BRFSS 2018 Justification Part A 01MAR2018X

Goal. To conduct Behavioral Risk Factor Surveillance System (BRFSS) surveys in collaboration with U.S. states, territories, and the District of Columbia. The surveys will produce state- or sub-state jurisdiction-level data about health-related risk behaviors, chronic health conditions, use of preventive services, and emerging health issues. CDC conducts an annual field test to prepare for the primary BRFSS information collection.
How data will be used. CDC and BRFSS partners use BRFSS data to plan for and evaluate public health programs at the (sub) state level. For most states, BRFSS data is the only source of health information that is targeted to state and local public health needs. CDC also creates a national level dataset that is used by HHS to evaluate its progress toward Healthy People 2020 goals and other policy needs.
Methods of collection. Each year, an independent sample of respondents is drawn for each participating state or jurisdiction. Information collection is conducted in a continuous, three-part telephone interview: screening, participation in the common BRFSS core, and participation in optional modules that states may use to customize survey content for their specific needs. Interviews are conducted on telephone landlines and mobile telephones. A few US territories with low telephone coverage conduct in-person interviews. Each state or territory coordinates BRFSS administration within its jurisdiction.
Respondent population. Adults > 18 years of age who live in private households or college housing.
Analysis techniques. Each state or territory submits a de-identified dataset to CDC for cleaning, weighting, and compilation. Because sample size and survey content vary by state or territory, CDC provides guidance on statistically appropriate uses of BRFSS data and technical assistance, as needed, on survey content and administration.

Revision Request

This request is to continue the data collection associated with the BRFSS for calendar years 2018, 2019, and 2020. The first two months of data collection for calendar year 2018 was covered by OMB control number 0920-1213, an emergency clearance approved consistent with the understanding that significant delays in the internal Departmental clearance process prevented the Agency from implementing the collection in a timely manner.

This request includes the most current even-numbered year core questionnaire (to be fielded in 2018 and 2020), and the most current odd-numbered year questionnaire (to be fielded in 2019 and 2021). Each year, however, CDC may make modifications to the questionnaire, based on emerging needs or improvements in data quality. CDC may continue to use the non-substantive change mechanism to modify existing questions, including those designed to add depth/additional detail and cycle in and out methodological modules that have been used in the past years (with modest updates). However, the addition of new question topics or modules or modifications to the sample will require that CDC submit a full revision of the ICR package to OMB.

Justification

Circumstances Making the Collection of Information Necessary

The National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP) was established at the Centers for Disease Control and Prevention (CDC) to provide coordinated leadership and support for prevention and control programs specifically related to chronic diseases or conditions. NCCDPHP works with partners from multiple sectors of society to create and disseminate expertise, information, and tools that reduce the burden of disease on the American public, and to identify populations that are most in need of strategies to promote physical and mental health and prevent disease. Key partners within CDC include the National Center for Immunization and Respiratory Diseases (NCIRD), the National Center for Environmental Health (NCEH), the National Institute of Occupational Safety and Health (NIOSH), the National Center for Injury Prevention and Control (NCIPC), the National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention (NCHHSTP), the Office of Public Health Preparedness and Response (OPHPR), and the National Center for Health Statistics (NCHS). Surveillance efforts that quantify disease and risk factors, and identify opportunities for prevention, are central to CDC’s planning and evaluation efforts. CDC’s authority to collect information for this purpose is provided by the Public Health Service Act (Attachment 1).

Scientific research has identified personal health behaviors that play a major role in premature morbidity and mortality. Patterns of behavior that affect health or predict adverse effects on health are called behavioral risk factors. For example: lack of physical activity is a behavioral risk factor for obesity, type 2 diabetes, cardiovascular disease, and other diseases and conditions; and alcohol consumption (including binge drinking) is a behavioral risk factor for injury, unplanned or unprotected sexual behavior, and a contributor to numerous chronic conditions. Risk behaviors are thus key targets for prevention. Timely and accurate information about risk behaviors is needed to plan, initiate, support, and evaluate public health programs designed to prevent, control, or mitigate disease and disability.

Due to the geographic, cultural, demographic and economic diversity of the U.S., the methods used to produce national estimates provide only limited insight into regional, or state-specific variability in health status and health-related behaviors. State and local health agencies are focal points for public health program planning and implementation. Nationally representative datasets that do not have the ability to be geographically disaggregated due to sample size limitations may not be informative about conditions found in any given state, or detailed enough to assist federal, state, and local health agencies inform effective allocation of public health resources. Surveys that sample at the state level are needed to monitor state-level prevalence of the major behavioral risks for premature morbidity and mortality.

In 1984, CDC and 15 states collaboratively initiated behavioral risk factor surveys designed to address the states’ information and public health needs. Since that time, CDC has provided technical assistance and funding to an increasing number of states and territories, and survey content and methods have become increasingly sophisticated. The surveys are now known as the Behavioral Risk Factor Surveillance System (BRFSS), managed by NCCDPHP’s Division of Population Health (DPH). Under the current BRFSS cooperative agreement program, CDC provides funding and technical assistance to 57 awardees including 50 states, the District of Columbia, and six territories (collectively referred to as “states” in this document). A list of current awardees is provided in Attachment 2. The primary purpose of the BRFSS is to provide the information that these awardees need in order to plan and evaluate public health programs at the state and sub-state level.

BRFSS information collection is conducted annually. The CDC and BRFSS awardees produce a new set of state-specific BRFSS questionnaires each calendar year (i.e., 2018 BRFSS series, 2019 BRFSS series, etc.). Information is collected through telephone interviews except in a limited number of jurisdictions with low telephone coverage, where interviews are conducted in-person. The methods for selecting respondents and determining questionnaire content reflect the system’s origins. An independent sample of adult, non-institutionalized respondents is drawn for each state and is based on the state’s parameters for state-level or sub-state analysis. The questionnaire is based on modular design principles allowing rapid customization of content. Each state’s annual questionnaire includes a common core for even-numbered years (see Attachment 3a) and a common core for odd-numbered years (see Attachment 3b) that is administered by all states, and standardized optional modules (see Attachment 4) that may be fielded at the states’ discretion. The number of optional modules available to states in any single year is dependent on state health departments’ needs and funding. Data sets are submitted to CDC for cleaning and weighting, and returned to the state of origin for its use. In keeping with its historical role in promulgating standards that strengthen data quality and comparability, CDC, in collaboration with the states, provides standard protocols for BRFSS data collection which all states are encouraged to adopt. The BRFSS collaboration among CDC and the states also develops and validates questions for both the core component(s) and the optional modules, and provides technical assistance to states on methodological issues such as sample selection, data quality, weighting, and the interpretation of findings. CDC makes state-level BRFSS data, as well as national datasets, broadly available through the BRFSS Web site along with guidance on statistically appropriate uses of the data.

The interactive partnership between CDC and BRFSS awardees results in the collection of surveillance information that is adaptive to both state and national needs, and uniform where states have common interests. Over time the BRFSS has developed into an important source of information for federal agencies and the public as well as state and local health departments. In recognition of the fact that HHS uses BRFSS as a unique and influential source of public health information developed with federal assistance, CDC applied for and received OMB approval for BRFSS annual surveys beginning with the 2015 cycle of data collection. CDC also seeks to ensure comparability across data sets by reviewing questions on similar topics on other surveys. The process by which questions are adopted for use on the BRFSS is provided in Attachment 12. The process includes reviews of questions on other federal surveys. In 2018 the Sexual Orientation and Gender Identity (SOGI) optional module has been revised to adhere to the question format used by the National Health Interview Survey.

CDC supports periodic updates of BRFSS content and/or the data collection system. Two types of routine updates are based on similar survey methods and are associated with the current Revision request (OMB No. 0920-1061). CDC submits each update to OMB on an annual basis. Although the updates are related, each will be submitted as a separate request to OMB.

Annual field test of proposed changes. Each year’s full-scale survey is fielded from January 1 – December 31. Approximately 5-8 months before the January 1 launch date, CDC conducts a limited field test to identify issues that may affect BRFSS implementation or data quality. The specifics of each year’s field testing plan are submitted to OMB through the Change Request mechanism as an Annual Field Test Supplement (see Attachment 13 for an example of a Field Test Questionnaire).

Annual update of the BRFSS core survey and optional modules.

Approximately 3-5 months before the January 1 launch date of the annual survey, and after reviewing the results of field test, CDC submits a Change Request to OMB that includes the updated BRFSS core survey and optional modules for the upcoming year. When the changes are limited to modifing existing questions, including those designed to add depth/additional detail and cycle in and out methodological modules that have been used in the past years (with modest updates), CDC will submit these changes as non-substantive changes. However, the addition of new question topics or modules or modifications to the sample will require that CDC submit a full revision of the ICR package to OMB. This Change Request outlines all adjustments to the estimated number of respondents, estimated burden, or other issues.

BRFSS content and methods also are refreshed through information collection activities that involve different methods and are not part of this OMB number. These activities, which are separately submitted to OMB for approval (using the CDC/ATSDR Formative Research and Tool Development generic package (OMB #0920-1154) include methodological studies to strengthen data collection procedures and the utility of the information collected, for example, cognitive testing of questions, pilot tests, or other protocol enhancements or callback surveys (conducted to collect detailed information on a single health topic, such as asthma). Any and all additional information collection activities will be submitted to OMB in separate requests, and may be requested under current generic approvals.

Purpose and Use of Information Collection

The BRFSS data will be used for a number of purposes by a diverse set of users. The primary uses of the data are listed below:

BRFSS data is used by states to identify specific program needs and track health status over time.
States use BRFSS data to help them establish and track state and local health objectives, plan health programs, implement disease prevention and health promotion activities, and monitor trends.
BRFSS data is used to compare weighted state-level behavioral health risk and health status information taken from persons residing within all U.S. states, D.C., the U.S. Virgin Islands, Guam, and Puerto Rico.
CDC disseminates a publicly available annual BRFSS dataset (see the BRFSS website at www.cdc.gov/brfss). This dataset is frequently used by public health officials in government at the national, state and local level as well as researchers at university and non-profit organizations. In the past, these data have been used for program evaluation, research and reporting related to health status, chronic disease indicators and health risk and risk preventive behaviors. Data have also been used for trend analyses, tests of differences among (demographic or jurisdictional) subpopulations, multivariate analyses of health outcomes and other statistical processes.
State health department websites as well as a CDC website will be used as platforms to illustrate differences in health and behaviors using mapping and charting software. CDC is committed to improving its webbased tool by making the strengths and weakness of such comparisons more transparent.
BRFSS data informs a variety of data resources, programs and organizations which use the data as a basis for smaller area estimation. These include but are not limited to the Health Indicators Warehouse, County Health Indicators, the Robert Wood Johnson Foundation, and HealthyPeople 2020.
BRFSS data may be used to draw comparisons from data taken from identical and/or similar questions on other surveys using other modes thereby creating a means for validation and comparisons across population samples.
The BRFSS data collection system is also used by states as a vehicle to add state specific questions exclusively for state health department needs. Although these data are not part of the public use dataset, they are essential to state health departments for planning and evaluation of public health resources.

Data collection based on state-level sampling also permits the analyses of data at the local level when sample sizes within county or MSAs are large enough for statistical interpretation. The ability to identify state and sub-state differences optimizes program interventions designed by state health departments.

The annual field test has distinct objectives. Field testing is the final check of changes in the questionnaire which have occurred in the preceding year. Field testing is conducted in a manner that mimics the full-scale project protocol, to the degree that is feasible. Field testing is the final means by which changes are made in data collection methods and data collection software is tested. Field tests are used to identify problems with instrument documentation or instructions, problems with conditional logic (e.g., skip patterns), software errors or other implementation and usability issues. Field tests are conducted in a single state, using only those parts of the questionnaire which have been substantively changed or sections of the extant questionnaire which lead into new or updated questions. In some instances, extant sections of the questionnaire may be field tested if they are topically related to new items on the questionnaire. For example, if a new question on disability is added, extant disability questions are included in the field testing to ensure that respondents do not feel that the questions are redundant or overlapping. Field testing is not intended to replace cognitive testing, it is only to check to be sure that questions which have already been thoroughly vetted are appropriately placed on the BRFSS. Sections of the questionnaire which are unchanged and unrelated to new or modified sections of the questionnaire are not field tested, although the demographic sections of the core are included in the field test. Results of the field test are used to inform development of the upcoming year’s BRFSS questionnaire(s) and the technical assistance and implementation guidance that CDC provides to BRFSS partners. Field test data are not incorporated into the analytic BRFSS datasets.

Use of Improved Information Technology and Burden Reduction

The BRFSS data will be collected using list-assisted random digit dialing (RDD) landline and cell phone telephone samples. Given the need for state-level samples that are large enough for statistical analyses, telephone surveys offer a cost effective method of data collection. In addition to their cost advantages, telephone surveys are especially desirable at the state and local level, where the necessary expertise and resources for conducting area probability sampling for in-person household interviews are available in many state health departments. Interviewers will use Computer Assisted Telephone Interview (CATI) software to enter data directly into a database. Use of CATI software promotes efficiency in two ways: skip patterns can be programmed to route respondents only to questions that they are eligible to answer, and real-time quality control checks can be used to eliminate some errors which may have been caused by manual data entry procedures.

Efforts to Identify Duplication and Use of Similar Information

For most states and territories, the BRFSS provides the only sources of data amenable to state and local level health and health risk indicators. Extant data on these topics are available at the national level, but do not include sufficient sample size to determine whether there are measureable changes/trends in health risk behaviors at lower geographic levels. National surveys such as the National Health Interview Survey (NHIS, OMB No. 0920-0214, exp. 12/31/2016), the National Tobacco Survey (OMB No. 0920-0828, exp. 7/31/2015), the National Health and Nutrition Examination Survey (NHANES; OMB No. 0920-0950, exp. 11/30/2016), among others offer data for prevalence estimates at the national level. The BRFSS differs in that it samples at (sub) state levels, and produces direct, not modeled, estimates for state and local geographic jurisdictions. It also provides a state level public use dataset on a broad range of topics, many of which are not included in national surveys.

In some cases, state prevalence may be modeled by other data collections. The National Adult Tobacco Survey and the National Health Interview Survey have both been used to model prevalence estimates at the state level. However they may not provide sufficient data from which direct state estimates can be derived nor do they allow for direct local area estimation. Moreover, in most instances state level data modeled from national surveys use national level control totals for weighting, while the BRFSS uses (sub) state control totals for all post-data collection raking weights. National surveys use modeled estimates to obtain state and local prevalence estimates, however, these modeled estimates cannot be used to evaluate interventions that public health at the state and local level may have implemented. Many states use BRFSS data to evaluate their public health interventions.

Data from the BRFSS provide a means by which states can tailor data collection to their unique needs. Optional modules which are important in one state may not be as salient in others. For example, the excess sun exposure optional module may be more critical in coastal states than in other jurisdictions. The addition of state-added questions is also a means by which specifically targeted information critical to a particular state may be obtained, using the infrastructure of the BRFSS.

Impact on Small Businesses or Other Small Entities

There will be no impact on small business.

Consequences of Collecting the Information Less Frequently

Annual data collection allows more detailed trend analyses than less frequent data collection. The BRFSS minimizes the number of questions included in the annual survey by including a rotating core (in even and odd numbered years) for a select number of indicators which have more detailed response sets.

Special Circumstances Relating to the Guidelines of 5 CFR 1320.5

There are no special circumstances.

Comments in Response to the Federal Register Notice and Efforts to Consult Outside the Agency

A 60-day Notice was published in the Federal Register on October 16, 2017, Vol. 82, No. 198, pp. 48087 (Attachment 6a). A total of three public comments were received during the 60 day public notification of intent to collect data using the BRFSS. Attachment 6b includes the comment and CDC’s response The first comment was from the Truth Initiative. The comment focuses on the inclusion of new questions on tobacco use to include asking respondents about the use of flavors in hookah use, e-cigarette use and cigars. The comment also requests that the current e-cigarette module be made into a core component of the BRFSS questionnaire. The comment in its entirety will be forwarded to the Office of Smoking and Health (OSH) which sponsors all tobacco use questions on the BRFSS for their review. The OSH is currently requesting a vote by the state health departments to include e-cigarette use in the core for the 2019 version of the questionnaire. Therefore this portion of the recommendation by the Truth Initiative will be considered by the states in the Spring of 2018 as part of the questionnaire development process (as described in Attachment 12).

The second comment was from the Academy of Nutrition and Dietetics. This comment focused on the length of the survey, some concerns about validity and reliability and the use of rotating core and optional module questions. The suggestions of the comment were to reduce the length of the questionnaire to 15-20 questions which are standard across states, and eliminate the optional modules. The suggestions also included removing questions that were redundant with national surveys. The suggestions to remove the state specificity of the questionnaire is contrary to the purpose of the BRFSS, which has as its mission to provide state-level data that can be tailored to the needs of each state. The response also directs the Academy to the literature on reliability and validity of the BRFSS, which is an ongoing activity within the Population Health Surveillance Branch. The suggestions provided by the Academy of Nutrition and Dietetics Academy of Nutrition and Dietetics will be forwarded to the CDC programs which sponsor questions on nutrition and physical activity.

The third comment was provided by the Williams Institute of the UCLA School of Law. This comment focused on the need for continued information related to SOGI which has been provided as an optional module on the BRFSS since 2014. The authors expressed a preference for the SOGI questions to be moved to the core questionnaire. The response focused on the process by which questions are adopted for the core, and the need to maintain flexibility for the states in the determination of their questionnaires.

Explanation of any Payment/Gift to Respondents

Not applicable.

Protection of the Privacy and Confidentiality of Information Provided by Respondents

Overview of the Data Collection System

Random digit dialing (RDD) telephone samples will be delivered to the states on a monthly or quarterly schedule. Information collection will be implemented by state health departments or their designees. States will administer the core/rotating/emerging core questions without change. States will determine which of the optional modules will be included in their jurisdictions and whether the state will split the sample to allow for the inclusion of a large number of optional modules. Field operations are managed by state health departments and/or their contractors following The Data Collectors’ Protocol provided by the BRFSS (see Attachment 10). States submit data to CDC for final cleaning, weighting, the production of analysis datasets, and other technical assistance as needed. Computer-assisted telephone interviewing (CATI) programming is provided by the CDC to states to convert the BRFSS questionnaire into a CATI interface from which interviewers will read and record answers to each question. States may opt to use their own CATI programming software. States run edit checking programs against the data and submit to the CDC on a monthly/quarterly basis. CDC then conducts additional data quality processes and summarizes the data in YTD reports provided to the states. At the end of each calendar year, data are finalized and weighted.

The datasets provided to the states at the end of the year include a large number of variables on calling attempts, final calling outcomes, questionnaire item responses and calculated variables. A subset of the data set provided to the states is published on the BRFSS website for public use.

Items of Information to be Collected

The BRFSS core questionnaire (see Attachment 3a for even-number year core questionnaire and Attachment 3b for odd-numbered year core questionnaire) includes information on health status, health risk and risk preventive behaviors, as well as basic demographic information. Optional modules which are selected by individual states, based on their information needs, must also be implemented as written without changes in wording. Optional modules cover a range of health topics (see Attachment 4). Other than phone numbers, which are part of the original sample files sent to the states, no information in individually identifiable form (IIF) will be collected from respondents during the telephone interviews. On occasions when states make appointments to call selected respondents back after portions of the survey have been completed, first names may be given to ensure that the remainder of the interview is conducted with the same individual. The BRFSS uses the HHS Safe Harbor guidelines (https://www.hhs.gov/hipaa/for-professionals/privacy/special-topics/de-identification/index.html) to determine suppression of variables from public use information. Researchers who request access to information not provided in the public use dataset may use the Research Data Cebter (RDC) hosting agreement with the BRFSS. States must develop and maintain procedures to ensure respondents’ privacy, assure and document the quality of the interviewing process, and supervise and monitor trained interviewers. The CDC provides states with guidelines for training interviewers and standard procedures for monitoring a minimum of 10% of all interviews. The Data Collectors’ Protocol (Attachment 10) provides instruction to states about data storage and confidentiality of responses as well as data quality.

How Information Will Be Shared and For What Purpose

Since state health departments and/or their designees are the data collectors for the BRFSS, information will originate with the states. States may determine whether and how their data will be released to third parties. The CDC maintains an upload website by which data are submitted monthly/quarterly. CDC does not transmit data from one state to any other, with the exception of cell phone interviews of persons who have an area code from one state, but who actually live in another state. Telephone numbers are not linked to respondents. Files containing RDD telephone samples are kept separately from files which include responses to questionnaire items. CDC receives only de-identified records. Sample files contain sequence numbers which are provided by the sampling vendor and used by data collectors (the states or their designees) to determine calling outcomes for each phone number. The CDC does not receive full phone numbers in the sample file. Sample files received by the CDC and states which have contracted data collection include only area code and prefixes of phone numbers which are associated with sequence numbers. States which have internal data collection systems and contracted data collectors have sole access to both sequence numbers and full phone numbers during the data collection process. States keep responses to the BRFSS questionnaire separately from sample files. After data collection, sequence numbers are recoded to prevent subsequent links of sample files and responses to questions by any person or organization involved in data collection. State level data sets are owned by individual states. A subset of state data sets is provided for public use. Public use data sets have been stripped of a number of variables which provide locational information on the respondents including zip codes, and county identifiers for counties with adult populations of less than 10,000, occupational information, uncategorized ages of respondents, and detailed race. CDC may provide data with locational information for internal users to produce small area estimates of health indicators.

Impact of the Proposed Collection on Respondents’ Privacy

BRFSS sample files include phone numbers, and some addresses. Addresses are only available for telephone numbers in the landline sample where the sample contractor has been able to match phone numbers and addresses. These addresses are used by some states to send advance letters to households in the landline sample. No address matching is done for cell phone numbers. If a state does not send advance letters, addresses are not included in the sample files. Approximately 40 states use advance letters, with about 60% of the landline sample having a matching address.

Since sample files are separate from datasets, no phone numbers or addresses are included in the datasets. No dates of birth, last names, or email address are obtained. Information that details race/ethnicity, occupation and small geographic residence (such as county or zip code) is suppressed in the public use dataset. In order to determine which variables to supporess, the BRFSS uses the HHS Safe Harbor guidelines (https://www.hhs.gov/hipaa/for-professionals/privacy/special-topics/de-identification/index.html).

How Individuals Are Informed That Providing Information Is Voluntary or Mandatory

Individuals participating in the BRFSS are informed that they do not have to participate and that they may refuse to answer any question.

Opportunities to Consent

Verbal consent is obtained during the initial contact and screening process (see Attachment 8). The introductory script, including the voluntary nature of the survey, precedes the survey questions.

How Information Will Be Secured

Access to state data sets will be limited to the states themselves and CDC contractors and staff who conduct weighting and data cleaning procedures. Security measures include: 1) Physical controls: CDC facilities are secure, ID accessed buildings. Data will not be stored in hard copy formats; and 2) Technical controls: All electronic data are stored on secured servers protected with firewalls and passwords. All employees are trained on data security measures by taking appropriate HHS courses online. All data collection and records management practices and systems adhere to HHS and CDC IT policies and procedures.

Privacy Act Determination

The Privacy Act does not apply. The BRFSS data will be collected using list-assisted random digit dialing (RDD) landline and cellular telephone samples. No IIF will be collected, filed or retrieved by the name of the individual or other unique respondent identifier such as social security number.

Annual Field Test

The information collection procedures and privacy safeguards for the annual field test are modeled on procedures and privacy safeguards for the main BRFSS survey. Any departures from standard procedures will be identified in the annual field test IC request.

Institutional Review Board (IRB) and Justification for Sensitive Questions

CDC has determined that the BRFSS information collection is exempt from the requirements of 45 CFR 46. IRB approval is not required.

The BRFSS includes standard demographic questions (such as race and income category) which may be considered sensitive. In addition, some questions regarding preventive behaviors, such as HIV testing, may be considered sensitive. Participating states have used these standardized questions on state behavioral risk surveys for many years. Sensitive questions are necessary to identify changes in behaviors and/or self-assessments of health status when aggregated to local or state levels. In order to assess changes in health risk behaviors, it is necessary to ask questions of a sensitive nature. This sensitive information would be pertinent to determine state needs for health programs and services.

Estimates of Annualized Burden Hours and Costs

Respondents are adults > 18 years of age. Information collection is conducted in a continuous, three-part telephone interview: screening, participation in the common BRFSS core, and participation in optional modules. Respondent burden is estimated separately for each step.

Although the number of interviews varies from state to state--based on the needs, population size and diversity of each state--a minimum number of interviews is set by the state in the sampling process. The estimated population for each U.S. state or territory, and the size of the BRFSS state-level sample for 2016 is provided in Attachment 7a. Approximately 55% of interviews were conducted on landlines and 45% on cell phones in 2016. Each year the proportion of interviews conducted by cell phone has increased in response to the use of cell phones by the US populations (Blumberg & Luke, 2016). Therefore for the 2018-2020 BRFSS, a larger proportion (50%) of cell phone interviews is anticipated. The screening questions for the two groups are slightly different and are presented in Attachment 8. Since the cooperation rate (based on AAPOR cooperation rate #2) for all BRFSS in 2015 was 64% for landline respondents and 82% for cellphone respondents, it is estimated that 375,000 landline respondents and 292,682 cell phone respondents will complete the screening questions. The estimated burden per response is one minute.

After completing the screening interview, respondents who are eligible and agree to participate will proceed to the core section of the BRFSS questionnaire for even- or odd-numbered years. Questions in the core may be edited from one year to the next but topics remain standard. Editing occurs as a result of an outdated question (for example questions on preventive testing which refer to testing which is no longer recommended) or due to a minor format change. Large scale question changes are rare in the core and would be reported to the OMB as part of the annual review of the questionnaire. The rotating (even/odd numbered year) core includes standard topics and questions which may also include editorial changes from one administration to the next. We estimate that 240,000 respondents screened on landline phones and 240,000 respondents screened on cell phones will participate in the core data collection (total of 480,000 respondents). The estimated burden per response for the core is 15 minutes.

Each state participating in the BRFSS may customize its survey by appending additional questions to the core survey. These questions will be drawn from the set of approved questions for optional modules (see Attachment 4). A complete 2018 BRFSS questionnaire (including screening scripts, even numbered year core and annual list of available optional modules) is provided in Attachment 5. Each state will administer one or more state-specific versions of the instrument in English. Spanish language versions are also routinely administered. The CDC does not provide requirements for administration of the survey in a second language. A standard Spanish translation is provided to the states, which may or may not adopt it for use. States may opt to modify the translation in order to match Spanish dialects of their populations. For example, states with larger number of Cuban residents may have populations with different word usage than states larger proportions of Spanish speaking residents with other backgrounds. States may translate the questionnaire into other languages as budgets permit, although this is rare and is not anticipated in 2018. The pre-call status of the landline sample alerts the states to the potential of a language barrier and in some cases bilingual interviewers are assigned to that segment of the sample. In other instances, states use a callback system of bilingual interviewers to return calls after interim dispositions indicate that the household or cell phone may have a language barrier. Differential weighting of surveys conducted in Spanish or other languages is not included in the post data collection methods, although race and Hispanic ethnicity are included in weighting margins.

A summary of optional modules fielded in 2016 was used as an initial estimate of burden for 2018 (see Attachment 9 for a list of optional modules by state). States may split their samples in order to include a wider range of topics in the questionnaire, without lengthening the time for each interview. The number of optional questions/modules varies from state to state and year to year, but states typically limit the total length of the BRFSS interview to no more than 30 minutes on average. We estimate an average burden across the states of 15 minutes for their choice of optional modules. The BRFSS experiences a very low drop off rate for respondents who complete the core but do not complete the optional modules. However, some optional modules are only asked of subsets of the respondents. For example, only those who indicate that they have been diagnosed with diabetes will be asked questions from the diabetes module. We estimate that 440,000 respondents will complete the optional modules, after completing the core questionnaire.

A field test is conducted each year prior to the implementation of new or changed portions of the questionnaire as a final check on their usability within the survey. See Attachment 13 for an example of a field test questionnaire. Portions of the questionnaire which immediately precede and follow changed or new sections and those extant questions which are topically related are also included. The burden hours for the field test are included in the table below. The total estimated annual burden includes up to 500 respondents who will qualify through screening and then participate in the field test. The estimated burden per response for these respondents is 45 minutes, which includes burden for screening, all demographic questions in the core survey, and the questions being field tested. Burden varies for each year’s field test survey, depending on the number of proposed changes to the core survey, changes to the optional modules, and testing of questions that are related to new or modified questions. The estimated burden per response of 45 minutes is a generous estimate, which should accommodate all cases. In some years, the actual burden of the field test may be lower.

The burden estimate also includes 400 respondents who will be determined ineligible after screening, or decline to participate in the field test. The estimated burden per response for these respondents is 1 minute.

Table A.12-1. Estimated Annualized Burden to Respondents

Type of Respondents	Form Name	No. of Respondents	No. of Responses per Respondent	Avg. Burden per Response (in hr)	Total Burden (in hr)
U.S. General Population	Landline Screener	375,000	1	1/60	6,250
	Cell Phone Screener	292,682	1	1/60	4,879
	Field Test Screener	900	1	1/60	15
Annual Survey Respondents (Adults >18 Years)	BRFSS Core Survey	480,000	1	15/60	120,000
Annual Survey Respondents (Adults >18 Years)	BRFSS Optional Modules	440,000	1	15/60	110,000
Field Test Respondents (Adults >18 Years)	Field Test Survey	500	1	45/60	375
Total					241,519

Annualized burden costs are summarized in Table A.12-2 below. These calculations assume the average hourly wage of $26.00 for all jurisdictions included in the BRFSS. Hourly rates were taken from the most recent publically available Current Employment Statistics of the Bureau of Labor Statistics and are based upon the average hourly earnings for January 2017 from the Current Employment Statistics survey conducted by the Bureau of Labor Statistics (available at https://www.bls.gov/news.release/empsit.t19.htm.).

Table A.12-2. Estimated Annualized Cost to Respondents

Type of Respondents	Form Name	Number of Respondents	Total Burden Hours	Average Hourly Wage Rate	Total Cost Burden
U.S. General Population	Landline screener	375,000	6,250	$26.00	$162,500
	Cell phone screener	292,682	4,879	$26.00	$126,854
	Field test screener	900	15	$26.00	$390
Annual Survey Respondents (Adults > 18 Years)	BRFSS core survey	480,000	120,000	$26.00	$3,120,000
Annual Survey Respondents (Adults > 18 Years)	BRFSS optional modules	440,000	110,000	$26.00	$2,860,000
Field Test Respondents (Adults ≥18 Years)	Field test	500	375	$26.00	$9,750
Total					$6,279,494

Estimates of Other Total Annual Cost Burden to Respondents or Record Keepers

There are no maintenance or capital costs to respondents.

Annualized Cost to Federal Government

Costs that are presented below include data collection, weighting and sampling as well as data distribution (i.e. websites and production of data sets). These are based on the funds provided to states for data collection as well as internal BRFSS costs.

Annualized Estimated Cost to the Federal Government

Estimated funds provided to states	$14,000,000
Estimated CDC BRFSS budget	$3,500,000
Total	$17,500,000

Explanation for Program Changes or Adjustments

The BRFSS Questionnaire is updated annually. This request includes the most current even-numbered year core questionnaire (to be fielded in 2018 and 2020; see Attachment 3a) and the most current odd-numbered year questionnaire (to be fielded in 2019 and 2021; see Attachment 3b). Changes to the 2019, 2020 and 2021 questionnaires will be submitted in the Annual Change Requests for their corresponding years.

The total number of BRFSS interviews has declined in recent years. The estimated 480,000 completed interviews may be higher than the number actually conducted annually from 2018-2020. The proportion of the sample dedicated to cell phones has also increased to reflect phone usage of potential respondents.

Plans for Tabulation and Publication and Project Time Schedule

Data collection for the BRFSS is scheduled to begin on January 1 annually. Data will be submitted monthly to CDC for editing and cleaning. Quarterly data quality reports are provided to states by the CDC. Final data sets for each year will be received from the states by February. Editing, cleaning and weighting of the data will take place until July. Final weighted data sets (see description below) will be returned to the states by September. Datasets and supporting technical documentation will be available for public use by October of the following year.

A. 16 – 1 Project Time Schedule
Activity	Approximate Time Schedule
Annual CDC/BRFSS partner meeting	Spring
Field testing of new or modified portions of the BRFSS core or optional modules	May - August of preceding year
Data collection	January 1 – December 31 of current calendar year
Monthly data submission	February- January of current calendar year
Quarterly data quality reports	March, June, September, December
Data cleaning and editing	March - July of current calendar year
Weighting	January - July of following calendar year
Final data sets to states	By September of following calendar year
Final public use datasets with supporting documentation	By October of following calendar year

The CDC assists the states by weighting each state’s dataset annually. Once all data are received from the states, CDC staff members apply individual respondent weights to ensure that the persons interviewed most accurately reflect the population of each state. Weighting is completed in two steps: a design weight to correct for the probability of selection and a raking (iterative proportional fitting) weighting process to match the demographic characteristics of the respondent to those of the population.

Design weights are based on the number of phones and eligible respondents in each household for landline phone numbers for each of the geostrata defined in the states’ samples. Cell phone respondents are treated as single adult households in the design weights. The formulae for the design weights are:

Stratum weight (_STRWT) = (number of records in the strata)/(number of records selected)

Design Weight = _STRWT* (1/number of phones within the household) * (number of eligible adults within the household)

Raking weights are based on population totals obtained from a private vendor (Neilson), the National Health Interview Survey (NHIS), and the American Community Survey (ACS). The vendor provides updated, county-level estimates and race/ethnicity totals which are not available from the Census in a timely manner. The following variables are used in the raking process: age, race, Hispanic ethnicity, home ownership, sex, phone ownership, (sub) state region, marital status and education. While the same variables are used for weighting for each state, in some instances the categories are collapsed differently from one state to the next. For example, if the state has a very small number of Asian residents, then there might not be sufficient numbers of respondents to include Asian as a separate race category. In these cases, smaller groups may be collapsed into a single category. Once raking weights are assigned a final weight is provided for each respondent using the formula:

Total Weight = Design weight * Raking weight

Because states may ask optional modules to only portions of their samples (split samples) several sets of weights are calculated for each state. These weights are also provided to the public on the BRFSS website (at www.cdc.gov/brfss) with technical documentation on the appropriate use of each weight assignment.

Reason(s) Display of OMB Expiration Date is Inappropriate

N/A. The expiration date will be displayed.

Exceptions to Certification for Paperwork Reduction Act Submissions

There are no exceptions to the certification.

File Type	application/vnd.openxmlformats-officedocument.wordprocessingml.document
File Title	How to Write and Submit
Author	CMS
File Modified	0000-00-00
File Created	2021-01-21

BRFSS 2018 Justification Part A 01MAR2018X

BRFSS 2018 Justification Part A 01MAR2018.DOCX

Behavioral Risk Factor Surveillance System (BRFSS)

OMB: 0920-1061

Contents

Attachments List

Revision Request

Justification

Circumstances Making the Collection of Information Necessary

Purpose and Use of Information Collection

Use of Improved Information Technology and Burden Reduction

Efforts to Identify Duplication and Use of Similar Information

Impact on Small Businesses or Other Small Entities

Consequences of Collecting the Information Less Frequently

Special Circumstances Relating to the Guidelines of 5 CFR 1320.5

Comments in Response to the Federal Register Notice and Efforts to Consult Outside the Agency

Explanation of any Payment/Gift to Respondents

Protection of the Privacy and Confidentiality of Information Provided by Respondents

Institutional Review Board (IRB) and Justification for Sensitive Questions

Estimates of Annualized Burden Hours and Costs

Estimates of Other Total Annual Cost Burden to Respondents or Record Keepers

Annualized Cost to Federal Government

Explanation for Program Changes or Adjustments

Plans for Tabulation and Publication and Project Time Schedule

Reason(s) Display of OMB Expiration Date is Inappropriate

Exceptions to Certification for Paperwork Reduction Act Submissions