A Modelling Analysis of Financial Incentives for Hepatitis C Testing and Treatment Uptake Delivered Through a Community-based Testing Campaign

Anna Y. Palmer; Kico Chan; Judy Gold; Chloe Layton; Imogen Elsum; Margaret Hellard; Mark Stoove; Joseph S. Doyle; Alisa Pedrana; Nick Scott

Disclosures

J Viral Hepat. 2021;28(11):1624-1634. 

In This Article

Methods

Design and Methods Summary

We used costing data from the Eliminate Hepatitis C (EC) Testing Campaign held during July and August 2019 to estimate the overall cost of the program. Using costs associated with hosting the campaign event, we estimated the unit cost per RNA-positive person completing testing. RNA-positive attendees were considered to have completed testing if they returned to the clinic to receive and discuss their test results (thus, receiving their RNA diagnosis). Using costs associated with follow-up and linkage to care, we estimated the unit cost per RNA diagnosed person initiating treatment. RNA diagnosed attendees were considered to have initiated treatment upon receiving a script for DAA treatment (as identified through clinic file audit three months after the campaign event).

We then created two mathematical models to estimate, for different hypothetical monetary incentive thresholds, what improvement in testing completion and treatment initiation would be required for the unit cost to remain the same as in the baseline case calculated using the EC Testing Campaign data. Figure 2 shows the target points in the hepatitis C care cascade for the modelled incentives.

Figure 2.

Hepatitis C cascade of care incentive target points. Incentives for undergoing testing were given out during the campaign event, while incentives for returning for a test result and initiating treatment (among RNA-positive people) were modelled in the analysis

The EC Testing Campaign

The main costing data were taken from a program report for the EC Testing Campaign. For full methods of the campaign event see Chan et al.[16] Briefly, the campaign was initiated by the Burnet Institute as part of the EC Partnership in 2019 around World Hepatitis Day (28th July), in collaboration with community partners and primary healthcare services. Over a total of nine days across four primary health services with a high caseload of PWID, clinic staff and partner organizations actively promoted hepatitis C testing, through peer-based outreach and promotional activities (including merchandise, door prizes and catering). An onsite clinical nurse was available during the event for hepatitis C consultations and blood collection for testing and attendees who self-reported not recently being tested for hepatitis C were given a small incentive (in the form of cash or gift cards A$10–20) for being tested, or for referring friends/family members to be tested. Attendees who knew they were antibody positive were offered RNA testing with pre-treatment work up to initiate treatment, while attendees who did not know their antibody status were offered antibody testing, with reflexive RNA testing and pre-treatment work up if antibody positive (meaning that pathology laboratories automatically performed RNA testing and pre-treatment work up if a positive antibody result was obtained). Reflexive genotype testing was also performed with those who tested RNA positive. Clinic staff followed up on serological laboratory testing initiated during the event and offered treatment to those with current hepatitis C infection through the clinic after the event.

Ninety-one participants underwent testing, of which 16 (18%) already knew they were antibody positive while 75 (82%) were antibody tested during the campaign. Of the 75 that were tested for antibodies, 53 were positive giving a total of 69 (76%) antibody-positive attendees. Due to reflexive testing, all 69 antibody-positive attendees had an RNA test and 24 (35%) tested RNA positive. Fifteen (63%) of the RNA-positive participants received their results, 10 (67%) of whom initiated treatment (Appendix, Figure S1).

Cost Calculation

Line-listed costs from the EC Testing Campaign event, and post-event follow-up were used as the main costing data source for the analysis (Table 1).

Costs associated with completing testing included pre-event costs, event costs, testing-specific costs and a small amount of research and evaluation costs. Pre-event and event costs were included in the cost of completing testing to emphasize the amount of effort required to find and link people with hepatitis C infection to care. Therefore, the cost of completing testing also included the cost of case finding in our analysis. To capture the realistic opportunity costs associated with the program, we also included both costs accumulated by RNA positive and RNA/antibody-negative individuals. Since some people already knew their antibody status, only a proportion of those tested were attributed the cost of an antibody test. In addition, as following up people with a positive RNA result was generally deemed more important by nurses compared to those with a negative result, more time to follow-up the result was costed for people testing RNA positive. The cost per RNA-positive person completing testing was calculated by dividing the total cost associated with testing and the campaign event by the number of RNA-positive attendees who completed testing.

Costs associated with initiating treatment included pre-treatment work up (needed to initiate hepatitis C treatment), hepatitis C genotyping and nurse consultation time to initiate treatment. The cost per RNA diagnosed person initiating treatment was calculated by dividing the total cost associated with post-event follow-up by the number of RNA diagnosed people who initiated treatment through the campaign.

The Reference Case for Global Health Costing was used as the framework for the costing analysis.[24] We used a micro-costing, activity-based approach and included all costs paid by the government healthcare system and partner institutions. Pathology costs were derived directly from the Australian Medicare Benefits Schedule (MBS); a listing of consultations, tests and procedures covered by the Australian healthcare system.

Baseline Cost Sensitivity Analyses

Campaign costs and loss to follow-up rates may vary when conducted in other settings, and RNA positivity rates are likely to decline over time as hepatitis C prevalence declines. However, since the cost data are line-listed and separated into categories, it is possible to use sensitivity analyses to estimate costs under different conditions. We used one-way sensitivity analyses to estimate the cost per RNA-positive person completing testing and RNA diagnosed person initiating treatment under different scenarios:[1] costs were 10% higher or lower,[2] the percentage of RNA-positive attendees completing testing was 40% or 80% (compared to 63% during the campaign),[3] the percentage of RNA diagnosed attendees initiating treatment was 40% or 80% (compared to 67% during the campaign), and[4] RNA test positivity rate among antibody-positive attendees was 10% or 50% (compared to 35% during the campaign).

Incentive Thresholds

Two modelling analyses were conducted to determine (1) how many more RNA-positive individuals would need to complete testing and (2) how many more RNA-positive individuals would need to initiate treatment to maintain the same unit cost (as calculated through the EC Testing Campaign) when a hypothetical incentive is offered. In model (1), we modelled additional incentive amounts (A$20+) given to RNA-positive individuals upon return to the clinic to receive and discuss their test results. For each additional incentive amount (modelled in A$20 increments), we determined the additional number of RNA-positive individuals that would need to complete testing so that the unit cost per RNA-positive individual completing testing would be equivalent to the baseline case (where no incentives were given). Incentive amounts were incremented until the corresponding number of RNA-positive individuals that would need to complete testing reached 100% (at this point, no additional incentive amount could retain the same unit cost). Similarly, in model (2), we modelled additional incentive amounts (A$20+) given to RNA-positive individuals who had completed testing, upon their treatment initiation. For each additional incentive amount (modelled in A$20 increments), we determined the additional number of RNA diagnosed individuals that would need to initiate treatment so that the unit cost per RNA diagnosed individual initiating treatment would be equivalent to the baseline case. Incentive amounts were incremented until the corresponding number of RNA diagnosed individuals that would need to initiate treatment reached 100%.

A simple equation model determined the outcome in both cases. The number of people that needed to complete testing so that the unit cost per RNA-positive person completing testing would be the same as in the baseline case was calculated according to the following formula:

Where: C test, Cost per person completing testing at baseline; E, Total campaign pre-event and event costs; T(ab+, rna+), Total testing-specific costs (dependent on the number of antibody-positive and RNA-positive attendees); x test, The number of people that complete testing; I testThe incentive amount given to each person that completes testing.

Likewise, the number of people that needed to initiate treatment so that the unit cost per RNA diagnosed person initiating treatment would be the same as in the baseline case was calculated according to the following formula:

Where, C treat, Cost per person initiating treatment at baseline; Tr(ab+, rna+), Total treatment initiation costs, besides staff costs (dependent on the number of antibody positive, the RNA-positive attendees); S treat, Cost of nurse time per person initiating treatment; x treat, The number of people that initiate treatment; I test, The incentive amount given to each RNA diagnosed person who initiates treatment.

Incentives Model Sensitivity Analysis

Since costs, RNA positivity and loss to follow-up rates may vary in other settings, we also performed a sensitivity analysis on the results of the two incentive models described above. We used one-way sensitivity analyses to estimate incentive thresholds for completing testing and initiating treatment under the same scenarios used for the baseline cost sensitivity analysis.

Cost calculations were conducted in Microsoft Excel,[25] and incentives modelling analyses were conducted in Python.[26]

processing....