Key PointsQuestion
What is the most clinically beneficial and cost-effective antibiotic treatment strategy for injection drug use–associated infective endocarditis (IDU-IE)?
Findings
In this decision analytical modeling study simulating 4 treatment strategies among 5 million individuals with IDU-IE in the US, a validated microsimulation model suggested that outpatient parenteral antimicrobial therapy was the most cost-effective strategy for the treatment of IDU-IE. A partial oral antibiotic treatment strategy was associated with the highest treatment completion rate and was most cost-effective when methicillin-resistant Staphylococcus aureus was not a causative pathogen.
Meaning
This study found that outpatient parenteral antimicrobial therapy and partial oral antibiotic therapy regimens were likely to be as clinically beneficial as and less costly than 6 weeks of inpatient intravenous antibiotic therapy for the treatment of IDU-IE.
Importance
Emerging evidence supports the use of outpatient parenteral antimicrobial therapy (OPAT) and, in many cases, partial oral antibiotic therapy for the treatment of injection drug use–associated infective endocarditis (IDU-IE); however, long-term outcomes and cost-effectiveness remain unknown.
Objective
To compare the added value of inpatient addiction care services and the cost-effectiveness and clinical outcomes of alternative antibiotic treatment strategies for patients with IDU-IE.
Design, Setting, and Participants
This decision analytical modeling study used a validated microsimulation model to compare antibiotic treatment strategies for patients with IDU-IE. Model inputs were derived from clinical trials and observational cohort studies. The model included all patients with injection opioid drug use (N = 5 million) in the US who were eligible to receive OPAT either in the home or at a postacute care facility. Costs were annually discounted at 3%. Cost-effectiveness was evaluated from a health care sector perspective over a lifetime starting in 2020. Probabilistic sensitivity, scenario, and threshold analyses were performed to address uncertainty.
Interventions
The model simulated 4 treatment strategies: (1) 4 to 6 weeks of inpatient intravenous (IV) antibiotic therapy along with opioid detoxification (usual care strategy), (2) 4 to 6 weeks of inpatient IV antibiotic therapy along with inpatient addiction care services that offered medication for opioid use disorder (usual care/addiction care strategy), (3) 3 weeks of inpatient IV antibiotic therapy along with addiction care services followed by OPAT (OPAT strategy), and (4) 3 weeks of inpatient IV antibiotic therapy along with addiction care services followed by partial oral antibiotic therapy (partial oral antibiotic strategy).
Main Outcomes and Measures
Mean percentage of patients completing treatment for IDU-IE, deaths associated with IDU-IE, life expectancy (measured in life-years [LYs]), mean cost per person, and incremental cost-effectiveness ratios (ICERs).
Results
All modeled scenarios were initialized with 5 million individuals (mean age, 42 years; range, 18-64 years; 70% male) who had a history of injection opioid drug use. The usual care strategy resulted in 18.63 LYs at a cost of $416 570 per person, with 77.6% of hospitalized patients completing treatment. Life expectancy was extended by each alternative strategy. The partial oral antibiotic strategy yielded the highest treatment completion rate (80.3%) compared with the OPAT strategy (78.8%) and the usual care/addiction care strategy (77.6%). The OPAT strategy was the least expensive at $412 150 per person. Compared with the OPAT strategy, the partial oral antibiotic strategy had an ICER of $163 370 per LY. Increasing IDU-IE treatment uptake and decreasing treatment discontinuation made the partial oral antibiotic strategy more cost-effective compared with the OPAT strategy. When assuming that all patients with IDU-IE were eligible to receive partial oral antibiotic therapy, the strategy was cost-saving and resulted in 0.0247 additional discounted LYs. When treatment discontinuation was decreased from 3.30% to 2.65% per week, the partial oral antibiotic strategy was cost-effective compared with OPAT at the $100 000 per LY threshold.
Conclusions and Relevance
In this decision analytical modeling study, incorporation of OPAT or partial oral antibiotic approaches along with addiction care services for the treatment of patients with IDU-IE was associated with increases in the number of people completing treatment, decreases in mortality, and savings in cost compared with the usual care strategy of providing inpatient IV antibiotic therapy alone.
Hospitalizations associated with infective endocarditis in the US increased from 16 per 100 000 adults in 2003 to 22 per 100 000 adults in 2016.1 Injection drug use–associated infective endocarditis (IDU-IE) currently accounts for 1 in 10 hospitalizations for infective endocarditis.2 This increase has largely been associated with the opioid epidemic, specifically the injection of heroin and fentanyl. If current patterns continue, more than 250 000 individuals in the US may die of IDU-IE between 2020 and 2030.3 There is a substantial need to define optimal treatment strategies given the increasing burden of IDU-IE in the US.
Standard treatment for IDU-IE comprises 4 to 6 weeks of intravenous (IV) antibiotic therapy.4 Emerging evidence exists for the use of oral antibiotic therapy for the treatment of many types of infective endocarditis and the use of outpatient parenteral antimicrobial therapy (OPAT) for at least part of the treatment course.5-9 However, patients with IDU-IE are often required to remain hospitalized until treatment completion.10 Almost 20% of patients admitted with IDU-IE have a patient-directed discharge (ie, leave the hospital against medical advice).11,12 Alternative antibiotic treatment strategies that shorten hospitalization and allow patients to complete treatment elsewhere could increase the likelihood of treatment completion and decrease costs.
Current US treatment guidelines state that partial oral antibiotic therapy may be a reasonable option for patients with IDU-IE associated with uncomplicated right-sided Staphylococcus aureus infection but recommend that this approach only be used when parenteral antibiotic therapy is problematic.13,14 However, a retrospective cohort study5 found that people with IDU-IE who received a complete course of IV antibiotic therapy had similar readmission rates to those who could not complete inpatient IV antibiotic therapy and were provided partial oral antibiotic therapy at the time of patient-directed discharge.
Another strategy, OPAT, is widely used to treat infections that require prolonged antibiotic therapy, and this treatment strategy has a proven safety record.15 However, clinicians’ concerns regarding the misuse of a peripherally inserted central catheter to inject drugs in addition to treatment nonadherence, unstable living situations, and stigma associated with substance use have typically excluded people who inject drugs from receiving OPAT.16,17 Despite these concerns, a 2018 systematic review8 found that OPAT may be safe and beneficial for treating IDU-IE. To our knowledge, no study to date has compared the long-term impact and cost-effectiveness of OPAT with IV and partial oral antibiotic treatment strategies.
Recent research18 has highlighted the role of addiction care services in improving outcomes among individuals hospitalized with IDU-IE. Addiction care services, which can include addiction counseling, opioid withdrawal management, long-term medication titration, and referral and linkage to outpatient addiction care, have been reported to increase the likelihood of receiving medication for opioid use disorder (MOUD) during and after treatment for IDU-IE and have been associated with reductions in mortality risk19 and decreases in the probability of reinfection.20,21 Despite these benefits, an analysis from 1 hospital found that fewer than 8% of patients admitted with IDU-IE were discharged with any plans to start MOUD,22 reflective of a gap in treatment which has since been confirmed in broader studies.19,20
We evaluated the likely long-term clinical outcomes and cost-effectiveness of different strategies for the treatment of IDU-IE. Given the emerging evidence and unanswered questions, we aimed to (1) compare the potential value of alternative antibiotic treatment strategies and (2) estimate the impact of addiction care services among patients with IDU-IE.
The study was approved by the Boston University Medical Campus Institutional Review Board, which reviewed the Reducing Infection Related to Drug Use Cost-Effectiveness (REDUCE) model used in the study and provided a waiver of informed consent because the study did not involve human participants. This study followed the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline for economic evaluations of health interventions (eMethods 1 in the Supplement).23
We used the REDUCE model, a validated Monte Carlo microsimulation model that simulated the natural history of injection opioid use, to compare the following treatment strategies for IDU-IE: (1) 4 to 6 weeks of inpatient IV antibiotic therapy along with opioid detoxification (usual care strategy), (2) 4 to 6 weeks of inpatient IV antibiotic therapy along with inpatient addiction care services that offered MOUD (usual care/addiction care strategy), (3) 3 weeks of inpatient IV antibiotic therapy along with addiction care services followed by OPAT (OPAT strategy), and 4) 3 weeks of IV antibiotic therapy along with addiction care services followed by partial oral antibiotic therapy (partial oral antibiotic strategy).
In 3 of the strategies (usual care/addiction care, OPAT, and partial oral antibiotic therapy), patients with IDU-IE could receive addiction care services while hospitalized. These strategies were based on the assumption that when addiction care services were implemented, hospitalized individuals would have an increased probability of receiving MOUD in addition to addiction counseling, opioid withdrawal management, long-term medication titration, and referral and linkage to outpatient addiction care at the end of hospitalization.
In the OPAT strategy, all hospitalized patients with IDU-IE transitioned to either home-based or outpatient OPAT after 3 weeks of hospitalization and the offer of addiction care services. We assumed that 50% of patients would have home infusion therapy and 50% would receive OPAT at a postacute care facility. This assumption was informed by unpublished data from Boston Medical Center (A. Hill, BA, email communication, June 3, 2021) suggesting that one-half of patients with IDU-IE were homeless and therefore could not be discharged home. For the partial oral antibiotic strategy, we assumed that only patients admitted with non–methicillin-resistant S aureus (non-MRSA) IDU-IE would be eligible to receive oral antibiotic therapy after 3 weeks of hospitalization; however, all patients would be eligible to receive addiction care services. The probability of treatment completion and costs differed for each strategy. Details of these parameters and parameter sources are available in eMethods 1 in the Supplement.
The REDUCE model simulated a closed cohort of people who injected short-acting opioid drugs. For this analysis, we simulated a cohort over a lifetime to estimate long-term outcomes, including the mean percentage of patients completing treatment for IDU-IE, deaths associated with IDU-IE, life expectancy (measured in life-years [LYs]), mean cost per person, and incremental cost-effectiveness ratios (ICERs). We compared costs using a payer system perspective and denominated currency in 2020 US dollars. We discounted all costs and benefits by 3% annually and expressed ICERs as cost per LY gained, with a willingness-to-pay threshold of $100 000 per LY.24 We evaluated LYs rather than quality-adjusted LYs because the interventions focused on mortality-based outcomes, and quality-adjusted LYs are intended to measure life expectancy among patients with diseases in which there is a measurable change in quality of life (eg, heart failure) and the experience with drug use is heterogeneous. Probabilistic sensitivity, scenario, and threshold analyses were performed to evaluate major findings.
Model Structure and Simulated Cohort
The REDUCE model has been described in detail elsewhere.3 The model simulated a closed cohort experiencing the natural history of injection opioid drug use. Individuals moved through time in weekly steps from model initialization until death. Each week, there was a probability of developing sequelae of injection opioid drug use (eg, overdose or IDU-IE), requiring hospitalization, receiving outpatient addiction care, and changing injection drug use behavior.
The simulated cohort was stratified by sex (male or female), age (0-99 years), and injection behavior profile, which included injection frequency (high, low, or not currently injecting drugs), sharing of injection equipment (yes, no, or never), and sterile injection technique (cleaning, no cleaning, or never) (eMethods 1 and eTable 1 in the Supplement).
We assumed that individuals with high-frequency injection drug use had a higher probability of both overdose and IDU-IE, and individuals who shared injection equipment or used unsterile injection techniques had a higher probability of IDU-IE (eMethods 1, eTable 2, and eTable 3 in the Supplement). Overdose and infection risks were stratified by age and sex.
We assumed that after developing IDU-IE or experiencing an overdose, individuals had a probability of hospitalization. While hospitalized, patients could receive MOUD and a consultation for addiction care services, if available. We also assumed that individuals receiving MOUD and addiction care services had a higher probability of linking to outpatient MOUD and addiction care, and both outcomes changed the probability of decreasing the frequency of injection drug use (eMethods 2 and eTable 4 in the Supplement). Not all patients receiving addiction care services began receiving MOUD. While hospitalized, patients had a probability of leaving before treatment completion. We assumed that patients who did not complete treatment for IDU-IE would remain infected until they were readmitted to the hospital or died.
Outpatient Treatment Services
We assumed that when individuals left the hospital, they had a probability of linking to outpatient addiction care services and MOUD (eMethods 2 and eTable 5 in the Supplement). Linkage could be increased through receipt of inpatient addiction care services and MOUD but could also occur spontaneously through a background mechanism reflecting outpatient addiction care uptake in the nonhospitalized population.
The OPAT and partial oral antibiotic strategies simulated the provision of outpatient antibiotic therapy to individuals with IDU-IE (eMethods 3 and eTables 6-12 in the Supplement). We incorporated a weekly probability of discontinuing treatment for IDU-IE.
We assumed that individuals had a probability of dying of overdose and IDU-IE in addition to age-, sex-, and drug use–adjusted mortality from competing causes of death (eMethods 2 in the Supplement). At hospitalization, individuals had an additional mortality risk applied to reflect inpatient mortality.
Each patient accrued costs associated with opioid use, hospitalization, and outpatient services. Care costs associated with opioid use varied by injection behavior profile. The cost analysis also accounted for health care services, stratified by age and sex, that were not associated with opioid use based on data from the Medical Expenditure Panel Survey.25
Model Data and Parameter Estimation
Population, sequelae of drug use, inpatient, outpatient, mortality, and cost parameters were included in the model. The parameters and data sources are summarized in Table 1,4-8,11,25-64 and full details are available in eTable 1 to eTable 12 in the Supplement.
Overdose and Hospitalization
We derived rates of fatal and nonfatal overdose from state-level data.36-38 Rates of IDU-IE were derived from the published literature.4,65-67
Data from the published literature and expert opinion (H. Englander, MD, and C. King, PhD, email communication, October 20, 2019) informed the rates of hospitalization, the probability of initiating MOUD while an inpatient, and the association of addiction care services and MOUD with injection frequency. We assumed that 26% of individuals accepted addiction care services while inpatients based on unpublished data from Boston Medical Center addiction care services (Z.M. Weinstein, MD, email communication, March 12, 2019).
Outpatient Treatment Services
We assumed that individuals receiving inpatient addiction care services and MOUD had a 70% probability of linking to outpatient MOUD compared with individuals receiving inpatient MOUD alone (45% linkage), individuals receiving addiction care services alone (57% linkage), and individuals not receiving either inpatient MOUD or addiction care services (5% linkage). We estimated the conditional probabilities of linking to outpatient MOUD based on data from cohort studies and clinical trials.48,49 The probabilities of completing OPAT and partial oral antibiotic therapies were informed by data from the published literature.5-8
After accounting for fatal overdose, we derived age- and sex-adjusted mortality rates from the National Vital Statistics System to inform mortality associated with competing risks.54,55 To account for additional opioid drug use–associated harms not captured by fatal overdose or IDU-IE, we multiplied the resulting mortality rates by 1.2.54
We derived some of the costs from the 2020 Laboratory and Physician Fee Schedules from the Centers for Medicare and Medicaid Services63,64 and the Medical Expenditure Panel Survey25 (eTable 10 and eTable 11 in the Supplement).
Probabilistic, Scenario, and Threshold Analyses
For the main analysis, we performed probabilistic sensitivity analyses (eMethods 4 in the Supplement) using distributions around important model parameters. We performed 1000 simulations with 5 million individuals over a lifetime.
Deterministic sensitivity analyses were conducted to evaluate the extent of uncertainty in the input parameters (eMethods 4 in the Supplement). These analyses were performed with 500 000 individuals over a lifetime. We varied (1) the percentage of patients with IDU-IE who were eligible to receive partial oral antibiotic therapy (to reflect differences in the percentage of non-MRSA IDU-IE cases), (2) the percentage of patients leaving the hospital with patient-directed discharge, (3) the treatment uptake of OPAT and partial oral antibiotic therapy, (4) the rate of overdose within the community and outpatient settings, (5) the uptake of addiction care services and MOUD during hospitalization, and (6) the length of inpatient stay and uptake of partial oral antibiotic therapy. We also conducted threshold analyses to assess which values for selected parameters (eg, treatment uptake or treatment completion) changed our major findings (eMethods 4, eTable 13 in the Supplement).
The model was constructed using C++ programming language, and analyses were performed using R software, version 3.2.2 (R Foundation for Statistical Computing), and Excel software (Microsoft Corporation). No significance tests were performed for this simulation study.
We initialized the model with a cohort of 5 million individuals who reflected the age and sex of the US population who inject opioid drugs, with data informed by the US Census and published literature.26-28,30-32,68 At model initialization, the mean age of the cohort was 42 years (range, 18-64 years), 70% were male, 53% had high-frequency injection drug use, 11% had low-frequency injection drug use, and 36% had no current injection drug use.26-28,30-32,68 We assumed imperfect access to harm reduction services, with 66% of the cohort practicing unsterile injection techniques and 45% sharing injection equipment.69
Over a lifetime horizon within the usual care strategy, 685 637 individuals developed IDU-IE, 557 386 were hospitalized with IDU-IE, and 250 654 died of IDU-IE. The usual care strategy resulted in 18.63 LYs; 77.6% of hospitalized patients with IDU-IE completed treatment, and 5.01% of deaths in the population attributable to IDU-IE (Table 2). Life expectancy was extended by each alternative strategy (0.016 years with the usual care/addiction care strategy, 0.013 years with the OPAT strategy, and 0.024 years with the partial oral antibiotic strategy). The partial oral antibiotic strategy provided the highest treatment completion rate (80.3%) compared with the OPAT strategy (78.8%) and the usual care/addiction care strategy (77.6%). All strategies were attributable to a lower percentage of IDU-IE–associated deaths compared with the usual care strategy (4.86% with the usual care/addiction care strategy, 4.89% with the OPAT strategy, and 4.79% with the partial oral antibiotic strategy vs 5.01% with the usual care strategy) and overdose (15.70% with the usual care/addiction care strategy, 15.71% with the OPAT strategy, and 15.71% with the partial oral antibiotic strategy vs 15.73% with the usual care strategy).
The usual care strategy yielded a discounted lifetime mean cost of $416 570 per person (95% credible interval [CrI], $334 000-$482 780) whereas the OPAT strategy was the least expensive at $412 150 per person (95% CrI, $331 540-481 460) compared with the partial oral antibiotic strategy ($413 920 per person; 95% CrI, $333 220-$483 000) and the usual care/addiction care strategy ($416 990 per person; 95% CrI, $334 580-483 530). The usual care strategy was dominated by (ie, cost more and had worse clinical outcomes) all other strategies. Compared with the OPAT strategy, the partial oral antibiotic strategy had an ICER of $163 370 per LY. The usual care/addiction care strategy was dominated by the partial oral antibiotic strategy.
In the scenario analyses, the partial oral antibiotic strategy was preferred (ie, performed best) when patients with MRSA-associated IDU-IE were assumed to be eligible to receive partial oral antibiotic therapy, when treatment uptake of partial oral antibiotic therapy or OPAT was held equal, and when the inpatient stay was decreased and treatment uptake of partial oral antibiotic therapy was increased (Table 3). Incremental discounted LYs gained with the partial oral antibiotic strategy ranged from 0.020 (treatment uptake equal to OPAT treatment uptake) to 0.025 (MRSA-associated IDU-IE eligible for treatment), and incremental discounted costs ranged from −$4450 to −$1250. In a scenario analysis that assumed addiction care services reduced patient-directed discharge from 5.0% to 2.5% per week, the OPAT strategy was the most cost-effective, with a gain of 0.250 LYs and incremental mean discounted cost of −$4073. Increasing the uptake of addiction care services and MOUD from 25% to 75% yielded greater cost for each strategy but similar conclusions (mean discounted costs increased from $412 150 to $412 420 for the partial oral antibiotic strategy and from $413 920 to $414 300 for the OPAT strategy).
Clinicians have expressed concern regarding the possibility of overdose while receiving outpatient antibiotic therapy. In a scenario quadrupling the rate of overdose in the community, our findings regarding improved outcomes with partial oral antibiotic and OPAT regimens did not qualitatively change. The OPAT strategy was the least expensive at $312 670 per person compared with the partial oral antibiotic strategy ($313 930 per person) and the usual care/addiction care strategy ($316 250 per person) and resulted in 0.059 additional LYs. The partial oral antibiotic strategy had an ICER of $167 410.
We performed several threshold analyses. First, because uncertainty remained regarding the comparative benefit of IV vs partial oral antibiotic therapies, we performed a threshold analysis of the minimum benefit of both partial oral antibiotic and OPAT strategies, lower than which the usual care strategy provided the best outcomes (Figure 1; eFigures 1 and 2 in the Supplement). We found that the usual care strategy provided the best outcome when treatment completion was lowered from the base case of 87% to 83% for the OPAT strategy and from the base case of 90% to 80% for the partial oral antibiotic strategy. When treatment completion was lowered to 83% for the OPAT strategy and 80% for the partial oral antibiotic strategy, there was no longer a gain in LYs compared with the usual care strategy. When completion of partial oral antibiotic therapy increased to 92%, the partial oral antibiotic strategy was preferred to the OPAT strategy. Partial oral antibiotic therapy was cost-effective compared with OPAT at the $100 000 per LY threshold.
Next, we explored the rate at which patients accepted a given therapy. When OPAT uptake decreased from 100% to 79%, OPAT was no longer the preferred strategy because the mean discounted cost of OPAT ($413 860) became equivalent in cost to the partial oral antibiotic treatment strategy ($413 920). When partial oral antibiotic therapy uptake increased from 79% to 86%, partial oral antibiotic therapy was the preferred strategy, with an ICER of $72 182 per LY (eFigure 1 in the Supplement). In a threshold analysis assessing cost, when OPAT cost was $26 000 per week (compared with $1590 per week in the base case model), the OPAT strategy no longer met the $100 000 per LY willingness-to-pay threshold compared with the usual care strategy (eFigure 2 in the Supplement).
Our major findings did not qualitatively change in the probabilistic sensitivity analyses (Table 2), in which the percentages of patient-directed discharge and treatment uptake were held constant while almost all other parameters were varied (Table 1). Probabilistic sensitivity analyses were used to calculate CrI s for discounted LYs for the usual care (18.63 LYs; 95% CrI, 17.28-18.67 LYs), OPAT (18.75 LYs; 95% CrI, 17.32-18.70 LYs), partial oral antibiotic therapy (18.66 LYs; 95% CrI, 17.34-18.74 LYs), and usual care/addiction care (18.65 LYs; 95% CrI, 17.30-18.70 LYs) strategies. A cost-effectiveness acceptability curve (Figure 2) using output from the probabilistic sensitivity analyses revealed that either the partial oral antibiotic or OPAT strategy yielded the greatest net monetary benefit 100% of the time. Up to a willingness-to-pay threshold of $60 000, the OPAT strategy was preferred, and at a willingness-to-pay threshold higher than $60 000, the partial oral antibiotic strategy was preferred.
In this microsimulation modeling study, treatment of IDU-IE with partial oral antibiotic therapy or OPAT was associated with similar or improved long-term clinical outcomes compared with usual care while also being cost-effective. Within our base case model, we assumed that patients with IDU-IE associated with MRSA infection were not eligible to receive partial oral antibiotic therapy and, as a result, the OPAT strategy was found to be the most cost-effective. Without the exclusion of MRSA infection, the partial oral antibiotic strategy was optimal.
Up to 1 in 4 patients with IDU-IE die within 1 year after hospital admission.70 Challenges associated with long periods of hospitalization may be justified if hospital stays improve outcomes; however, establishment of the inferiority of alternative approaches is necessary. An increasing body of evidence suggests that OPAT and partial oral antibiotic strategies are feasible for the treatment of IDU-IE in this population, producing similar or improved clinical outcomes.5,6 Informed by these existing studies, we modeled the long-term outcomes associated with offering alternative antibiotic strategies paired with addiction care services and found that both the OPAT and partial oral antibiotic strategies were associated with improved outcomes compared with the usual care strategy. Our results suggest that OPAT and partial oral antibiotic regimens may be as clinically beneficial and less costly than the usual care regimen, and these findings support expanding opportunities to research and implement these options among patients with IDU-IE. Our findings also reinforce the importance of addiction care services and contribute to increasing evidence suggesting the necessity of addiction care services for the treatment of individuals with opioid use disorder.18
Concerns regarding the efficacy of oral antibiotic medications have hindered efforts to expand the use of partial oral antibiotic therapy for the treatment of IDU-IE. Within our main analysis, we assumed that patients with MRSA-associated IDU-IE were ineligible to receive partial oral antibiotic therapy but that otherwise the modeled antibiotic therapies had similar treatment completion rates if the treatment course was completed. Adherence to and completion of antibiotic treatment are important parameters to consider when assessing potential administration of partial oral antibiotic regimens to patients with IDU-IE. Previous studies on the implementation of care for hepatitis C viral infection among individuals who inject drugs and are receiving MOUD have reported high rates of adherence to antiviral treatment regimens that were similar to the rates of antiretroviral treatment adherence among people with HIV infection who inject drugs.71,72 Our threshold analysis revealed that when 80% or more of the patients receiving partial oral antibiotic therapy and 83% or more of the patients receiving OPAT successfully completed treatment, these regimens would continue to improve life expectancy compared with usual care. Although the intention of the usual care approach is universal treatment completion, the reality of noncompletion of treatment is likely underappreciated when weighing the risks and benefits of treatment strategies. Within the model, potential differences in the benefits of regimens were overcome by large differences in treatment completion. The model also assumed that a full 6 weeks of therapy was needed before treatment completion and that a mean inpatient stay of 3 weeks was needed before initiation of partial oral antibiotic or OPAT regimens. Therefore, our estimate was likely conservative.
Access to postacute care facilities for administration of OPAT may limit the ability of institutions to offer this treatment regimen.10,22 Postacute care facilities often refuse to accept patients with histories of active substance use despite the fact that these practices violate the Americans with Disabilities Act.16 However, we found within a scenario analysis that even when individuals had a very high probability of overdose after leaving the hospital, alternative antibiotic regimens were associated with improvement in outcomes compared with the usual care regimen. This finding suggests that the opportunity to complete treatment and link to MOUD through addiction care services may prevent more overdose fatalities than an extended hospital stay. These results can be used as an advocacy tool for agencies such as Medicaid to work with postacute care facilities to improve access to OPAT.
We accounted for some socioeconomic challenges, such as homelessness, by assuming that only one-half of patients could receive at-home OPAT. There are circumstances in which hospitalization may be preferable to the alternative (eg, no housing), but administration of OPAT within in a postacute care facility rather than a hospital may be preferable to both approaches. However, the high rate of adverse events associated with OPAT, including peripherally inserted central catheter line infection and thrombosis, will need to be considered when discussing alternative antibiotic therapy strategies.9 Clinicians may consider engaging in patient-centered decision-making when offering these treatment strategies, with housing not used as the sole determining factor when selecting an antibiotic treatment strategy.
This study has several limitations. First, we relied on a single published study to inform parameters on partial oral antibiotic treatment completion and regimen costs. However, a prospective cohort study73 examining the efficacy of a partial oral antibiotic regimen for patients with early patient-directed discharge is currently being conducted, and model parameters are within the currently observed range. Second, although important model parameters were informed by studies of the target population, unmeasured confounders may have impacted the results of these studies. Despite these limitations, our findings did not qualitatively change in sensitivity analyses and when varying assumptions were used, which may enable clinicians and hospital staff to consider these findings within their local context.
Results from this decision analytical modeling study suggest that, if implemented, the strategies could save the health care system a substantial amount of money in lifetime hospitalization costs alone for the estimated 750 000 individuals currently injecting drugs in the US.26 Those savings could be shifted to programs that specifically address the opioid epidemic, such as initiatives to improve access to MOUD, promote safer injection techniques, and provide multidisciplinary outpatient support systems, including peer navigators and case managers, to decrease the future incidence of IDU-IE and support patient retention in substance use disorder care programs.
Accepted for Publication: January 10, 2022.
Published: February 28, 2022. doi:10.1001/jamanetworkopen.2022.0541
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022 Adams JW et al. JAMA Network Open.
Corresponding Author: Joshua A. Barocas, MD, Division of General Internal Medicine, Anschutz Medical Campus, University of Colorado, 12631 E 17th Ave, 8th Fl, Academic Office 1, Mailstop B180, Aurora, CO 80045 (joshua.barocas@cuanschutz.edu).
Author Contributions: Drs Barocas and Adams had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Adams, Savinkina, Jawa, Linas, Kimmel, Barocas.
Acquisition, analysis, or interpretation of data: Adams, Hudspeth, Gai, Marks, Linas, Hill, Flood, Kimmel, Barocas.
Drafting of the manuscript: Adams, Jawa, Hill, Barocas.
Critical revision of the manuscript for important intellectual content: Savinkina, Hudspeth, Gai, Jawa, Marks, Linas, Flood, Kimmel, Barocas.
Statistical analysis: Adams, Savinkina, Gai, Kimmel, Barocas.
Obtained funding: Barocas.
Administrative, technical, or material support: Savinkina, Jawa, Marks, Flood, Barocas.
Supervision: Jawa, Linas, Barocas.
Conflict of Interest Disclosures: Dr Kimmel reported receiving personal fees from Abt Associates and the American Academy of Addiction Psychiatry outside the submitted work. No other disclosures were reported.
Funding/Support: This work was supported by grants K01DA051684 (Dr Barocas and Ms Savinkina), DP2DA051864 (Dr Barocas), R01DA046527 (Dr Linas), and P30DA040500 (Dr Linas) from the National Institute on Drug Abuse and grant T32-AI052074 (Drs Adams and Jawa) from the National Institute of Allergy and Infectious Diseases.
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
2.Deo
SV, Raza
S, Kalra
A,
et al. Admissions for infective endocarditis in intravenous drug users.
J Am Coll Cardiol. 2018;71(14):1596-1597. doi:
10.1016/j.jacc.2018.02.011 3.Barocas
JA, Eftekhari Yazdi
G, Savinkina
A,
et al. Long-term infective endocarditis mortality associated with injection opioid use in the United States: a modeling study.
Clin Infect Dis. 2021;73(11):e3661-e3669. doi:
10.1093/cid/ciaa1346PubMedGoogle Scholar 4.Miller
AC, Polgreen
PM. Many opportunities to record, diagnose, or treat injection drug–related infections are missed: a population-based cohort study of inpatient and emergency department settings.
Clin Infect Dis. 2019;68(7):1166-1175. doi:
10.1093/cid/ciy632PubMedGoogle ScholarCrossref 5.Marks
LR, Liang
SY, Muthulingam
D,
et al. Evaluation of partial oral antibiotic treatment for persons who inject drugs and are hospitalized with invasive infections.
Clin Infect Dis. 2020;71(10):e650-e656. doi:
10.1093/cid/ciaa365PubMedGoogle ScholarCrossref 6.Fanucchi
LC, Walsh
SL, Thornton
AC, Nuzzo
PA, Lofwall
MR. Outpatient parenteral antimicrobial therapy plus buprenorphine for opioid use disorder and severe injection-related infections.
Clin Infect Dis. 2020;70(6):1226-1229. doi:
10.1093/cid/ciz654PubMedGoogle ScholarCrossref 7.D’Couto
HT, Robbins
GK, Ard
KL, Wakeman
SE, Alves
J, Nelson
SB. Outcomes according to discharge location for persons who inject drugs receiving outpatient parenteral antimicrobial therapy.
Open Forum Infect Dis. 2018;5(5):ofy056. doi:
10.1093/ofid/ofy056PubMedGoogle Scholar 8.Suzuki
J, Johnson
J, Montgomery
M, Hayden
M, Price
C. Outpatient parenteral antimicrobial therapy among people who inject drugs: a review of the literature.
Open Forum Infect Dis. 2018;5(9):ofy194. doi:
10.1093/ofid/ofy194PubMedGoogle Scholar 9.Spellberg
B, Chambers
HF, Musher
DM, Walsh
TL, Bayer
AS. Evaluation of a paradigm shift from intravenous antibiotics to oral step-down therapy for the treatment of infective endocarditis: a narrative review.
JAMA Intern Med. 2020;180(5):769-777. doi:
10.1001/jamainternmed.2020.0555PubMedGoogle ScholarCrossref 10.Kimmel
SD, Rosenmoss
S, Bearnot
B, Larochelle
M, Walley
AY. Rejection of patients with opioid use disorder referred for post-acute medical care before and after an anti-discrimination settlement in Massachusetts.
J Addict Med. 2021;15(1):20-26. doi:
10.1097/ADM.0000000000000693 11.Kimmel
SD, Kim
JH, Kalesan
B, Samet
JH, Walley
AY, Larochelle
MR. Against medical advice discharges in injection and non-injection drug use–associated infective endocarditis: a nationwide cohort study.
Clin Infect Dis. 2021;73(9):e2484-e2492. doi:
10.1093/cid/ciaa1126PubMedGoogle Scholar 13.Baddour
LM, Wilson
WR, Bayer
AS,
et al; American Heart Association Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease of the Council on Cardiovascular Disease in the Young; Council on Clinical Cardiology; Council on Cardiovascular Surgery and Anesthesia; Stroke Council. Infective endocarditis in adults: diagnosis, antimicrobial therapy, and management of complications: a scientific statement for healthcare professionals from the American Heart Association.
Circulation. 2015;132(15):1435-1486. doi:
10.1161/CIR.0000000000000296PubMedGoogle ScholarCrossref 14.Iversen
K, Ihlemann
N, Gill
SU,
et al. Partial oral versus intravenous antibiotic treatment of endocarditis.
N Engl J Med. 2019;380(5):415-424. doi:
10.1056/NEJMoa1808312 15.Mitchell
ED, Czoski Murray
C, Meads
D, Minton
J, Wright
J, Twiddy
M. Clinical and cost-effectiveness, safety and acceptability of community intra
venous antibiotic service models: CIVAS systematic review.
BMJ Open. 2017;7(4):e013560. doi:
10.1136/bmjopen-2016-013560PubMedGoogle Scholar 16.Fanucchi
L, Leedy
N, Li
J, Thornton
AC. Perceptions and practices of physicians regarding outpatient parenteral antibiotic therapy in persons who inject drugs.
J Hosp Med. 2016;11(8):581-582. doi:
10.1002/jhm.2582PubMedGoogle ScholarCrossref 17.Rapoport
AB, Beekmann
SE, Polgreen
PM, Rowley
CF. Injection drug use and infectious disease practice: a national provider survey.
Open Forum Infect Dis. 2017;4(suppl 1):S340. doi:
10.1093/ofid/ofx163.810 19.Kimmel
SD, Walley
AY, Li
Y,
et al. Association of treatment with medications for opioid use disorder with mortality after hospitalization for injection drug use–associated infective endocarditis.
JAMA Netw Open. 2020;3(10):e2016228. doi:
10.1001/jamanetworkopen.2020.16228PubMedGoogle Scholar 20.Barocas
JA, Morgan
JR, Wang
J, McLoone
D, Wurcel
A, Stein
MD. Outcomes associated with medications for opioid use disorder among persons hospitalized for infective endocarditis.
Clin Infect Dis. 2021;72(3):472-478. doi:
10.1093/cid/ciaa062PubMedGoogle Scholar 21.Marks
LR, Munigala
S, Warren
DK, Liang
SY, Schwarz
ES, Durkin
MJ. Addiction medicine consultations reduce readmission rates for patients with serious infections from opioid use disorder.
Clin Infect Dis. 2019;68(11):1935-1937. doi:
10.1093/cid/ciy924PubMedGoogle ScholarCrossref 24.Neumann
PJ, Sanders
GD, Russell
LB, Siegel
JE, Ganiats
TG, eds. Cost Effectiveness in Health and Medicine. 2nd ed. Oxford University Press; 2016.
26.Lansky
A, Finlayson
T, Johnson
C,
et al. Estimating the number of persons who inject drugs in the United States by meta-analysis to calculate national rates of HIV and hepatitis C virus infections.
PLoS One. 2014;9(5):e97596. doi:
10.1371/journal.pone.0097596PubMedGoogle Scholar 27.Martins
SS, Sarvet
A, Santaella-Tenorio
J, Saha
T, Grant
BF, Hasin
DS. Changes in US lifetime heroin use and heroin use disorder: prevalence from the 2001-2002 to 2012-2013 National Epidemiologic Survey on Alcohol and Related Conditions.
JAMA Psychiatry. 2017;74(5):445-455. doi:
10.1001/jamapsychiatry.2017.0113PubMedGoogle ScholarCrossref 28.Degenhardt
L, Peacock
A, Colledge
S,
et al. Global prevalence of injecting drug use and sociodemographic characteristics and prevalence of HIV, HBV, and HCV in people who inject drugs: a multistage systematic review.
Lancet Glob Health. 2017;5(12):e1192-e1207. doi:
10.1016/S2214-109X(17)30375-3PubMedGoogle ScholarCrossref 31.Tan
S, Makela
S, Heller
D,
et al. A bayesian evidence synthesis approach to estimate disease prevalence in hard-to-reach populations: hepatitis C in New York City.
Epidemics. 2018;23:96-109. doi:
10.1016/j.epidem.2018.01.002 32.Buresh
M, Genberg
BL, Astemborski
J, Kirk
GD, Mehta
SH. Recent fentanyl use among people who inject drugs: results from a rapid assessment in Baltimore, Maryland.
Int J Drug Policy. 2019;74:41-46. doi:
10.1016/j.drugpo.2019.08.006 38.Hedegaard
H, Minino
AM, Warner
M. Drug overdose deaths in the United States, 1999-2017. NCHS Data Brief. 2018;329:1-8.
39.N’Guyen
Y, Duval
X, Revest
M,
et al; AEPEI Study Group. Time interval between infective endocarditis first symptoms and diagnosis: relationship to infective endocarditis characteristics, microorganisms and prognosis.
Ann Med. 2017;49(2):117-125. doi:
10.1080/07853890.2016.1235282PubMedGoogle ScholarCrossref 40.Hope
VD, Ncube
F, Parry
JV, Hickman
M. Healthcare seeking and hospital admissions by people who inject drugs in response to symptoms of injection site infections or injuries in three urban areas of England.
Epidemiol Infect. 2015;143(1):120-131. doi:
10.1017/S0950268814000284PubMedGoogle ScholarCrossref 42.Alagna
L, Park
LP, Nicholson
BP,
et al. Repeat endocarditis: analysis of risk factors based on the International Collaboration on Endocarditis—prospective cohort study.
Clin Microbiol Infect. 2014;20(6):566-575. doi:
10.1111/1469-0691.12395PubMedGoogle ScholarCrossref 44.Priest
KC, Lovejoy
TI, Englander
H, Shull
S, McCarty
D. Opioid agonist therapy during hospitalization within the Veterans Health Administration: a pragmatic retrospective cohort analysis.
J Gen Intern Med. 2020;35(8):2365-2374. doi:
10.1007/s11606-020-05815-0PubMedGoogle ScholarCrossref 45.Murphy
SMM, McCollister
KE, Leff
JA,
et al. Cost-effectiveness of buprenorphine-naloxone versus extended-release naltrexone to prevent opioid relapse.
Ann Intern Med. 2019;170(2):90-98. doi:
10.7326/M18-0227PubMedGoogle ScholarCrossref 47.Rodger
L, Glockler-Lauf
SD, Shojaei
E,
et al. Clinical characteristics and factors associated with mortality in first-episode infective endocarditis among persons who inject drugs.
JAMA Netw Open. 2018;1(7):e185220. doi:
10.1001/jamanetworkopen.2018.5220PubMedGoogle Scholar 51.Larochelle
MR, Bernson
D, Land
T,
et al. Medication for opioid use disorder after nonfatal opioid overdose and association with mortality: a cohort study.
Ann Intern Med. 2018;169(3):137-145. doi:
10.7326/M17-3107PubMedGoogle ScholarCrossref 52.Morgan
JR, Schackman
BR, Leff
JA, Linas
BP, Walley
AY. Injectable naltrexone, oral naltrexone, and buprenorphine utilization and discontinuation among individuals treated for opioid use disorder in a United States commercially insured population.
J Subst Abuse Treat. 2018;85:90-96. doi:
10.1016/j.jsat.2017.07.001PubMedGoogle ScholarCrossref 53.Wakeman
SE, Metlay
JP, Chang
Y, Herman
GE, Rigotti
NA. Inpatient addiction consultation for hospitalized patients increases post-discharge abstinence and reduces addiction severity.
J Gen Intern Med. 2017;32(8):909-916. doi:
10.1007/s11606-017-4077-zPubMedGoogle ScholarCrossref 56.Verhagen
DWM, Vedder
AC, Speelman
P, van der Meer
JTM. Antimicrobial treatment of infective endocarditis caused by
Viridans streptococci highly susceptible to penicillin: historic overview and future considerations.
J Antimicrob Chemother. 2006;57(5):819-824. doi:
10.1093/jac/dkl087PubMedGoogle ScholarCrossref 59.Hill
EE, Herijgers
P, Claus
P, Vanderschueren
S, Herregods
MC, Peetermans
WE. Infective endocarditis: changing epidemiology and predictors of 6-month mortality: a prospective cohort study.
Eur Heart J. 2007;28(2):196-203. doi:
10.1093/eurheartj/ehl427PubMedGoogle ScholarCrossref 61.Jiang
Y, McDonald
JV, Koziol
J, McCormick
M, Viner-Brown
S, Alexander-Scott
N. Can emergency department, hospital discharge, and death data be used to monitor burden of drug overdose in Rhode Island?
J Public Health Manag Pract. 2017;23(5):499-506. doi:
10.1097/PHH.0000000000000514PubMedGoogle ScholarCrossref 65.Jackson
KA, Bohm
MK, Brooks
JT,
et al. Invasive methicillin-resistant
Staphylococcus aureus infections among persons who inject drugs—six sites, 2005-2016.
MMWR Morb Mortal Wkly Rep. 2018;67(22):625-628. doi:
10.15585/mmwr.mm6722a2PubMedGoogle ScholarCrossref 67.Kadri
AN, Wilner
B, Hernandez
AV,
et al. Geographic trends, patient characteristics, and outcomes of infective endocarditis associated with drug abuse in the United States from 2002 to 2016.
J Am Heart Assoc. 2019;8(19):e012969. doi:
10.1161/JAHA.119.012969PubMedGoogle Scholar 71.Crawford
S, Bath
N. Peer support models for people with a history of injecting drug use undertaking assessment and treatment for hepatitis C virus infection.
Clin Infect Dis. 2013;57(suppl 2):S75-S79. doi:
10.1093/cid/cit297PubMedGoogle ScholarCrossref 72.Litwin
AH, Harris
KA
Jr, Nahvi
S,
et al. Successful treatment of chronic hepatitis C with pegylated interferon in combination with ribavirin in a methadone maintenance treatment program.
J Subst Abuse Treat. 2009;37(1):32-40. doi:
10.1016/j.jsat.2008.09.009PubMedGoogle ScholarCrossref 73.Lewis S, Liang SY, Schwarz ES, et al. Patients with serious injection drug use related infections who experience patient directed discharges on oral antibiotics have high rates of antibiotic adherence but require multidisciplinary outpatient support for retention in care.
Open Forum Infect Dis. 2022;9(2):ofab633. doi:
10.1093/ofid/ofab633