Examining Substance Use Disorder Treatment Demand and Provider Capacity in a Changing Health Care System: Final Report. II. DEMAND FOR SUBSTANCE USE DISORDER TREATMENT

07/31/2018

In this section, we examine recent trends in receipt of SUD treatment services by service type and geography. Then we analyze the relationship between the prevalence of SUDs and use of treatment services by type of SUD and geographic area. Last, we look specifically at the relationship between Medicaid coverage expansion and receipt of SUD treatment.

A. How Many People are Receiving SUD Treatment Services and What Services are They Receiving?

Here we analyze information on the number of people receiving SUD treatment services and the type of services they receive as derived from multiple data sources. Based on the National Survey on Drug Use and Health (NSDUH), we begin by looking at whether individuals used any services in the past year and the type of services they used. Then we analyze the number of clients in care at a given point-in-time by service type, based on the N-SSATS. Finally, we assess trends in the distribution of admissions by primary substance.

FIGURE II.1. Number of Individuals Who Received Any Treatment or Specialty Treatment, NSDUH 2004-2014
FIGURE II.1, Line Chart: There are two series displayed. A dark blue line shows the number of individuals receiving any treatment for substance use disorders has remained relatively constant between 2004 and 2014 at about 4 million individuals. A light blue line indicates the number of Individuals receiving specialty treatment has also remained relatively constant between 2004 and 2014, varying slightly between 2.2 and 2.6 million individuals in a given year.
SOURCE: NSDUH 2004-2014.

1. Trends in Receipt of Any or Specialty SUD Treatment in Past Year

We use data from the NSDUH to analyze trends in SUD treatment use in the community-based population in the United States. According to the NSDUH, the number of individuals receiving any SUD treatment in the past year was relatively constant between 2004 and 2014, at about 4 million individuals (Figure II.1).

TABLE II.1. Number of Individuals Who Received Any Treatment or Specialty Treatment, NSDUH 2015-2016
Type of Treatment 2015 2016
Any treatment in past 12 months 3.7 3.8
Specialty treatment in past 12 months 2.3 2.2
SOURCE: NSDUH 2015 and 2016.

About 60 percent of the individuals who received any treatment (2.2-2.6 million individuals per year) received treatment in a specialty setting, defined by SAMHSA as any of the following types of facilities: hospitals (inpatient only), drug or alcohol rehabilitation facilities (inpatient or outpatient), or mental health centers.8

Because of changes in the methodology of the survey between 2014 and 2015 survey estimates from 2015 and later may not be comparable to earlier years. Thus, we present estimates for 2015 and later separately from those in the earlier period. In 2015 there were 3.7 million and 2.3 million individuals receiving any and specialty treatment according to the NSDUH (Table II.1). Between 2015 and 2016 there was no significant change in the number of individuals receiving any and specialty treatment in the past year.

TABLE II.2. Number of Individuals Receiving Any Specialty SUD Treatment by the Settings in Which They Received Care, NSDUH 2012-2014
Type of Care Number (in thousands) Percentage
2012 2013 2014 2012 2013 2014
Total 2,496 2,466 2,606 100 100 100
Specialty settings
   Hospital inpatient 861 879 921 34 36 35
   Rehabilitation facility--inpatient 1,010 1,042 1,076 40 42 41
   Rehabilitation facility--outpatient 1,505 1,753 1,731 60 71 66
   Mental health center--outpatient 1,000 1,176 1,157 40 48 44
Non-specialty setting
   Emergency room 557 574 499 22 23 19
   Private doctor's office 470 522 561 19 21 22
   Self-help group 1,461 1,505 1,554 59 61 60
   Prison or jaila 340 189 280 14 8 11
SOURCE: NSDUH 2012-2014.
NOTE: The counts only include individuals who received care in a specialty setting during the year; however, the counts indicate the number of these individuals receiving care in non-specialty settings. Counts do not sum to the total and percentages do not sum to 100% because individuals may receive care in multiple settings.
  1. NSDUH surveys individuals living in the community. Individuals living in an institutional setting are excluded. Therefore counts of individuals receiving treatment in a prison or jail only include individuals who have been released from those settings and are living in the community at the time of the survey.

The distribution of the number of people receiving treatment by treatment setting also remained relatively constant from 2012 to 2014 (Table II.2) and 2015 to 2016 (Table II.3). Outpatient rehabilitation and self-help groups were the most common settings of care. About one-third of individuals who received specialty treatment received some services in an inpatient hospital; about 20 percent received emergency room care.

TABLE II.3. Number of Individuals Receiving Any Specialty SUD Treatment by the Settings in Which They Received Care, NSDUH 2015-2016
Type of Care Number (in thousands) Percentage
2015 2016 2015 2016
Total 2,346 2,229 100 100
Specialty settings
   Hospital inpatient 702 732 30 33
   Rehabilitation facility--inpatient 974 918 42 41
   Rehabilitation facility--outpatient 1,524 1,446 65 65
   Mental health center--outpatient 1,093 1,054 47 47
Non-specialty setting
   Emergency room 429 489 18 22
   Private doctor's office 445 540 19 24
   Self-help group 1,389 1,183 59 53
   Prison or jaila 221 202 9 9
SOURCE: NSDUH 2015-2016.
NOTE: The counts only include individuals who received care in a specialty setting during the year; however, the counts indicate the number of these individuals receiving care in non-specialty settings. Counts do not sum to the total and percentages do not sum to 100% because individuals may receive care in multiple settings. The 2015 and 2016 estimates are not comparable to estimates from prior years due to methodological changes in the survey.
  1. NSDUH surveys individuals living in the community. Individuals living in an institutional setting are excluded. Therefore counts of individuals receiving treatment in a prison or jail only include individuals who have been released from those settings and are living in the community at the time of the survey.

2. Trends in Point-in-Time Clients in Care, by Care Setting

In contrast to the consistency of NSDUH findings, analysis of N-SSATS indicates notable shifts between 2013 and 2015 in client counts and the distribution of clients by service type (Table II.4).9 Overall, N-SSATS client counts indicate a small increase in clients in care between 2013 and 2015 (4.5 percent). This increase was driven by large increases in several service types: outpatient pharmacotherapy for opioid use disorders (14.8 percent), residential detoxification (34.2 percent) and short-term care (34.8 percent), and hospital inpatient detoxification (114.9 percent) and treatment (33.7 percent). There was little change in the number of clients in regular outpatient care.

The observed increases in clients receiving pharmacotherapy reflect national efforts to improve quality of care by increasing access to these evidence-based treatments. Pharmacotherapy use is associated with more consecutive weeks of abstinence from illicit opioids (Fiellin et al. 2014) and reduced mortality due to overdose (Brugal et al. 2005; Clark et al. 2011; Cousins et al. 2016; Degenhardt et al. 2009; Pierce et al. 2016) .

TABLE II.4. Services Provided by Setting of Care, N-SSATS 2013 and 2015
Type of Care Number of Clients Percentage of All Clients in Care
2013 2015 % Change 2013 2015 % Change
Total 1,249,629 1,305,647 4.5% 100.0 100.0 0.0%
Outpatient 1,127,235 1,161,456 3.0% 90.2 89.0 -1.4%
   Regular 603,315 604,819 0.2% 48.3 46.3 -4.1%
   Intensive 147,162 128,536 -12.7% 11.8 9.8 -16.4%
   Detoxification 13,839 14,457 4.5% 1.1 1.1 0.0%
   Day treatment/partial hospitalization 22,828 23,138 1.4% 1.8 1.8 -3.0%
   Methadone/buprenorphine maintenance or injectable naltrexone 340,091 390,506 14.8% 27.2 29.9 9.9%
Residential (non-hospital) 107,727 119,900 11.3% 8.6 9.2 6.5%
   Detoxification 10,244 13,748 34.2% 0.8 1.1 28.4%
   Short-term 27,184 36,651 34.8% 2.2 2.8 29.0%
   Long-term 70,299 69,501 -1.1% 5.6 5.3 -5.4%
Hospital inpatient 14,667 24,291 65.6% 1.2 1.9 58.5%
   Detoxification 5,768 12,394 114.9% 0.5 0.9 105.7%
   Treatment 8,899 11,897 33.7% 0.7 0.9 28.0%
Clients receiving methadone, buprenorphine, or injectable naltrexone treatment 382,237 439,602 15.0% 30.6 33.7 10.1%
   Clients receiving methadone 330,308 356,843 8.0% 26.4 27.3 3.4%
   Clients receiving buprenorphine 48,148 75,724 57.3% 3.9 5.8 50.5%
   Clients receiving injectable naltrexone 3,781 7,035 86.1% 0.3 0.5 78.1%
SOURCE: N-SSATS 2013 and 2015.
NOTE: N-SSATS surveys the universe of specialty SUD treatment facilities. In 2013 and 2015, respectively, the survey had a 94% and 92% response rate. Estimates are not adjusted for facility or item non-response. For inpatient and residential services counts indicate the number of clients in treatment on the last working day in March of each survey year. For outpatient services counts indicate the number of clients receiving services during March who are still enrolled in treatment on the last working day in March. Care categories defined to align with the American Society of Addiction Medicine levels of care.

Discrepancies between trends in the N-SSATS and NSDUH are expected due to differences in the scope of the surveys and measures of service use (Batts et al. 2014). Although both NSDUH and N-SSATS collect information on the number of individuals in care at specialty SUD treatment facilities, they differ in how they measure this population. The NSDUH measures the number of individuals reporting any receipt of treatment in the past year in a specialty setting, whereas for N-SSATS specialty facilities report the number of clients in treatment on a single day in each year (the last working day in March of each survey year). Outpatient client counts in N-SSATS include individuals receiving services during March who are still enrolled in treatment on the last working day in March. If individuals experience a longer duration of care or repeat admissions to the same type of care, NSDUH will show no change in the number of individuals with service use; N-SSATS client counts, on the other hand, will increase under these circumstances. The lack of change in the population with service use in the past year in NSDUH, paired with the increases in client counts observed in the N-SSATS, suggests that the increase in outpatient clients observed in N-SSATS stems from a longer duration of care or repeated admissions rather an increase in the total number of individuals receiving treatment in the course of a year. Analysis of the distribution of length of stay in the Treatment Episode Data Set (TEDS) Discharge file in 2012 relative to 2014 (Table II.5) supports a small 2 percentage point decline in the number of discharges with length of stay 30 days or less and corresponding 2 percentage point increase stays greater than 180 days.

TABLE II.5. Distribution of Discharges by Length of Stay, TEDS 2012 and 2014
Length of Stay 2012 2014
1 to 30 days 47.8 45.9
31 to 45 days 6.8 6.7
46 to 60 days 5.3 5.3
61 to 90 days 9.0 9.0
91 to 120 days 7.4 7.5
121 to 180 days 8.9 9.2
181 to 365 days 9.8 10.5
More than a year 4.9 6.0
SOURCE: TEDS 2012 and 2014 (CBHSQ 2017a and 2017b).
NOTE: Individual states report discharges to SUD treatment facilities within their state to TEDS. The scope of SUD treatment providers included in each state's data may vary over time and based on differences across states in state licensure, certification, accreditation, and disbursement of public funds. At a minimum, facilities receiving federal substance abuse treatment block grant funds are included. The following states did not report usable data for the year 2012: Kansas, Mississippi, and New Mexico. The following states did not report usable data for the year 2014: Mississippi, Florida, Georgia, Kansas, West Virginia, and New Mexico.

Differences in the scope of the two surveys can explain why the N-SSATS reports show increases in residential and hospital inpatient care, whereas the NSDUH results show constant use of these services. NSDUH surveys residents in households with a fixed address and individuals in non-institutional group quarters. It excludes individuals who are institutionalized or homeless and not in a shelter from its respondent pool. Thus, the NSDUH will not accurately assess the number of individuals receiving institutional services. Specifically, NSDUH will not count hospital services provided to individuals who enter a hospital but who do not re-enter the community due to drug overdose death. Thus, N-SSATS is a more accurate source of trends in institutional service use.

3. Trends in the Primary Substance for Treatment Admission

According to the data collected in TEDS, over the last decade the primary substance for which individuals receive SUD treatment has shifted. Alcohol use disorders as a primary substance accounted for the highest proportion of clients in care in 2004 (40 percent) and 2014 (36 percent), but the proportion represented by these admissions has declined (Table II.6). The proportion of admissions for a disorder related to cocaine as a primary substance also has declined, from 14 percent to 5 percent. In contrast, the proportion of admissions for heroin and non-heroin opiates and synthetics as a primary substance rose from 18 percent to 30 percent.

TABLE II.6. Number and Percentage of Specialty SUD Treatment Admissions by Primary Substance, TEDS 2004 and 2014
  Number Percentage
2004 2014 2004 2014
Total 1,808,469 1,614,358 100 100
Alcohol 729,366 585,024 40 36
Marijuana 285,193 247,461 16 15
Heroin 262,518 357,293 15 22
Cocaine 248,492 87,510 14 5
Methamphetamine/amphetamine 142,510 143,659 8 9
Non-heroin opiates/synthetic 62,895 134,401 3 8
Other or not reported 77,495 59,010 4 4
SOURCE: TEDS 2004 and 2014 (SAMHSA 2016).
NOTE: Individual states report admissions to SUD treatment facilities within their state to TEDS. The scope of SUD treatment providers included in each state's data may vary over time and based on differences across states in state licensure, certification, accreditation, and disbursement of public funds. At a minimum, facilities receiving federal substance abuse treatment block grant funds are included. Alaska, Arkansas, and District of Columbia reported either no data, or less than a full calendar year of data for 2004. South Carolina did not report usable data for the year 2014.

B. How Does Service Receipt Vary Geographically, by Level of Urbanicity and by Facility Operation?

Geographically. SUD prevalence and treatment use varies based on geography. Geographic variation results from cultural and environmental influences on disorder prevalence as well as differences in jurisdictional policies, treatment funding and availability, and availability of other social services. Below, we discuss the geographic variations in service receipt, reflected in Table B.1 through Table B.4.b of Appendix B.

National and regional average changes in clients by type of care between 2013 and 2015 mask substantial variation by state. Across all regions, there was a substantial increase in clients in inpatient care (65.6 percent). We also observed substantial increases for each of the four regions (Table B.1), but the increase in the Midwest was much lower than for the other regions (16.8 percent). Within each region, however, changes in inpatient care varied substantially by state (Table B.2). Residential clients increased by 11.3 percent nationally but, as was true for inpatient care, results varied by region. At the extremes, the number of residential clients in the Midwest declined by 9.6 percent, whereas the number in the South increased by 28.5 percent. Outpatient client counts increased modestly in each region.

Use of pharmacotherapies targeted to alcohol and opioid dependence substantially increased in all regions (Table B.3.a and Table B.3.b). Buprenorphine and injectable naltrexone had higher percentage increases, but these medications were less commonly used in 2013 than methadone. Rates of change varied dramatically across states, with some states seeing declines in pharmacotherapy use (particularly for methadone), whereas others saw a surge in use (Table B.4.a and Table B.4.b).

Urbanicity. Variation in treatment use based on the level of urbanicity may be expected due to access barriers for individuals in more rural areas. Jackson and Shannon (2012) reviewed the literature on barriers to treatment access for rural residents and found: (1) rural residents are less likely to have access to health insurance; (2) there is a shortage of providers in rural areas; and (3) people in need of treatment in rural areas must travel longer distances to facilities. Cummings and colleagues (2014) used the 2009 N-SSATS and the Area Resource File to look at access to outpatient SUD treatment for Medicaid enrollees. This study found that rural counties are less likely than urban counties to have at least one outpatient SUD facility that accepts Medicaid. Lenardson and Gale (2007) compared SUD treatment offered in rural and urban counties using variables in the 2004 N-SSATS. Comparing the number of facilities and treatment beds to population size revealed that rural areas actually had a larger number of treatment facilities, but the facilities had fewer inpatients beds available per population. In addition, few facilities in rural counties not adjacent to a metropolitan area provided detoxification, transitional housing services, or intensive outpatient care. Nearly all opioid treatment programs (OTPs) were located in urban areas.

FIGURE II.2. Number of Clients by Urbanicity, N-SSATS 2013 and 2015
FIGURE II.2, Bar Chart: Each bar shows the number of clients (in thousand) who received treatment in specialty substance use disorder treatment facilities. The first bar displays that in 2013 facilities in urban areas served 305 thousand clients receiving pharmacotherapy and 650 thousand clients receiving other services. The second bar displays that in 2015 facilities in urban areas served 404 thousand clients receiving pharmacotherapy and 700 thousand clients receiving other services. The third bar displays that in 2013 facilities in rural areas served 69 thousand clients receiving pharmacotherapy and 211 thousand clients receiving other services. The fourth bar displays that in 2015 facilities in rural areas served 31 thousand clients receiving pharmacotherapy and 160 thousand clients receiving other services.
SOURCE: N-SSATS 2013 and 2015.
NOTE: Pharmacotherapy is limited to methadone, buprenorphine or injectable naltrexone. Urbanicity is assigned based on the HHS National Center for Health Statistics urbanicity classification scheme. Facilities in rural areas include those in micropolitan areas with an urban core population of at least 10,000 but less than 50,000, as well as those in non-core areas. Facilities in a central or fringe urban core with a population of 50,000 or more are considered urban. Information on urbanicity was not available for all facilities; urban and rural client counts are only reported for facilities with known urbanicity.

The number of clients in each care setting declined substantially in rural areas and increased in urban areas between 2013 and 2015 (Figure II.2). The number of clients receiving pharmacotherapy also declined substantially in rural areas and increased in urban areas with the exception of clients receiving buprenorphine which stayed fairly constant in rural areas (Table B.3.a). According to the U.S. Department of Agriculture (2016) the population living in rural areas was fairly constant in this period while the population living in urban areas has increased steadily at approximately 1 percent annually. Given the treatment access barriers for individuals living in rural areas that pre-date this period and the consistent size of the population in these areas the substantial declines in treatment use in rural areas warrant further investigation.

Facility operation. Facility operation may affect the characteristics of clients served and types of services offered as facilities that are publicly owned or non-profits may have distinct missions to provide charitable care or act as the provider of last resort. Between 2013 and 2015 clients served in public facilities (Table B.1) declined substantially for outpatient care (13.7 percent) and somewhat for residential care (4.3 percent). Meanwhile clients served in private for-profit and private non-profit facilities expanded in these settings. This shift may be related to increased rates of insurance coverage. Inpatient clients increased substantially for facilities of all operation types.

C. What is the Relationship between the Prevalence of SUDs and Demand for Care? How Does the Relationship Vary Geographically?

According to NSDUH, the number of individuals with SUDs was relatively constant between 2004 and 2014 (Figure II.3). The aggregate estimates, however, mask substantial shifts in the substances with which the disorders are associated (Table B.5). Cocaine/crack-related, hallucinogen-related, inhalant-related, and alcohol-related disorders have declined over the last decade, whereas heroin, non-medical use of psychotherapeutics, and use of pain relievers have increased.

The direction of trends in the percent of the population with use disorders was similar across age groups (Table II.7). Between 2002 and 2015, the proportion of the population with an SUD declined for all age groups for alcohol and cocaine and increased for heroin. In contrast, the proportion of the population with marijuana use disorders remained constant among individuals 26 and older, but declined for individuals 12-17 and 18-25. Due to the survey sample size and prevalence of disorders, it is difficult to detect trends in disorder prevalence among more detailed subgroups within the 26 and older group with a single year of NSDUH data. Han et al. (2017) pooled two years of NSDUH data to compare the proportion of individuals 50 and older with alcohol use disorders in 2005-2006 to 2013-2014. In contrast to the results for the 26 or older group, they found the proportion of individuals 50 and older with an alcohol use disorder increased from 3.0 percent to 3.7 percent (a 23.3 percent increase). There is concern that SUD prevalence may increase among older age groups over time as the baby boomers age given their higher rates of substance use relative to previous generations (Elinson 2015).

FIGURE II.3. Number of Individuals Age 12 and Older with Abuse of or Dependence on Alcohol or Illicit Drugs in the Past Year, NSDUH 202-2014
FIGURE II.3, Line Chart: Three series are displayed. Each series is displayed as a line in a different shade of blue indicating the number of individuals, in thousands, with a substance use disorder of the indicated type in the past year. The series never intersect. The lowest line is in dark blue. It indicates that between 2002 and 2014 the number of individuals with an illicit drug disorder remained relatively constant at about 7 million. The middle line is in a medium shade of blue. It indicates that between 2002 and 2010 the number of individuals with an alcohol disorder disorder remained relatively constant at about 18 million. The number of individuals with an alcohol disorder declined between 2010 and 2014. In 2014 the figure individuals about 16 million individuals with an alcohol disorder. The highest line in light blue indicates that between 2002 and 2014 the number of individuals with an illicit drug disorder or an alcohol remained relatively constant at about 22 million.
SOURCE: NSDUH 2002-2014.

According to the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition ("DSM-IV-TR" 2017), which was used to develop the diagnostic criteria in the NSDUH for having an SUD an individual must have serious negative consequences to qualify as having a disorder. For substance use dependence an individual must have three or more symptoms of dependence such as withdrawal symptoms, increased tolerance, repeated unsuccessful attempts to quit, having given up social, occupational or recreational activities or using the substance in larger amounts and for longer periods of time than intended. For abuse the individual must continue use despite having at least one negative consequence due to use including failure to fulfill a major work, school or home role, recurrent use in hazardous situations, recurrent legal issues, or social and interpersonal problems caused by use. Individuals meeting criteria for a disorder continue substance use despite serious negative consequences in their personal lives.

TABLE II.7. Percentage of Population with an SUD by Type of Substance and Age Group, NSDUH 2002 and 2015
  Age 12-17 Age 18-25 Age 26 or Older
2002 2015 2002 2015 2002 2015
Alcohol use 5.9* 2.5 17.7* 10.9 6.2* 5.4
Marijuana use 4.3* 2.6 6.0* 5.1 0.8 0.8
Cocaine use 0.4* 0.1 1.2* 0.7 0.6* 0.3
Heroin use 0.1 0.0 0.2* 0.4 0.1* 0.2
SOURCE: NSDUH 2002 and 2015 (CBHSQ 2016b).
* Estimate is significantly different from 2015 estimate at the 0.5% level.

Based on the NSDUH survey, there is a substantial gap between the number of people with an SUD and the number of individuals who receive specialty treatment in a given year. In 2014, an estimated 20.3 million United States residents aged 12-64 met criteria for an SUD in the past year. Among this group, less than 10 percent of individuals abusing or dependent on alcohol only received specialty SUD treatment in the past year (Figure II.4). The treatment rate was higher (about 20 percent) among individuals abusing or dependent on illicit drugs only. About 15 percent of those abusing or dependent on both illicit drugs and alcohol received treatment. Differences between the 2013 and 2014 rates are not statistically significant. According to a review by Foster (2014), treatment rates among individuals with SUDs are substantially lower than those for common health conditions, such as hypertension (77 percent), diabetes (73 percent), and major depression (71 percent). However, an individual's need for professional support to address an SUD may depend on several factors, such as the severity of the disorder, comorbid health conditions, personal coping skills, the individual's environment, and available sources of informal specialty support (Mechanic 2003). Treatment rates vary little across states (Table B.6).

FIGURE II.4. Percentage of Individuals Age 12-64 with Abuse or Dependence Who Received Specialty SUD Treatment, 2013-2014
FIGURE II.4, Bar Chart: The chart demonstrates the percentage of individuals with substance abuse or dependence disorders who received specialty SUD treatment. There are 3 groups of blue bars and each group represents a disorder type. The darker bar in each group represents the year 2013. The lighter bar in each group represents the year 2014. 6% of individuals with abuse or dependence on alcohol received treatment in 2013 and 8% received treatment in 2014. 20% of individuals with abuse or dependence on illicit drugs only received treatment in 2013 and 21% received treatment in 2014. 13% of individuals with abuse or dependence on both illicit drugs and alcohol received treatment in 2013 and 17% received treatment in 2014.
SOURCE: Mathematica analysis of NSDUH 2013 and 2014 public use files.

Individuals who meet the criteria for an SUD but do not receive treatment fall into three groups: (1) those who do not feel they need treatment; (2) those who feel they need treatment but do not seek it; and (3) those who feel they need and seek treatment but do not receive it. Based on the 2015 NSDUH, Lipari et al. (2016) found that 95.4 percent of people who met the criteria for an SUD but did not receive specialty treatment did not feel they needed treatment (19.3 million people). Among the remaining small percentage (4.6 percent, or 880,000 people) who felt they needed treatment but did not get it, 64.4 percent (about 567,000 people) reported making no effort to get treatment, whereas 35.6 percent (about 313,000 people) reported they did make such efforts. Figure II.5 displays trends in the number of individuals who received specialty SUD treatment, felt they needed treatment but did not seek it, and felt they needed treatment and sought it but did not receive it. These numbers were relatively constant from 2004 through 2014.

FIGURE II.5. Number of Individuals Who Received Specialty SUD Treatment or Felt They Needed but Did Not Receive Treatment, NSDUH, 2004-2014
FIGURE II.5, Wave Chart: The chart demonstrates the number of individuals who received specialty SUD treatment or felt that they needed it between 2004 and 2014. There are 3 waves, each a varying shade of blue. The darkest blue wave represents individuals who received treatment. The number of individuals receiving treatment stayed constant between 2004 and 2014 at about 2.5 million individuals. The wave that is medium blue is relatively constant from 2004 to 2014 at about 300,000 individuals who perceived a need for treatment and sought treatment, but did not receive treatment. The top wave that is light blue is relatively constant from 2004 to 2014 at about 700,000 individuals who perceived a need for treatment, but made no effort to obtain and did not receive treatment.
SOURCE: NSDUH 2004-2014.
NOTE: The figure depicts treatment receipt within the past 12 months.

D. What Evidence Exists about How Increases in Medicaid Enrollment Have Impacted Demand?

In Phase 1 of this project we reviewed the literature on the relationship between Medicaid insurance coverage and SUD treatment use. We found only a few studies in the literature that have explicitly examined how health insurance coverage impacts demand for SUD treatment services; furthermore, many of these studies did not employ experimental designs, so the findings may be confounded by other factors. Although a rigorous study with experimental design found that insurance coverage has a positive effect on the use of general health services (Newhouse and the Insurance Experiment Group 1993), findings for SUD treatment could differ for several reasons. First, SUD treatment is typically provided outside of the general health sector, and insurance coverage for these services may be less comprehensive, have a limited network of providers, and require greater out-of-pocket costs for the client, thereby deterring treatment use. The social stigma attached to SUDs and SUD treatment may also limit treatment seeking despite insurance coverage. In addition, states and the Federal Government (through block grants) fund SUD treatment directly, particularly for those who are uninsured. Thus, although other types of care may be more affordable for those who are insured, the availability of publicly funded SUD treatment for individuals without insurance may mean that access to Medicaid coverage has less impact on SUD treatment use than use of other types of health care services.

The findings from the limited studies we identified on the relationship between Medicaid coverage and SUD treatment use indicate that individuals with Medicaid coverage are more likely to use SUD treatment than those with private insurance or who are uninsured (Bouchery et al. 2012; Epstein et al. 2004; Larson et al. 2005). This finding may be due to out-of-pocket expenses being lower under Medicaid. It may also be due to differences in the care management and benefit packages provided through Medicaid and private insurance plans. Since these studies did not use an experimental design the findings may be due to characteristics of the Medicaid population that were not controlled for in the models. In particular, individuals who are eligible for Medicaid may be enrolled in Medicaid by a treatment provider.

For Phase 2 of this study data from the NSDUH on Medicaid enrollment and treatment use prior to (2009-2013) and following ACA implementation (2014) was available for analysis. We used these data to estimate how increased rates of Medicaid enrollment influenced SUD treatment use. First, among individuals with SUDs we estimated changes in Medicaid enrollment rates and the number of individuals with SUD who gained Medicaid coverage as a result of increased enrollment rates. Then we estimated treatment use rates among those with SUDs and assessed how access to Medicaid coverage likely affected treatment use among individuals who gained Medicaid coverage.

According to our analysis of the NSDUH, the proportion of individuals ages 12-64 with SUDs who were enrolled in Medicaid rose from 13.4 percent in the five years from 2009 to 2013 to 18.1 percent in 2014--a statistically significant change (Table II.8). There was a corresponding decline in the percentage uninsured from 24.8 percent in 2009-2013 to 20.0 percent in 2014. This change may be related to expansion of Medicaid eligibility under the ACA. The opioid epidemic and efforts to increase treatment use for individuals affected may also have contributed to increased Medicaid enrollment among individuals with SUDs.

Because of the higher Medicaid enrollment rate observed in 2014, we estimate that approximately 944,000 more individuals with SUDs were enrolled in Medicaid in 2014 than would have been expected, given the Medicaid enrollment rates observed between 2009 and 2013 (Table II.9). This represents a 34 percent increase in the size of the Medicaid population with SUDs. We estimate this by projecting what Medicaid enrollment would have been among individuals with SUDs given the average enrollment rate in 2009-2013 compared to the observed enrollment rate in 2014. The steps of this calculation are presented in Table II.9. The first step was to determine the actual number of individuals enrolled in Medicaid with SUDs in 2014. Then we calculated the ratio of the observed average Medicaid enrollment rate in 2009-2013 relative to that for 2014 based on the estimates in Table II.8. We multiplied these ratios by the actual number of Medicaid enrollees with an SUD in each diagnostic category in 2014 to calculate projected Medicaid enrollment for 2014 given the average enrollment rate between 2009 and 2013. We then subtracted the projected enrollment levels for 2009-2013 from the actual enrollment levels in 2014 to estimate the increase in the number of enrollees.

TABLE II.8. Percentage of Individuals Ages 12-64 with SUDs Who Were Enrolled in Medicaid or Uninsured in 2009-2013 versus 2014, by SUD Type
Type of Substance 2009-2013 2014
Mean Lower Bound
for 95% CL
Upper Bound
for 95% CL
Mean Lower Bound
for 95% CL
Upper Bound
for 95% CL
Medicaid Enrolled
   Total 13.4 12.8 14.1 18.1 16.7 19.6
   Alcohol dependence 11.1 9.9 12.3 16.6 14.0 19.2
   Other alcohol and marijuana disorders 11.7 11.0 12.5 14.9 13.4 16.4
   Other drug abuse or dependence disorders 23.1 21.4 24.7 30.9 27.3 34.5
Uninsured
   Total 24.8 23.7 25.9 20.0 18.3 21.7
   Alcohol dependence 25.4 23.9 26.8 23.6 20.2 27.0
   Other alcohol and marijuana disorders 22.7 21.4 24.0 17.0 14.9 19.1
   Other drug abuse or dependence disorders 30.4 28.1 32.8 22.5 18.5 26.6
SOURCE: NSDUH 2009-2014.

 

TABLE II.9. Estimated Increase in Medicaid Enrollment Associated with Medicaid Expansion for Individuals Ages 12-64 with an SUD, by SUD Type
Type of Substance 2014 Medicaid Enrollmenta 2009-2013 Medicaid Enrollment
Rate as a Percentage
of 2014 Rateb
Projected 2014 Medicaid
Enrollment Based on
2009-2013 Enrollment Ratesc
Estimated Increase in
Medicaid Enrollment Associated
with Medicaid Expansiond
Total 3,684,517 74 2,740,333 944,184
Alcohol dependence 1,042,102 67 699,424 342,678
Other alcohol and marijuana disorders 1,571,584 79 1,241,446 330,138
Other drug abuse or dependence disorders 1,070,831 75 799,463 271,368
SOURCE: Mathematica analysis of NSDUH 2009-2014 public use files.
  1. These counts are estimated based on the NSDUH survey sample.
  2. This is the mean percentage of individuals ages 12-64 with an SUD enrolled in Medicaid in 2009-2013 divided by that for 2014. Means for these periods are those presented in Table II.4.
  3. This number is the 2014 Medicaid enrollment times the 2009-2013 Medicaid enrollment rate as a percentage of the 2014 rate.
  4. This number is the difference between 2014 Medicaid enrollment and the projected Medicaid enrollment based on 2009-2013 enrollment rates. The estimated increase in Medicaid enrollment may result from Medicaid eligibility expansion under the ACA, or other changes such as new policies implemented to address the opioid epidemic.

According to the NSDUH, in 2014 most Medicaid-enrolled individuals with an SUD, 85 percent, did not receive any specialty treatment. Assuming their treatment use rates are similar to those of the Medicaid population overall, most of the individuals whose enrollment is associated with Medicaid expansion (about 798,000, or 85 percent of the 944,000) did not receive any specialty SUD treatment in 2014 (Figure II.6). Individuals who are uninsured access treatment at a lower rate than those on Medicaid; a logit model predicting specialty treatment use for the NSDUH using data from 2009 to 2014 indicated that the likelihood of specialty SUD treatment for someone who was uninsured was 60 percent of that for an individual who was Medicaid insured. Thus, we would expect 60 percent of those whose enrollment in Medicaid was associated with Medicaid expansion who received SUD treatment in 2014 would have received specialty SUD treatment even if they were uninsured. The remaining approximately 59,000 (40 percent) would not have received specialty treatment in the absence of Medicaid enrollment.10

FIGURE II.6. Estimated Rate of Treatment Receipt among Medicaid Expansion Population Ages 12-64 with an SUD, 2014
FIGURE II.6, Pie Chart: The chart demonstrates the estimated rate of treatment receipt among the Medicaid expansion population with an SUD in the year 2014. 84.5% of the population received no treatment, represented by the darkest blue sector of the pie chart. 9.3% of the population, represented by the sector of the pie chart that is medium blue, received treatment and would have been expected to receive treatment if they were uninsured. 6.2% of the population, represented by the sector of the pie chart that is light blue, received treatment and would not have been expected to receive treatment if they were uninsured.
SOURCE: Mathematica analysis of NSDUH 2009-2014 public use files.
NOTE: We estimated the percentage receiving any specialty treatment (15.5%) based on the observed rate of specialty SUD treatment receipt in the NSDUH 2009-2014 for Medicaid-enrolled individuals ages 12-64 with an SUD. We estimated the share of individuals who would have received treatment without Medicaid expansion based on findings from a logit model using NSDUH 2009-2014 data, which indicated the likelihood of specialty treatment use for uninsured individuals was 60% of that for Medicaid-enrolled individuals.

A limitation to this analysis is that we assume that the SUD treatment use rate for the Medicaid expansion population with SUDs is the same as that for other Medicaid enrollees with SUDs. To assess the importance of this limitation, we compared treatment use rates by disorder type in 2014 to the average for 2009-2013. Despite the 34 percent increase in the number of individuals enrolled in Medicaid with SUDs specialty treatment use rates were unchanged between the two periods suggesting that the expansion population had similar rates of treatment use.

TABLE II.10. Percentage of Individuals Ages 12-64 with SUDs Who Were Enrolled in Medicaid Who Used Specialty Treatment in 2009-2013 versus 2014, by SUD Type
Type of Substance 2009-2013 2014
Mean Lower Bound
for 95% CL
Upper Bound
for 95% CL
Mean Lower Bound
for 95% CL
Upper Bound
for 95% CL
Total 14.5 12.7 16.3 15.5 11.8 19.2
Alcohol dependence 10.7 7.3 14.0 11.1 5.5 16.8
Other alcohol and marijuana disorders 8.3 6.3 10.3 9.6 4.4 14.9
Other drug abuse or dependence disorders 28.2 24.2 32.2 28.2 20.1 36.4
SOURCE: Mathematica analysis of NSDUH 2009-2014 public use files.

Another limitation of this analysis is that it included only individuals with SUDs as identified by responses to questions in the NSDUH survey. Some individuals receive treatment for substance use, but do not meet criteria for an SUD in the past year. These may be individuals who previously met criteria for a disorder and are continuing to receive treatment to reduce the likelihood of relapse. According to the NSDUH, on average between 2009-2013 and in 2014, respectively, about 292,000 and 261,000 individuals enrolled in Medicaid who did not meet criteria for an SUD in the past year received specialty treatment (Table II.11). The difference between these estimates is not statistically significant suggesting the Medicaid expansion did not substantially change the number of individuals in this population receiving specialty treatment.

TABLE II.11. Number of Individuals Ages 12-64 Not Meeting Criteria for an SUD Who Received Specialty SUD Treatment in 2009-2013 versus 2014, by Medicaid Enrollment
Type of Substance 2009-2013 2014
Mean Lower Bound
for 95% CL
Upper Bound
for 95% CL
Mean Lower Bound
for 95% CL
Upper Bound
for 95% CL
Medicaid enrolled 291,898 243,132 340,664 261,103 179,849 342,356
Not Medicaid enrolled 672,861 588,171 757,551 728,473 540,615 916,332
SOURCE: Mathematica analysis of NSDUH 2009-2014 public use files.