1,803
Views
1
CrossRef citations to date
0
Altmetric
Original Research

Five-year budget impact of a prescription digital therapeutic for patients with opioid use disorder

ORCID Icon, , & ORCID Icon
Pages 599-607 | Received 13 Apr 2021, Accepted 06 Dec 2021, Published online: 24 Jan 2022

ABSTRACT

Background

Opioid use disorder (OUD) is associated with high healthcare resource utilization (HCRU) and costs. reSET-O is an FDA-cleared prescription digital therapeutic that delivers neurobehavioral therapy as an adjunct to treatment-as-usual (TAU; buprenorphine, face-to-face counseling, and contingency management).

Methods

A budget impact model was developed to evaluate reSET-O as an adjunct to TAU in OUD for a 1 million-member US mixed health plan over a 5-year time horizon. Model inputs included treatment costs and medical costs of hospitalizations, partial hospitalizations, intensive care unit stays, and emergency department visits.

Results

The base-case results and the alternative scenario analysis showed the addition of reSET-O was projected to result in consistently lower total yearly costs vs TAU and no treatment. The estimated total and per member per month (PMPM) budget impact over 5 years was -$763,026 and -$0.0116, respectively. When the upper range of cost estimates was used, the total and PMPM budget impacts over 5 years were -$2,481,563 and -$0.0378, respectively. Sensitivity analysis showed results were most sensitive to the proportion of patients untreated.

Conclusion

The introduction of reSET-O in addition to TAU for OUD has the potential to reduce healthcare resource utilization and costs from 12 weeks up to 5 years.

1. Introduction

Opioid use disorder (OUD) is considered a chronic, relapsing disease. It is currently fueled by the widespread availability, low cost, and greater potency of illicit opioids such as heroin and fentanyl. An early goal of therapy is harm reduction, or the implementation of patient-centric measures to reduced exposure to, and harm from, illicit opioids [Citation1]. Although 40% to 60% of patients relapse following withdrawal from opioids, relapse rates are significantly reduced by 50% or more in patients who stabilize on an opioid agonist medication [Citation2–6]. OUD is a serious national crisis that affects public health and is responsible for a substantial social and economic burden [Citation7–9]. In the United States (US) in 2019, it was estimated that 10.3 million individuals misused opioids and 1.6 million were diagnosed with OUD [Citation10]. With the disruption to OUD treatment due to the COVID-19 pandemic and its impact on patient cognition and behavior, a recent sharp increase in opioid-related mortality has been reported [Citation11]. In the 12-month period ending January 2021, 70,456 individuals died from an opioid-related overdose, which translated to approximately 193 deaths every day, and was approximately 20,000 additional deaths, or 40% higher than the 12-month period ending January 2020 [Citation11]. In addition to lost workplace productivity, criminal activity, and premature mortality associated with OUD, the annual direct medical costs were recently estimated to be $90 billion annually, which, across the 1.6 million patients in the US, represents a per-patient per-year cost of $56,000 [Citation9].

Current standard of care for OUD includes medications for opioid use disorder (MOUDs) such as buprenorphine, used in conjunction with counseling, and behavioral therapies [Citation12,Citation13]. Buprenorphine therapy is associated with reductions in opioid overdose deaths (38% over a 12-month period), healthcare resource utilization (eg, hospitalizations are twice as likely in patients nonadherent to buprenorphine therapy), and total cost of care (to about $2,500 to $11,000 per-patient per-year) if used appropriately and without early discontinuation [Citation6,Citation14–16]. However, the retention rates of MOUDs are low overall and range widely across settings.

A systematic review reported median 1- and 3-year retention rates of 57% and 38.4%, respectively, for opioid substitution treatment [Citation17]. Other literature reported that approximately 20% of dropouts from treatment programs occurred within the first month [Citation18]. In one study, patients retained on buprenorphine for 15 to 18 months experienced significantly better outcomes compared to patients retained for 6 to 9 months, including a lower risk of ED visits, inpatient hospitalizations, and filling of opioid prescriptions [Citation19]. The combination of counseling and behavioral therapies (e.g. cognitive behavioral therapy [CBT] and contingency management [CM]) has been recognized as being useful in increasing patient engagement, improving adherence, and facilitating behavior change [Citation20]. The combination has also been shown to be superior to treatment-as-usual (TAU) in increasing retention in treatment, and the odds of testing negative for opioids and cocaine increased during the third month of treatment when delivered as a prescription digital therapeutic [Citation21].

Prescription digital therapeutics (PDTs) have the potential to facilitate and expand access to OUD treatment by delivering evidence-based neurobehavioral treatment in a convenient, familiar, and confidential form. reSET-O is a 12-week Food and Drug Administration (FDA)-cleared PDT that combines CBT with fluency training (FT) and CM through 67 on-demand text, audio, and video lessons. Lessons are designed based on the community reinforcement approach (CRA) that teaches patients to master specific skills (e.g. communication skills, problem-solving, drug refusal, etc.) to achieve their recovery goals and increase satisfaction with personal, social, and vocational aspects of their lives, among others [Citation22]; lessons are designed to replicate a 30-minute face-to-face counseling session. Following each lesson, FT in the form of a simple quiz is presented to the user in order to increase the retention and understanding of positive adaptive behaviors. Patients are recognized and rewarded via CM with merit badges or gift cards of modest value upon the completion of each lesson and quiz [Citation23–26]. In the pivotal randomized controlled trial for reSET-O (NCT00929253), its precursor (therapeutic education system [TES]) in combination with TAU (buprenorphine, counseling, and CM) demonstrated a statistically significant higher retention rate than TAU alone (80% vs 64%, respectively; odds ratio for completing the 12-week treatment was 2.30 favoring TES; P = 0.018) in patients with OUD [Citation27]. In addition, patients treated with reSET-O + TAU were twice as likely to test negative for substances (opioids and cocaine) during weeks 9 to 12 compared to TAU [Citation21].

A budget impact analysis (BIA) of a new health technology is an important part of a comprehensive economic evaluation. A BIA provides guidance for budget forecasting, planning, and estimating the impact of adoption and diffusion of a new healthcare intervention on premiums in health insurance schemes given inevitable resource constraints. A previous study estimated the 12-week incremental cost per member per month (PMPM) of adding reSET-O to a commercial US healthcare plan as $0.0012 to $0.006 based on a 10% to 50% market uptake assumption [Citation28]. In the current study, the budget impact of reSET-O in OUD up to 5 years was assessed from a US third-party payer perspective incorporated recently available real-world HCRU data, re-treatment rates, and updated cost scenarios reflecting low- and high-cost treatment episodes to better account for different payer perspectives.

2. Methods

The analysis was developed following the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) guidelines for budget impact models [Citation29]. The model aimed to estimate the budget impact of reSET-O as an adjunct to TAU for OUD in which reSET-O + TAU was compared to TAU only and to no treatment. The budget impact on treatment-related costs and medical costs due to changes in treatment utilization was assessed from a US third-party payer perspective of a 1-million-member health plan that serves commercial and Medicare patients. The budget impact model structure is shown in . The differences in costs between the current and new markets for the total plan, per-treated-member per-month (PTMPM), and per-member per-month (PMPM) for each year were assessed over a 5-year time horizon. The PTMPM and PMPM budget impact results were calculated by dividing the total plan budget impact results by the treated population size and target population size, respectively, and further dividing by 12 months. Budget impact model findings are traditionally presented in this fashion, as PTMPM and PMPM.

Figure 1. Budget impact model structure.

Key: OUD – opioid use disorder; TAU – treatment-as-usual.
Figure 1. Budget impact model structure.

The patient population included adults aged ≥18 years with OUD who enrolled in outpatient treatment with sublingual buprenorphine and CM under clinician supervision. The model used epidemiological inputs to determine the target population and assign treatment based on market share rates (). Market share rates prior to the introduction of reSET-O were obtained from the 2018 National Survey on Drug Use and Health [Citation30]. In the new market, utilization of reSET-O was assumed to increase 1.5% annually from 0% to 7.5% at year 5, with market share proportionally taken from all other comparators.

Table 1. Plan and market assumptions for base-case model

Untreated patients incurred no treatment cost. For treated patients, each treatment cycle was 12 weeks. Total net treatment cost to the payer over 12 weeks was calculated for each comparator as shown in . For TAU, average patient cost-share [Citation31] was subtracted from the wholesale acquisition cost (WAC) to calculate the net cost to payers. Given that reset-O delivers counseling content, patients in the reSET-O group were assumed to receive 50% fewer counseling sessions but the same frequency of urinalyses as patients in the TAU group. Average numbers of treatment cycles for the 2 treatment groups were based on an analysis of real-world claims [Citation33, Citation62]. The total per-patient annual costs were calculated by multiplying the assumed average number of treatments and total net cost per cycle of reSET-O and TAU, respectively. Treatment retention was accounted for in the calculations of total per-patient cost after the first year [Citation33, Citation34]. Starting at Year 2, 6.4% of patients were assumed to receive subsequent reSET-O therapy and 71.25% were assumed to continue TAU in the reSET-O group. For the TAU only group, 57% were assumed to continue treatment ().

Table 2. Treatment costs for reSET-O + TAU and TAU only groups

2.1. Medical service utilization rates

For medical service utilization rates for reSET-O + TAU and TAU alone, the lower bound of the 95% confidence interval (CI) represents the incidence of abstinent patients and the higher bound of the 95% CI represents nonabstinent patients. All utilization rates were obtained from published data as shown in [Citation38, Citation39, Citation41, Citation42] For each year, an annual growth of 11% in medical service utilization rates for nonabstinent patients in each comparator group was assumed; different medical resource utilization rates were assumed for nonabstinent patients who continued treatments and those who stopped treatments. All treated patients incur the associated medical costs regardless of whether they had received treatment in the given year or had stopped subsequent treatment. Distributions of patients by abstinence and treatment status for each comparator group were informed by internal data from Pear Therapeutics, Inc. For reSET-O + TAU, 61.5% of patients were abstinent, 24.2% continued treatment but were nonabstinent, and 14.3% were nonabstinent and dropped out of treatment. For TAU alone, the percentages were 54.4%, 13.9%, and 31.7%, respectively. Patients in the no treatment group were assumed to be untreated and nonabstinent.

Table 3. Medical service utilization rates

The base-case and high-cost scenario inputs for hospitalizations, intensive care unit (ICU) stays, partial hospitalizations (PH), and emergency department (ED) visits are outlined in . The costs of fatal and nonfatal overdoses are captured in these cost inputs for both abstinent and nonabstinent patients. All costs were adjusted and reported in 2020 US dollars, with no discounting performed, per ISPOR guidelines for budget impact models.

Table 4. Medical costs per episode

2.2. Sensitivity analyses

To evaluate parameter uncertainties, a one-way sensitivity analysis and a scenario analysis were performed to test the robustness of the results. Parameters tested in the one-way sensitivity analysis included population characteristics, market shares, treatment costs, treatment cycles, and medical services utilization rates and costs, and each was varied by 10%. In the scenario analysis, an alternative set of costs per episode for medical service was used in the model.

3. Results

3.1. Base-case analysis

The projected total cost of treating OUD increased over 5 years. With the introduction of reSET-O into the market, total yearly costs were projected to be consistently lower than those of the current market (), with the magnitude of savings increasing steadily over the 5 years. The base-case total budget impact of reSET-O over 5 years was estimated as −$763,026, with a PTMPM budget impact of −$106.68 and a PMPM budget impact of −$0.0116 (). The projected total costs for each treatment comparator and the detailed breakdown by cost component are presented in .

Figure 2. Treatment costs, medical costs, and total costs with and without reSET-O by year.

Figure 2. Treatment costs, medical costs, and total costs with and without reSET-O by year.

Table 5. Budget impact results summary

3.2. Sensitivity analyses

The one-way sensitivity analysis revealed that results were most sensitive to the proportion of patients in the untreated state. Other influential variables in the model were total plan size and proportion of patients aged 18 to 64 diagnosed or treated for OUD (). In the scenario analysis, where a higher set of per-episode costs for medical resource utilization (hospitalizations, ICU stays, partial hospitalizations, and ED visits) was assumed, the magnitude of savings with reSET-O over 5 years increased by an additional −$1,718,537 over the base case, with a PTMPM budget impact of −$346.94 and a PMPM budget impact of −$0.0378 (, ). The budget impact results were consistent with the base case, with greater savings with the introduction of reSET-O.

Figure 3. Tornado diagram of base case one-way sensitivity analysis.

Key: BUP – buprenorphine; OUD – opioid use disorder; PH – partial hospitalization; TAU – treatment-as-usual.
Figure 3. Tornado diagram of base case one-way sensitivity analysis.

4. Discussion

From a US third-party payer perspective, reSET-O is a cost-saving PDT from 12 weeks up to a cumulative 5-year period when added to TAU in the management of OUD in adults treated in an outpatient setting. This BIA suggested that the additional treatment costs from reSET-O were fully offset by the reduction in costs associated with HCRU. The savings start with the first treatment cycle due to a reduction in hospitalizations and ED visits, the inclusion of CM, and a 50% reduction in counseling sessions, and the cost savings increase as the proportion of patients treated with reSET-O increases. Findings were consistent across various sensitivity analyses. The one-way sensitivity analysis showed that the model was most sensitive to the proportion of patients not receiving treatment, while the scenario analysis showed additional cost reductions over the base case through the avoidance of high-cost treatment episodes.

In the previously published BIA conducted by Wang and colleagues (2021) [Citation28], reSET-O was budget neutral over a 12-week time horizon, which was consistent with the findings in the current analysis [Citation35]. With a hypothetical 1-million-member health plan, the incremental total cost of adding reSET-O ranged from $3,563 at 10% market share uptake to $17,815 at 50% market share uptake, which translated to $0.0012 PMPM and $0.006 PMPM, respectively. Under similar settings, the current analysis showed that 10% market share uptake resulted in a −$1,377 difference in total cost and −$0.0004 PMPM, while 50% market share uptake resulted in a −$6,884 difference in total cost and −$0.0022 PMPM. The Wang model showed minimal budget impact to US payers, which increased with the expansion of reSET-O [Citation28]. On the other hand, findings from the current analysis suggest increased cost savings associated with reSET-O with larger market shares. The different trends of budget impact between the models are driven by differences in the estimation of medical costs. For instance, in the analysis conducted by Wang, 12-week medical and pharmacy costs of $3,886 and $11,218 were reported for adherent and nonadherent patients, respectively [Citation28]. Cost estimates were combined with the retention rates reported by Christensen and colleagues (2014) to calculate the weighted-average 12-week per-patient HCRU costs for reSET-O + TAU ($5,323) and TAU only ($6,519) groups [Citation27]. In the current analysis, the data source and approach utilized estimates of HCRU reported by Velez and colleagues (2020) by site of care [Citation42]; in addition, unit cost for each healthcare type, which provided extra information to better understand the driver of medical costs, was also incorporated [Citation44]. Patients were also distinguished by abstinence status, realizing that nonabstinent patients incurred more HCRU and costs (Reutsch) [Citation6]. As such, the current model accounted for both retention and abstinence for each comparator and provided a more comprehensive picture of OUD treatment.

The per-episode cost of HCRU across sites of care may be an important factor affecting budget impact. A wide variation exists in unit costs for HCRU and thus a scenario analysis was conducted with a set of higher cost estimates from published literature. The results suggest that payers can potentially save more in circumstances where medical costs are higher, since patients treated with reSET-O + TAU reported lower medical resource utilization rates (Velez) [Citation42]. One-way sensitivity analysis revealed limited effect of variation (10%) in medical costs. Regardless of this variation, reSET-O + TAU remained a cost-saving regimen in the current analysis.

The one-way sensitivity analysis found that the model is most sensitive to the proportion of untreated patients. Untreated patients with OUD incur significantly higher total healthcare costs compared to treated patients ($17,477 per patient per year in 2008 US dollars) Lynch [Citation14]. Based on the 2018 National Survey on Drug Use and Health, 81.9% of patients with OUD received no treatment SAMHSA [Citation10]; therefore, targeting the large population of untreated patients can be an opportunity to reduce healthcare costs. For example, a model constructed by the Institute for Clinical and Economic Review suggested that each dollar spent on expanding maintenance treatment would return $1.80 in savings ($2.6B in savings from an additional 505 patients in treatment) [Citation50]. Cost reductions have also been observed in a real-world expansion of treatment in Texas [Citation46].

Several barriers have been identified for the delivery of psychosocial support services, which can potentially be overcome with PDTs such as reSET-O. Workforce shortages of clinical and psychosocial providers – including behavioral counselors and buprenorphine prescribers – can decrease the availability of TAU and patient access [Citation45]. According to the Bureau of Labor Statistics Occupational Employment Statistics, there were 283,540 behavioral counselors in the workforce in 2019 [Citation47], which represented a shortage of 357,821 counselors; an estimated 641,361-person workforce is needed to support patients with serious emotional disturbance, serious mental illness, or substance use disorder [Citation30]. The shortages in rural areas were even worse, as 17% of counties that were not part of a metropolitan or micropolitan area had no behavioral health providers [Citation51]. Psychosocial support can help patients manage the challenges of transitioning from illicit opioid to a partial agonist such as buprenorphine [Citation52]. Delivered on mobile devices, PDTs can help decrease barriers to treatment and increase patient access to OUD treatments, improve patient engagement with the treatment process, and motivate patients for longer [Citation53, Citation54]. The addition of reSET-O may increase the variety of treatment options available to patients, which could possibly increase the overall number of treated patients, while reducing the incidence of high-cost hospitalizations and emergency department visits.

Longer treatment with MOUD is associated with positive outcomes on mortality, abstinence, and healthcare utilization [Citation55–60]. In previous studies, reSET-O demonstrated cost-effectiveness, dominant cost-utility, improved treatment retention, abstinence, and decreased real-world HCRU [Citation42, Citation61, Citation62]. In a long-term real-world data analysis, Velez et al. (2021) demonstrated continued benefits for reSET-O in reducing HCRU in OUD, with inpatient stays and ED visits decreased by 50% and 27% at 9 months, respectively [NaN]. To capture the long-term benefits associated with MOUD, the time horizons in this BIA were 12 weeks up to 5 years [Citation49]. Our study provided additional evidence showing that the addition of reSET-O provides a cost-saving opportunity to US payers for a longer period.

As other modeling studies, this BIA was limited by key assumptions made around market shares, treatment patterns, and HCRU due to uncertainties. Specifically, a conservative annual market share increase of 1.5% was used and was taken proportionally from TAU only and no treatment. This assumption was based on the potential increase in access to OUD treatment due to the convenience of PDTs delivered on mobile devices. As discussed, increased market share of reSET-O + TAU has the potential for greater cost savings by having more patients initiate treatment. Due to the lack of definitive abstinence data associated with real-world HCRU, the current model assumed that abstinent patients incurred HCRU at the lower bound of the 95% CI of the reported utilization rate, while nonabstinent patients utilized HCRU at the upper bound of the 95% CI. The rationale for this was based on the patient cohort reported by Velez and colleagues (2020) [Citation42], which consisted of a mix of abstinent and nonabstinent patients, where nonabstinent patients incurred higher HCRU based on the existing literature [Citation6]. Despite best efforts to select the most accurate model inputs from existing studies, parameters and assumptions were varied through sensitivity analyses to assess the model’s robustness to these assumptions and uncertainties, and the cost-savings of reSET-O were still realized.

5. Conclusion

The results of this model suggest initial healthcare cost savings after only 1 treatment cycle. Additionally, healthcare cost will decrease with reSET-O and this decrease is greater with higher adoption of the technology. The findings align with current evidence that the increased availability and expanded access to recovery services is associated with cost savings [Citation32]. reSET-O represents a potential opportunity to avoid the substantial costs and use of inpatient and ED services by retaining patients who might otherwise drop out of therapy and return to the use of illicit opioids.

Declaration of Interests

F Velez is an employee of Pear Therapeutics. Pear Therapeutics contracted Xcenda to assist in the research and completion of this study, and D Huang and L Mody are employees of Xcenda. DC Malone is an employee of Strategic Therapeutics and is a consultant for both Pear Therapeutics and Xcenda. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial relationships or otherwise to disclose.

Additional information

Funding

This study was funded by Pear Therapeutics (US), Inc.

References