Care&WorkMOD: An Australian microsimulation model projecting the economic impacts of early retirement in informal carers

Informal carers provide care to others in need of support, and make a significant contribution to a nation’s productivity by replacing paid care of people with disability, mental illnesses, chronic conditions, and the frail aged (Nguyen & Connelly, 2014). Informal carers are important as they provide a service to the government that would otherwise need to be paid for by employed carers (Diminic et al., 2017). Similarly, informal carers allow for ‘ageing in place’, with the aged population remaining in their homes instead of moving to aged care facilities (Commonwealth Department of Health and Ageing, 2014). Government spending on ‘aged care’, which is separate to informal carers, was $16.2 billion in 2016 (Aged Care Financing Authority, 2017). The majority of informal carers are female (Nepal, Brown, Ranmuthugala, & Percival, 2011; Stacey, Gill, Price, Warmington, & Taylor, 2016), aged 25–64, and the burden of informal care often leads to reduced labour force participation (Access Economics, 2005; Deloitte Access Economics, 2015). Informal carers were estimated to be providing 1.9 billion hours of care in 2015, with each carer providing an average of 13 hours per week. The replacement cost of informal care in 2015 in Australia was estimated to be A$60.3 billion, equivalent to 3.8% of the gross domestic product and 60% of the health and social work industry (Deloitte Access Economics, 2015).

The Australian Government is seeking ways to enable workers to remain in the workforce longer, in order to fund health and other essential services needed by the ageing populations, and manage labour shortages. At the same time, the need for informal carers is increasingly taking on greater relevance as the burden of disease transitions from acute fatal diseases to long-term chronic diseases (Stacey et al., 2016). Indeed, it has been argued that if this care was not provided informally to people with mental illness, the Australian National and state/territory governments would need to spend A$13.2 billion to provide the same level of health and social services (Diminic et al., 2017). Many studies have shown that informal carers that have reduced work hours are more likely to leave the labour force and have lower incomes compared to non-carers (Bittman, Hill, & Thomson, 2007; Gray & Edwards, 2009; Adams, 2010). Therefore, the ability to project future numbers of informal carers and their labour force participation is important for national governments to provide long-term care and manage health care costs.

Microsimulation models are increasingly being used to estimate various aspects of health economics, including health expenditure, the cost-effectiveness of health interventions and programs, and the impact of various health conditions on the labour force (Brown & Harding, 2002; Dieleman et al., 2016; Doble, Schofield, Roscioli, & Mattick, 2016; Schofield, Shrestha & Cunich, 2016). A number of microsimulation models have been developed to estimate the demand for care (Percival & Kelly, 2004; Chin & Harding, 2006; Pickard, Wittenberg, Comas-Herrara, Kink, & Mallay, 2007; Lymer, Brown, Harding, & Yap, 2009; Nepal, Brown, Kelly, Percival, Anderson, & Hancock, 2011). An Australian study, based on the 1998 Australian Bureau of Statistics (ABS) Survey of Disability, Ageing and Carers (SDAC) and the 2001 Census data, produced small area estimates of disability levels and the need for aged care across age groups, in order to allow the execution of social policy and elderly care services for older Australians (Chin & Harding, 2006; Lymer et al., 2009). These studies reported that as age increases both the presence and severity of disability increase, indicating a higher requirement for care, and particular geographic areas had higher needs for provision of care (Lymer et al., 2009). Recent microsimulation studies have been used to quantify how much policies cost, rates of growth of costs, the distribution of expenditure, regional differences in health, and whether the health labour force will be sufficient to match the future needs for health care (Schofield, Shresta, Kelly et al., 2014). Microsimulation models are useful at a national level, by modelling the records of individuals from national surveys, with the ability to examine the effects of policy changes for well-defined ranges of individuals or demographic groups, as well as for national aggregates.

A microsimulation study projected the demand for informal carers of disabled people to 2031 in England (Pickard et al., 2007). The projections for informal carers to 2031, as well as other forecasts for the next ten years, show that the demand for informal carers is growing at a faster rate than the supply, meaning the ratio of carers “supply to demand” is likely to substantially decrease in future decades (Deloitte Access Economics, 2015). Reasons for this include couples having fewer children, increasing life expectancy, increasing geographic distance between parents and children, and a higher retirement age (Pickard et al., 2007), with trends reported to be similar in Australia (The Treasury, 2015a, 2015b). In Australia, the Australian Government’s commitment to caring for the elderly at home (‘ageing at home’), rather than in care (Commonwealth Department of Health and Ageing, 2014), is likely to increase the proportion of the population who are informal carers. Financial support through the National Disability Insurance Scheme (NDIS) may also influence the proportion of informal carers, however, it is too early to tell impacts of the NDIS on the number of carers. Similarly, microsimulation has shown that the substantial increase in years lived with disability (YLDs) over the past 20 years has also increased the demand for informal carers (Schofield, 2016). Although there have been several microsimulation models projecting the demand for carers, there have been none estimating the long-term economic impact of the provision of informal care on both carers themselves, as well as on the government.

In this paper we describe the methods, data sources and assumptions used in constructing a new microsimulation model, Care&WorkMOD, which has been developed to estimate the long-term economic costs of lost labour force participation of informal carers due to their caring responsibilities. Care&WorkMOD can model the economic costs and estimate economic impacts such as lost Productivity Life Years (PLYs) due to informal caring, lost personal incomes, lost taxation revenue, increase in welfare payments, lost gross domestic product (GDP) and economic hardship. The model is able to capture the long-term projections of

• demographic changes;

• changes labour force participation rates (for all genders, and across age groups);

• real wage growth;

• real growth in other economic variables such as the value of assets and savings;

• number of informal carers.

Care&WorkMOD is a static microsimulation model designed to project socio-demographic and economic profiles of Australian informal carers aged 15–64 years every five years from 2015 to 2030. It uses static ageing methods to project population profiles. The model was written and run in SAS V9.4 software (SAS Institute, Cary, NC, USA). The output datasets of Care&WorkMOD include four datasets containing similar information on Australians aged 15–64 years, at different points in time (2015, 2020, 2025, and 2030). Each dataset contains individuals’ socio-demographic and economic information such as their carer status, labour force status, reason for not being in the labour force, earned income, government support payments, and wealth. The datasets also include information on their care recipients such as their age, gender, chronic health conditions, if they are primary carers, and if their main care recipient is living with them in the same household.

Care&WorkMOD has four main components: i) the base population — the microdata on which the model is built; ii) population and labour force projections from the 2015 Australian Intergenerational Report; iii) output datasets of the Australian Population and Policy Simulation Model (APPSIM); and iv) output datasets of Static Incomes Model (STINMOD). We assumed that these four component datasets were representative of the Australian population and can be linked. Figure 1 shows the schematic diagram of Care&WorkMOD and how each component and its linkage.

Figure 1

Download asset Open asset

2.2 Population and labour force projections

To ensure that the model accounted for the growth in population and the trends in labour force participation, we statically aged the data using the population and labour force projections from the 2015 Australian Intergenerational Report (IGR) (The Treasury, 2015a), with the reweighting described in Section 2.6. We extracted population projections and projected full-time and part-time employment rates to 2030 for Australians aged 15 to 64 years by five-year age group and gender.

Care&WorkMOD covers only the Australians living in households and excludes those living in cared accommodation and correctional institutions whereas the 2015 IGR population projections are for the whole nation. For that reason, we estimated the proportions of individuals (by age and sex) living in households (excluding cared-accommodation and correctional institutions) both private and non-private dwellings using the Australian Census of Population and Housing data for 2011 (Australian Bureau of Statistics, 2011). These estimated proportions from the 2011 Australian Census for Population and Housing were then applied to the 2015 IGR population projections to estimate the projected number of individuals living in households and were used as benchmarks to align the projected population coverage of Care&WorkMOD. Model input data disaggregated by age and sex is shown in Figure 2.

Figure 2

Download asset Open asset

2.4 Static incomes model (STINMOD)

STINMOD is a static microsimulation model of Australia’s income tax and transfer (welfare) system (Percival, Abello, & Vu, 2007). It plays an important role in conducting assessment of the distributional impacts of government’s tax and transfer policies, independent from the Commonwealth of Australia. The model was developed and is now regularly updated by NATSEM and was previously used by other Commonwealth departments. The model simulates the impact of Australian Government’s major reforms on cash transfers, income tax and the Medicare levy on individuals and families and identifies the winners and losers of the reforms. STINMOD is based on the ABS Surveys of Income and Housing data and covers Australians living in households excluding those living in cared accommodation, similar to the Care&WorkMOD base population. It is updated each year by incorporating recent changes to the Australian Government’s tax and transfer payment system announced in each year’s federal budget.

We extracted a STINMOD output dataset for 2015, which includes the relevant Australian Government social policies, rates of payment and eligibility criteria for that year. STINMOD uses income unit level unit record data from the ABS Survey of Income and Housing as a base data file and simulates changes in income, tax and welfare payments at income unit level. We expanded the unit record data at income unit level of STINMOD output dataset to person level unit record data to use in the Care&WorkMOD. The STINMOD output dataset has socio-demographic and economic information such as income, income tax paid, and transfer payments received assigned to individual persons. Savings and assets information are at the income unit level. We shared savings and assets of the income unit equally among the two oldest people in a couple family, and to the oldest person in a single parent family, and assigned them to person record data.

Figure 3

Download asset Open asset

2.7.1 Synthetic matching results

About 95.5% of concatenated SDAC unit records were synthetically matched with STINMOD unit records of the same category for each of the nine matching variables (Figure 4). Records were matched in the same categories for all records for at least five matching variables. An analysis of how closely SDAC unit records matched with STINMOD unit records for each matching variable showed that all variables were matched within 3% accuracy (Table 2). Out of nine matching variables, age group had the highest rate of matching error at 2.9%; however, most of these errors were due to records in one age group matching with the records from one step above or one step below age groups (e.g. records from 45–49 years age group matching with the records from either 41–44 or 50–54 years age group).

Figure 4

Download asset Open asset

Table 2

Matching variables	No of SDAC records matched with STINMOD record of different category	%
Labour force status	0	0.0%
Income unit type	0	0.0%
Income quintiles	6	0.0%
Sex	135	0.1%
Age group	3217	2.9%
Hours worked per week	1007	0.9%
Highest level of education	344	0.3%
Home ownership	308	0.3%
Receiving carer payment	395	0.4%

We have developed a new model, Care&WorkMOD, to project the economic impacts of informal care of people with chronic disease, leading to people aged 15 to 64 years leaving the labour force early, over the years 2015 to 2030. This model simulates the impacts of an ageing population, different labour force participation patterns, and income and savings trends. Care&WorkMOD can be used to calculate (i) the number of people who left the labour force due to caring for people with chronic disease; (ii) the lost income and savings of people who have left the labour force due to caring for those with ill health, compared to those who continue in paid employment; and (iii) the extent of these losses due to demographic, disease, or economic trends across 2015, 2020, 2025 and 2030. Care&WorkMOD can also be used to estimate possible long-term economic productivity gains from informal carers returning to the labour force as a result of interventions that address individual health conditions, including mental health disorders and heart disease.

The application of microsimulation to social and economic modelling was pioneered in the late 1950’s (Orcutt, 1957). Since then, microsimulation modelling has evolved to use current computing power and available datasets, allowing sophisticated dynamic microsimulation modelling. The use of a dynamic microsimulation model for long-term projections is now wide spread and well- established (Li & O’Donoghue, 2013). In a dynamic microsimulation model, individuals transit from one stage to another in regular unit of time, commonly every year, based on transition probabilities. With this, the model can simulate a life trajectory of an individual, such as informal carers’ entry to labour market, exit from the labour market, re-entry into the market and permanently exit from the labour market. This allows the simulation of working years of every individual in the model. However, building a dynamic microsimulation model is resource intensive, time consuming and can be costly (O’Donoghue, Li, & Dekkers 2014). In this instance, we chose the less complex and labour intensive model structure. We therefore built a static microsimulation model, Care&WorkMOD, which uses static ageing through reweighting data and synthetic matching methods instead of building it as a dynamic microsimulation model. Static ageing is defined here as a model that simulates time indirectly through uprating or ‘reweighting’ (Dekkers, Keegan, & O’Donoghue, 2014). In contrast, dynamic modelling alters the content of the dataset itself over time. We also have utilized the output datasets of two well-established Australian microsimulation models APPSIM and STINMOD in the process. The main advantage of building a microsimulation model this way is that it is faster to assemble, less computationally intensive and can take advantage of existing microsimulation models. However, as Care&WorkMOD depends upon outputs of APPSIM and STINMOD, its robustness depends upon the strength of APPSIM and STINMOD simulations and any underlying limitations in APPSIM and STINMOD will be carried over to Care&WorkMOD. For example, APPSIM does not have geographical subdivisions information and as a result, Care&WorkMOD has no geographical division information in its datasets. Thus, it limits the use of Care&WorkMOD to estimate the economic costs of informal carers’ early exit from the labour force only for individuals (i.e. informal carers) and the Commonwealth Government at a national level. The model does not provide the estimates at a state or regional level, which are outside the scope of Care&WorkMOD project. Since both APPSIM and STINMOD are well-established Australian models and are widely used (Lambert, Percival, Schofield, & Paul, 1994; Percival et al., 2007; Keegan & Kelly, 2009; Lymer, 2009), this gives confidence in the robustness of their output.

Care&WorkMOD created synthetic datasets by matching individual records of SDACs with records from a STINMOD output dataset based on matching variables through a synthetic matching process as opposed to the exact matching of individual records through record linkage. It is not possible to directly match SDAC records with STINMOD records because they use different samples, so the individuals in each survey are not the same people. Most importantly, the ABS does not allow an exact matching based on nine matching variables to create a synthetic population dataset, which is nationally representative sample of the Australian population. We therefore linked records from SDAC and STINMOD using synthetic matching based on nine matching variables to create a synthetic dataset which includes information from both original datasets. This method is commonly used in microsimulation models to create synthetic datasets (Hynes, Morrissey, & O’Donoghue, 2006; Von Randow, Davis, Lay Yee, & Pearson, 2012) and we have previously used this method in our earlier microsimulation models Health&WealthMOD and Health&WealthMOD20330 (Schofield et al., 2011; Schofield, Shrestha, Kelly et al., 2014).

One of the main limitations of synthetic matching is that records of one dataset can be matched with the records of different categories of matching variables in another dataset creating a mismatch in the categories of the same matching variables. The potential of this mismatch was limited in Care&WorkMOD by matching the closest records between two datasets using a standardized Euclidean distance method based on the nine matching variables. Care&WorkMOD has all SDAC records matched with STINMOD records of the same categories for up to five matching variables, with about 95% records matched for each of the nine matching variables. When looking at individual matching variables, each of the variables had a matching error rate of less than 1% except for age group, which had a matching error rate of 2.9%.

As the synthetic matching was based on the closest matching records using an unconstrained matching process, it was possible that multiple SDAC records have been synthetically linked with one STINMOD record or some of STINMOD records were not matched with any SDAC records. This means that there may be inherent differences in the distributions of economic variables in STINMOD and Care&WorkMOD, and as a result there is the potential for the distributions of economic variables not used as synthetic matching variables in Care&WorkMOD to be different from the original distributions in STINMOD. We note that one limitation of this study is that female carers aged 35–64 years were a higher proportion in actual SDAC 2015 data than projected (from the 2012 data). This indicates that these estimates are conservative and the number of informal carers may be higher.

The analysis of Care&WorkMOD output datasets will provide the results that will help fill gaps in current evidence on the costs of chronic diseases (e.g. pain, neurodegenerative diseases, mental illness, heart disease, and diabetes, and childhood onset cognitive and physical disability) and to estimate reductions in labour productivity of carers due to informal caring responsibilities, at a national and individual level. Findings from this microsimulation model may be used internationally, where other countries have the appropriate data available. Given that the population is ageing and the demand for informal caring is increasing in many countries, estimates of the economic costs of informal caring may be used to inform health polices across the globe. Economic impacts of informal carers leaving the workforce early for both individuals in terms of lost incomes and the National Governments in terms of additional welfare costs and lost tax revenues are likely to grow in future given most countries have a growing population of informal carers. Indeed, around 13% of people aged 50 and over report providing informal care at least weekly, and more than 20% in the Czech Republic and Belgium (Basu, Seligman & Bhattacharya 2013). It is noteworthy that the lowest rates of daily informal provision are found in Sweden, Switzerland, Denmark and the Netherlands, where the formal long-term care network is well developed and government provided coverage is thorough and comprehensive (Basu, Seligman, & Bhattacharya, 2013). Results from microsimulation on lost income and welfare payments can be used to inform government policymakers about the likely impacts on national productivity, and ways to increase workforce participation and standard of living of informal carers of chronic disease.

Here we describe the model structure, datasets used, and assumptions of a microsimulation model called Care&WorkMOD. The results from Care&WorkMOD, such as costs of welfare and lost taxes due to informal carers being out of the workforce, will enable policy makers to use accurate and reliable estimates of long-term costs of informal caregiving, leading to exit from the labour market (from both individual and government perspectives).

1. 1
   The economic value of informal care. Report to Carers Australia

   1. Access Economics

   (2005)

   The economic value of informal care. Report to Carers Australia.

   • Google Scholar
2. 2
   Scoping the Australian Care Economy: A Gender Equity Perspective Security

   1. V Adams

   (2010)

   Scoping the Australian Care Economy: A Gender Equity Perspective Security, 4Women Incorporated.

   • Google Scholar
3. 3
   Fifth report of the funding and financing of the aged care sector

   1. Aged Care Financing Authority

   (2017)

   Fifth report of the funding and financing of the aged care sector, Canberra.

   • Google Scholar
4. 4
   Census of Population and Housing. Canberra

   1. Australian Bureau of Statistics

   (2011)

   Census of Population and Housing. Canberra.

   • Google Scholar
5. 5
   Disability, Ageing and Carers, Australia: Summary of Findings, 2015. Canberra

   1. Australian Bureau of Statistics

   (2016)

   Disability, Ageing and Carers, Australia: Summary of Findings, 2015. Canberra.

   • Google Scholar
6. 6
   Nutritional policy changes in the supplemental nutrition assistance program: A microsimulation and cost-effectiveness analysis

   1. S Basu
   2. H Seligman
   3. J Bhattacharya

   (2013)

   Medical Decision Making 33:937–948.

   • Google Scholar
7. 7
   GREGWT and TABLE macros - users guide

   1. P Bell

   (2000)

   Canberra: Australian Bureau of Statistics.

   • Google Scholar
8. 8
   The impact of caring on informal carers’ employment, income and earnings: A longitudinal approach

   1. M Bittman
   2. T Hill
   3. C Thomson

   (2007)

   Australian Journal of Social Issues 2:255–272.

   • Google Scholar
9. 9
   “Its Not Just About the Dollars: Measuring and Reporting the Economic and Social Benefits of Standards under the Disability Discrimination Act” - a Forum held by the Disability Studies and Research Institute (DSaRI) in conjunction with the Social Relations of Disability Research Network (SRDRN)

   1. L Brown
   2. A Harding

   (2002)

   University of Technology, Sydney.

   • Google Scholar
10. 10
    Regional Dimensions: Creating Synthetic Small-area Micro Data and Spatial Microsimulation Models. (NATSEM Technical Paper No. 33)

    1. SF Chin
    2. A Harding

    (2006)

    NATSEM. Canberra.

    • Google Scholar
11. 11
    Ageing in place: Quality care for older Australians

    1. Commonwealth Department of Health and Ageing

    (2014)
12. 12
    New pathways in microsimulation

    1. G Dekkers
    2. M Keegan
    3. C O’Donoghue

    (2014)

    • Google Scholar
13. 13
    The economic value of informal care in Australia in 2015. Report to Carers Australia

    1. Deloitte Access Economics

    (2015)

    The economic value of informal care in Australia in 2015. Report to Carers Australia, Canberra.

    • Google Scholar
14. 14
    National spending on health by source for 184 countries between 2013 and 2040

    1. JL Dieleman
    2. T Templin
    3. N Sadat
    4. P Reidy
    5. A Chapin
    6. K Foreman
    7. C Kurowski

    (2016)

    The Lancet 387:2521–2535.

    • Google Scholar
15. 15
    The economic value of informal mental health caring in Australia

    1. S Diminic
    2. E Hielscher
    3. YY Lee
    4. M Harris
    5. J Schess
    6. J Kealton
    7. H Whiteford

    (2017)

    Mind Australia and University of Queensland. Queensland.

    • Google Scholar
16. 16
    The promise of personalised medicine

    1. B Doble
    2. DJ Schofield
    3. T Roscioli
    4. JS Mattick

    (2016)

    The Lancet 387:433–434.

    • Google Scholar
17. 17
    Determinants of the Labour Force Status of Female Carers

    1. M Gray
    2. B Edwards

    (2009)

    Australian Journal of Labour Economics 12:5–20.

    • Google Scholar
18. 18
    Building a static farm level spatial microsimulation model: Statistically matching the Irish National Farm Survey to the Irish Census of Agriculture

    1. S Hynes
    2. K Morrissey
    3. C O’Donoghue

    (2006)

    46th Congress of the European Regional Science Association.

    • Google Scholar
19. 19
    APPSIM - Dynamic microsimulation modelling of social security and taxation

    1. M Keegan
    2. S Kelly

    (2009)

    Canberra: NATSEM.

    • Google Scholar
20. 20
    An Introduction to STINMOD: A Static Microsimulation Model

    1. S Lambert
    2. R Percival
    3. D Schofield
    4. S Paul

    (1994)

    Canberra: NATSEM.

    • Google Scholar
21. 21
    A survey of dynamic microsimulation models: uses, model structure and methodology

    1. J Li
    2. C O’Donoghue

    (2013)

    International Journal of Microsimulation 6:3–55.

    • Google Scholar
22. 22
    APPSIM: Modelling health. National Centre for Social and Economic Modelling (NATSEM)

    1. S Lymer

    (2009)

    University of Canberra. Canberra.

    • Google Scholar
23. 23
    Predicting the need for aged care services at the small area level: The CAREMOD spatial microsimulation model

    1. S Lymer
    2. L Brown
    3. A Harding
    4. M Yap

    (2009)

    International Journal of Microsimulation 2:27–42.

    • Google Scholar
24. 24
    Projecting the need for Formal and Informal Aged Care in Australia: A Dynamic Microsimulation Approach

    1. B Nepal
    2. L Brown
    3. S Kelly
    4. R Percival
    5. P Anderson
    6. R Hancock

    (2011)

    Canberra: NATSEM.

    • Google Scholar
25. 25
    A comparison of the lifetime economic prospects of women informal carers and non-carers, Australia, 2007

    1. B Nepal
    2. L Brown
    3. G Ranmuthugala
    4. R Percival

    (2011)

    Australian Journal of Social Issues 46:91–108.

    • Google Scholar
26. 26
    The effect of unpaid caregiving intensity on labour force participation: Results from a multinomial endogenous treatment model

    1. HT Nguyen
    2. LB Connelly

    (2014)

    Social Science and Medicine 100:115–122.

    • Google Scholar
27. 27
    Dynamic Models

    1. C O’Donoghue
    2. J Li
    3. G Dekkers

    (2014)

    In: C O’Donoghue, editors. Handbook of Microsimulation Modelling. Emerald Publishing Ltd.. pp. 305–343.

    • Google Scholar
28. 28
    A new type of socio-economic system

    1. G Orcutt

    (1957)

    Review of Economics and Statistics 39:116–123.

    • Google Scholar
29. 29
    Modelling Our Future: Population ageing, health and aged care

    1. R Percival
    2. A Abello
    3. Q Vu

    (2007)

    STINMOD (Static Income Model), Modelling Our Future: Population ageing, health and aged care, Amsterdam, Elsevier B.V.

    • Google Scholar
30. 30
    Who’s going to care? Informal care and an ageing population

    1. R Percival
    2. S Kelly

    (2004)

    Canberra: NATSEM.

    • Google Scholar
31. 31
    Care by Spouses, Care by Children: Projections of Informal Care for Older People in England to 2031

    1. L Pickard
    2. R Wittenberg
    3. A Comas-Herrara
    4. D King
    5. J Malley

    (2007)

    Social Policy and Society 6:353–366.

    • Google Scholar
32. 32
    Statistical matching: A frequentist theory, practical applications, and alternative Bayesian approaches

    1. S Rässler

    (2002)

    New York: Springer-Verlag.

    • Google Scholar
33. 33
    RBA Statistical table E1 - Household and Business Balance Sheets

    1. Reserve Bank of Australia

    (2017)

    RBA Statistical table E1 - Household and Business Balance Sheets, Sydney.

    • Google Scholar
34. 34
    The economic impact of diabetes through lost labour force participation on individuals and government: evidence from a microsimulation model

    1. D Schofield
    2. M Cunich
    3. R Shrestha
    4. M Passey
    5. L Veerman
    6. E Callander
    7. R Tanton

    (2014)

    BMC Public Health, 14, 220.

    • Google Scholar
35. 35
    Modelling the cost of ill health in Health&WealthMOD (Version II): lost labour force participation, income and taxation, the impact of disease prevention

    1. D Schofield
    2. R Shrestha
    3. E Callander
    4. R Percival
    5. S Kelly
    6. M Passey
    7. S Fletcher

    (2011)

    The International Journal of Microsimulation 4:32–36.

    • Google Scholar
36. 36
    Counting the Cost: Part 2 Economic Costs: the current and future burden of arthritis

    1. D Schofield
    2. R Shrestha
    3. M Cunich

    (2016)

    Report to Arthritis Australia.

    • Google Scholar
37. 37
    Health&WealthMOD20330: A microsimulation model of the long term economic impacts of disease leading to premature retirements of Australians aged 45–64 years old

    1. D Schofield
    2. R Shrestha
    3. S Kelly
    4. L Veerman
    5. R Tanton
    6. M Passey
    7. E Callander

    (2014)

    International Journal of Microsimulation 7:94–118.

    • Google Scholar
38. 38
    Economic costs of chronic disease through lost productive life years (PLYs) among Australians aged 45–64 years from 2015 to 2030: results from a microsimulation model

    1. D Schofield
    2. RN Shrestha
    3. MM Cunich
    4. R Tanton
    5. L Veerman
    6. SJ Kelly
    7. ME Passey

    (2016)

    BMJ Open 6:e011151.

    • Google Scholar
39. 39
    Understanding calibration estimators in survey sampling

    1. AC Singh
    2. CA Mohl

    (1996)

    Survey Methodology 22:107–115.

    • Google Scholar
40. 40
    Unpaid informal caregivers in South Australia: Population characteristics, prevalence and age-period-cohort effects 1994–2014

    1. AF Stacey
    2. TK Gill
    3. K Price
    4. R Warmington
    5. AW Taylor

    (2016)

    PLoS ONE, 11, 9.

    • Google Scholar
41. 41
    Small area estimation using a reweighting algorithm

    1. R Tanton
    2. Y Vidyattama
    3. B Nepal
    4. J McNamara

    (2011)

    Journal of the Royal Statistical Society: Series A (Statistics in Society) 174:931–951.

    • Google Scholar
42. 42
    Intergenerational Report: Australia in 2055

    1. The Treasury (2015a)

    (2015)

    Intergenerational Report: Australia in 2055, Canberra.

    • Google Scholar
43. 43
    Budget Paper No 1: Budget Strategy and Outlook 2015–16

    1. The Treasury (2015b)

    Budget Paper No 1: Budget Strategy and Outlook 2015–16, Canberra.

    • Google Scholar
44. 44
    Data matching to allocate doctors to patients in a microsimulation model of the primary care process in New Zealand

    1. M Von Randow
    2. P Davis
    3. R Lay Yee
    4. J Pearson

    (2012)

    Social Science Computer Review 30:358–368.

    • Google Scholar

Author details

1. Rupendra N Shrestha

   GenIMPACT: Centre for Economic Impacts of Genomic Medicine, Macquarie University, Australia

   For correspondence

   rupendra.shrestha@mq.edu.au

2. Deborah Schofield

   GenIMPACT: Centre for Economic Impacts of Genomic Medicine, Macquarie University, Australia

   For correspondence

   deborah.schofield@mq.edu.au

3. Melanie J B Zeppel

   GenIMPACT: Centre for Economic Impacts of Genomic Medicine, Macquarie University, Australia

   For correspondence

   melanie.zeppel@mq.edu.au

4. Michelle M Cunich

   The Boden Institute of Obesity, Nutrition, Exercise and Eating Disorders, Faculty of Medicine and Health and Sydney Health Economics, Sydney Local Health District, The University of Sydney, Australia

   For correspondence

   michelle.cunich@sydney.edu.au

5. Robert Tanton

   National Centre for Social and Economic Modelling, University of Canberra, Australia

   For correspondence

   robert.tanton@canberra.edu.au

6. Simon J Kelly

   National Centre for Social and Economic Modelling, University of Canberra, Australia

   For correspondence

   simon.kelly@canberra.edu.au

7. Lennert Veerman

   Cancer Council NSW, Woolloomooloo, Australia and School of Medicine, Griffith University, Australia

   For correspondence

   l.veerman@griffith.edu.au

8. Megan E Passey

   University Centre for Rural Health, The University of Sydney, Australia

   For correspondence

   megan.passey@sydney.edu.au