2005 National Injury and Prevention and Control Conference

2005 National Injury and Prevention and Control Conference, Denver, CO Injury and Poisoning Estimates from the National Health Interview Survey -Why They Differ from Other Estimates Manon Boudreault MPH, Margaret Warner Ph. D, and Lois Fingerhut, MA Office of Analysis and Epidemiology National Center for Health Statistics Centers for Disease Control and Prevention U. S. DEPARTMENT OF HEALTH AND HUMAN SERVICES

Overview • National Health Interview Survey -Questionnaire History -Current data • Other Injury Data from NCHS -Comparison of data sources • Linking NHIS files and public-use data files

NHIS • All medically-attended injuries occurring to any family member during the 3 month period prior to the interview are asked about from an adult family member. • Uses a multistage sample designed to represent the U. S. civilian population non-institutionalized • Collected throughout the year, nearly 40, 000 households are visited (info persons) on over 90, 000

NHIS Strengths • More detailed information: circumstances, cause, and location of episode (64% of the adult episodes are self-reported) • Nearly complete external cause codes • Largest range of data available in one system • Linkable with other national data sets

Data prior to the redesigned 1997 NHIS Prior years are not directly comparable due to: • Placement, phrasing and number of questions • Recall period • Severity threshold 25% injuries due to these and other changes (1995 to 1997)

Analysis of Recall Bias • Questionnaire recall period …past 2 weeks? (before 1997) to… past 3 months? (current) “Recall Bias” Fewer minor injuries reported • More severe injuries (fractures and sprains/strains) showed similar numbers whether recall was 2 weeks or 3 months. • Further analyses showed drop in reporting minor injuries after 5 weeks.

Annual estimates (NHIS) of injury and poisoning episodes Episodes in millions Same questions Nearly same questions Detailed questionnaire history: Year http: //www. cdc. gov/nchs/about/otheract/injury/presentations. htm

“Packed” question 2000 -03: DURING THE PAST THREE MONTHS, that is since {91 days before today date}, {were/was} {you/anyone in the family} injured or poisoned seriously enough that {you/they} got medical advice or treatment? Three questions in one: 1) Were you injured? 2) Were you poisoned? 3) Did you seek medical advice?

“Unpacked” question- NHIS 2004 During the past three months, that is since {January 2}, did you have an injury where any part of your body was hurt, for example, with a broken bone, sprain, burn, wound, bruise, or animal or insect bite? During the past three months, that is since {January 2}, were you poisoned by swallowing or breathing in a harmful substance such as bleach, carbon monoxide, or too many pills or drugs? Did you talk to or see a medical professional about this {injury/poisoning}?

Annual estimates (NHIS) of injury and poisoning episodes Episodes in millions Same questions New questions Nearly same questions * * Year *Preliminary data, anticipated substantial increase based on revised questions.

Comparison of Other NCHS Data Sources with the NHIS

Important Factors Related to NHIS Estimates • Recall period effects reporting of minor injuries • Proxy reporting for all children and 36% of adult i/p episodes • Relatively few intentional injuries reported • Total treatment may not be reported • Institutionalized population not captured, e. g. , nursing homes • Acute injury episodes rather than chronic injury

Annual rate of injury-related ED visits for seniors by patient residence: 2001 -2002 Institution Age in years Community Number of visits per 100 persons 65 -79 41 8 80+ 37 14 Institution is defined as a nursing home or other institution, such as a prison, mental hospital, group home for the mentally or physically disabled. Source: NCHS, NHAMCS-ED From the NCHS 2004 Data Users Conference Presentation Overview of the NAMCS and NHAMCS by Linda Mc. Caig and David Woodwell http: //www. cdc. gov/nchs/ppt/ahcd/Sess_10_Intro_NAMCS_WOODWELL_&_Mc. CAIG. ppt

Injury poisoning episodes (all), pain/swelling* due to injury (adults), and some disability due to injury (adults): 1997 -2003 In millions i/p episodes pain disability 1997 1998 1999 2000 2001 2002 2003 Injury/poisoning episodes-all ages-(I/p Adults w/some disability due to injury-(sample adult file) Adults- *pain/ache/stiffness/swelling due to injury-(sample adult file)

Some disability due to injury: 1997 -2003 Number of adults in millions Injury condition duration: *Selected from a printed condition list as “fracture/bone/joint injury” and/ or “other injury” (sample adult section/file)

NCHS Injury Data Sources Medical/Administrative Records Household Interview

Injury-related treatment: NHIS/other NCHS surveys Household Interview NHIS episodes (2002) 23. 7 million episodes NHIS Treatment reported (2002) Medical Records-2002 Estimates NHAMCS NHIS 33. 0 million 13. 1 million 10. 2 million 3. 0 million NAMCS 99. 2 million 8. 2 million Dr. visits NHDS-discharges 1. 8 million 1. 0 million ED visits NHAMCS OPD visits

Treatment reported in the NHIS • All treatment received most likely not reported—only 16 percent of those hospitalized with a fracture also reported a visit to a doctor. • Compared with other NCHS treatment administrative databases, differences in frequency were greatest for physicianbased office visits. • Fracture ED visits and reported fracture NHIS ED visits were the most similar (1. 4 times the NHAMCS-ED).

Medical Records-Based Data Strengths Limitations • More accurate • Collection limited to accounting of total what’s entered on treatment / healthcare form/medical chart utilization • Treatment, procedures, and prescribed drug information available • Automated collection systems may differ (NHDS) • Injury mechanism/cause often unspecified

Percent with unspecified cause of injury: 2002 NHIS episodes NHAMCS 2. 6% ED visits 26% NHAMCS OPD visits 49% NAMCS Dr. visits NHDS-discharges 45% 36%

NHIS Public-Use Files and Linking

NHIS-Public Use Data Files • For estimated annual injury and poisoning episodes: -Person file -Injury and poisoning episode file -Injury verbatim file • Using SAS, STATA, SPSS

NHIS Links Expand Scope of Available Injury Data • More within NHIS -Sample adult -Sample child • External links to NHIS -Medical Expenditure Panel Survey (MEPS)/NHIS linkage -National Death Index (NDI)/NHIS linkage

NHIS Injury/Poisoning Episodes linked with NHIS Sample Adult File • Sample Adult Questionnaire -Cigarette smoking - physical activity -alcohol consumption - height and weight -gender specific questions -specific and chronic conditions -respiratory ailments -sensory impairment -mental health -health care access and utilization

NHIS Links with Other National Data Sets • MEPS/NHIS Link -1995 NHIS/1996 MEPS to 2002 MEPS/2001 and 2000 NHIS links -Health services used, cost of services, and how paid. • NDI/NHIS Link -2004 linked file includes NHIS survey years 1986 -2000 with mortality follow-up through 2002 -Age at NHIS interview, -underlying and multiple cause of death -Other variables by special request (e. g. , occupatio http: //www. cdc. gov/nchs/about/major/nhismep. htm http: //www. cdc. gov/nchs/r&d/nchs_datalinkage/nhis_data_linkage_activities. htm

National Health Interview Survey Strengths • Much more detailed info on cause /circumstances • Provides national estimates collected from persons not records • Largest range of data available in one system • Responses to specific prevention questions • Linking to other major data systems Limitations • Proxy reporting • Recall bias • Intentional injuries masked or underreported • Missing institutionalized population (e. g. , nursing homes)

Questions or additional information nchsinjury@cdc. gov mboudreault@cdc. gov NHIS: www. cdc. gov/nchs/about/major/nhis/quest_data_related_doc. htm Injury Data and Resources Website! http: //www. cdc. gov/nchs/injury. htm