Return on investment 2: evaluating the cost-effectiveness of needle and syringe programs in Australia 2009

Simulating the past if NSPs had not been in place

Page last updated: 2009

It is estimated that approximately 10-15% of syringes used for injecting drugs are purchased from pharmacies.24 We assume that if NSPs were not in place, the number of syringes in circulation would decrease to 15% of the current distribution. The population transmission model was used to simulate the expected epidemiological trends under conditions that no NSPs existed over the period from 2000 to 2009 (Figure 3). Based on the model, it is estimated that if NSPs were not in place, the incidence of HIV would have increased substantially (Figure 3). A large expansive epidemic of HIV among IDUs could have been expected if NSPs were not in place, with more than 3000 HIV infections per year after ten years of no needle and syringe program. High prevalence levels are common in other international settings where NSPs are not in place.4, 8-16 The model also predicts that HCV incidence would have been substantially greater if NSPs were not in place (Figure 3). According to the model, in the first year without NSPs there would be a large increase in incidence as susceptible IDUs become infected. This would be followed by a period of decreased incidence, as the pool of susceptible people decreases. However, incidence would then return to near-current levels.

It is estimated that over the ten year period 2000-2009, the cumulative incidence of HIV and HCV infections averted due to NSPs is ~32,050 (median; 20,765–42,211 interquartile range) and ~96,667 (92,465–103,055, IQR) respectively (Figure 4, Table 2; note that in the conservative scenario, steady state levels are assumed from the outset, which is why incidence is slightly higher following the cessation of NSPs for this scenario). Furthermore, it is estimated that the cumulative incidence of other disease outcomes have also decreased substantially due to NSPs (Table 2). It should be noted that there are only small changes in the long-term serious outcomes, such as HCC, liver failure and liver transplants, because only a ten-year timeframe was considered. The benefits in these outcomes become more marked over a longer time period as the effect of infections averted filters through to aversions of these clinical and disease-related outcomes. The modelling also suggests that NSPs have significantly reduced the potentially high prevalence of HIV and HCV that would have resulted had NSPs not been in place. In Figure 4, the cumulative number of HIV and HCV cases is shown with and without NSPs; the red curve represents the level suggested by the model as realistic numbers of cases without NSPs. For HIV, a very conservative case is also shown, where change in expected incidence is immediate due to fewer syringes in circulation but prevalence does not change with the number of new cases (that is, newly infected cases no longer share injecting equipment and are removed from the population).

Figure 3: Expected HIV and HCV cases among Australian IDUs from 100 model simulations under conditions of no NSPs (1999-2009)

Text equivalent below for Figure 3: Expected HIV cases among Australian IDUs from 100 model simulations under conditions of no NSPs (1999-2009)Text equivalent below for Figure 3: Expected HCV cases among Australian IDUs from 100 model simulations under conditions of no NSPs (1999-2009)Top of page

Text version of Figure 3

Figures in this description are approximate as they have been read from the graph.

Figure 3 consists of two graphs:
  • Expected annual HIV incidence among Australian IDUs from 100 model simulations under conditions of no NSPs (1999-2009)

    • The 100 model simulations rise dramatically from a range of 0-100 in 1999 to 500-3,000+ in 2004. By 2009, few simulations are below 3,000, with the lowest simulation rating at 1,800.

    • The median of the model simulations rises dramatically from 50 in 1999 to 2,600 in 2005 and surpassing 3,000 between 2005 and 2009.

    • The lower quartile of the model simulations rises dramatically from 50 in 1999 to 2,400 in 2006 and surpassing 3,000 between 2006 and 2009.

    • The upper quartile of the model simulations rises dramatically from 50 in 1999 to 2,300 in 2004 and surpassing 3,000 between 2004 and 2009.

    • The notification data for annual incidence of HIV among Australian IDUs is consistently below 100 for all years between 1999 and 2007 inclusive.

  • Expected annual HCV incidence among Australian IDUs from 100 model simulations under conditions of no NSPs (1999-2009)

    • The 100 model simulations begins at a range of 12,000-23,000 in 1999 before spiking to 35,000-58,000 in 2000. The simulated incidence rate then drops dramatically to 10,000-17,000 in 2002, before rising gradually to a range of 12,500-24,000 in 2009.

    • The median of the model simulations begins at 17,000 in 1999 before spiking to 48,000 in 2000. The median then drops dramatically to 13,000 in 2002, before rising gradually to 18,000 in 2009.

    • The lower quartile of the model simulations begins at 16,000 in 1999 before spiking to 44,000 in 2000. The lower quartile then drops dramatically to 12,000 in 2002, before rising gradually to 17,000 in 2009.

    • The upper quartile of the model simulations begins at 19,000 in 1999 before spiking to 51,000 in 2000. The upper quartile then drops dramatically to 14,000 in 2002, before rising gradually to 20,000 in 2009.

    • The notification data for annual incidence of HCV among Australian IDUs is:
      • 1999 - 13,000
      • 2000 - 13,100
      • 2001 - 10,500
      • 2002 - 9,100
      • 2003 - 9,000
      • 2004 - 8,900
      • 2005 - 8,800
Top of page

Figure 4: Model-based estimate of the cumulative number of HIV and HCV incident cases (median) among Australian IDUs with and without NSPs (2000-2009).

Text equivalent below for Figure 4: Model-based estimate of the cumulative number of HIV incident cases (median) among Australian IDUs with and without NSPs (2000-2009).Text equivalent below for Figure 4: Model-based estimate of the cumulative number of HCV incident cases (median) among Australian IDUs with and without NSPs (2000-2009).

Text version of Figure 4

Figures in this description are approximate as they have been read from the graph.

Figure 4 consists of two graphs:
  • Model-based estimate of cumulative HIV incidence among Australian IDUs with and without NSPs (2000-2009)

    • The cumulative number of HIV incident cases (median) among Australian IDUs with NSPs (current coverage) remains consistently below 100 during the period 1999 to 2009.

    • The model-based estimate of the cumulative number of HIV incident cases (median) among Australian IDUs without NSPs rises dramatically from less than 100 in 1999 to 16,000+ by 2007.

  • Model-based estimate of cumulative HCV incidence among Australian IDUs with and without NSPs (2000-2009)

    • The cumulative number of HCV incident cases (median) among Australian IDUs with NSPs (current coverage) rises steadily from 18,000 in 1999 to 120,000 in 2009.

    • The model-based estimate of the cumulative number of HCV incident cases (median) among Australian IDUs without NSPs rises dramatically from less than 18,000 in 1999 to 99,000 in 2001 and then rises steadily to 190,000+ by 2008.

Table 2: Estimated HIV and HCV related outcomes with and without NSPs (medians)

Table 2 is separated into 2 smaller tables in this HTML version for accessibility reasons. It is presented as one table in the PDF version.
Top of page

Estimated HIV related outcomes with and without NSPs (medians)

Outcome (2000-2009)With NSPsWithout NSPsCases averted
Prevalence of HIV among IDUs (2009)
0.1%
14.0%
Cumulative incidence of HIV infections
305
32,355
32,050
Cumulative number of HIV-related deaths
383
2,574
2,191

Estimated HCV related outcomes with and without NSPs (medians)

Outcome (2000-2009)With NSPsWithout NSPsCases averted
Prevalence of HCV among IDUs (2009)
65.1%
87.1%
Cumulative incidence of HCV infections
103,124
199,791
96,667
Number of cirrhosis cases (2009)
4,337
5,035
698
Cumulative incidence of HCC
1,854
1,859
5
Cumulative incidence of liver failure
2,704
2,720
16
Cumulative number of liver transplants
4,277
4,278
1
Cumulative number of liver-related deaths
4,084
4,088
4