Change Request_6_5_2013

OMB Change Request: The National Violent Death Reporting System (NVDRS)

OMB No. 09200-0607 exp. Date 12/31/2015

Justification

The proposed changes will enhance the security, reliability, scalability and quality of NVDRS by: 1) replacing the current distributed software system operated by funded states with a Center for Disease Control and Prevention (CDC) maintained server accessed via the web, 2) streamlining data abstraction and 3) adding a small number of data elements identified by state users or CDC scientists. The changes focus on how information is entered, but does not modify the data sources or variables substantially.

• Security: The new web system will comply with national security standards (i.e., NIST 800-53A standard). In contrast, current state-owned servers have not been evaluated for compliance. Compliance ensures effective identity verification, access control, vulnerability scanning and backup and disaster recovery. Variables with person identifying information (e.g., victim’s last name) that previously were just collected at the state level, but not received nationally, will not be collected in the national web system to protect confidentiality.

• Reliability: The CDC-owned web platform enhances database stability and reliability by replacing varying state maintenance and use policies with a centralized backup system, centralized software maintenance and simultaneous implementation of updates across states.

• Scalability: The current system requires extensive technical support, including a site visit, to install, upgrade and maintain software. The proposed web system can cheaply, quickly and securely add new users if they have web access. Also, the new system enhances the capacity of funded states to enter information from multiple locations in a secure fashion.

• Efficiency: The current system enters the same variable three times, once from each source (e.g., death certificate, coroner/medical examiner reports and law enforcement reports). Analyses found high agreement among sources (e.g., same race/ethnicity reported by three sources). Consequently, each variable will only be entered once and states will reconcile inconsistencies across data sources. Also, variables that were not available in documents (i.e., rarely or never coded) were dropped.

• Quality: Data quality will be improved in the following ways: 1) the web interface is easier to learn for first-time users, 2) the strategy for collecting information on crises precipitating violent deaths (e.g., received an eviction notice on the day of the suicide) now includes check boxes versus the current approach of writing down each crisis (e.g., change is equivalent to a wording change in a survey), 3) toxicology information is enhanced by using a more comprehensive standardized substance list and capturing more categories of drugs, and 4) the web import function enhances the ability of states to import electronic data easily and correctly.

Project Description

Violence is a major public health problem with 16,671 homicides and 38,364 suicides occurring in 2010. A key to preventing violent deaths is to understand and target their circumstances (the “who”, “when”, “where”, and “how”). Receiving initial OMB approval in November 2004, NVDRS addresses this need by funding state health departments to combine information from three primary data sources, death certificates, coroner or medical examiner reports (i.e., some states have coroner systems and other have medical examiner or combined systems) and law enforcement reports, into a single description of violent deaths. Currently, NVDRS employs a distributed software system that allows standardized coding and data entry in each state health department, but requires each state to own and operate a server.

Proposed Changes

• Data will be collected through a single CDC-owned served accessed via the web.

• Data abstraction will be streamlined to improve efficiency and accuracy.

  • State abstractors will enter most variables once instead of entering variables three times or once for each data source (i.e., law enforcement, coroner/medical examiner, and death certificate).

  • Variables which were not available in the documents (e.g., abstractors never coded these variables) and personal identifying information will be dropped (See Attachment 1 for variables impacted).

  • Checkboxes will be used to determine if a violent death involved a crisis (e.g., a suicide occurred on the day the decedent was to be evicted) instead of having the abstractor write down each crisis. Similar to a question wording change, the same information is being collected in a more accurate manner (See Attachment 1 for variable impacted).

  • A few new variables were added or renamed in response to state abstractor requests or CDC scientists (See Attachment 1 for variables impacted).

  • Collection of toxicology information was enhanced by expanding the list of substances and collecting information on more classes of drugs (e.g., benzodiazepines and muscle relaxants).

Change to Burden and/or Cost

This non-substantive change requests does not change the currently approved burden and/or cost (See Table 1). For the two burden estimates, we explain why the changes will not impact burden.

1. Funded State Health Departments – In the current request, 27 health departments are estimated to incur an average of 2.0 hours per death in identifying the deaths from death certificates, contacting the law enforcement, coroners and medical examiners to get copies of or to view the relevant records, abstracting all the records, various data processing tasks, various administrative tasks, data utilization, training, communications, etc. The average number of violent deaths is 1,000. Thus, the cost can be estimated as 2,000 hours x 27 states x $15/hour = $810,000. The vast majority of time involves securing and reviewing the records. This process remains the same. While data entry using the web system will be more efficient than the current system due to fewer variables and an improved import, saving and navigation with a web system is slower than the desktop. This will result in the burden and cost estimates remaining the same (See below).

2. Public Agencies – States using the web system will request the same information from the same public agencies that they are currently requesting resulting in no burden or cost changes. Public Agencies who retrieve and refile records estimate costs at [500 hours x 27 states x $15/hour] = $202,500 (See below).

Estimated Annualized Respondent Burden Hours

Type of Respondent	Form Name	No. of Respondents	No. Responses per Respondent	Average Burden per Response (in hours)	Total Burden Hours
State Health Departments	Completion of case abstraction	27	1,000	2.0	54,000
Public Agencies	Retrieving and refile records	27	1,000	0.5	13,500
Total					67,500

Estimated Annualized Respondent Burden Costs:

Type of Respondent	No. of Respondents	No. Responses per Respondent	Average Burden per Response (in hours)	Total Burden Hours	Hourly Wage Rate	Total Respondent Cost
State Health Departments	27	1000	2.0	54,000	$15	$810,000
Public Agencies	27	1000	30/60	13,500	$15	$202,500
Total						$1,012,500

3