Edward J. Sondik, PhD
Director National Center for Health Statistics (NCHS), CDC

Elizabeth Ward, RN, MN
Assistant Secretary, Epidemiology, Health Statistics, and Public Health Laboratory, Washington

1. Standards
   
• Confidentiality
• HIPAA
• Integrated data systems
• Noncompatibility of the different TB data sets illustrates the need for integration.
• Race/ethnicity
• How and when to collect?
• There is great difficulty in collecting this type of data; sites for collection of data include point of service at hospitals and funeral homes, which present difficult circumstances. Self-reporting may achieve higher accuracy.
2. Analysis Capacity Threat
   
• At the state and local level, we still do not understand all data collected because of the lack of professionals such as epidemiologists who can analyze such data.
3. State and Local Measurement
   
• Need to clearly define data sources for the HP2010 developmental objectives
• The lack of resources to conduct surveys at the local level (e.g. BRFSS) presents the challenge of making HP2010 a local document.
• The State health department is forced to make the difficult decision between maintaining personnel and conducting such surveys.

Ron Bialek
Executive Director, Public Health Foundation

From: Measuring Health Objectives and Indicators: 1997 State and Local Capacity Survey

1. What we have
   
• Consensus set of 18 health status indicators (developed from Objective 22.1 of Healthy People 2000)
• Usefulness of health status indicators for Healthy People 2000?
• Found that data is important in driving objectives; availability of baseline data influenced the selection of objectives
• Vital statistics most frequently used for measuring objectives
2. What we know
   
• Gaps in Healthy People 2010 Objectives
• Baseline data were generally not available at local level
• Timeliness of tracking data improved over baseline data
• Lack of data systems and staff perceived as leading barriers to tracking state objectives
• Staffing needs, data sharing/linking, and new data collection systems dominated ‘states’ wish lists
• Private sector data was a significant source of data for some states
3. What we need
   
• Refine and promote consensus set of health status indicators.
• Enhance existing information systems to provide valid and reliable data to local jurisdictions.
• Identify and communicate current efforts so that models will be identified.
• Training of staff.
• Enhance sharing of data.

In summary

• Consensus set of indicators make a big difference.
• Need to improve behavioral, chronic disease, access, and environmental health data.
• Need to better use what already exists.

Richard Klein, MPH
Chief, Data Monitoring and Analysis Branch, NCHS

Three critical data issues:

1. Implementation of ICD-10 (International Statistical Classification of Diseases and Related Health Problems) for mortality
   
• Change in code categories for diseases and other health problems.
2. Implementation of a new population standard for age-adjusted death rates from 1940 to the projected 2000 standard.
   
• New policy of DHHS changes the population weights from 1940 to projected 2000
• This is in response to 1) the need to update, 2) recognizing that the 1940 population does not reflect that of today, and 2) elimination of the current use of multiple standards.
• Because of the different characteristics between the 1940 population and projected 2000 population, there will be changes in statistical rates simply because of the change to the projected 2000 standard (e.g. starting in 1999, there will be an Aapparent@ doubling of deaths related to heart disease). This highlights the critical need to clearly communicate the effects of switching to a new standard.
• This also has implications for race/ethnicity differentials - need to be prepared to communicate the effects of this change (e.g. numerical disparities between races may change because of the change in population standard).
• Would have to readjust Healthy People goals.
• This is concurrent with the implementation of ICD-10.
3. Implementation of new guideline for collecting and tabulating race statistics
   
• Allows for the selection of more than one race and Hispanic origin for the 2000 Census.
• New guideline may affect comparability among races.
• Important for revised birth and death certificates: How and when to collect information on vital records according to OMB Directive 15? Quality and completeness of information?

Olivia Carter-Pokras, PhD
Director, Division of Policy and Data, Office of Minority Health

• Racial data presented for many but not all people-specific objectives.
• Racial data only shown for those with greatest disparities.
• If data were missing, not clear whether data 1) were not available or 2) did not show a disparity.
• Latest data show many widening gaps.
• For Healthy People 2010, recommended that baseline and monitored data be presented for each people-specific objective for at least 5 major race groups; when possible, SES data would be presented but not broken out by race.
• 2 approaches: 1) For 3 types of objectives, targets will be set that reported improvements for all segments of the population (better than the best); 2) For one type of objective, targets will be selected that represent improvements for the majority of the population.

Discussion:

• Rural/urban also shows marked disparity.
• Keep in mind how much resources are needed to do a survey.
• Although the local level knows its own demographics, data by race and ethnicity is still necessary for the translation into action--action-driven epidemiology.
• GIS down to the school district level - users of the data need to make it understandable to the public.
• Does every local health department have the capacity to assess and analyze data collected?
• One major data challenge is to translate data to meet the need for communities to understand themselves, giving them the data to drive action necessary to have healthy communities.

• Communities need sufficient data for action at the community level, however, Asufficient@ is not adequately defined. Technical expertise is needed to work directly with communities.
• Need to reexamine the respective roles of state and national data collection systems.
• Suggested to hold regional workshops to focus on data and analysis issues in communities.
• Data and programs are two circles that do and should overlap, but are often thought of as non-overlapping.

Seattle Transcripts and Summaries