The aim of this monograph is to estimate the costs of treatment and health care related to cannabis use disorders in New South Wales (NSW) for the year 2007. It describes the costs associated with treatment for cannabis use disorder and the health care costs attributable to cannabis use. The rationale for this study was to identify these costs for policy and other research purposes as comprehensive studies on cannabis treatment costs have not been undertaken.
Specifically, the types of treatment are:
Treatment in drug treatment agencies:

  • Counselling
  • Withdrawal management (detoxification)
  • Residential rehabilitation
  • Assessments (a precursor to treatment)
  • Information and education
Treatment in general practice (provided by general practitioners (GP)):
  • GP consultations
Treatment in hospitals:
  • Cannabis-related hospital admissions
Treatment of health consequences attributable to cannabis use:
  • Schizophrenia/ psychotic disorders
  • Low birth weight (LBW) babies in hospitals
  • Road traffic accident casualties
There are two types of costs estimated in this study:
  1. Costs associated with cannabis dependence/use disorder (in drug treatment agencies, general practice and hospitals); and
  2. Costs associated with providing health care for health consequences attributable to cannabis use (schizophrenia/psychotic disorders, LBW babies and road traffic accident casualties).
The costs of treatment in drug treatment agencies and general practice (in 1) were estimated by identifying the resources used and then multiplying their amount required (e.g. number of counsellor hours) for the treatment by their associated unit costs. For residential rehabilitation and hospital costs, the average costs (per day, separation) were taken from existing costing studies and applied to the number of separations or length of stay. The costing method for (1) is further outlined below:
  • The number of episodes, consultations, or separations were identified and counted;
  • A unit cost was then applied to these counts. These are discussed more in detail in each sub-section of treatment and health care types. Examples of unit costs are the Medicare Benefits Scheme (MBS) fee, Pharmaceutical Benefits Scheme (PBS) cost per pack, cost per day, hourly wage cost, and, hospital cost weights;
  • Completion rates as a proportion of all treatment episodes for withdrawal management and counselling episodes were used to calculate the number of completed and non-completed episodes. Costs were calculated for all reported treatment episodes and also for those completed and not completed;
  • For residential rehabilitation, an average cost per day was obtained from other costing studies for services in NSW. This was then multiplied by the average length of stay from the literature and the number of episodes;
  • For cannabis hospital separations, both as principal and secondary diagnoses were taken into account when calculating the costs of treatment. Specifically, admissions where cannabis was the secondary diagnosis and alcohol the primary has been found to influence hospital costs significantly (Pacula, Ringel, Dobkin, & Truong, 2008). The additional cost of the secondary diagnosis, in addition to the primary diagnosis, was added onto the hospital cost;
  • Where appropriate, an on-cost of 28% was added to the estimated wage costs for treatment services provided in the community setting. Additionally, in some instances, such as for counselling and withdrawal management, overhead costs of 28% were also added to ensure comparability to hospital and GP consultations costs;
  • Not included in the cost calculations were support and case management, and counselling under the Magistrates Early Referral into Treatment (MERIT) scheme. These costs are included elsewhere in a project examining criminal justice costs.
For (2), costs associated with providing health care to treat health consequences of cannabis use, a comprehensive literature review was conducted to include health consequences of cannabis use that had the strongest and most convincing evidence. The costing method was slightly different: the number of persons (for psychotic disorders and road traffic accident casualties) were multiplied by respective average costs from costing studies (Andrews & Tolkien II Team, 2006; BITRE, 2009), while for LBW babies, the number of hospital separations were multiplied by NSW hospital cost weights. Attributable fraction methods were derived from data in the literature and applied to general data to obtain the numbers attributable to cannabis use. These numbers were then applied to the respective costs.

The focus of this study was on cost of treatment for cannabis borne by health departments (State and Commonwealth) in NSW in 2007; and a one-year costing for the 2006-07 period was adopted. Costs were adjusted to reflect the reference year, 2007. This removes the effect of inflation and allows costs from different years to be compared on an equal dollar-for-dollar basis (ABS, 2008b; Riddell, Shanahan, Roxburgh, & Degenhardt, 2007). As costs are only calculated for one year, no discounting was applied (Drummond, Sculpher, Torrance, O'Brien, & Stoddart, 2005).

For the three largest costs, one-way sensitivity analyses were conducted on parameters that were considered sensitive to changes to the impact on costs. Data for the ‘sensitive’ parameters were obtained from the literature.
The main estimate of total treatment and health care costs of cannabis use for NSW in 2007 was $16.9 million with a range from $16.9 - $22.0 million obtained from sensitivity analysis:
Treatment and health care costs, NSW, 2007
Cannabis treatment

No. of episodes/ separations

Cost (2007 $) % cost Cost per occasion
Residential rehabilitation (episodes*) 431 $2,898,684 17.1% $6,725
Hospital (separations) 902 $1,307,610 7.7% $1,450
Withdrawal management (detoxification) (episodes*) 1,127 $1,083,124 6.4% $961
Counselling (episodes*) 2,451 $1,072,308 6.3% $437
GP (consultations) 3,018 $217,170 1.3% $72
Assessment only (episodes*) 1,727 $163,674 1.0% $95
Information and education only (episodes*) 113 $35,098 0.2% $310
Subtotal cannabis treatment  - $6,777,668 40.1%  


Treating health consequenes of cannabis use No. persons/ separations Cost (2007 $) % cost Cost per occasion
Psychotic disorders/schizophrenia (persons) 916 $6,220,049 36.8% $6,790
Road traffic accident casualties (persons) 443 $2,309,115 13.7% $5,212
Low birth weight (separations) 90 $1,605,291 9.5% $17,837
Subtotal health consequences treatment  - $10,134,454 59.9%  
Grand total  - $16,912,123 100.0%  

Note: Totals may not sum due to rounding
* All episodes, including complete and incomplete.

Sensitivity analysis conducted showed that the largest impact on treatment costs was for the proportions of ‘cannabis drivers’ killed and injured: there was a two-fold increase (200%) in treatment costs for road traffic accident casualties when these variables were varied at 11% and 7.1% respectively, compared to assuming that the rate of cannabis-related accident fatality and injury was the same (2.39%). The next largest impact was when the length of stay in residential rehabilitation treatment varied from 53.1 days to 147.3 days. This resulted in an increase in treatment costs by about 177%. However, these parameters did not have a large impact on the overall total costs. Total costs ranged from $16.4 million to $22.1 million.

While this study’s total cost estimate is comparable to another Australian study (Collins & Lapsley, 2008), that estimated the total health care costs from drug abuse at $16.6 million (adjusted for NSW and 2007 prices) Collins & Lapsley costs did not include drug treatment costs. On the other hand, Moore (2007) used a burden of disease approach (DALY), which resulted in an estimate of $421 million (2007 dollars) for the NSW total burden related to cannabis use.

As with any study, there are limitations and we outline some below.

  1. This study was not intended to assess the outcome of treatment or health care services for cannabis use.
  2. This study did not include intangible costs, e.g. opportunity costs, pain and suffering, etc.
  3. The lack of data meant assumptions had to be made in cost estimations.

Due to lack of data some treatments and their costs were not included. The key missing costs are those provided by private psychologists and psychiatrists.

Policy makers or researchers wishing to estimate the cost of providing treatment to cannabis, or other users, may use the estimation methods in this study as a guide. For example, they might adopt the same or similar methodology to estimate counsellor cost or the number of counselling sessions, attribution fraction methods, and, also use the costs and amount of resources as a guide in their own costing study. As other studies have not detailed the number of resources and unit costs for health care and treatment, this study would provide guidance in terms of such costing exercise. The results of this study would be useful in informing decision and policy making, particularly in funding or allocating resources to drug treatment agencies in NSW. It could also be used as a guide in facilitating economic or clinical evaluation of drug treatment programs, across treatment settings (drug treatment centres, GPs and hospitals).


Date Commenced
01 Dec 2010
Resource Type
DPMP Monographs