We all know that a lot of resources are funnelled into the development of health infrastructure all around Australia at certain times of the year… But how do governments and organisations decide what the Australian healthcare sector needs? This is where health forecasting comes in!
On the surface, health forecasting is a fairly simple concept: try to accurately predict what Australia will need in the short-term and long-term future of healthcare, and try to pre-emptively build infrastructure to support those needs. Sounds relatively straightforward.
However, the actual act of forecasting is incredibly complex! Whole teams and organisations full of specialised professionals work together to try to give Australia the best possible chance of predicting future healthcare trends. It really does take a village.
With this complexity in mind, we’ve broken down the whole show into just a few simple points: what health forecasting is, who does it, how it’s done and why they do it.
What is health forecasting?
Health forecasting is the act of analysing data – from various sources, such as surveys, public statistics and privately owned organisation reports – in order to predict trends in the healthcare industry.
The results of health forecasting are usually presented in an official report with key takeaways, suggestions and points of interest for policymakers. When it comes to healthcare, a lot of these reports are publicly available, as opposed to industries that are more commonly privately owned and run such as mining or agriculture.
Why do organisations perform health forecasting?
Health forecasts are largely used by policymakers and healthcare executives, as these forecasts help them make informed decisions about what they should invest money, resources and people into. For example, Deloitte’s 2022 Global Health Care Outlook report predicts that by 2030 ‘mental illness costs are expected to reach more than $6 trillion annually’ (Allen et al, 2022). With this in mind, a hospital executive may table a plan to build a specific wing dedicated to mental health support and counselling within the next 10 years, as there is a forecasted need for it.
Health forecasts also help other industries stay stable in the face of growing trends while leaving enough room for flexibility and adaptability. For example, if a state government knew that it needed to build a new major public hospital within the next four years to deal with a predicted growth in a major health issue, then money and resources will have to be put aside for that project. This may impact other sectors, such as the commercial construction industry which would presumably boom – and also become preoccupied – as a result of such a huge project.
How is health forecasting actually done?
There isn’t a single agreed-upon approach to forecasting. According to global health experts Dr Ireneous Soyiri and Prof. Daniel Reidpath (Soyiri et al, 2013), there are four main principles of healthcare forecasting:
The measure of uncertainty and errors
The nature of data aggregation and how it affects accuracy (in other words, whether saying something like ‘most people with broken ribs develop pneumonia’ instead of ‘x% of people with broken ribs develop pneumonia’ changes the accuracy of the study)
The horizon of the forecast (in other words, the length of time the prediction is intended to cover)
When it comes to actually using data, information and analytics to produce an accurate forecast, the element of healthcare that you’re focusing on will determine which approach you use. For example, predicting trends around a specific illness or condition, such as diabetic retinopathy, would be completely different to predicting trends around ED admissions or ICU deaths per capita. In this example, one is specific and one is aggregate: not only would the range of data be incredibly disparate, but the type of data required to produce a trend forecast would be very different.
Do forecasters ever get it wrong?
Like everyone, sometimes forecasts turn out to be incorrect. Unfortunately, forecasters can’t see the future: if they could, the COVID-19 pandemic would have been a lot easier to work with.
Aside from incorrect predictions, forecasting organisations can also occasionally miss the mark when it comes to demographic inclusivity. In 2019, an algorithm built using historical data and machine learning reportedly produced healthcare predictions that favoured white people above Black people in the US (Jee, 2019). The algorithm used over 50,000 records from a hospital in order to predict how much certain groups would likely cost the US healthcare system, but failed to acknowledge – and subsequently adjust to the fact – that black communities generally incur lower healthcare costs due to lower socio-economic living conditions compared to white communities (Jee, 2019).
While this issue was reported to the owners of the software, Optum, and the issue has been worked through and rectified, there are hundreds of different software companies around the world that may be succumbing to the same diversity-blind pitfalls in their own forecasting algorithms (Jee, 2019).
What developers should be asking is: what effect can incidental biases have upon forecasting, policy, and infrastructure development? And even, eventually, human lives?
Who does health forecasting for Australia?
Perhaps unsurprisingly, the organisations that do most of the health forecasting in Australia are the ‘big four' consulting firms; Deloitte, Ernst & Young, PricewaterhouseCoopers and KPMG. While each of these organisations brings a slightly different specialty to their forecasts and outlooks, they mostly report the same four key predictions from 2022 onwards:
The immeasurable pressure placed on the healthcare workforce during the COVID-19 pandemic will force fundamental changes to the way workforce planning occurs (Allen et al, 2022).
In-home care, such as telehealth, will ideally remove a lot of excess presentations from non-urgent hospital care (Schlesinger et al, 2021).
The cost of mental health in the healthcare industry, as well as the Australian government, is going to increase exponentially in the next few years (Allen et al, 2022).
Personalisation of healthcare is going to become a top priority for both consumers and providers, meaning equitable, affordable and accessible healthcare will be closer to the new normal (EY India, 2022).
As you can see, all four of these are quite long-term, large-scale predictions that relate to huge demographics of Australians, such as health workers and older adults.
When it comes to predictions about more specific areas of health, such as specific health conditions or demographics within the greater population, there are generally far fewer organisations focusing on those areas and therefore fewer contesting predictions.
Where do healthcare practitioners come in?
It’s absolutely integral that healthcare practitioners are active participants in the forecasting process: healthcare is more than just numbers and data points. Providing experiences and opinions is one of the most valuable things you can do to help guide Australia’s healthcare industry into the future.
Healthcare practitioners are able to make their voice heard throughout the whole process: your experiences working in the healthcare sector are used as qualitative data points for these forecasts. It’s not difficult to have something to offer to forecasters – every single person that works in healthcare has a trove of important and valid stories and experiences from working in health. The hard part is finding a way to provide this information to the people who need it.
Keep an eye on surveys sent out by health-focused organisations, such as the Australian Commission on Safety and Quality in Health Care, and provide your opinions and experiences accordingly. Additionally, look out for more specialised organisations that may pertain to more specific aspects of your experience, such as LGBTIQ+ Health Australia. If you can’t find a survey to submit to, you can also send your opinions to the head of your union.
Where can you learn more?
A great, in-depth journal publication about different health forecasting techniques and options is ‘An overview of health forecasting’ by Dr Ireneous Soyiri and Dr Daniel Reidpath, which was published in Environmental Health and Preventive Medicine in 2012.
If you’re interested in following Australian health forecasts and predictions, keep an eye on media outlets and publishing forums such as the MIT Technology Review, PwC Australia: Health Matters, and CSIRO: Research. All three are reputable sources of information and have shown interest in reporting on forecasting and health in the past.
Allen, S. & Baxby, L., 2022. ‘2022 Global health care sector outlook.’ Deloitte Australia: Life Sciences and Health Care reports. Accessed 30 May 2022 via https://www2.deloitte.com/au/en/pages/life-sciences-and-healthcare/articles/global-health-care-sector-outlook.html
Dempster, A.; Kuek, A.; Makay, T.; & Peters, A., 2020. ‘COVID-19: Opportunities in mental health through virtual care.’ KPMG: Insights. Accessed 30 May 2022 via https://home.kpmg/au/en/home/insights/2020/09/covid-19-mental-health-virtual-care.html
EY India, 2022. ‘The future of health lies in unlocking the power of data to deliver personalized experience and improved outcomes: EY report.’ EY India: Press Releases. Accessed 31 May 2022 via https://www.ey.com/en_in/news/2022/02/the-future-of-health-lies-in-unlocking-the-power-of-data-to-deliver-a-personalized-experience-and-improved-outcomes
Huang, Y.; Xu, C.; Ji, M; Xiang, W; & He, D., 2020. ‘Medical service demand forecasting using a hybrid model based on ARIMA and self-adaptive filtering method.’ BMC Medical Informatics and Decision Making. Vol. 20, no. 237. Accessed 31 May 2022 via https://bmcmedinformdecismak.biomedcentral.com/articles/10.1186/s12911-020-01256-1
Jee, C., 2019. ‘A biased medical algorithm favored white people for health-care programs.’ MIT Technology Review. Accessed 30 May 2022 via https://www.technologyreview.com/2019/10/25/132184/a-biased-medical-algorithm-favored-white-people-for-healthcare-programs/
Moore, K.D. & Coddington, D.C., 2019. ‘Why Healthcare Forecasting is a Combination of Art and Science.’ Healthcare Financial Management Association: Benchmarking and Forecasting. Accessed 30 May 2022 via https://www.hfma.org/topics/hfm/2019/april/why-healthcare-forecasting-is-a-combination-of-art-and-science.html
Schlesinger, N.; Fielding, B.; Wilhelm, E.; & Pringle, N., 2021. ‘Where next for healthcare.’ PWC: Health Matters. Accessed 30 May 2022 via https://www.pwc.com.au/health/health-matters/where-next-for-healthcare.html
Soyiri, I. & Reidpath, D.D., 2013. ‘An overview of health forecasting.’ Environmental Health and Preventive Medicine. Vol. 18, no. 1. Accessed 30 May 2022 via https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3541816/
Unit4 Communications, 2021. ‘Modern techniques for forecasting in healthcare.’ Unit4. Accessed 31 May 2022 via https://www.unit4.com/blog/modern-techniques-forecasting-healthcare
Wharam, J.F. & Weiner, J.P., 2012. ‘The Promise and Peril of Healthcare Forecasting.’ The American Journal of Managed Care. Vol. 18, no. 3. Accessed 31 May 2022 via https://www.ajmc.com/view/the-promise-and-peril-of-healthcare-forecasting