Ssa_0604_savd_2019

School-associated violence, particularly homicides and suicides that occur in schools, has been a significant public concern for several years. Despite the important role of schools as a setting for violence research and prevention interventions, relatively little scientific or systematic work has been conducted to describe the nature and level of fatal violence associated with schools. Public health and education officials have had to rely on limited local studies and estimated numbers to describe the extent of school-associated violent death. As a result, the U.S. Department of Education (DOE) requested assistance from the Division of Violence Prevention (DVP)/National Center for Injury Prevention and Control (NCIPC) in establishing an ongoing surveillance system of school-associated violent deaths (SAVD) in the United States with the goal of tracking and monitoring the extent of this problem on an ongoing basis. The purpose of this revision request is to continue collecting surveillance data. SAVD is planned as an ongoing surveillance project, with recurring survey preparation and design, data collection, and preparation and analysis of survey results. In this revision request, CDC seeks to eliminate the use of telephone interviews with both school officials and law enforcement officers. The intent is to minimize the burden on these respondents by transitioning data collection to abstracting information from law enforcement reports and entering it into the Data Abstraction Tool. The overall burden will be reduced by 59 hours. The instrument used previously for the Law Enforcement Interview has been amended for data abstraction instead of data collection. Study staff will no longer conduct telephone interviews with school and law enforcement officials. Instead, data will be abstracted from the law enforcement report and entered into the Law Enforcement Data Abstraction Tool. CDC study investigators will review public records and published press reports concerning each SAVD; for each identified case, investigators will contact local law enforcement officials who are knowledgeable about the case in question to confirm information about each identified case. Additionally, researchers will request that law enforcement agencies provide their investigative report for each confirmed case.

Intended use of the resulting data:

The SAVD surveillance system remains the only systematic effort to document school-associated violent deaths on a national basis. Data from the SAVD surveillance system is intended to contribute to the understanding of fatal violence associated with schools, guide further research in the area, and help direct ongoing and future prevention programs.

Methods to be used to collect:

CDC study investigators will review public records and published press reports concerning each SAVD; for each identified case, investigators will contact local law enforcement officials who are knowledgeable about the case in question to confirm information about each identified case. Additionally, researchers will request that law enforcement agencies provide their investigative report for each confirmed case.

The subpopulation to be studied:

The study population will include the victims and offenders from all identified events in which there was a school-associated violent death in the U.S.

How the data will be analyzed:

Data will be entered into EpiInfo, Microsoft Access and Excel software packages, and analysis will be performed with either the SPSS or SAS statistical/database packages; more complex analyses may be performed with JoinPoint, SUDAAN or MPLUS statistical software programs or other software as necessary. Most analyses will be restricted to simple descriptive statistics--frequencies and univariate analysis. Case-finding methods will be compared using capture-recapture calculations, which can also be used to estimate the proportion of cases not identified by a particular case-finding technique. To calculate a rate of school-associated violent death in the United States, the DOE and Current Population Survey will provide national school enrollment data; in computing the rate, it may be necessary to restrict the numerator to those cases that concern the death of a student on school property during regularly scheduled hours of operation.

JUSTIFICATION

A.1. Circumstances Making the Collection of Information Necessary

This is a revision request for the currently approved “School Associated Violent Deaths Surveillance System” (0920-0604, expiration 05/31/2019). Additionally, CDC seeks to eliminate the use of telephone interviews with both school officials and law enforcement officers. The intent is to minimize the burden on these respondents by transitioning data collection to abstracting information from law enforcement reports and entering it into the Data Abstraction Tool. The overall burden will be reduced by 59 hours.

Background

As a leading cause of death among young people, violence is increasingly recognized as an important public health and social issue. In 2016, over 3,600 school aged youth (5 to 18 years old) in the United States died violent deaths (due to suicide, homicide, and unintentional firearm injuries)¹. The vast majority of these fatal injuries were not school associated. However, whenever a homicide or suicide occurs in or around school, it becomes a matter of particularly intense public interest and concern.

A number of studies of violent behavior and risk factors for violent injury have been conducted in school-based populations. Furthermore, schools have been the sites for many interventions to prevent suicide and inter-personal violence among young people^{2, 3}.

Despite the important role of schools as a setting for violence research and prevention interventions, relatively little scientific or systematic work has been done to describe the nature and level of fatal violence associated with schools. Public health and education officials have had to rely on limited local studies and estimated numbers to describe the extent of school-associated violent death^{4, 5}. As a result, the U.S. Department of Education (DOE) requested assistance from the Division of Violence Prevention (DVP)/ National Center for Injury Prevention and Control (NCIPC) in establishing an ongoing surveillance system of school-associated violent deaths in the United States. This surveillance system remains the only ongoing systematic effort to document school-associated violent deaths on a national basis.

The surveillance system will continue to contribute to the understanding of fatal violence associated with schools, guide further research in the area, and help direct ongoing and future prevention programs.

Currently, only limited injury risk factor data are collected by a few existing national surveillance systems, e.g., the CDC’s National Violent Death Reporting System (OMB no. 0920-0607), DOJ’s National Crime Victimization Survey (OMB no. 1121-0111), and the National Incident Based Reporting System (FBI’s NIBRS; OMB no. 1110-0002). These systems are primarily focused on and intended for purposes other than injury prevention and school-associated violence. Further, they face other constraints – for instance NIBRS is not yet fully national. While their vision is to collect data from across the U.S. in the future, in 2017 (the most recent year for which data are available), only 43% of U.S. law enforcement agencies that participated in the Uniform Crime Reporting program submitted data via NIBRS. Further, while the National Violent Death Reporting System expanded to all 50 states in 2019, some states are not yet collecting data statewide. Thus, for several reasons, these systems are limited in their ability to provide national data on SAVDs. Further, because these systems must cover large numbers of mandated topics, time constraints preclude adequate coverage of the gamut of injury risk factors. Moreover, these systems have varying methods, definitions, and timeliness of data availability, and gaps exist in addressing data needs for tracking the Healthy People 2020 injury objectives. Thus, the School-Associated Violent Death Reporting System serves as an alternative that allows for the monitoring of the occurrence of fatal school violence and risk factors associated with it at the victim, perpetrator, and school levels. The availability of these data can allow the federal government, communities, and schools to understand the context of fatal school violence and can serve as data to help evaluate programs and inform policy.

The public health importance of the school-associated violent death problem is such that there should be a dedicated means of rapidly collecting national data about the prevalence of risk factors for violent death and defining which population groups are most affected.

Data from this ongoing surveillance effort have been used in a variety of settings. For instance, the US Department of Education has used the data extensively in developing their programs. Data are published yearly in the Indicators of School Crime and Safety Report. Researchers from Harvard University have used these data in preparing a report entitled, Rampage: The Social Roots of School Shooting. CDC staff have also written and published reports that were presented in the Journal of the American Medical Association (JAMA) and CDC's Morbidity and Mortality Weekly Report (MMWR).

The following authorizing legislation permits this data collection:

1) Section 301 of the Public Health Service Act (42 USC 241) (Attachment A) authorizes CDC to conduct research relating to the prevention and control of disease.

2) Section 391 of the Public Health Service Act (42 USC 280b) (Attachment A) authorizes CDC to conduct research relating to the causes and prevention of injuries and assist the States in activities for the prevention of injuries. This survey is intended to define the prevalence of risk factors for injury in the U.S. as a whole and in specific subgroups. These data will help to identify populations with the greatest need for interventions to reduce risk factors and suggest specific behaviors to be targeted by intervention programs.

3) Section 42 USC 242(k), and 42 USC 242(m) (Attachment A) The Confidentiality Assurance under this law protects the privacy of people and organizations taking part in this study. It keeps their names and other facts that can identify them from anyone who is not on the study staff.

This request is a revision to eliminate interviews with school staff and law enforcement officials who investigated identified SAVD cases. The specific changes are more thoroughly described in section A.15. There are no changes to the statistical design/analysis.

A.2. Purpose and Use of Information Collection

The School-Associated Violent Deaths Surveillance System (SAVD) has been conducted since 2013. Data collection for the 2019-2020 cycle will consist of abstractions from law enforcement reports, and the data will be entered into the Law Enforcement Data Abstraction Tool (Att D). We are requesting to eliminate interviews with school staff and law enforcement officials who investigated identified SAVD cases. The decision to end data collection via school interviews is being made primarily because it is exceedingly difficult to obtain interviews from school officials, particularly as those associated with the schools are often hesitant to comment on the deaths of victims affiliated with the school. This could be the result of a number of reasons – e.g., concern over liability issues, emotional distress that may result from the interview. Additionally, CDC staff recognize that very little, if any, additional information is gained from school interviews that was not able to be obtained from the law enforcement reports. Thus, to reduce the burden of potential respondents, we are requesting to end the collection of data through the school official interviews. Additionally, because interviews with law enforcement officials were occurring after much time had passed between the date of the incident in question and the date of the interview, and the IRB interview protocol requires that law enforcement officials not review their investigative report prior to participating in the interview, law enforcement’s responses to interview questions is likely to be subjected to recall bias. Further, in some instances, the law enforcement officers who participated in the interviews were not the same ones who engaged in the investigation, resulting in an excessive number of “don’t know” responses and ultimately requiring that CDC staff review and manually code information from the original law enforcement report. Thus, CDC proposes to end data collection through law enforcement interviews, thereby reducing burden to law enforcement respondents. Overall, the intent is to minimize the burden on these respondents by transitioning to data collection to abstracting information from law enforcement reports and entering it into the Law Enforcement Data Abstraction Tool (Attachment D). More information on the proposed revisions is included in Section A.15.

Data from the School-Associated Violent Deaths Surveillance System (SAVD) will be collected for the purposes of furthering understanding of fatal violence associated with schools, guiding further research in this area, and helping to direct ongoing and future prevention programs. There is a need to continue the surveillance system to gather data to inform general perceptions that school-associated violence, particularly fatal violence, is increasing, but also to inform evaluation of ongoing school violence programs and to provide guidance in the development of new school violence prevention programs. A possible negative consequence of not conducting SAVD would be perpetuating fear that fatal school violence is increasing and spending money on ineffective prevention programs because of inadequate data for program evaluation. Another important negative consequence would be continued high morbidity and mortality from school violence because of inaction resulting from inadequate knowledge about preventable risk factors. Lastly, this system addresses the Healthy People 2020 focus area of Injury and Violence Prevention along with its goal of reducing injuries, disabilities, and deaths due to violence.

As mentioned in the previous section, data from the surveillance system have been used extensively to inform public officials, researchers, and the public in general. These data have appeared in several published reports that have been used to guide programmatic activities and evaluate interventions. A list of publications using data from the School-Associated Violent Deaths Surveillance System (SAVD) is presented in Attachment F.

Data collected through the surveillance system will be reviewed and used by CDC and the US Department of Education, as well as other outside agencies and organizations that rely on the information to understand trends in fatal school violence. This information will be used in concert with other sources of information, such as National Center for Education Statistics school-level data, which provide context regarding schools where SAVDs take place (e.g., % free/reduced lunch, school locale [urban, rural, suburban]). Further, the data collected under this revised ICR will be analyzed along with data collected through this surveillance system in previous years (approved under OMB# 0920-0604) in order to conduct trend analyses related to SAVDs over an extended period of time. There is no budgetary line item, contractor, or grantee associated with this surveillance system; all work is conducted by CDC staff.

A.3. Use of Improved Information Technology and Burden Reduction

We request approval to eliminate the telephone surveys, thus reducing the respondent burden. CDC Investigators will enter data from law enforcement reports into Excel and EpiInfo databases, which can then be used for further analyses.

A.4. Efforts to Identify Duplication and Use of Similar Information

There are no systems of comparable scope currently in existence. Our previous discussions with some agencies (e.g., the U.S. Department of Justice and the U.S. Secret Service) and ongoing interactions and discussions with violence prevention researchers and practitioners throughout the country (e.g., representatives of the CDC-funded Injury Control and Research Centers, the CDC-funded Youth Violence Prevention Centers, the State and Territorial Injury Prevention Directors Association, the U.S. Department of Education), have identified no plans for a national surveillance system of comparable scope. The ongoing discussions occur semi-annually at reverse site visits for CDC grantees and cooperative agreement principal investigators and on regular teleconference calls and at in-person meetings with other governmental agencies. Further, given the lack of data of this nature available from other resources, the CDC worked together with the U.S. Department of Education and the U.S. Department of Justice years ago to develop the questionnaire and methods for the School-Associated Violent Death Surveillance System. Through this collaboration, the CDC provides the National Center of Education Statistics data annually on the frequency of school-associated violent deaths.

We have identified several efforts designed to systematically collect information on school-associated violent deaths. However, these projects are time-limited, as opposed to ongoing, and focus only on a small subset of cases, e.g., United States Secret Service Safe Schools Initiative (USSS-SSI).

No federally funded, truly national system like this one currently exists. USSS-SSI is limited to a select number of “targeted violence” events where school-based attacks were premeditated and often resulted in multiple deaths and injuries. Thus, since the USSS-SSI does not collect data on every school-associated violent death as the SAVD does, using the data collected by the U.S. Secret Service, it is not possible to produce national trends and risk estimates.

As indicated above, the National Violent Death Reporting System (NVDRS) recently expanded its funding to include 50 states and 2 territories. NVDRS collects and maintains data on homicides, suicides, legal intervention, accidental firearm, and undetermined deaths. Given the system has experienced rapid growth and not all states have the same capacity to report on every violent death within the state, the system is not yet capturing every violent death in the U.S. For this reason, SAVD-SS is still needed to ensure a truly national surveillance system of SAVDs. However, because of the labor-intensive and time-consuming process for collecting SAVD data and because there are limitations in the methods implemented to collect the data (e.g., the media scan captures only those cases reported in the media), NCIPC’s goal is to transition SAVD data collection efforts to NVDRS over the course of the next 18 months. Unfortunately, this is not a rapid process, as an SAVD module must be developed and programmed into the current NVDRS platform, and state-based NVDRS data abstractors must be trained on completing the module to ensure that the data are collected in a consistent manner across the U.S. SAVD data collection via NVDRS is expected to begin taking place in January 2021.

A.5. Impact on Small Businesses or Other Small Entities

No small businesses or small non-profit organizations will be involved in this study. The only small government jurisdiction that may be affected by this system is a local law enforcement agencies, whose employees may be asked to participate in the study by providing a law enforcement investigative report if a case occurred at a school within their jurisdiction. As described in more detail below, this impact should be minimal, involving on average 20 minutes of a law enforcement officer’s time.

A.6. Consequences of Collecting the Information Less Frequently

This is an ongoing data collection effort. If this information is not collected in a timely manner, it will not be possible to accurately assess trends in school-associated violent deaths. Without these data it will be difficult to determine the impact of federally funded programs to reduce school related violence. Since there is no other source for data on school-associated violent deaths, researchers, policy makers, and the general public will be dependent upon the media to supply this information. Due to the rarity of these events, it is unlikely that data sources would be contacted more than once. There are no legal obstacles to reduce the burden.

A.7. Special Circumstances Relating to the Guidelines of 5 CFR 1320.5

The request fully complies with the regulation 5 CFR 1320.5.

A.8. Comments in Response to the Federal Register Notice and Efforts to Consult Outside the Agency

A.8.a) Federal Register Notice

A 60-day Federal Register Notice was published in the Federal Register on February 7, 2019, vol. 84 No. 26, pp. 2514 (Attachment B). CDC received four public comments (Attachment B1).

A.8.b) Efforts to Consult Outside the Agency

The following persons reviewed the survey instrument and study design, including components related to the availability of data, the frequency of data collection, the clarity of instructions and record keeping, and the specific data elements to be collected:

Lisa Barrios, DrPH, Division of Adolescent and School Health, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention (NCHHSTP), CDC. 770-488-6172, [email protected]

Nancy Brener, PhD, Division of Adolescent and School Health, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention (NCHHSTP), CDC. 770-488-6184, [email protected]

William Modzeleski, MA, formerly of Safe and Drug Free Schools Program, U.S. Department of Education. 202-245-7831, [email protected]

Paul Kesner, Director, Drug-Violence Prevention – State Programs, Office of Safe and Drug-Free Schools, [email protected]

Phelan Wyrick, PhD, Division Director, Crime and Crime Prevention Research Division, Director, Comprehensive School Safety Initiative, National Institute of Justice, U.S. Department of Justice, 202-353-9254, [email protected]

Deborah Stone, ScD, Division of Violence Prevention, Research and Evaluation Branch, National Center for Injury Prevention and Control, 770-488-3942, [email protected]

Sarah Bacon, PhD, Division of Unintentional Injury Prevention, National Center for Injury Prevention and Control, 770-488-0520, [email protected]

Lloyd Potter, PhD, Department of Demography and Organization Studies
The University of Texas at San Antonio, 210-458-5730, [email protected]

Kenneth Powell, MD, MPH, Georgia State Department of Health. 404-657-2578, [email protected]

The consultation did not reveal any major problems that could not be resolved. Consultation with representatives of those from whom information is obtained is not possible given the specific nature of the events about which data are collected and the rarity of these occurrences. Contacts with respondents with knowledge regarding these rare events are limited to those to facilitate data collection in order to reduce burden. In addition, once a case has been confirmed, a case identification number is assigned, and all links to any information that can identify the school, the individuals involved, or the locations involved are destroyed or stored separately in a password-protected file within a directory on the NCIPC DVP LAN. This password protected file can only be accessed under extraordinary circumstances.

A.9. Explanation of Any Payment or Gift to Respondents

Respondents will not be compensated for their participation.

A.10. Protection of the Privacy and Confidentiality of Information Provided by Respondents

The Office of the Chief Information Officer at the CDC has determined that the Privacy Act does not apply. The Privacy Impact Assessment (PIA) is attached (Attachment I). Law enforcement officials will be contacted and asked to confirm whether a case they investigated meets SAVD case inclusion criteria (i.e., the case was a homicide, suicide, or legal intervention in which the fatal injury occurred 1) on the campus of a functioning public or private elementary or secondary school in the United States, 2) while the victim was on the way to or from regular sessions at such a school, or 3) while the victim was attending or traveling to or from an official school-sponsored event). Cases include deaths of students as well as non-students (e.g., faculty, school staff, family members, or community residents). Next, law enforcement officials will be asked to provide a copy of their investigative report from which CDC study investigators will abstract relevant data. Information in Identifiable Form (IIF) is not collected from law enforcement officials who respond to these inquiries. However, IIF related to SAVD decedents is collected. This information includes:

Name (for victims)
Date of Birth (for victims and offenders)
Other:

Surveillance data regarding school-associated violent deaths are collected to enable CDC and its partners to establish the magnitude of these problems and their public health burden (overall and across subgroups), discern their epidemiologic characteristics, and examine longitudinal trends in their occurrence.

NCIPC has, on an annual basis, used and will continue to use collected data, to:

Identify common features of school-associated violent deaths;
Measure the prevalence of risk factors for school-associated violent deaths;
Define which population groups are most affected;
Estimate the rate of school-associated violent death in the United States;
Monitor the impact of interventions and help direct interventions and resources toward the highest risk subgroups of the population.

Information from this ongoing surveillance effort is only shared in aggregate form, and no personally identifiable data or individual-level data are ever shared outside of the National Center for Injury Prevention and Control. Historically, data from the SAVD Surveillance System have been used in a variety of settings. For instance, the US Department of Education has used the data extensively in developing their programs. Data are published yearly in the Indicators of School Crime and Safety Report. Researchers from Harvard University have used these data in preparing a report entitled, Rampage: The Social Roots of School Shooting. CDC staff have also written and published reports that were presented in the Journal of the American Medical Association (JAMA) and CDC's Morbidity and Mortality Weekly Report (MMWR).

The proposed data collection will have little or no effect on the respondent’s privacy. Respondents are law enforcement officials who will provide information about cases based on their roles and their knowledge of case specifics. They will be contacted for two reasons: 1.) to confirm a possible SAVD case, and 2) to provide a copy of the official law enforcement report for confirmed SAVD cases. The study’s request for information is thus limited to that pertaining to the cases of interest and does not request the respondent’s personal information.

NCIPC applied for and received an Assurance of Confidentiality (see Attachment F). This was done to further safeguard the information collected. Under the provisions of the Assurance of Confidentiality, all identifiable information that CDC gathers in this surveillance system will be kept confidential. This is specifically assured under Section 308(d) of the Public Health Service Act (42 U.S.C. 242 m(d)). The Confidentiality Assurance under this law will protect the privacy of people and organizations taking part in this system.

In addition, once a case has been confirmed, a case identification number is assigned, and all links to any information that can identify the school, the individuals involved, or the locations involved are destroyed, creating de-identified databases. These databases are stored separately in a password-protected file within a dedicated, secure directory on the NCIPC DVP LAN. The principal investigator and study coordinator will have “read/write” privileges to all files; other project staff will have limited “read” privileges to de-identified files.

After case identification numbers are assigned, they are placed in an independent linkage database, and used as an administrative means for the study principal investigator and project manager to coordinate case related data management, and the administration of blinded interviews. This number is the only means of linking info collected in the study’s four core processes. The linkage database is password-protected and stored within a second dedicated secure directory on the NCIPC DVP LAN. This second LAN is wholly separate from the first LAN containing the de-identified databases. Only the principal investigator and study coordinator can access this file. All files are routinely audited to assure preservation of measures employed to assure data integrity, availability, and relevancy.

Respondents are informed about the voluntary nature of their participation. This is done using language in paragraph two of the Letter to Law Enforcement Officials (Attachment E).

A.11. Institutional Review Board (IRB) Justification for Sensitive Questions

IRB Approval

CDC has received IRB approval for this project thorough CDC IRB, which meets all of the Federal requirements as specified in 45 CFR 46, registered with the Office for Human Research Protections and with Federal Wide Assurance (FWA00001413 ). This ensures that this project comply with Federal regulations (Attachment C).

Justification for Sensitive Questions

When possible, SAVD study staff will be abstracting “sensitive data” from law enforcement reports and entering into the Law Enforcement Data Abstraction Tool (Attachment D). The Data Abstraction Tool contains some questions that are sensitive (e.g., questions regarding drug use/abuse, alcohol use/abuse, intimate partner/interpersonal violence, history of sexual violence, and demographic data on race/ethnicity). No social security numbers or other individually identifiable data will be collected.

Alcohol use (V15, V15a, V16, V17, O12, O12a, O13, O14) Alcohol use is an important risk factor for violence. Information on alcohol consumption is relevant for interpretation of both risk-taking and risk-avoidance behaviors.

Demographic data Information on race and ethnicity (V08-09, O05-06) is needed because, as noted in Healthy People 2020, there are important disparities in rates and types of violent injuries in different population subgroups. These differences may be due to differences in the prevalence of injury risks and/or injury prevention measures in populations that have different educational levels or income levels, for which racial or ethnic composition may be a marker.

Family history of violence, alcohol/drug abuse, child maltreatment (V39, O36) Chronic fear of violence has psychosocial consequences including increased risk for suicide.

Self-harm and Suicidal Behavior (V42-43, O38-39) A history of self-harm and suicidal behavior are significant risk factors for subsequent completed suicide, and the number of previous suicide attempts is related to subsequent suicide outcomes and other health problems.

Criminal Activity (O10, O10a, O10b) History of criminal activity is an important risk factor for subsequent violent behavior.

Psychiatric History (V36-38, O33-35) History of depression is a leading risk factor for suicidal activity. Other emotional and mental health problems may also play roles in shaping involvement in violence.

Sexual Orientation (V44, O41) Important to examine if sexual orientation is a risk factor for victimization and/or suicidal activity.

A.12. Estimates of Annualized Burden Hours and Costs

The estimated number of respondents is 50 per year. This is based on an estimated 50 events per year and 1 telephone confirmation conversation per event (Attachment H). The estimated total annual hour burden on respondents is 17 hours (Table 1). The estimates are based on the average time to complete a brief telephone conversation to confirm a case (Attachment H), plus the amount of time to read the official letter requesting the law enforcement report (Attachment E) and time to locate said report and send to the SAVD investigator. The hour burden will differ for individual respondents because some reports may be larger than others or may take longer to locate.

Table 1. Estimated Annualized Burden Hours

Type of Respondent	Form Name	No. of Respondents	No. of Responses per Respondent	Response Burden (Hours)	Total Burden Hours
Law Enforcement Officer	Law Enforcement Case Confirmation Script (Att H)	50	1	5/60	4
Law Enforcement Officer	Letter to Local Law Enforcement Officials (Att E)	50	1	15/60	13
Total					17

The only cost to respondents will be time spent on the telephone confirming case information, reading the official letter requesting law enforcement reports, and time spent locating and sending said report.

Table 2. Estimated Annualized Burden Costs

Type of Respondent	Form Name	Total Burden Hours	Hourly Wage Rate	Total Respondent Costs
Law Enforcement Officer	Law Enforcement Case Confirmation Script (Att H)	4	$29.29	$117.00
Law Enforcement Officer	Letter to Local Law Enforcement Officials (Att E)	13	$34.06	$443.00
Total				$560.00

A.13. Estimates of Other Total Annual Cost Burden to Respondents or Record Keepers

Respondents will incur no capital and maintenance costs.

A.14. Annualized Cost to the Government

SAVD is planned as an ongoing surveillance project, with recurring survey preparation and design, data collection, and preparation and analysis of survey results. The government costs are the personnel costs of federal staff involved in oversight, design, and analysis. No outside contractors will be used. There will be no printing or publication costs for the government.

Table 3. Estimated Annualized Government Cost

Position	Tasks	Avg time / yr	Avg. cost/yr
Lead behavioral scientist	oversight and supervision	10%	$ 3,570
Principal investigator	oversight; coordination of data collection; management of study information; quality assurance implementation	100%	$ 87,219
Project analyst	case identification; data collection; data analysis and interpretation	100%	$ 68,597
Annualized federal costs:			$159,566

Funds for this project are transferred to the CDC budget from the Department of Education via an Interagency Agreement. This amount is approximately $85,000.

A.15. Explanation for Program Changes or Adjustments

Changes to data collection instruments:

The instrument used previously for the Law Enforcement Interview has been amended for data abstraction instead of data collection. Study staff will no longer conduct telephone interviews with school and law enforcement officials. Instead, data will be abstracted from the law enforcement report and entered into the Law Enforcement Data Abstraction Tool (Att D). The elimination of telephone interviews will significantly reduce the respondent burden.

In the past, to obtain as much detailed information as possible concerning each SAVD case, CDC staff have sought to obtain information from multiple sources, including 1) the initial law enforcement investigative report, and interviews with 2) a law enforcement official, and 3) a school official. This multiple source approach was initially designed to allow CDC researchers to triangulate data obtained from the different sources. However, over the course of the project, obtaining the interviews in particular has proven difficult. For instance, of a total of 890 SAVD cases that occurred between the 1994/95 and 2017/18 school years, law enforcement interviews have been completed for a total of 480 (53.9%) cases. The ability to obtain data through school official interviews has proven even more difficult, with these interviews having been completed for only 255 (28.7%) cases. Further, because the interviews are typically conducted long after the incident occurred, the interviews are likely impacted by recall bias, which could result in incorrect responses, responses influenced by media coverage of the incident as opposed to personal knowledge of the incident, and missing data. For these reasons, the law enforcement investigative reports have been relied upon as the premiere data source to be cited in cases where discrepancies exist between the three data sources. Analyses indicate that law enforcement investigative reports are relied upon to fill in missing data or to correct discrepant information from interviews for almost all of the cases for which both at least one interview (i.e., either the school official or law enforcement official) and a law enforcement report exists.

Fortunately, efforts to obtain data through law enforcement reports have been more successful. A total of 622 (69.9%) law enforcement reports have been received for the 890 SAVD cases contained in the system. Some law enforcement reports are unable to be obtained for a number of reasons, such as cases not yet being adjudicated and involving juvenile victims and/or perpetrators, excessive payment being required for printing reports, and general nonresponse from law enforcement agencies contacted. Cases are closed out after five years and no further efforts to obtain the reports are made.

Given that law enforcement investigations are conducted with the intent to determine the motive and circumstances surrounding the incidents in question, they make ideal data sources for the SAVD Surveillance System. Not only do they include demographic information on all of the victims for each case, but they often include demographic data for alleged perpetrators (if known). Additionally, they often include information from a number of incident bystanders and witnesses, as well as victims’ acquaintances, friends, teachers, and family members, who are asked to describe the circumstances that may have contributed to the incident (e.g., victims’ mental health circumstances, peer group affiliations, relationship or school-related issues). Indeed, these are the data used to assess risk factors associated with SAVDs.

A.16. Plans for Tabulation and Publication and Project Time Schedule

It is anticipated that there will be multiple publications from the survey. All data will be received, reviewed, analyzed, published, and disseminated by CDC.

The analysis plan follows the objectives of the SAVD System, which are to:

Identify common features of school-associated violent deaths;
Measure the prevalence of risk factors for injury;
Define which population groups are most affected;
Estimate the rate of school-associated violent death in the United States;
Monitor the impact of interventions and help direct interventions and resources toward the highest risk subgroups of the population.

The analysis plan has four parts:

Describe the study population;
Estimate the prevalence of injury risk factors by demographic characteristic;
Estimate crude odds ratios for injury outcomes by risk factor (where outcome questions are available); and
Build logistic regression models to better describe the association between risk and demographic characteristics, and outcomes.

All analyses will be conducted using complex survey software that takes into account the complex nature of the survey design when computing variance estimates. In bivariate analyses (parts 2 and 4, above), the relative standard error (RSE) of the point estimate will be assessed. Estimates with RSEs ranging from 23-30% will be flagged as possibly unreliable while those with RSEs > 30% will be suppressed, or if presented, flagged as unstable. Where reasonable, categories will be collapsed to improve the stability of estimates. Estimates that are unstable in bivariate analyses will not be further analyzed in multivariate analyses.

Describing the study population

This step in the analysis includes a comparison of the distribution of the study population to the distribution of the US population of elementary and secondary school students as a means of evaluating the characteristics of the study population.

Prevalence analysis of injury risk factors:

This descriptive analysis will produce prevalence estimates and NCIPC will use these data to identify potential interventions and target populations.

Multivariable analysis: The purpose of the multivariable analysis is to clarify the relationships among preventable injury risk factors and outcomes after adjusting for potential confounders that may modify associations between these risk factors and outcomes.

Multivariable analyses will be presented in terms of adjusted odds ratios. Adjusted odds ratios and 95% confidence intervals will be calculated by using logistic regression to adjust for potential confounders identified in bivariable analyses. Possible effect modification of risk by selected demographic variables and other potential confounders will be identified based on evidence in the literature, and assessed using a likelihood ratio test.

Table 4: Project Time Schedule

Task	Schedule for Completion
Identification of School-Associated Violent Deaths through systematic media scans	Ongoing, to continue upon OMB approval
Data collection regarding SAVDs through law enforcement and school official interviews	Ongoing, to continue upon OMB approval
Data entry in SAVD interview databases	Ongoing, to begin 2 weeks after OMB approval and to continue throughout approval period
Data analysis for student homicide MMWR publication	4 weeks after OMB approval
Provide SAVD frequency data to the Bureau of Justice Statistics	4 weeks after OMB approval, and annually thereafter
Publish results in MMWR regarding student homicides	12 weeks after OMB approval
Conduct other analyses for publications	Beginning 12 weeks after OMB approval and ongoing thereafter
Publish results from additional analyses	Beginning 16 weeks after OMB approval and ongoing thereafter

A.17. Reason(s) Display of OMB Expiration Date is Inappropriate

The display of the OMB expiration date is not inappropriate.

A.18. Exceptions to Certification for Paperwork Reduction Act Submissions

There are no exceptions to the certification.

REFERENCES

1. Office of Statistics and Programming. Data Source: NCHS Vital Statistics System for numbers of deaths. Bureau of Census for population estimates: National Center for Injury Prevention and Control, CDC; 2013.

2. Substance Abuse and Mental Health Services Administration. Preventing Suicide: A Toolkit for High Schools. HHS Publication No. SMA-12-4669. Rockville, MD: Center for Mental Health Services, Substance Abuse and Mental Health Services Administration, 2012.

3. David-Ferdon C, Simon TR. Preventing Youth Violence: Opportunities for Action. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, 2014.

4. Violence-related attitudes and behaviors of high school students--New York City, 1992. MMWR Morb Mortal Wkly Rep. Oct 15 1993;42(40):773-777.

5. Geiger K. A safe haven for children: curbing violence in schools. The Washington Post, February 21, 1993.

File Type	application/vnd.openxmlformats-officedocument.wordprocessingml.document
File Modified	0000-00-00
File Created	0000-00-00