HealthLumen in conversation with Professors Deborah Schofield and Rupendra Shrestha
Deborah Schofield is a Distinguished Professor at the Macquarie Business School, Macquarie University, where she is the Director of the Centre for Economic Impacts of Genomic Medicine (GenIMPACT). An international leader in the economics of genomic medicine, Professor Schofield has pioneered the application of microsimulation modelling in health. Her research program focuses on the productivity impacts of illness, underpinning GenIMPACT’s work on the health, social and economic impacts of genetic disorders and the costs and benefits of genomic testing and targeted interventions.
Professor Schofield has over two decades of experience working as an economist and spent over a decade holding senior governmental roles in The Treasury, the Department of Health and Ageing, and the Department of Family and Community Services in Australia.
Rupendra Shrestha is an Associate Professor at GenIMPACT. He is an applied statistician and specialises in the application of statistical methods in large-scale complex microsimulation models for health economics research. His research focuses on the costs of chronic illness, and particularly on the societal costs and economic impacts of productivity loss. His recent research interests focus on the development of economic models to analyse the economic impacts of genetic disorders, and cost-effectiveness analysis of genomic sequencing and precision medicine – from both the healthcare and societal perspectives.
In 2018, a seminal review paper was published by Professor Deborah Schofield, Associate Professor Rupendra Shrestha, and colleagues, entitled: “A brief, global history of microsimulation models in health” [1].”
The review discussed how microsimulation models in health have evolved over the past three decades and what the future holds.
With great prescience, the paper highlighted genomics and precision medicine as key emerging health technology trends, especially in view of the decreasing costs and increasing accuracy of whole-genome sequencing, enabling it to become part of routine clinical practice.
In alignment with this trend, the authors went on to state that “Genomics requires a form of modelling that can capture these individual genetic differences and responses to therapy – a purpose to which microsimulation is eminently suited.”
Professors Schofield and Shrestha act as advisors on HealthLumen’s genomic modelling development program and the HealthLumen team recently sat down with them to discuss the advantages of applying microsimulation modelling techniques to precision medicine, and learn more about their recent research projects.
Our conversation, supplemented with quotes and references from recent key papers, provided fascinating insights. A summary is provided below, with further details in the full publications (see References).
Why is microsimulation so well-suited to genomic modelling and the field of precision medicine?
Precision medicine is an approach to patient care that allows interventions to be tailored specifically according to the patient’s genetic makeup. The essence of microsimulation lies in its granularity, as the modelling can simulate individual level characteristics and behaviours within a population and can therefore accommodate different treatments and outcomes for each individual. Consequently, this modelling approach is well suited for handling the heterogeneity in genetic differences and therapeutic pathways involved in precision medicine.
This is of course dependent on the availability of high quality genomic and socioeconomic data in order to produce robust, credible and accurate model results. Genomic data quality has increased dramatically over the past few years, with rapid development of Next Generation Sequencing (NGS), coupled with the decreasing costs required to perform sequencing. Many more institutions now undertake Whole Exome Sequencing (WES) or Whole Genome Sequencing (WGS), providing powerful data repositories which can provide inputs to microsimulation models. However there is still limited socioeconomic data on many genetic conditions.
Can you share some recent examples of how microsimulation modelling has been applied to precision medicine in specific conditions?
Treatment of paediatric cancer is a very good example of application of this methodology. There have been huge advances over the past 50 years or so in terms of survival rates for some cancer types. New treatment approaches such as precision medicine are still needed to further improve survival rates in some cancer types such as high risk childhood cancers, where survival rates are less than 30%.
There has been significant progress in this domain; due to the advances in genome sequencing technology and an expansion in the range of ‘omics’ approaches which facilitate the potential application of precision medicine.
However, cost is of course also a major consideration, as it is critical for evaluating the feasibility of implementing genome sequencing and other omics as part of the treatment strategy. Consequently, there is a need to develop a modelling approach to assess the cost-effectiveness of prospective precision medicine programs in many areas of medicine including childhood cancer.
With this remit in mind, our recent PhD graduate, Dr Owen Tan, recently developed a microsimulation model, Paediatric Cancer MOD (PeCanMOD), to simulate the costs and benefits of precision medicines [2], the first such model of this kind applied to precision medicine and paediatric cancer.
For the data inputs of the model, we used records from the New South Wales (NSW) Cancer Registry linked to mortality and hospital datasets, and simulated the genomic landscape of the cancer registry population. Cost of drugs, and hospital admissions were also included in the model. The model simulated the number of individuals eligible for precision medicine, and the incremental cost of treatment per life year gained if precision medicine was introduced for late-stage cancer patients as a final treatment option.
The model demonstrates the application of microsimulation modelling to simulate the impacts of NGS and precision medicine on costs and health outcomes for childhood cancer, and the potential of this approach to inform resource planning strategies.
We have taken a similar approach to mitochondrial diseases. These are serious conditions, caused by mutations either in the mitochondrial DNA or nuclear DNA, resulting in impaired production of cellular energy.
A key characteristic of this condition is its genetic heterogeneity which, as mentioned earlier in our discussion, requires modelling approaches that operate at the individual-level, for which microsimulation is ideally suited.
Our aim was to develop a microsimulation model – MitoMOD – to estimate the costs of mitochondrial diseases using a cohort of clinically diagnosed adult patients with mitochondrial diseases as the base population [3].
In the model, we considered a wide range of economic and social impacts of these conditions at the patient, family, health service, and government level. MitoMOD can be used to report the health and social costs of mitochondrial diseases and was used to estimate the cost-effectiveness of implementing whole genome sequencing (WGS) to diagnose mitochondrial patients in comparison to using current diagnostic tests, which supported recent listing of Medicare item numbers for public funding of WGS and/or WES for the diagnosis of mitochondrial diseases in Australia.
We have also built a microsimulation model for inherited retinal diseases (IRD) [4], which have significant cost implications, both for people with these disorders and for their families. This model enabled us to estimate the overall lifetime cost per person for individuals with an IRD, and the related societal costs. We found that societal costs accounted for 87% of all costs, and health care costs only 13%, which demonstrates that when assessing the cost-effectiveness of IRD treatments, including genomic testing and gene therapies, the substantial societal costs of IRDs must be taken into account.
How can these techniques be applied at a population-level?
The flexibility and granularity of the microsimulation-based approach means it can be used effectively to quantify intervention benefits to individuals or populations.
Microsimulation models can be used for conditions with many different underlying genetic causes, such that for some molecular diagnoses, the sample size may be very small or even a sample size of 1, because its granularity means it can be used to examine individual characteristics, behaviours, and interactions between them. This granularity can provide insights that are obscured in aggregate models. Microsimulation can reflect the unique attributes of a single entity (in this case a patient), which can be crucial for understanding specific rare diseases.
At the other end of the scale, microsimulation can be used to quantify the impact of interventions at a national level before real-world implementation. This versatility makes microsimulation a powerful tool for both industry and government applications.
One such example is our work on familial intellectual disability (ID) [5]. People with ID require a wide range of support from both government, and their family. Understanding the costs of ID is important for policy makers to decide on resources required to support families affected by the condition. We developed a microsimulation model – IDMOD – to provide a holistic perspective of the economic costs of ID so that cost-effectiveness studies could be undertaken relating to the application of genomic testing and precision medicine for familial ID.
Leveraging on the results of this study, the Australian government decided to roll out a AUD $20-million preconception intervention screening pilot.
How do you address the data requirements for genomic modelling?
A major development over the past few years has been the increasing availability of good quality genomic data, and we expect this to continue to develop rapidly.
However, it is still the case that for many conditions, especially those that are often misdiagnosed, primary data collection is necessary given the lack of other relevant data. This is often an exercise we undertake ourselves.
For example, for the MitoMOD model mentioned above [3], we used data from participants attending a tertiary referral centre dedicated to supporting adults clinically diagnosed with mitochondrial diseases who had been recruited into our Economic and Psychosocial Impacts of Caring for Families Affected by Mitochondrial Disease (EPIC-MITO) study, with many of the participants also having a molecular diagnosis as a result of genomic testing. Each study participant, which included carers of patients with mitochondrial diseases, was asked to complete a questionnaire covering a wide range of health and socio-economic issues. The data inputs for the MitoMOD model thus consisted of the survey responses obtained from the participants of this study.
What does the future hold and what are the main challenges?
It’s a very exciting time to be in the field, as microsimulation, with its ability to simulate individual behaviours and interactions, can play a pivotal role in the evolving landscape of genomics and precision medicine.
Much of our focus is on cost-effectiveness analysis, assessing the economic implications of genomic-based treatments to evaluate the impact of precision medicine interventions on population health.
However, other key application areas include pharmacogenomics, i.e. understanding the role of genomic testing in drug response; where a microsimulation-based approach can model individual responses to treatments based on the genetic profile of patients, which has huge value for the pharmaceutical and biotech companies developing therapies.
The main challenges continue to be data quality and availability for genetic disorders. However, we are confident that with the rapid pace of development of genomic technologies, access to high-quality, comprehensive genomic data will continue to improve.
HealthLumen extends our sincere thanks to Prof. Schofield and Associate Prof. Shrestha for their insights and for participating in this interview.
References:
- A brief, global history of microsimulation models in health: Past applications, lessons learned and future directions
- Modelling the Economic Impact of Next Generation Sequencing and Precision Medicine on Childhood Cancer Management—a Microsimulation Approach
- The Development of a Microsimulation Model (MitoMOD) to Estimate the Economic Impact of Mitochondrial Disease in Adults
- The health care and societal costs of inherited retinal diseases in Australia: a microsimulation modelling study
- IDMOD: An Australian microsimulation model of lifetime economic and social factors in familial intellectual disability