Prevalence of Buprenorphine Providers Requiring Cash Payment from Insured Women Seeking Opioid Use Disorder Treatment

Abstract

Context:

Medications for opioid use disorder (OUD) are known to be effective, especially for reducing the risk of overdose death. Yet, many individuals suffering from OUD are not receiving treatment. One potential barrier can be patients’ ability to access providers through their insurance plans.

Methods:

We used an audit (simulated patient) study methodology to examine appointment granting behavior by buprenorphine prescribers in 10 different US states. Trained callers posed as women with OUD and were randomly assigned Medicaid or private insurance status. Callers request an OUD treatment appointment and then asked if they would be able to use their insurance to cover the cost of care, or alternatively, if they would be required to pay fully out-of-pocket.

Findings:

We found that Medicaid and privately insured women were often asked to pay cash for OUD treatment—40% of the time over the full study sample. Such buprenorphine provider requests happened more than 60% of the time in some states. Areas with more providers or with more generous provider payments were not obviously more willing to accept patients’ insurance benefits for OUD treatment. Rural providers were less likely to require payment in cash in order for the woman to receive care.

Conclusions:

State-to-state variation was the most striking pattern in our field experiment data. The wide variation suggests that women of reproductive age with OUD in certain states face even greater challenges to treatment access than perhaps previously thought; however, it also reveals that some states have found ways to curtail this problem. Our findings encourage greater attention for this public health challenge and possibly opportunities for shared-learning across states.

Introduction

The opioid crisis continues to take a devastating toll on communities across the US. In 2020, more Americans died of an overdose than any year on record, mostly from opioids.¹ For individuals with opioid use disorder (OUD), medications for OUD (e.g., buprenorphine) reduce the risk of relapse and overdose deaths.² While improving access and uptake of OUD medications is a public health priority,³ the majority of individuals with the disorder are not receiving them.^3–5 Buprenorphine is currently the most widely available OUD medication, partly due to its accessibility through outpatient settings and recent policy initiatives relaxing patient limits and waiver requirements; however, a multitude of access barriers exist even for this treatment option.^3–10

For example, in a recent study, we found that private as well as Medicaid insurance was often not accepted for women seeking treatment with buprenorphine, and scheduling an appointment was only successful after agreeing to pay cash.¹¹ A similar study of Medicaid and self-pay patient callers found better appointment success with cash payment.¹² With median out of pocket costs of up to $250 for the initial appointment^11,12 and nearly half of individuals with OUD below 200% of the federal poverty level,¹³ a healthcare market heavily relying on cash-based, self-financing can create additional challenges for individuals seeking effective and timely OUD treatment. Consequently, it also risks driving further health inequities for an already vulnerable population.

While there have been smaller-scale studies reporting preferences among buprenorphine prescribers for receiving payment in cash,^14–16 little research has characterized this issue with a wider geographic scope as well as been able to examine contextual features that may amplify or restrain this particular provider behavior. In what follows, we leverage unique and detailed data from an audit study of OUD treatment provider appointment granting behavior across 10 US states to better understand the prevalence and potential drivers of women’s inability to use their health insurance to receive needed OUD care. Doing so can shed more light on financial barriers to treatment access and may inform strategies to improve health among those suffering from OUD.

Data and Methods

We conducted a randomized field experiment using a simulated patient (“audit study”) methodology among ten states representing a range of opioid-related complications and state policies^11,17,18 (FL, KY, MA, MI, MO, NC, TN, VA, WA, and WV) from March through September 2019. The field experiment was specifically interested in the treatment access experiences of reproductive age women seeking care for OUD among buprenorphine (BUP) and opioid treatment program (OTP) providers.

Providers were randomly drawn from and contacted using the Substance Abuse and Mental Health Services Administration’s (SAMHSA) publicly available OUD treatment provider information. 6,062 unique BUP provider practices and 385 unique OTP practices spanning the ten states were included in the audit study. Each drawn provider was then paired with a trained female caller, who was randomly assigned pregnancy status and private or Medicaid insurance status. The caller sought to be scheduled for the next available new patient appointment for OUD treatment. Practice responses to this request were then translated into population-level measures of OUD treatment access. Full methodological details and related results can be found in Patrick et al. (2020).¹¹ We also provide the caller script and post-call data collection instrument in the supplementary appendix for this article.

For this study, our primary interests were documenting the prevalence of cash payment requests by BUP providers as well as how this provider behavior varied over different geographies and healthcare markets. We therefore limited analyses to completed and experimentally valid BUP provider practice audit calls from Patrick et al. (2020) that resulted in a new patient appointment being scheduled (2,312 calls in total). We defined our outcome of interest as a binary indicator equal to one when the insured caller was told she would need to pay out-of-pocket for the appointment and zero otherwise — meaning that her insurance (Medicaid or private) would be billed for the appointment. We subsequently used this outcome to calculate the rates of requiring cash for new OUD patients among BUP providers across different environments as well as the two insurance types. Because callers only offered to pay cash after they were told their insurance was not accepted, and the fixed cash-rate they were quoted for the appointment had an interquartile range of $160–$300, it is most likely the requested amount was to cover a visit rather than expected patient cost-sharing.

Our first analysis examined if BUP providers requiring out of pocket payments for treatment was more pronounced among certain states. We then stratified the analytic sample by the county-level BUP provider density (per 10,000 population). The underlying rationale was to investigate if greater competition (proxied by provider density) was associated with greater willingness to participate in public or private insurance plans for OUD care. To do so, we used our provider list to calculate the number of unique BUP providers within a given county per 10,000 population. We then divided the resulting distribution into five quintiles. We note that Patrick et al. (2020) found a markedly high rate of incorrect contact information on the BUP provider public list, which could challenge our provider density measure if counties with lots of entries also have disproportionately higher provider contact failure rates. However, in Supplemental Appendix Figure 1, we observe that is not the case. Failure rates were not correlated with the number of BUP providers. Our third stratification was based on a county-level measure of rurality. We considered a county to be a rural area if the percent of its population classified as rural by the 2010 census was in the top quartile of the distribution (effectively 33% or higher). Given the known OUD hardships facing much of rural America,^3,19–21 this seemed like a potentially important source of heterogeneity to examine. And finally, we used transaction data from a commercial claims database (IBM^® MarketScan^® Research Databases) to assess if cash payment requests tracked with local (Metropolitan Statistical Area (MSA)-level) median prices for standard evaluation and management (E&M) physician services among the ten field experiment states in 2018. Plausibly, more generous payments by insurers could lower the likelihood BUP providers require out-of-pocket payment from prospective OUD patients. For each empirical exercise described above, we also pooled pregnant and non-pregnant audit calls. Analyses that stratified on pregnancy status did not demonstrate clear differences, so we consolidated the two patient profiles within an insurance status in order to preserve cell sizes within our data stratifications of interest and to streamline results presentations.

Limitations

Although our analytic data benefit from their novelty, they are also limited to the selected ten states and a cross-sectional view. Accordingly, our findings may not generalize to other states or time periods, nor to patients outside of women of reproductive age. We additionally relied on SAMHSA’s public database of providers to simulate how patients would access treatment information—given that government officials frequently point to these data to connect patients to treatment. However, providers can opt out of the SAMHSA database, and we cannot assess whether providers opting out might respond differently to the audit experiment. We are also relying on tractable proxy measures of provider density (and hence competition) as well as insurance payment generosity. Such measures are inherently imperfect. Next, our analysis focused on provider acceptance of insurance. We have no information on the frequency on balance billing, in which patients seek some reimbursement from insurers for care delivered by out of network providers. Lastly, while we have intentionally explored the variation in this provider behavior in ways that are easily understood and visually displayed, we acknowledge that the findings in cannot be interpreted as causal or definitive. Instead, they are intended to draw greater attention to this healthcare access and public health issue and encourage more research focused on the phenomena we have described.

Results

Figure 1 illustrates the scope of the problem, showing that roughly 40% of buprenorphine (BUP) providers required that privately insured and Medicaid women pay out-of-pocket to receive an OUD treatment appointment across our 10 study states. However, these same rates varied wildly between states. Only one out of four BUP providers or less made such a request in Massachusetts, Washington, and West Virginia, but the majority (>60%) expected OUD patients to pay cash in Florida and Tennessee. Rates of requiring cash payment were often similar between privately insured and Medicaid callers in most states, although such provider requests were higher (with a statistically significant difference) for prospective Medicaid patients relative to privately insured callers in Florida and Michigan. Women relying on Medicaid insurance would be required to pay out-of-pocket to keep the available OUD treatment appointment with the contacted BUP provider approximately 16-ppt and 21-ppt more often in these two states, respectively.

Figure 1: State Level Estimates of Cash Requirements among Insured Women Granted an OUD Treatment Appointment.

^a “BUP” stands for buprenorphine provider. Analytic data were restricted to providers granting appointments to simulated patient callers (2,312 calls in total).

Figure 2 displays the percent of BUP providers requesting cash payment for OUD treatment visits for each payer type and across the distribution of waivered provider densities. Interestingly, rather than more plausibly competitive areas being more willing to accept the caller’s insurance benefits, the pattern is just the opposite until reaching the highest quintile of BUP providers per 10,000 population. Over 30% of calls to areas with the lowest BUP provider density led to a requirement for cash payment to keep the granted OUD treatment appointment; meanwhile, nearly 40–50% of calls in areas with the greater density made such a request, with the exception of counties with the greatest BUP provider density. Thus, there is no linear pattern between provider density quintiles and cash-payment request rates that could otherwise suggest greater competition restraining this particular provider behavior. If such a competitive force does exist, it would appear limited to areas with relatively high levels of BUP provider availability. Of note, defining provider density by unique practices (rather than unique clinicians) generates the same pattern of results as found in Figure 2.

Figure 2: Variation in Cash Requirements for Accepted New Patients by Provider Density.

^c “BUP” stands for buprenorphine provider. Analytic data were restricted to providers granting appointments to simulated patient callers. Provider density distribution was derived from the public SAMHSA OUD provider list and is per 10,000 county population.

Figure 3 demonstrates the rural to non-rural comparison for each payer type. Consistent with lower provider density areas showing a weaker propensity to request that patients pay out-of-pocket (Figure 2), BUP providers located in more rural areas are 9- to 17-percentage points less likely to engage in this behavior for privately insured and Medicaid callers, respectively (Figure 3). These differentials are also statistically significant (p < 0.01) at conventional levels. We also note that the more rural areas in Figure 3 most commonly fall in the first quintile of the provider density measure used in Figure 2.

Figure 3: Variation in Cash Requirements for Accepted New Patients by Rurality.

^c “BUP” stands for buprenorphine provider. Analytic data were restricted to providers granting appointments to simulated patient callers. A county is considered to be a rural area if the percent of its population classified as rural by the 2010 census was in the top quartile of the distribution (effectively 33% or higher).

Finally, in Figure 4, we examine the association between the share of BUP providers requiring cash payment for OUD treatment and our measure of the prevailing private insurance reimbursement rates in the local healthcare market. Neither the pattern for privately insured callers (top panel) nor the pattern for Medicaid callers (bottom panel) offers a clear relationship between these variables. From median price levels of $75 to $150 (where most of the data localize), a full spectrum is evident in terms of the share of calls requiring cash payment. These simple data depictions are, at the very least, inconsistent with an expectation that more generous payments by insurers would eliminate this specific BUP provider behavior.

Figure 4: Variation in Cash Requirements for Accepted New Patients by Privately Insured Payment Rates.

^d “BUP” stands for buprenorphine provider. Analytic data were restricted to providers granting appointments to simulated patient callers. Each bubble represents a unique MSA from the ten audit study states. Bubbles are weighted by the number of calls made to the specific MSA.

Discussion

Using a unique combination of data, we show that many insured individuals with OUD are having to navigate and finance their treatments as though they are uninsured. For a patient population known to commonly face severe resource constraints and weak support structures, requiring cash payment for services risks compounding their treatment struggles and disease burden.

Importantly, this issue is not isolated to those relying on Medicaid insurance plans, which are known to regularly face provider participation challenges.^22–25 Privately insured callers often experience requests to pay out-of-pocket for their OUD care at similar rates to those insured by Medicaid in the same areas. Potential mitigators, such as more buprenorphine prescribers or higher reimbursement rates, are not strongly associated with lower rates of required cash payments. BUP providers in more rural areas engage in this behavior less, which also aligns with previous research demonstrating that rural practices tend to offer greater access to more disadvantaged patient populations.^26–28

However, the widest discrepancies in the rates of requiring cash payment occur between states in our study sample (Figure 1). Whether (and which) policy levers and/or provider norms drive such marked differences between states remains unclear, but the data patterns at least suggest that a reliance on self-financing for OUD treatment does not have to be the prevailing experience for these individuals. OUD-related public health efforts to address the issue may involve better tracking of patient access, comparable to what we have done in this study, and accompanying policy initiatives by private insurance regulators and state policymakers. Otherwise, efforts to expand insurance coverage with OUD treatment benefits may have limited practical impact for many hoping to seek care that can put an end to their opioid addiction.

As the nation grapples with an escalating opioid crisis, with record setting rates of overdose, there is an urgent need for policy solutions that facilitate receipt of medications for OUD which have been shown to reduce risk of overdose and improve myriad clinical outcomes. In our analysis, the surprisingly limited effect of provider density and reimbursement on patient acceptance suggest that novel approaches may be needed to improve receipt of these medications.