Social cost or cost of illness studies attempt to monetarize all the costs and harms arising from a single disease or condition, and are mainly used for public advocacy and to inform policies. Given the complexity of the task, a national team of experts was assembled from the National Drug Research Institute, the National Drug and Alcohol Research Centre, the National Centre for Education and Training on Addiction, the South Australian Centre for Economic Studies, and Curtin University, covering economics, health economics, criminology, employment, health, epidemiology, data linkage and methamphetamine interventions, with additional advice from stakeholders (e.g. police, environmental health) (4). The aim was to estimate the social and economic costs of methamphetamine for the financial year 2013/14.
There is an extensive body of Australian research on the cost of illness of due to alcohol, tobacco, and, to a lesser extent, some illicit drugs, that provided an excellent guide on the available data and the domains most likely to be affected. For 2004/05, Collins and Lapsley estimated the costs for alcohol use at $15.3 billion (5). The majority of these costs were tangible ($10.5 billion) with the only included intangible costs arising from loss of life, and pain and suffering from road accidents ($4.5 billion). However, a 2010 report focusing on the harm alcohol causes to others estimated $13.4 billion in tangible costs and $6.4 billion in intangible costs to these people, but emphasized that this total should not be added to the earlier report given the potential for substantial overlap in costs (6). The Collins and Lapsley report found the costs of tobacco use to be $31.5 billion in 2004/05, including the cost to those impacted by involuntary smoking (5). Their analysis combined the costs for illicit drugs, with an overall total of $8.2 billion.
A 2007 analysis by Moore (7) separated the costs for the major illicit drugs (cannabis, cocaine, opiates and amphetamine) and together with a major report on methamphetamine from the USA (8) provided a starting point for our analysis. It should be noted that since the Moore analysis there has been an increase in crystal methamphetamine as the preferred form of methamphetamine (9) and a shift in the route of administration (increased use by inhalation / smoking) together with increased purity (10). The extent of this change is open to debate, with self-report data indicating stable prevalence between 2010-2013, but treatment data showing a marked increase in demand for services (9, 11).
What did we do?
The approach chosen for the analysis was a prevalence based method, that is, all cases or events in the target year were eligible, not just new cases. Having reviewed the available data and acknowledging the changing nature of methamphetamine use, 2013/14 was the most recent year with reasonably complete data. The domains identified for attention were: criminal justice (including police, courts, prisons and victims of crime), premature mortality, health (e.g. hospitals, ED, GP, specialist treatment) workplace costs, road crashes, child maltreatment, prevention, and clandestine laboratories. In addition, separate costs were estimated for methamphetamine users (internalities) and harms to others (partners and resident children). These latter two were segregated due to the controversy of including internalities (12) and the dearth of information on harms to others.
We used the Degenhardt and colleagues (11) estimate of the number of dependent users of methamphetamine (n=268,000 or 2.1% of those aged 15-54 years). In calculating the cost per statistical life year lost a range of values were used. The low range was set by the implied value ($45,000) from decisions made by the Pharmaceutical Benefits Advisory Board in approving new medications. The best or central estimate was based on a 2007 valuation produced for the Australian Government, inflated by the growth rate in the gross domestic product since that time ($281,798) (13). The upper value of $841,393 was derived from a USA based analysis (14).
Each of the domains required data from different sources (e.g. Steering Committee for the Review of Government Service Provision (15), Independent Hospital Pricing Authority (16), Medicare Australia (17), National Coronial Information System (NCIS)). Each domain also required that a proportion of total costs for an event be allocated to methamphetamine, typically based on an overall administrative budget. For example, the Drug Use Monitoring Australia (DUMA) study was used to estimate the number of crimes by category due to methamphetamine use (18). This provided a fraction to estimate subsequent costs: to police; magistrates; higher courts; and prison costs. Clearly each step requires stronger assumptions about the stability of costs through the judicial system. Further, there were domains where there was no standard data source (e.g. clandestine laboratories) or data that had to be extrapolated from a single study (e.g. general practice (19) or jurisdiction (e.g. child protection (20)). Finally, even where data were theoretically available, there could be delays in finalising cases (e.g. NCIS).
What did we find?
The overall cost of methamphetamine use to Australia was $5.0 billion in 2013/14 (Table 1). As with previous studies (7, 8), crime and the criminal justice system was the largest single cost domain at $1.9 billion, before adding the cost to victims of crime ($1.3 billion). Notably, we were unable to estimate some costs in this area. Thus, this total does not include Federal police or Federal courts or border protection. In addition, we did not include juvenile crime as we considered the youth equivalent to DUMA too dated (21). Parenthetically, juvenile crime accounts for about 12% of police cases (22) with 19% reporting that they were using methamphetamine at the time of their last offence (21).
Table 1: Estimated social and economic costs of methamphetamine use
Domain
|
Best estimate ($million)
|
Judicial system (e.g. police, courts, prisons & victims of crime)
|
3,244.5
|
Mortality
|
781.8
|
Employment
|
289.4
|
Health (e.g. hospital, ED, treatment, GPs)
|
270.8
|
Child maltreatment
|
260.4
|
Road traffic accident
|
125.2
|
Prevention harm reduction measures
|
40.0
|
Clandestine laboratories
|
11.7
|
Total
|
5,023.8
|
BBV = Blood borne viruses: ED = Emergency Department: GP = general practitioners: Column may not sum due to rounding
Although the number of deaths attributed to methamphetamine via the NCIS (n=170) was considerably lower than those attributed to opioids at the peak of the heroin crisis (1999, n=1116) (23, 24) it was the second leading cost. The full report also provided an upper estimate (not shown here) with the costs and offsetting savings over a 30 year horizon for premature mortality (4). It should be noted that many deaths where methamphetamine use was identified (n=291) had multiple potential causes, such as poly-drug use: in these cases only a “fraction” of a death was added to the total.
Workplace costs were also a major contributory factor, even though workplace deaths and traffic accident cost were listed in other domains. The largest component was excess absenteeism by methamphetamine users ($251 million) followed by excess injuries ($38 million). We were unable to quantify the costs associated with “presenteeism” (impaired performance at work due to prior or current drug use), staff turnover due to methamphetamine use and potential stress on co-workers.
While media attention has focused on the implications of methamphetamine induced psychosis in emergency departments, inpatient costs accounted for 45% ($90 million) of the total health costs with GP services ($28 million) and treatment of blood borne diseases (HIV/AIDS $30 million: hepatitis C $23 million) also being substantial contributors. The rapid change in the treatment options for these latter diseases mean that their costs may need to be regularly revisited.
The most contentious areas of the report are the costs arising to dependent drug users and the estimate of harms to other people, so these were listed separately from the main total. Two approaches to evaluating the harms to dependent users were adopted. Firstly, through the costs that they incurred, including drug purchases and lost income due to premature mortality or detention ($2.1 billion) and secondly via the disability adjusted life years (DALYs) (25) attributed to dependent drug use ($8.6 billion).
Estimating the costs to partners and dependent children were also separated from the main costing because of the lack of prior information on the topic and the uncertainly of the number of individuals affected. The number of partners was projected at 45,728 by applying the Degenhardt estimate of dependent users to the number of partners reported by the National Drug Strategy Household Survey (NDSHS) (9, 11). The number of resident children is not precisely specified in the NDSHS, with our estimate ranging from 30,161 to 120,854 children. To these figures we applied DALYs and the cost per DALY to arrive at the harms to partners being $860 million and the harms to children being $2.3 billion – $3.3 billion, depending on the actual number of children involved.
What does it mean?
Although cost of illness studies are popular for advocacy and to inform policy, their use in implementing change is more complex. For example, they do not identify the best or most effective ways of reducing costs. However, they do allow high cost areas to be identified and targeted for policy change or intervention. For example, in the current data, the major cost area is criminal justice, but any change in the legal status of methamphetamine would be driven by political imperative rather than factors relating to costs. Furthermore, even if its legal status were to be changed, it is not undisputed that there would be cost savings (26).
Where to next?
In conducting this analysis the most persistent problem across all the domains was the issue of poly-drug use, which is effectively ubiquitous in this population. Thus, attributing harm or a particular event to methamphetamine use alone will always be problematic. With some health conditions there is an established literature on attributable / aetiological fractions (25) but their development with respect to methamphetamine is currently in its infancy. The method we used with the DUMA data is conceptually the same as attributable fractions, but estimates across other domains had to be derived using less established methods or from more limited data ( e.g. a single state (ambulance data (27)) or from a single study, (ED costs (28)), which raises concerns when subsequently extrapolating to national costs. Therefore, developing consistent methods that can be applied across different substances is a priority.
There is increasing interest in the harms that substance use causes to others, with notable findings in relation to alcohol (29-31). Estimates for the number of children impacted by an adult’s “problem” or dependent drug use are in the range of 2% - 6% (32), with a USA based study finding that child engagement was the third leading cost due to methamphetamine use (8). However, the current study found few Australian data with which to estimate this cost, so the extent of these harms are effectively unknown at present.
A further area for research is the potential for unique costs for those living in more remote locations. Our report included a qualitative component, which involved interviewing key service providers in a rural and a remote location. National data suggests that the prevalence of methamphetamine use is higher outside of major cities and inner regional centres (33). However, we were unable to quantify additional costs for these people or services in rural and remote locations. Thus, there may be lower use of treatment services due to difficulties in accessing services, but high levels of severity of presentation when they are accessed. This deficit in our knowledge is compounded by the fact that there is likely to be greater diversity in remote than urban settings, both in the availability of services and the penetration of methamphetamine into a community.
The full report of this research is available to download from the NDRI website:
Whetton, S., Shanahan, M., Cartwright, K., Duraisingam, V., Ferrante, A., Gray, D., Kaye, S., Kostadinov, V., McKetin, R., Pidd, K., Roche, A., Tait, R.J. and Allsop, S. (2017). The Social Costs of Methamphetamine in Australia 2013/14. National Drug Research Institute, Curtin University, Perth, Western Australia. Link