INTRODUCTION
Medications are highly effective in treating opioid use disorder (OUD) but generally require long-term use to prevent relapse and fatalities.1,2 Health insurance disenrollment poses a significant challenge to sustaining treatment and can lead to high-risk treatment discontinuation and transitions in care.2–5 In this study, we examined disenrollment from membership in an integrated insurance plan and health care delivery system (Kaiser Permanente Colorado) and its association with all-cause mortality among patients prescribed buprenorphine for OUD.
METHODS
We conducted a retrospective cohort study of patients aged 18–64 who received buprenorphine for OUD between January 1, 2014 and December 31, 2020, and were enrolled in Kaiser Permanente Colorado, which serves more than 545,000 members. We required at least 45 days of enrollment prior to buprenorphine treatment initiation to collect baseline characteristics. Buprenorphine prescriptions (excluding transdermal patches indicated for pain); insurance enrollment and type (employer-sponsored commercial, individual pay, Medicaid, and other); race, ethnicity, gender, and age; mental health disorders; and drug, tobacco, and alcohol use disorders were identified from pharmacy and electronic health records data. Deaths were identified from state vital records for cohort members regardless of enrollment status. A buprenorphine treatment episode began on the date of first dispensing and ended on the date of final supply, with no more than 45 days between fills. Follow-up started from the first dispensing, could have included multiple treatment episodes, and ended 1 year after disenrollment or upon death. Disenrollment represents discontinuation of health insurance or care coverage from the health system. The primary exposure was a time-varying measure of enrollment and buprenorphine treatment status: enrolled and on treatment, enrolled and off treatment, and disenrolled. We calculated crude mortality rates (CMR) (per 100 person-years) by status. To assess the association of insurance disenrollment and mortality, we fitted a Cox proportional hazards model to calculate adjusted hazard ratios (aHRs) and 95% CIs. Statistical significance was based on 2-sided tests with a threshold of P < .05.
This data-only study was approved by the Kaiser Permanente Colorado Institutional Review Board with a waiver of informed consent.
RESULTS
The study cohort consisted of 1,338 patients who received buprenorphine for OUD (mean age [SD], 34.1 [11.1] years; 62.6% male; 66.2% white; 65.3% commercially insured, 10.2% on Medicaid only) (Table 1). There were 850 (63.5%) patients with disenrollment during follow-up, and 332 (24.8%) with disenrollment during a treatment episode. Cumulative days spent on treatment averaged 308.7, and the average number of treatment episodes was 1.4. There were 46 observed deaths (CMR, 1.14 per 100 person-years) (Table 2), of which 14 (30.4%) occurred within 1 year after disenrollment and 21 (45.7%) were opioid overdose-related deaths. The mortality rate was higher during enrollment and off treatment (CMR, 1.70 per 100 person-years) and within 1 year after disenrollment (CMR, 1.59 per 100 person-years) compared to during enrollment and on treatment (CMR, 0.27 per 100 person-years). Time during disenrollment was independently associated with a more than fivefold increase in mortality risk (aHR, 5.22; 95% CI, 1.70–16.08). The magnitude of this association was comparable to time during enrollment and off treatment (aHR, 5.31; 95% CI, 1.83–15.33).
Table 1.
Characteristics of Study Population
| Characteristic | Overall | No disenrollment | Any disenrollment |
|---|---|---|---|
| N=1,338 (100%) | N=488 (36.5%) | N=850 (63.5%) | |
| No. (%) | No. (%) | No. (%) | |
| Age at study entry (mean [SD]) | 34.1 (11.1) | 36.2 (11.5) | 32.9 (10.6) |
| Male sex | 837 (62.6) | 297 (60.9) | 540 (63.5) |
| Race | |||
| White | 886 (66.2) | 335 (68.6) | 551 (64.8) |
| Non-White* | 131 (9.8) | 52 (10.7) | 79 (9.3) |
| Missing | 321 (24.0) | 101 (20.7) | 220 (25.9) |
| Ethnicity† | |||
| Hispanic | 252 (18.8) | 92 (18.9) | 160 (18.8) |
| Non-Hispanic | 911 (68.1) | 359 (73.6) | 552 (64.9) |
| Missing | 175 (13.1) | 37 (7.6) | 138 (16.2) |
| Insurance type at study entry‡ | |||
| Commercial | 874 (65.3) | 308 (63.1) | 566 (66.6) |
| Medicaid | 137 (10.2) | 52 (10.7) | 85 (10.0) |
| Individual pay | 141 (10.5) | 48 (9.8) | 93 (10.9) |
| Other | 186 (13.9) | 80 (16.4) | 106 (12.5) |
| Alcohol use disorder | 195 (14.6) | 69 (14.1) | 126 (14.8) |
| Tobacco use disorder† | 428 (32.0) | 140 (28.7) | 288 (33.9) |
| Drug use disorder, other than opioid use disorder | 510 (38.1) | 170 (34.8) | 340 (40.0) |
| Mental health disorders | 558 (41.7) | 219 (44.9) | 339 (40.0) |
| Total treatment episodes (mean [SD])§ | 1.4 (0.8) | 1.4 (0.9) | 1.4 (0.8) |
| Total days on treatment (mean [SD])‖ | 308.8 (441.2) | 309.2 (443.7) | 298.6 (355.9) |
Abbreviation: SD, standard deviation
*Includes Black/African American, American Indian or Alaska Native, Asian, Native Hawaiian or other Pacific Islander, and other race
†Ethnicity (P<0.001) and tobacco use disorder (P<0.05) were significantly different between the no and any disenrollment groups
‡Commercial plans are employer-sponsored; individual pay indicates plans where individuals are responsible for all or nearly all the cost of premiums; and other represents coverage by multiple plans
§A buprenorphine treatment episode began on the date of first dispensing and ended on the date of final supply, with no more than 45 days between fills
‖Total days accumulated across all buprenorphine treatment episodes
Table 2.
Crude Mortality Rates and Adjusted Relative Risks by Insurance Enrollment and Buprenorphine Treatment Status
| Person-years (PY) | Deaths | Rate per 100 PY (95% CI) | aHR† (95% CI) | |
|---|---|---|---|---|
| Enrolled and on treatment | 1501 | 4 | 0.27 (0.10, 0.71) | 1.00 [reference] |
| Enrolled and off treatment | 1647 | 28 | 1.70 (1.17, 2.46) | 5.31 (1.83, 15.33) |
| Disenrolled* | 883 | 14 | 1.59 (0.94, 2.68) | 5.22 (1.70, 16.08) |
| Total | 4031 | 46 | 1.14 (0.85, 1.52) | — |
Abbreviation: aHR, adjusted hazard ratios
*Includes up to 1 year of follow-up after insurance disenrollment
†Model adjusted for age, gender, race, ethnicity, the number of buprenorphine treatment spells, insurance type, mental health disorders, tobacco use disorder, alcohol use disorder, and drug use disorder other than opioid use disorder
DISCUSSION
In this cohort of patients in one site, disenrollment from the health system was common. Disenrollment amid an episode of buprenorphine treatment occurred in a quarter of the cohort. Time during disenrollment was associated with heightened mortality risk, along with time during enrollment and off treatment. While we do not know whether individuals disenrolled due to insurance loss or changes in insurance, results suggest that treatment discontinuation is high-risk and that insurance disenrollment may impede buprenorphine treatment retention3 by inducing high-risk gaps or transitions in care. Disenrollment may also be a marker for increased drug use, which may contribute to volatile employment4, subsequent insurance instability, treatment disengagement, and mortality.
Limitations of the study include lack of capture of insurance coverage and care utilization outside the health system among Medicaid recipients and after disenrollment and buprenorphine prescriptions paid by cash. There is also potential confounding due to differential risk of disenrollment that may also be associated with heightened mortality.
Growing evidence demonstrates deleterious effects of insurance instability6, but more research is needed to understand its burden and associated health consequences in patients receiving OUD treatment compared to other patient populations.
Funding
This study was supported by an internal departmental grant from the Institute for Health Research at Kaiser Permanente Colorado and a Research Diversity Supplement from the National Institute on Drug Abuse of the National Institutes of Health under parent grant award number R01DA042059. The National Institutes of Health did not contribute to the conception of this commentary, writing, or the decision to submit this article for publication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Declarations
None
Conflict of Interest
Dr. Binswanger receives royalties for educational content on the health of incarcerated persons from UpToDate and is employed by Colorado Permanente Medical Group. All other authors are employed by Kaiser Permanente.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.National Academies of Sciences, Engineering, and Medicine . Medications for Opioid Use Disorder Save Lives. Washington, DC: National Academies Press; 2019. [PubMed] [Google Scholar]
- 2.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. [DOI] [PMC free article] [PubMed]
- 3.Manhapra A, Agbese E, Leslie DL, Rosenheck RA. Three-year retention in buprenorphine treatment for opioid use disorder among privately insured adults. Psychiatr Serv. 2018;69(7):768–776. doi: 10.1176/appi.ps.201700363. [DOI] [PubMed] [Google Scholar]
- 4.Becker W, Fiellin D, Merrill J, et al. Opioid use disorder in the United States: insurance status and treatment access. Drug Alcohol Depend. 2008;94(1-3):207–2013. doi: 10.1016/j.drugalcdep.2007.11.018. [DOI] [PubMed] [Google Scholar]
- 5.Williams AR, Samples H, Crystal S, Olfson M. Acute care, prescription opioid use, and overdose following discontinuation of long-term buprenorphine treatment for opioid use disorder. Am J Psychiatry. 2020;177(2):117–124. doi: 10.1176/appi.ajp.2019.19060612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yabroff KR, Zhao J, Halpern MT, et al. Health insurance disruptions and care access and affordability in the U.S. Am J Prev Med. 2021;61(1):3–12. doi: 10.1016/j.amepre.2021.02.014. [DOI] [PubMed] [Google Scholar]