Key Points
Question
Among privately insured adults using buprenorphine, is switching to a high-deductible health plan (HDHP) associated with changes in buprenorphine dispensing or health care visits for opioid use disorder (OUD)?
Findings
In this repeated cross-sectional study of 14 801 patient claims from a national commercial claims database, switching to an HDHP was associated with decreased buprenorphine dispensing but not with changes in OUD-related health care visits. Among patients experiencing above-median deductible increases, switching to an HDHP was associated with decreased buprenorphine dispensing and OUD-related outpatient visits.
Meaning
Results of this study suggest that patients who switch to an HDHP may experience heightened barriers to buprenorphine use.
Abstract
Importance
Buprenorphine can prevent opioid overdose deaths among patients with opioid use disorder (OUD). Among privately insured patients using buprenorphine, it is unclear whether switching from a non–high-deductible health plan (HDHP) to an HDHP is associated with changes in buprenorphine dispensing or OUD-related health care visits.
Objective
To evaluate the association between switching to an HDHP, buprenorphine dispensing, and OUD-related health care visits.
Design, Setting, and Participants
Repeated cross-sectional difference-in-differences analysis of the 2010 to 2023 Optum Labs Data Warehouse, a longitudinal clinical database with deidentified claims. Analyses included privately insured adults who were continuously enrolled throughout a baseline and follow-up year and had buprenorphine dispensing in the baseline year. The treatment group included patients who switched from a non-HDHP to an HDHP. The control group included patients who remained in a non-HDHP. Data were analyzed from January 1, 2025, through February 1, 2026.
Exposure
Switching to an HDHP.
Main Outcomes and Measures
Annual number of days with active buprenorphine prescriptions; annual number of OUD-related outpatient visits; annual number of OUD-related emergency department visits or hospitalizations. Linear models with patient fixed effects compared changes in outcomes among the treatment and control groups. In a subgroup analysis, the treatment group was limited to patients experiencing a deductible increase exceeding the median of $1250.
Results
Among 14 801 included patients (9419 [63.6%] male), switching to an HDHP was associated with a differential decrease in the number of days with active buprenorphine prescriptions (−29.0; 95% CI, −35.0 to −22.9) but not with changes in OUD-related outpatient visits (−0.5; 95% CI, −1.0 to 0.05) or OUD-related emergency department visits and hospitalizations (4.0 events per 100 patients; 95% CI, −0.5 to 8.4 events per 100 patients). In the subgroup analysis, switching to an HDHP was associated with a differential decrease in the number of OUD-related outpatient visits (−1.0; 95% CI, −1.8 to −0.3). Results for other outcomes were similar to the overall analysis.
Conclusions and Relevance
In this repeated cross-sectional study using difference-in-differences analysis of privately insured patients using buprenorphine, those who switched to an HDHP were more likely to decrease buprenorphine use. When deductible increases were large, patients who switched to an HDHP were also more likely to decrease the number of OUD-related outpatient visits. Findings suggest that switching to an HDHP may be associated with heightened barriers to OUD treatment.
This repeated cross-sectional study assesses the association of switching to a high-deductible health plan with buprenorphine use and opioid use disorder-related health care visits.
Introduction
Buprenorphine is a medication that can reduce opioid-related mortality by 50% among patients with opioid use disorder (OUD).1 Despite this, national data indicate that only 1 in 5 US residents with OUD receive treatment with buprenorphine or any other medication for OUD.2 Moreover, even among patients who receive treatment with buprenorphine, discontinuation is common,3 raising the risk of opioid overdose.4 Addressing underuse of buprenorphine is a crucial clinical and public health goal, as the number of US opioid overdose deaths remains substantial despite recent decreases.5
In 2023, 42% of privately insured US residents younger than 65 years were enrolled in a high-deductible health plan (HDHP) – defined in 2023 as plans with a minimum deductible of $1500 for individual coverage and $3000 for family coverage.6 Several studies suggest that switching to an HDHP is associated with decreased use of high-value services among patients with chronic diseases, as well as with worsened health outcomes in some patients.7,8,9 These findings raise the possibility that switching from a non-HDHP to an HDHP may be associated with decreased buprenorphine dispensing and increases in opioid-related adverse events among privately insured patients using buprenorphine.
To date, this possibility has been incompletely evaluated. In a quasi-experimental study10 of commercial claims from 2007 to 2017, an employer’s offer of an HDHP was not associated with changes in buprenorphine dispensing or the number of OUD-related outpatient visits among patients initiating buprenorphine therapy. However, no study of which we are aware has evaluated the association between switching to an HDHP and these outcomes. Additionally, the association between switching to an HDHP and OUD-related emergency department (ED) visits and hospitalizations, a potential indicator of opioid-related adverse events, remains unclear. Finally, it is unknown whether patients who switch to an HDHP and experience large increases in deductibles have greater changes in buprenorphine dispensing or OUD-related health care visits compared with those who switch to an HDHP and experience small increases in deductibles. One might expect the changes to be greater among patients experiencing a large deductible increase (eg, patients switching from a plan with no deductible to an HDHP) compared with those experiencing a small deductible increase (eg, patients switching from a plan with a moderate deductible to an HDHP), as the former would experience greater exposure to the price of health care after the switch.
This study evaluated whether switching to an HDHP is associated with changes in buprenorphine dispensing and OUD-related outpatient visits, ED visits, and hospitalizations among privately insured adults with pre-existing buprenorphine use. Analyses were conducted both overall and among patients experiencing larger vs smaller increases in deductibles.
Methods
Data Source
We analyzed the 2010 to 2023 Optum Labs Data Warehouse, which reports fully adjudicated claims from more than 300 million commercial and Medicare Advantage enrollees. Only commercially insured patients were included in this analysis. Data elements include a unique identifier for patients, employers, and plans, as well as information on insurance benefit design, including the in-network plan deductible, the amount of co-insurance required for patients, and the annual out-of-pocket maximum. Because data were deidentified, the Institutional Review Board of the University of Michigan Medical School exempted this study from review; informed consent was not required. This manuscript follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies.11
Sample
We identified all 2-year spans of continuous enrollment that began on January 1 of a baseline year and extended through December 31 of the subsequent year (follow-up year). Spans were eligible if the baseline year occurred from 2010 through 2022. We excluded spans if patients lacked pharmacy coverage in either the baseline or follow-up year, were less than 18 years old in the baseline year, changed employers between the baseline and follow-up year, switched from an individual to family plan or vice versa, or had multiple employers or plans during either the baseline or follow-up year. Because analyses focused on patients with pre-existing buprenorphine use, we limited analyses to spans in which patients had at least 1 dispensed buprenorphine prescription during the baseline year. We further excluded spans in which patients remained in an HDHP in both years, leaving spans in which patients either stayed in a non-HDHP or switched from a non-HDHP to an HDHP. Finally, we limited to each patient’s earliest span, thus ensuring the unit of analysis was the patient. Patients were not excluded if they had buprenorphine dispensing prior to the baseline year.
HDHPs were identified based on whether the plan’s in-network general deductible was equal to or greater than the statutorily defined minimum for HDHPs (eMethods 1 in Supplement 1). We did not require the presence of a health savings account or health retirement account. The general plan deductible was the amount patients or families must pay out-of-pocket before coverage occurs.
Study Design
We conducted a difference-in-differences analysis. The treatment group included patients who switched from a non-HDHP to an HDHP on January 1 of the follow-up year. The control group included patients who remained in a non-HDHP. Analyses compared within-patient changes in outcomes between the baseline and follow-up years in the treatment group to the corresponding change in the control group.
Outcomes
The first primary outcome was the annual number of days covered by an active buprenorphine prescription. A buprenorphine prescription was considered active from the date of dispensing through the date of dispensing plus days supplied minus 1. If multiple buprenorphine prescriptions were active on a given date, the date only counted once toward the outcome. In the baseline year, only buprenorphine prescriptions dispensed during the year contributed to the count of days with active prescriptions. In the follow-up year, this count included buprenorphine prescriptions dispensed during the year, as well as prescriptions that were dispensed during the baseline year and extended through the follow-up year. We did not include prescriptions for buprenorphine formulations approved to treat pain as opposed to OUD.
The second primary outcome was the annual number of OUD-related outpatient visits, defined as the number of outpatient visits with an International Statistical Classification of Diseases and Related Health Problems 10th Revision diagnosis code for OUD in any position. This outcome was included to explore whether any findings for switching to an HDHP and changes in buprenorphine dispensing may have been associated with changes in the number of outpatient visits in which a patient with OUD might be prescribed buprenorphine. The third primary outcome was the annual number of OUD-related ED visits or hospitalizations, an indicator of opioid-related adverse events. The number of OUD-related ED visits and hospitalizations was expressed per 100 patients owing to the rareness of these events. Measurement of OUD-related use began on January 1 of each year.
Finally, to contextualize results, we calculated mean out-of-pocket spending (sum of deductibles and copays) per dispensed buprenorphine prescription, per OUD-related outpatient visit, and per OUD-related ED visit and hospitalization. We also calculated total annual out-of-pocket spending across all medical and pharmacy claims, a measure of overall out-of-pocket burden. eMethods 2 in Supplement 1 reports additional details on outcome measurement.
Statistical Analysis
We used descriptive statistics to characterize the sample and to assess changes in out-of-pocket spending among the treatment and control groups. To estimate differential changes in primary outcomes, we fitted linear difference-in-differences models (eMethods 3 in Supplement 1). Models included an indicator for treatment group status, the interaction between this indicator and an indicator for follow-up year (coefficient of interest), year fixed effects, and patient fixed effects. Patient fixed effects were included to mitigate potential bias from time-invariant patient-level confounders that may be associated both with the decision to switch to an HDHP and with study outcomes. We used robust standard errors clustered at the patient level. Analyses used 2-sided hypothesis tests with α = .05 and were conducted using Stata version 17.1/MP (StataCorp).
In subgroup analyses, we repeated analyses when limiting the treatment group to patients experiencing a deductible increase exceeding $1250 vs a deductible increase of $1250 or less. The $1250 threshold demarcated the 50th percentile of the deductible increase in the treatment group. The control group was the same as in the main analysis.
The main analysis included patients with buprenorphine dispensing in the baseline year, regardless of the timing of dispensing. If the treatment group were more or less likely than the control group to discontinue buprenorphine before the end of the baseline year, this could lead to a misleading conclusion that there was a true association between switching to an HDHP and the study outcomes when in fact the association was based on another factor. To explore this possibility, we conducted a sensitivity analysis in which the sample was limited to patients with at least 1 buprenorphine prescription dispensed in the last quarter of the baseline year. In a second sensitivity analysis, we limited to patients with a baseline year from 2010 to 2018, thus ensuring the follow-up year occurred before the COVID-19 pandemic.
Results
Sample Characteristics
After sample exclusions (Figure), the sample included 14 801 patients (5382 [36.4%] female; 9419 [63.6%] male; mean [SD] age, 38.4 [12.2] years), including 1587 patients in the treatment group (mean [SD] age, 38.3 [12.3] years) and 13 214 patients in the control group (mean [SD] age, 38.8 [11.6] years). Table 1 reports sample characteristics, including rates of missing data for demographic variables. In both groups, most patients were male, resided in the South, and resided in urban areas. In the treatment group, the mean (SD) in-network general deductible increased from $1462 ($695) to $3163 ($1486) between the baseline and follow-up years compared with $833 ($661) to $884 ($677) in the control group. The first dispensed buprenorphine prescription in the baseline year was more likely to occur in January for the treatment group (946 [59.6%]) compared with the control group (5735 [43.4%]) (eFigure in Supplement 1).
Figure. Flow Diagram of Sample Inclusion and Exclusion Criteria.
HDHP indicates high-deductible health plan.
Table 1. Sample Characteristics.
| Characteristic | Patients, No. (%)a | P value | ||
|---|---|---|---|---|
| Total | Control: remained in a non-HDHP | Treatment: switched to an HDHP | ||
| Sample size | 14 801 | 13 214 | 1587 | |
| Age, mean (SD), y | 38.4 (12.2) | 38.3 (12.3) | 38.8 (11.6) | .12 |
| Sex | ||||
| Male | 9419 (63.6) | 8390 (63.5) | 1029 (64.8) | .41 |
| Femaleb | 5382 (36.4) | 4824 (36.5) | 558 (35.2) | |
| Region of residence | ||||
| Northeast | 2242 (16.5) | 2283 (17.3) | 159 (10.0) | <.001 |
| Midwest | 2493 (16.5) | 2212 (16.7) | 281 (17.7) | |
| South | 6229 (42.1) | 5426 (41.1) | 803 (50.6) | |
| West | 2489 (16.8) | 2238 (16.9) | 251 (15.8) | |
| Unknown | 1148 (7.8) | 1055 (8.0) | 93 (5.9) | |
| Community | ||||
| Urban | 13 000 (87.8) | 11 577 (87.6) | 1423 (89.7) | .86 |
| Rural | 194 (1.3) | 172 (1.3) | 22 (1.4) | |
| Unknown | 1607 (10.9) | 1465 (11.1) | 142 (9.0) | |
| Median household income by zip code, quartile | ||||
| 1 (Lowest) | 1709 (11.6) | 1480 (11.2) | 229 (14.4) | <.001 |
| 2 | 2280 (15.4) | 2030 (15.4) | 250 (15.8) | |
| 3 | 2777 (18.8) | 2469 (18.7) | 308 (19.4) | |
| 4 (Highest) | 5871 (39.7) | 5285 (40.0) | 586 (36.9) | |
| Unknown | 2164 (14.6) | 1950 (14.8) | 214 (13.5) | |
| Plan type | ||||
| Individual | 3993 (27.0) | 3625 (27.4) | 368 (23.2) | <.001 |
| Family | 10 808 (73.0) | 9589 (72.6) | 1219 (76.8) | |
| Deductible, mean (SD), $ | ||||
| Year 1 | 900 (693) | 833 (661) | 1462 (695) | <.001 |
| Year 2 | 1128 (1069) | 884 (677) | 3163 (1486) | |
| Baseline year | ||||
| 2010 | 1886 (12.7) | 1784 (13.5) | 102 (6.4) | <.001 |
| 2011 | 1508 (10.2) | 1412 (10.7) | 96 (6.1) | |
| 2012 | 1576 (10.7) | 1270 (9.6) | 306 (19.3) | |
| 2013 | 1219 (8.2) | 1092 (8.3) | 127 (8.0) | |
| 2014 | 1132 (7.7) | 939 (7.1) | 195 (12.3) | |
| 2015 | 910 (6.2) | 807 (6.1) | 103 (6.5) | |
| 2016 | 866 (5.9) | 776 (5.9) | 90 (5.7) | |
| 2017 | 966 (6.5) | 841 (6.4) | 125 (7.9) | |
| 2018 | 988 (6.7) | 878 (6.6) | 110 (6.9) | |
| 2019 | 1051 (7.1) | 943 (7.1) | 108 (6.8) | |
| 2020 | 984 (6.7) | 927 (7.0) | 57 (3.6) | |
| 2021 | 836 (5.7) | 777 (5.9) | 59 (3.7) | |
| 2022 | 877 (5.9) | 768 (5.8) | 109 (6.9) | |
Abbreviation: HDHP, high-deductible health plan.
Percentages may not total 100 due to rounding.
A very small number of patients had missing data for sex. Due to cell size restrictions, a category for unknown sex was not included in this table. Per instructions from Optum Labs, patients with missing sex were assigned female sex.
Out-of-Pocket Spending
For each measure assessed, out-of-pocket spending increased in the treatment group and decreased in the control group between the baseline and follow-up years (Table 2). For example, mean (SD) out-of-pocket spending per dispensed buprenorphine prescription increased from $42.0 ($37.2) to $51.8 ($77.3) in the treatment group but decreased from $31.0 ($24.5) to $22.5 ($24.1) in the control group. Mean (SD) out-of-pocket spending per OUD-related outpatient visit increased from $28.6 ($85.5) to $34.1 ($67.3) in the treatment group but decreased from $24.1 ($57.3) to $21.4 ($68.2) in the control group. Mean (SD) annual out-of-pocket spending across all medical and pharmacy claims increased from $2721.0 ($2420.4) to $3273.2 ($3489.2) in the treatment group but decreased from $2531.7 ($3464.4) to $2322.2 ($3191.9) in the control group. Among patients in the treatment group, the magnitudes of the increases in out-of-pocket spending were greater when deductible increases exceeded $1250 compared with when deductible increases were $1250 or less.
Table 2. Changes in Unadjusted Out-of-Pocket Spending Among Patients Using Buprenorphine Who Did and Did Not Switch to a High-Deductible Health Plan.
| Outcome | Mean (SD) | |||
|---|---|---|---|---|
| Treatment | Control | |||
| Baseline year | Follow-up year | Baseline year | Follow-up year | |
| All patients | ||||
| OOP spending per dispensed buprenorphine prescription | 42.0 (37.2) | 51.8 (77.3) | 31.0 (24.5) | 22.5 (24.1) |
| OOP spending per ED visit or hospitalization for OUD | 82.2 (335.3) | 90.6 (506.5) | 102.8 (436.6) | 68.5 (354.6) |
| OOP spending per outpatient visit for OUD | 28.6 (85.5) | 34.1 (67.3) | 24.1 (57.3) | 21.4 (68.2) |
| Annual OOP spending across all health care services | 2721.0 (2420.4) | 3273.2 (3489.2) | 2531.7 (3464.4) | 2322.2 (3191.9) |
| Subgroup analysis | ||||
| Deductible increased by more than $1250a | ||||
| OOP spending per dispensed buprenorphine prescription | 35.4 (28.0) | 58.6 (93.0) | 31.0 (24.5) | 22.5 (24.1) |
| OOP spending per ED visit or hospitalization for OUD | 86.8 (357.8) | 98.5 (575.1) | 102.8 (436.6) | 68.5 (354.6) |
| OOP spending per outpatient visit for OUD | 27.1 (44.3) | 36.6 (71.7) | 24.1 (57.3) | 21.4 (68.2) |
| Annual OOP spending across all health care services | 2598.3 (2426.2) | 3560.1 (3992.6) | 2531.7 (3464.4) | 2322.2 (3191.9) |
| Deductible increased by $1250 or lessb | ||||
| OOP spending per dispensed buprenorphine prescription | 48.2 (42.9) | 45.3 (58.0) | 31.0 (24.5) | 22.5 (24.1) |
| OOP spending per ED visit or hospitalization for OUD | 77.9 (312.8) | 83.0 (432.0) | 102.8 (436.6) | 68.5 (354.6) |
| OOP spending per outpatient visit for OUD | 29.9 (111.2) | 31.7 (62.9) | 24.1 (57.3) | 21.4 (68.2) |
| Annual OOP spending across all health care services | 2836.9 (2410.9) | 3002.0 (2911.8) | 2531.7 (3464.4) | 2322.2 (3191.9) |
Abbreviations: ED, emergency department; OOP, out-of-pocket; OUD, opioid use disorder.
The first subgroup analyses included treatment group patients whose deductible increased by more than $1250 and all control group patients.
The second subgroup analysis included treatment group patients whose deductible increased by $1250 or less and all control group patients.
Main Results
Table 3 reports unadjusted values of the primary outcomes in the baseline and follow-up year among the treatment and control groups, as well as adjusted difference-in-differences estimates. In the treatment group, the mean (SD) annual number of days with active buprenorphine prescriptions decreased from 214.3 (121.7) to 193.2 (143,5) days, whereas it increased from 182.4 (126.4) days to 190.3 (146.3) days in the control group. Switching to an HDHP was associated with a differential decrease in this outcome (−29.0 days, 95% CI, −35.0 to −22.9 days). The mean (SD) annual number of OUD-related outpatient visits decreased from 5.5 (9.2) to 5.1 (8.9) in the treatment group but increased from 5.3 (9.9) to 5.5 (11.9) in the control group. Switching to an HDHP was not associated with a differential change in this outcome (−0.5 visits, 95% CI, −1.0 to 0.05 visits). The mean (SD) annual number of OUD-related ED visits and hospitalizations per 100 patients decreased from 21.9 (90.7) to 15.8 (68.9) in the treatment group and from 29.5 (87.3) to 19.4 (77.9) in the control group. Switching to an HDHP was not associated with a differential change in this outcome (4.0 events per 100 patients; 95% CI, −0.5 to 8.4 events per 100 patients).
Table 3. Difference-in-Differences Estimates.
| Outcome | No., mean (SD) | Adjusted difference-in-differences estimate (95% CI) | |||
|---|---|---|---|---|---|
| Treatment | Control | ||||
| Baseline year | Follow-up year | Baseline year | Follow-up year | ||
| All patients | |||||
| Sample size | 1587 | 1587 | 13 214 | 13 214 | NA |
| Annual days with active buprenorphine prescriptions per patient | 214.3 (121.7) | 193.2 (143.5) | 182.4 (126.4) | 190.3 (146.3) | −29.0 (−35.0 to −22.9) |
| Annual OUD-related outpatient visits per patient | 5.5 (9.2) | 5.1 (8.9) | 5.3 (9.9) | 5.5 (11.9) | −0.5 (−1.0 to 0.05) |
| Annual OUD-related ED visits or hospitalizations per 100 patients | 21.9 (90.7) | 15.8 (68.9) | 29.5 (87.30 | 19.4 (77.9) | 4.0 (−0.5 to 8.4) |
| Subgroup analysis | |||||
| Deductible increased by more than $1250a | |||||
| Sample size | 771 | 771 | 13 214 | 13 214 | NA |
| Annual days with active buprenorphine prescriptions per patient | 211.3 (120.9) | 183.9 (145.3) | 182.4 (126.4) | 190.3 (146.3) | −35.2 (−43.7 to −26.7) |
| Annual OUD-related outpatient visits per patient | 5.9 (10.2) | 5.0 (9.3) | 5.3 (9.9) | 5.5 (11.9) | −1.0 (−1.8 to −0.3) |
| Annual OUD-related ED visits or hospitalizations per 100 patients | 24.5 (83.6) | 19.6 (75.9) | 29.5 (87.3) | 19.4 (77.9) | 5.1 (−1.1 to 11.4) |
| Deductible increased by $1250 or lessb | |||||
| Sample size | 816 | 816 | 13 214 | 13 214 | NA |
| Annual days with active buprenorphine prescriptions per patient | 217.1 (122.4) | 202.0 (141.3) | 182.4 (126.4) | 190.3 (146.3) | −23.0 (−31.3 to −14.8) |
| Annual OUD-related outpatient visits per patient | 4.7 (7.1) | 5.1 (8.1) | 5.3 (9.9) | 5.5 (11.9) | 0.1 (−0.7 to 0.8) |
| Annual OUD-related ED visits or hospitalizations per 100 patients | 19.5 (97.0) | 12.3 (61.5) | 29.5 (87.3) | 19.4 (77.9) | 2.8 (−3.2 to 8.9) |
Abbreviations: ED, emergency department; NA, not applicable; OUD, opioid use disorder.
The first subgroup analyses included treatment group patients whose deductible increased by more than $1250 and all control group patients.
The second subgroup analysis included treatment group patients whose deductible increased by $1250 or less and all control group patients.
Subgroup Analyses
There were 771 treatment group patients with deductible increases exceeding $1250; 691 (89.6%) were in family plans. When limiting the treatment group to these patients, the magnitude of the differential decrease in the annual number of days with active buprenorphine prescriptions increased (−35.2 days; 95% CI, −43.7 to −26.7). Moreover, switching to an HDHP was associated with a differential decrease in the number of OUD-related outpatient visits (−1.0 visits; 95% CI, −1.8 to −0.3 visits) unlike in the overall analysis. Switching to an HDHP was not associated with a differential change in the annual number of OUD-related ED visits and hospitalizations per 100 patients (5.1 events per 100 patients; 95% CI, −1.1 to 11.4 events per 100 patients).
There were 816 treatment group patients with deductible increases of $1250 or less; 528 were in family plans (64.7%). When limiting the treatment group to these patients, the difference-in-differences estimate for the annual number of days with active buprenorphine prescriptions remained negative (−23.0 days; 95% CI, −31.3 to −14.8). Difference-in-differences estimates for the 2 other primary outcomes were not significant.
Sensitivity Analyses
Among the 1587 patients receiving treatment, 1234 had buprenorphine dispensing in the last quarter of the baseline year (77.8%), and among 13 214 control patients, 10 038 had buprenorphine dispensing in the last quarter of the baseline year (76.0%). When limiting analyses to these patients, findings were similar to the main analysis. The findings were also similar when limiting to prepandemic data (eTables 1-2 in Supplement 1).
Discussion
In this difference-in-differences analysis of a national commercial claims database, switching to an HDHP was associated with decreased buprenorphine dispensing but not with changes in the number of OUD-related outpatient visits, ED visits, or hospitalizations among privately insured adults with pre-existing buprenorphine use. Among patients experiencing above-median deductible increases, switching to an HDHP was associated with a differential decrease in both buprenorphine dispensing and the number of OUD-related outpatient visits.
Many studies have evaluated the association between switching to an HDHP and the use of medications for chronic diseases, with some showing reductions in use and others showing no change.8,12,13,14,15,16,17 The current study adds to this literature with the finding of an association between switching to an HDHP and decreased buprenorphine dispensing. Findings contrast with a prior analysis of commercial claims data, which showed that an employer offer of an HDHP was not associated with changes in buprenorphine dispensing.10 A potential explanation is that the current analysis examined patients who made the switch to an HDHP, not those who were newly offered the option to enroll in an HDHP.
Prior studies have reported seemingly contradictory findings regarding the degree to which buprenorphine dispensing changes as a function of cost-sharing, defined as the amount insurers require patients to pay to fill prescriptions. On the one hand, several groups of authors used commercial claims databases to demonstrate that the amount paid out-of-pocket for buprenorphine prescriptions – an indicator of cost-sharing – is negatively associated with retention in buprenorphine therapy.18,19 On the other hand, other authors used a national pharmacy transaction database to show that the association between cost-sharing and patients’ decisions on whether to fill a buprenorphine prescription was either null or small at most.20,21
These contradictory findings might be reconciled under 2 conditions: (1) high out-of-pocket spending for buprenorphine is an indicator that patients are enrolled in a nongenerous plan that requires high cost-sharing not just for buprenorphine but also for other OUD care, including the outpatient visits during which buprenorphine is prescribed; and (2) use of these outpatient visits is sensitive to cost-sharing. In support of the first condition, the current study found that out-of-pocket spending for buprenorphine and OUD-related outpatient visits both increased among patients who switched to an HDHP, particularly among those experiencing above-median deductible increases. In support of the second condition, the current study found that switching to an HDHP was associated with a differential decrease in both buprenorphine dispensing and in the number of OUD-related outpatient visits among patients experiencing above-median deductible increases. Thus, the finding in prior studies that high out-of-pocket spending for buprenorphine is associated with reduced treatment retention may reflect sensitivity of OUD-related outpatient visits to cost-sharing, as opposed to sensitivity of buprenorphine dispensing to cost-sharing.18,19
In this study, switching to an HDHP was not associated with a differential change in the annual number of OUD-related ED visits and hospitalizations. However, the point estimate (4.0 events per 100 patients) was positive and the lower bound of the 95% CI was −0.5. This finding suggests that a better-powered study may have detected a significant increase.
This study has policy implications. Findings suggest that patients using buprenorphine who switch to an HDHP are at greater risk of experiencing barriers to OUD treatment compared with those who do not switch. To mitigate this risk, employers and private insurers might consider providing first-dollar coverage for buprenorphine and other OUD-related services, meaning that coverage would be provided even if the deductible has not yet been met. This approach has precedent, as the US Internal Revenue Service allows first-dollar coverage of preventive services for chronic conditions in HDHPs.22
Limitations
This study has limitations. First, there were differences in the observed characteristics of patients who did and who did not switch to HDHPs in this study, including differences in the baseline intensity of buprenorphine use and level of deductibles. This raises the possibility that the 2 groups differed on unobserved factors associated with the study’s outcomes. Bias could occur if these differences change to a greater or lesser degree among 1 group of patients compared with the other. For example, if the severity of OUD improved in the treatment group but not the control group, this could result in a differential decrease in buprenorphine dispensing that is unrelated to the switch to an HDHP. Notably, however, the sign of the difference-in-differences estimate for the number of OUD-related ED visits and hospitalizations was positive, which does not suggest that OUD severity improved differentially in the treatment group.
Second, the database lacked information on prescription indication. Thus, analyses could not be restricted to patients using buprenorphine for OUD as opposed to pain. Third, because each patient only contributed data from 2 time points, the parallel trends assumption of difference-in-differences analysis could not be evaluated. Fourth, claims data do not capture cash-pay buprenorphine prescriptions, which may have been more common in the follow-up year among patients in the treatment group. Fifth, findings are limited to patients with existing buprenorphine use. Additional research is needed to assess whether switching to an HDHP is associated with initiation of buprenorphine therapy. Sixth, owing to small sample sizes, analyses could not be limited to patients whose employer forced them to switch into an HDHP. Seventh, small sample sizes also precluded the ability to analyze whether changes differed among patients enrolled in individual vs family plans.
Conclusions
In this study, switching to an HDHP was associated with decreased buprenorphine dispensing among privately insured adults with pre-existing buprenorphine use, as well as with decreased OUD-related outpatient visits among patients experiencing above-median deductible increases. Findings suggest that switching to an HDHP may be associated with heightened barriers to OUD treatment.
eMethods 1. Definition of high-deductible health plans
eMethods 2. Details on measurement of opioid use disorder-related health care
eMethods 3. Details on model specification
eTable 1. Sensitivity analysis limiting to patients with buprenorphine dispensing in the last quarter of the baseline year
eTable 2. Sensitivity analysis limiting to patients with baseline years from 2010-2018
eFigure. Month of first dispensed buprenorphine prescription in the baseline year
Data Sharing Statement
References
- 1.Chan B, Gean E, Arkhipova-Jenkins I, et al. Retention strategies for medications for opioid use disorder in adults: a rapid evidence review. J Addict Med. 2021;15(1):74-84. doi: 10.1097/ADM.0000000000000739 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Jones CM, Han B, Baldwin GT, Einstein EB, Compton WM. Use of medication for opioid use disorder among adults with past-year opioid use disorder in the US, 2021. JAMA Netw Open. 2023;6(8):e2327488. doi: 10.1001/jamanetworkopen.2023.27488 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chua KP, Nguyen TD, Zhang J, Conti RM, Lagisetty P, Bohnert AS. Trends in buprenorphine initiation and retention in the United States, 2016-2022. JAMA. 2023;329(16):1402-1404. doi: 10.1001/jama.2023.1207 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sordo L, Barrio G, Bravo MJ, et al. Mortality risk during and after opioid substitution treatment: systematic review and meta-analysis of cohort studies. BMJ. 2017;357:j1550. doi: 10.1136/bmj.j1550 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Centers for Disease Control and Prevention . Provisional drug overdose death counts. Accessed October 1, 2023. https://www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data.htm
- 6.Cohen RA, Briones EM. Enrollment in high-deductible health plans among people younger than age 65 with private health insurance. National Center for Health Statistics . December 5, 2024. Accessed October 1, 2025. https://www.cdc.gov/nchs/data/nhsr/nhsr214.pdf [DOI] [PMC free article] [PubMed]
- 7.Wharam JF, Wallace J, Zhang F, et al. Association between switching to a high-deductible health plan and major cardiovascular outcomes. JAMA Netw Open. 2020;3(7):e208939. doi: 10.1001/jamanetworkopen.2020.8939 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wharam JF, Zhang F, Eggleston EM, Lu CY, Soumerai S, Ross-Degnan D. Diabetes outpatient care and acute complications before and after high-deductible insurance enrollment: a natural experiment for translation in diabetes (NEXT-D) study. JAMA Intern Med. 2017;177(3):358-368. doi: 10.1001/jamainternmed.2016.8411 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wharam JF, Zhang F, Eggleston EM, Lu CY, Soumerai SB, Ross-Degnan D. Effect of high-deductible insurance on high-acuity outcomes in diabetes: a natural experiment for translation in diabetes (NEXT-D) study. Diabetes Care. 2018;41(5):940-948. doi: 10.2337/dc17-1183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kennedy-Hendricks A, Schilling CJ, Busch AB, et al. Impact of high deductible health plans on continuous buprenorphine treatment for opioid use disorder. J Gen Intern Med. 2022;37(4):769-776. doi: 10.1007/s11606-021-07094-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Equator Network . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. Accessed March 5, 2019. http://www.equator-network.org/reporting-guidelines/strobe/
- 12.Agarwal R, Mazurenko O, Menachemi N. High-deductible health plans reduce health care cost and utilization, including use of needed preventive services. Health Aff (Millwood). 2017;36(10):1762-1768. doi: 10.1377/hlthaff.2017.0610 [DOI] [PubMed] [Google Scholar]
- 13.Brot-Goldberg ZC, Chandra A, Handel BR, Kolstad JT. What does a deductible do? the impact of cost-sharing on health care prices, quantities, and spending dynamics. Q J Econ. 2017;132(3):1261-1318. doi: 10.1093/qje/qjx013 [DOI] [Google Scholar]
- 14.Fendrick AM, Buxbaum JD, Tang Y, et al. Association between switching to a high-deductible health plan and discontinuation of type 2 diabetes treatment. JAMA Netw Open. 2019;2(11):e1914372. doi: 10.1001/jamanetworkopen.2019.14372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Galbraith AA, Ross-Degnan D, Zhang F, et al. Controller medication use and exacerbations for children and adults with asthma in high-deductible health plans. JAMA Pediatr. 2021;175(8):807-816. doi: 10.1001/jamapediatrics.2021.0747 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lu CY, Busch AB, Zhang F, et al. Impact of high-deductible health plans on medication use among individuals with bipolar disorder. Psychiatr Serv. 2021;72(8):926-934. doi: 10.1176/appi.ps.202000362 [DOI] [PubMed] [Google Scholar]
- 17.Wharam JF, Busch AB, Madden J, et al. Effect of high-deductible insurance on health care use in bipolar disorder. Am J Manag Care. 2020;26(6):248-255. doi: 10.37765/ajmc.2020.43487 [DOI] [PubMed] [Google Scholar]
- 18.Dunphy C, Peterson C, Zhang K, Jones CM. Do out-of-pocket costs influence retention and adherence to medications for opioid use disorder? Drug Alcohol Depend. 2021;225:108784. doi: 10.1016/j.drugalcdep.2021.108784 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Leech AA, McNeer E, Roberts AW, et al. Buprenorphine out-of-pocket costs and discontinuation in privately insured adults with opioid use disorder. JAMA Intern Med. 2023;183(9):1023-1026. doi: 10.1001/jamainternmed.2023.2826 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chua KP, Conti RM, Lagisetty P, Bohnert ASB, Nuliyalu U, Nguyen TD. Association between cost-sharing and buprenorphine prescription abandonment. J Gen Intern Med. 2024;39(12):2160-2168. doi: 10.1007/s11606-024-08819-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Nguyen TD, Conti RM, Lagisetty P, Bohnert ASB, Nuliyalu U, Chua KP. Cost-sharing and buprenorphine prescription dispensing. JAMA Health Forum. 2025;6(7):e251913. doi: 10.1001/jamahealthforum.2025.1913 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Internal Revenue Service . Publication 969 (2025), health savings accounts and other tax-favored health plans. Accessed October 7, 2025. https://www.irs.gov/publications/p969
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods 1. Definition of high-deductible health plans
eMethods 2. Details on measurement of opioid use disorder-related health care
eMethods 3. Details on model specification
eTable 1. Sensitivity analysis limiting to patients with buprenorphine dispensing in the last quarter of the baseline year
eTable 2. Sensitivity analysis limiting to patients with baseline years from 2010-2018
eFigure. Month of first dispensed buprenorphine prescription in the baseline year
Data Sharing Statement
