigms have characterised changes in the field of epidemiology over the last century. Accordingly, epidemiological theory and methods have developed in-step with this changing focus leading to the modern epidemiology movement which gained considerable traction in the second half of the 20th century.
Exponential advances in technology over recent decades have unquestionably created new and exciting avenues for health research. The advent of big data, data linkage, geo-spatial technology, machine learning, natural langue processing, real-time analytics and the expanding 'omics universe, to name but a few, provide many opportunities, and also possess a range of risks, for the field of epidemiology as a standalone discipline. We likely stand on the cusp of the next era for our discipline. However, for epidemiology to thrive, new methodologies need to be conceptualised and embraced with fresh approaches to adopt new technologies as they continue to develop. This session will overview the potential opportunities and challenges ahead for the discipline of epidemiology. An expert panel will also bring together leading thinkers and practitioners across health, technology and community sectors to discuss implications and possible solutions for the future of epidemiology.
Dr Maree Toombs
How working with Aboriginal communities towards health improvements will close the gap in health outcomes: Evidenced based research done Mob way
Historically, Indigenous Australian are largely obliged to rely on a health system and services that they have not developed and that they do not control. Through successful collaboration with local Indigenous community leaders and stakeholders over the past 5 years, developing co-designed evidenced based programs that translate into best practice are becoming the norm for Indigenous health research and in South West and Central Queensland. I will present 3 NHMRC funded projects that demonstrate Participatory Action research shows great efficacy and translates into data that leads to better outcomes. Emphasis and resourcing of targeted, evidence-based community engagement at the front end of research, to increase health literacy and knowledge ultimately enhances input into research and improvements of service delivery in the catchment area of these projects