OMB Ss-b-n-ssats 3.6.19_final

^National Survey of Substance Abuse Treatment Services (N-SSATS)

SUPPORTING STATEMENT

B. COLLECTIONS OF INFORMATION EMPLOYING STATISTICAL METHODS

1. Respondent Universe and Sampling Methods Respondent Universe and Response Rates

N-SSATS: The universe for the N-SSATS includes all known substance use treatment facilities in the United States, District of Columbia, and territories. The universe, derived from the I-BHS inventory of facilities, is shown in the table below as of April 30, 2018. The substance use treatment universe is subdivided into two categories: (1) facilities that have state- agency licensing or other approval as substance use treatment facilities, and (2) non-state- approved treatment facilities. Prisons, jails, detention centers, and solo practitioners are not included in the N-SSATS universe.

SUBSTANCE USE TREATMENT FACILITIES	TOTAL SUBSTANCE USE TREATMENT FACILITIES ON I-BHS (as of April 30, 2018)	STATE- APPROVED	NON-STATE- APPROVED
I-BHS substance use treatment facilities	20,527	16,697	3,830

The overall response rates for the 2016 N-SSATS was 92 percent, for the 2017 N-SSATS was 90 percent, and for the 2018 N-SSATS was 92 percent.

Sampling

There is currently no sampling in N-SSATS. A complete census is needed because

N-SSATS is the only source of information for SAMHSA’s Behavioral Health Treatment Services Locator and the National Directory of Drug and Alcohol Abuse Treatment Programs. Without doing a census, SAMHSA would not have information for all the facilities for the public who are seeking treatment. A sample of treatment facilities would not be feasible for the locator because information for all of the facilities would not be listed. Additionally, a complete census for client counts every other year is necessary for the following reasons.

1. The substance use field and states are in need of not only national, but state-level and sub-state statistical data for comparative analyses of access to care (needs assessment) and service utilization across the various components of the substance use treatment service delivery system.

The delivery system is comprised of many various types of treatment through many different operation structures. Types of treatment include, and are not limited to, hospital inpatient (both detoxification and treatment), residential (detoxification, short-term treatment, and long-term treatment), and outpatient (detoxification, regular outpatient treatment, intensive outpatient treatment, partial hospitalization/day treatment, and medication-assisted treatment). Within those basic types of treatment are specific types of facilities that serve specific needs or populations. For example, serving clients with opioid use disorder is done through programs such as Opioid Treatment Programs (OTPs), facilities that prescribe buprenorphine, facilities that offer naltrexone, facilities that themselves don’t offer medication but accept clients who receive those medications from other sources, and facilities that treat opioid use disorder without medication. Similar breakouts occur for facilities that treat alcohol use disorder. Some facilities treat exclusively certain types of client, such as adolescents only, women only, men only, opioid clients, DUI/DWI clients, active military, active military and their dependents, veterans, and members of recognized tribes, etc. Facilities differ in the numbers of clients they can serve, and the designations of “small, medium, or large” vary based on the specific types of clients they serve and type of treatment offered. For example, outpatient OTPs and outpatient DUI/DWI facilities typically have greater numbers of clients than regular outpatient treatment, so a medium outpatient OTP would have substantially larger numbers of clients than a medium regular outpatient facility. Some facilities are in metropolitan areas and others are in rural areas. Some facilities are in established networks where resources are shared and others are stand-a-lone facilities. Also, among these many various treatment types and offerings, there are the various operating structures: private for profit, private nonprofit, state government, local/community/county government, tribal government, and federal government, with breakouts of Department of Defense, Veteran’s Affairs, Indian Health Service, and other federal government.

The table below illustrates a portion of the types of care and services offered that would need to be considered in each county, state, and U.S. Jurisdiction/Territory. Approximately 94 percent of counties in the U.S. have 25 or fewer facilities and 84 percent have 10 or fewer facilities. Below is the data for the largest county in the survey, Los Angeles county in California. These data below would need to be broken out again by operation type (private for profit, private nonprofit, and various governments), major service types, and size indication, further reducing the number of facilities in each cell.

Location: California, Los Angeles County

Categories are not mutually exclusive. Further breakouts would need to be done to arrive at mutual exclusivity in order to determine exact size of cells. These numbers would also need to be broken out by each operation category: Private for profit, Private nonprofit, Local/community government, State government, Tribal government, Federal government (Department of Defense, Department of Veterans Affairs, Indian Health Service, and Other Federal), size indication, and service type combinations.

Specific Types	General Types
	HI Detox	HI Treat	RC Detox	RC Short Term	RC Long Term	OP Regular	OP Intensive	OP Day/part Hosp	OP Detox
Not OTP	6	5	52	84	116	172	131	63	22
OTP	1	.	.	2	2	38	4	3	41
Not Bup, not OTP	3	2	7	35	72	137	102	39	5
Bup, not OTP	3	3	45	49	44	35	29	24	17
Not Naltrexone, not OTP	2	1	17	40	75	131	93	36	9
Naltrexone, not OTP	4	4	35	44	41	41	38	27	13
No MAT	2	1	6	29	65	125	89	31	5
DUI/DWI only	.	.	.	.	.	7	.	.	.
Offer DUI/DWI	1	.	.	1	1	26	9	1	.
No DUI/DWI	6	5	52	85	117	184	126	65	63
Adolescent only	.	.	7	12	13	13	7	.	6
Male only	.	.	6	15	28	7	6	5	.
Female only	.	.	3	10	25	9	8	2	.
Both Male and Female	7	5	43	61	65	194	121	59	63
Adult only	7	5	42	70	101	157	95	48	55
Both Adol and Adult	.	.	5	6	6	46	34	18	8
Treat Opioid	7	5	52	80	100	162	107	54	62
Treatment only	.	.	.	34	73	146	105	40	.
Detox only	1	.	1	.	.	.	.	.	1
Both Treatment and Detox	6	5	51	52	45	64	30	26	62
*HI = Hospital Inpatient, RC = Residential Care, OP = Outpatient Care
Bup = Buprenorphine MAT=Medication Assisted Treatment
DUI/DWI = Driving under the influence/Driving while intoxicated
Adol = Adolescent

Because of the small number of facilities that comprise each of the cells identified in the table above (especially when mutual exclusivity is removed through additional breakouts), it would not be possible to take a sample of facilities from each of the cells and still generate state or sub-state or county level estimates, even in the largest county in the country. The ability to generate county- and state-level figures for which the field is in need of for the comparative analyses described previously, would be lost using a sampling method. There are many different combinations of types of treatment and attributes of the treatment system that would need to be considered.

Numbers of clients are needed in small areas (counties and communities) to aid in assessing and forecasting needs and allocation of resources. For example, when state and county information was requested for earthquakes and hurricanes, recently one state was planning safety procedures/disaster preparation in the event of a major catastrophe (earthquake) in a neighboring highly populated area; they needed to know numbers of opioid clients from that area that they might expect to see coming into their area so that they could put procedures into place to be able to accommodate the influx. General numbers would not have been beneficial because they needed accurate and location-specific data. During recent hurricanes, SAMHSA was asked for numbers of substance use clients in affected counties so states could provide appropriate and sufficient care in evacuation areas. Again accurate, locale specific numbers were needed. A sample could likely not provide sufficiently accurate numbers for disaster relief in small, defined areas.

2. Accurate numbers are needed by the White House, Congress, States, and counties for trends projections so adequate resources can be allocated. For example, numbers of clients on the various opioid use disorder Medication Assisted Treatments (MATs) demonstrate the acceptance and use of those treatments. Because reporting of buprenorphine usage for opioid use disorder was not required with the FDA approval process for buprenorphine, N-SSATS is the only survey that collects those data, both nationally and locally. N-SSATS is the only source that collects data on the growing adaptation of buprenorphine. States and counties use the N-SSATS numbers of clients to make funding decisions and to educate their communities as to the need for increased treatment.

A common reason for sampling is to obtain information about population parameters with lower cost and potentially lower respondent burden. The implementation of a sampling strategy that provides the opportunity for unbiased estimates assumes that:

• Population parameters can be known; and

• Data come from a probability sample—a sample in which each case in the population has a known and non-zero probability of being selected into the sample---only the laws of chance are included in the sample

The use of strategies included in non-probability sampling such as convenience samples, quotas, or snowballing, among many others, are unlikely to capture a relevant set of data because of the uniqueness of each facility based on the many different types and operations of treatment. Probability samples have a better track record for capturing representative samples than non-probability samples.

Among the various probability sampling methods, the only appropriate method to begin to be able to provide appropriate estimates for each of these very different types of service delivery would be Stratified Random Sampling which requires information about each case and thus may only be used if basic census information is available followed by sampling among groups of interest. In other words, a census would be required in order to be able to provide an appropriate sampling frame. Given that approximately 10 percent of facilities turnover (close/stop providing treatment or start providing treatment) in any given year, a census from any given year would be outdated with no size characteristics available for sampling for many facilities.

A systematic sample would need to be taken of facilities that fall within the above categories so that a representative sample could be drawn from facilities of different sizes and other characteristics. A “measure of size” of facilities would need to be included to sample facilities of different sizes (small, medium, and large) so as not to bias the sample selection process in selecting more facilities of a certain size. The measure of size for facilities would be based on the reported counts of persons in treatment within a treatment type as of the survey reference date.

If only a sample of facilities collected client caseload data in the full-scale 2019 N-SSATS, for example, then the ability to use measure of size calculations for the next full-scale N-SSATS would be lost, especially with an approximate 10 percent turnover of facilities. The variables used for measure of size (i.e., reported counts of persons in treatment within a type of treatment as of the survey reference date) would only be available from the 2019 sampled facilities that remain in the universe and not the universe of substance use treatment facilities.

The N-SSATS is the only source that can provide state and sub-state-level data on both government operated/funded facilities and facilities that are privately owned/operated (for profit and non-profit) which encompass the full spectrum of organized service providers providing services in at least one type of care (inpatient, residential, outpatient, detoxification), and the characteristics of the facility’s current caseload as of a single day (survey reference date).

In efforts to keep burden at a minimum and to collect only the absolutely necessary numbers, only basic client counts are requested (total numbers of clients in each of the major categories of treatment, numbers under age 18, numbers receiving each of the opioid and alcohol medications as of a survey reference day, and annual admissions). No demographics (age breakouts [other than under 18], race/ethnicity, gender, substances used, etc.) are collected. The survey reference day is requested to keep burden at a minimum. The last working day of March is used. In clinical settings, people refer to the number of patients at any given time as the “patient census.” The last working day of March is a “typical” day in that it is not at the end of a fiscal year, when school is either just starting up or on vacation, or directly around a major holiday or leading into a major holiday when many people may be out of family vacation. Additionally, facilities requested a spring date rather than an autumn date. A one-day count is relatively easy to pull from records (from what respondents indicate) and requires no de-duplicating or monthly adding of unique clients as annual counts would require. The numbers asked for are numbers that facilities can easily pull from their electronic health records; it would not be necessary to pull individual records to obtain the numbers. This method of client counts is the same as what the Census Bureau uses for their decennial census, the number of persons on the survey reference day. It provides the number of clients on a specific day. When these number are used, in publications or in data requests asking about numbers of clients, it is always indicated that the numbers are one-day counts and that they are not annual counts. It is indicated that they provide a “snap-shot” of what the treatment numbers look like on that particular day, which is understood to be a “typical” day (meaning not a fiscal count or a holiday count).

The overall response rate in 2018 N-SSATS was 92.4 percent and by state/territory, ranged from 80 percent (Virgin Islands where one of five facilities did not respond) to 100 percent (where all facilities responded in Palau, Micronesia, Guam, Marshall Islands, and Northern Mariana Islands.) Among the 50 states (and DC), the response rate ranged from 85.3 percent (DC) to 98.2 percent (WV). Only 11 states plus DC had response rates below 90 percent. Overall, most of the small counties achieved a 100 percent response rate. Many other counties appeared to be missing only one facility. A total of 39 counties (out of 2,382 counties) had no responders, most of which included only one facility, and were spread among 20 states/territories. Some of those non-responding facilities were not directory eligible facilities. There appeared to be no systematic reason or locale for the non-responders. While not a 100 percent response from facilities, considering the overall high response, numbers of clients for most communities and counties would provide an accurate accounting of numbers of clients in treatment on the survey reference day and numbers provided to counties and states would be accurate.

SAMHSA’s Inventory of Behavioral Health Services (I-BHS, OMB No. 0930-0106), which is updated and maintained with the data collected by the N-SSATS and the National Mental Health Services Survey (N-MHSS, OMB No. 0930-0119), is used by SAMHSA, other Federal agencies, and researchers as a sampling frame for other surveys and studies of mental health and substance use treatment facilities. No alternative comprehensive listing of such facilities exists. An important stratification variable in drawing a representative sample of facilities is facility size as measured by the number of clients served. Without this information for all facilities, the usefulness of the I-BHS as a sampling frame would be severely diminished.

Estimation Procedures

Selected N-SSATS data items are imputed for missing values using generally accepted methodologies. While imputation is used, there are generally very few numbers that need to be imputed. For example, in 2017, only 4 client count numbers for facilities were imputed. These four facilities were residential facilities and the imputed numbers were based on the facilities’ reported designated bed counts. Overall, the effect of imputation was negligible. There is no imputation of missing data for any of the services (or locator) data.

2. Information Collection Procedures

1. I-BHS On-line, Facility Applications, and Augmentation

I-BHS Online: The I-BHS is designed to be continuously updated by states as they license or certify facilities, decertify or cancel licenses for facilities, and learn of facilities that have gone out of business or moved. The update process is online, so states can easily update information on the facilities in their states. The I-BHS Online update forms used by state representatives to enter or change facility registration are included as Attachment A1.

I-BHS Facility Applications: New facilities can request to be included in I-BHS through an online facility application form found on the locator (https://findtreatment.samhsa.gov). I-BHS staff will verify all facility requests to avoid duplication or the addition of inappropriate facilities to I-BHS. The information on new facilities will also be passed to the cognizant state agency for possible designation as state-approved. The I-BHS facility registration form is provided at Attachment A2.

I-BHS Augmentation: The facility information provided by states is augmented by SAMHSA through searches of directories and other data bases. In 2017, the data bases searched for substance use treatment facilities included the ABI (American Business Information) file and the American Hospital Association (AHA) directory and for mental health treatment facilities included the ABI and AHA, as well as NASMPHD-National Association of State Mental Health Program Directors, American Residential Treatment Association, , and the National Council for Behavioral Health. All potential treatment facilities identified from these sources are matched to the I-BHS to identify duplicates. In addition, a processing step matches the potential new facilities against augmentation runs from prior years, to eliminate facilities that had been identified and screened out in earlier augmentation efforts. The remaining unmatched facilities are then screened by phone to identify those that provide substance use treatment or mental health treatment services. These screening phone calls and the N-SSATS and N-MHSS surveys often generate reports of additional facilities, because respondents will volunteer that their parent organization has treatment facilities at several sites. The facilities identified in this way are also matched against the I-BHS, and the questionable matches and non-matches are screened by phone. There will be an augmentation each year, several months prior to the start of the N-SSATS and N-MHSS surveys. The augmentation screener questionnaire used to screen the questionable matches and non-matches is included as Attachment A3.

2. N-SSATS

The 2019 N-SSATS will be conducted through an online web survey, with a mail questionnaire option and telephone follow-up of non-respondents. An advance letter will be mailed to the facility director six weeks before the March 29 reference date, notifying/reminding them of the survey. (Many Directors have participated in the N-SSATS for years.) A cover letter, a set of on-line access instructions, and a list of frequently asked questions (FAQs), will be mailed on or about March 29. (See Attachment B6 for a copy of the advance letter, Attachment B7 for a copy of the cover letter, online questionnaire instructions, and FAQs, Attachment B2 for a copy of the 2019 questionnaire, Attachment B4 for a copy of the screens for the online response option, and Attachment B13 for a copy of the web pages for the N-SSATS information website.)

Approximately four weeks after the initial mailing, a thank you/reminder letter will be faxed or mailed to all facilities (see Attachment B8 for a copy of the thank you/reminder letter). Facilities that have not responded by the last week in May will be sent a second packet including a cover letter, Frequently Asked Questions, a state profile, and instructions for completing the questionnaire on the web. (See Attachment B9 for a copy of the second mailing cover letter.) Facilities that have not responded by the middle of July will be sent a third packet including a cover letter, web flyer, Frequently Asked Questions, state profile, and paper version of the questionnaire (See Attachment B10 for a copy of the third mailing cover letter.)

Reminder calls will begin in mid-June. During the initial reminder call, respondents will be encouraged to respond by mail or web, but they may also respond by telephone. After every facility has received one reminder call, all telephone efforts will be directed toward completing the questionnaire by telephone through a CATI interview (see Attachment B5 for a copy of the 2019 N-SSATS CATI questionnaire.) The CATI follow-up will continue through the end of October. These procedures resulted in a response rate of 92 percent in 2016 and 90 percent in 2017.

At the end of data collection, a final letter will be sent to all responding facilities thanking them for their completion of the N-SSATS questionnaire (see Attachment B11 for a copy of the completion thank you letter).

So that state-approved facilities identified after the N-SSATS survey do not have to wait a full year to be added to the online Locator, the N-SSATS BC survey will be conducted during the year, using a subset of the N-SSATS questions. An advance letter will be sent to the new facility describing the Locator and inviting the facility to call a toll-free number to schedule a brief interview. (See Attachment B12 for a copy of the N-SSATS BC advance letter and Attachment B3 for a copy of the N-SSATS BC) If the facility does not call, the N-SSATS survey contractor will make one attempt to contact the facility by telephone. Facilities that complete the N-SSATS BC, and those that do not, will be included in the next full N-SSATS survey.

3. Methods to Maximize Response Rates

I-BHS: The universe of behavioral health treatment facilities is not static. Experience with the N-SSATS and I-BHS has shown that in a 12-month period, approximately 12 to 15 percent of facilities close and roughly the same number of “new” facilities are identified. Additionally, another 20 percent of facilities change their basic contact information (name, address, telephone number) each year. Aware of this turnover, SAMHSA takes all reasonable measures to ensure that the I-BHS is as complete as possible. Since no other comprehensive listing of treatment facilities exists against which to judge the completeness of the Inventory, the only avenues available are to collaborate with state agencies to maintain the listings and to do regular augmentations to identify new facilities that state agencies may not have authority over. Facilities remain on the I-BHS until SAMHSA receives evidence that the facility is no longer providing treatment services or is otherwise ineligible. Thus a facility that does not respond to the N-SSATS stays on the I-BHS until there is evidence that it no longer exists.

N-SSATS: The methods to maximize response rates will be those that proved successful in the 2017 N-SSATS. They include:

• advance letters to alert facility directors of the upcoming N-SSATS mailing;

• state letters of support mailed with the N-SSATS packet;

• an online survey with the option to request a paper questionnaire if preferred;

• pre-filled responses in the online survey for selected questions that have little year-over-year change (e.g., public versus private ownership, hospital type, etc.) The pre-filled responses help reduce burden and improve survey response without impairing the integrity of the data;

• A second mailing packet again providing instructions on how to complete the survey on the web;

• a third mailing packet providing a paper questionnaire to all non-respondents;

• reminder telephone calls, e-mails, faxes, and re-mailings as needed;

• a toll-free N-SSATS hotline that facilities may call with questions about the survey;

• an N-SSATS e-mail address that enables facilities to e-mail questions about the survey;

• an N-SSATS informational website that provides N-SSATS history and other material to respondents;

• tracing and locating efforts to determine whether a facility is still in business, closed, or has merged with another facility; and

• telephone interviews to collect the information from those not responding online or by mailed paper questionnaire.

4. Tests of Procedures

No large-scale pretests (more than 9 respondents) of N-SSATS are proposed for the next three years. Most items in the questionnaire have been in place for many years. If rewording or small changes in questions are required during the period of approval, the N-SSATS questionnaire would be tested on a small number of facilities (9 or fewer facilities), and the respondents would be debriefed by telephone to verify that they were interpreting the items as intended. The new Electronic Health Record questions are currently being developed and under review through OMB; However, it is not expected that they will be finalized in time for addition into the 2020 N-SSATS. If they are finalized, a new package will be submitted.

5. Statistical Consultants

The data are collected under a contract with Eagle Technologies which has a subcontract with Mathematica Policy Research, Inc., for the N-SSATS forms design, field work, and data entry and cleaning. The project directors for the two contractors are:

Eagle Technologies Mathematica Policy Research

Doren Walker Karen CyBulski

571-382-6738 609-936-2797

Eagle Technologies is also responsible for the management of the I-BHS systems, the statistical aspects of the N-SSATS (primarily imputation of missing data), and preparation of the Locator/National Directory and the annual N-SSATS report and state profiles.

The SAMHSA Project Officer and Co-Project Officer are:

Cathie Alderks Nichele Waller

Statistician Statistician

240-276-1269 240-276-0547

LIST OF ATTACHMENTS

Attachment A1 I-BHS Online State Add/Update Forms

Attachment A2 I-BHS Facility Application Form

Attachment A3 Augmentation Screener Questionnaire

Attachment B1 N-SSATS 2019 Questionnaire (Version A)

Attachment B2 N-SSATS 2020 Questionnaire (Version B)

Attachment B3 Between Cycles N-SSATS (N-SSATS BC)

Attachment B4 N-SSATS 2019 screens for online questionnaire

Attachment B5 N-SSATS 2019 CATI Questionnaire

Attachment B6 N-SSATS 2019 Advance Letter

Attachment B7 N-SSATS 2019 Cover Letter, Online Questionnaire Access Instructions, and Frequently Asked Questions sheet

Attachment B8 N-SSATS 2019 Thank You/Reminder Letter

Attachment B9 N-SSATS 2019 Second Mailing Cover Letter

Attachment B10 N-SSATS 2019 Third Mailing Cover Letter

Attachment B11 N-SSATS 2019 Completion Thank You Letter

Attachment B12 Between Cycles N-SSATS (N-SSATS BC) Advance Letter

Attachment B13 Web pages for the 2019 N-SSATS information website

19