Centers for Disease Control and Prevention Household Survey Comparison Melissa Heinen, M.P.H. Margaret Warner, Ph.D. Lois Fingerhut, M.A. Montreal 2002
Dec 24, 2015
Centers for Disease Control and Prevention
Household Survey Comparison
Melissa Heinen, M.P.H.
Margaret Warner, Ph.D.
Lois Fingerhut, M.A.
Montreal 2002
Importance of Household Surveys
•Not limited to medical records
•Health professionals description
•More detail about the event
•Costs and resources
National Household Injury Surveys
Surveys reviewed
•Canada – Canadian Community Health Survey (CCHS) and the National Population Health Survey (NPHS)
•New Zealand – National Health Survey
•Scotland – Scottish Health Survey
•South Africa – South African Demographic and Health Survey
•United States – National Health Interview Survey
SOURCE: CDC/NCHS
National Household Injury Surveys
Items compared
•Recall Period
•Severity Threshold
•Survey Methodology
•Unit of Analysis
•Injury Definitions
•Survey Frequency
•Probing / Screening Question
SOURCE: CDC/NCHS
Recall Period
Items to consider
•Memory decay
•Telescoping
•Heaping
•Sample size / relative standard error
SOURCE: CDC/NCHS
Annual estimates of injury and poisoning conditions using different recall periods,
NHIS 1997
30
35
40
45
50
55
0 0.5 1 1.5 2 2.5 3
Months of recall
Mil
lio
ns
Injury andpoisoningconditions
Recall Period
SOURCE: CDC/NCHS
Canada 12-months
New Zealand 12-months
Scotland 12-months
South Africa 1-month
United States 3-months
Severity Threshold
Measurements
•Limitation in daily activities
•Medical advice or treatment
•Hospitalization status
SOURCE: CDC/NCHS
Severity Threshold
Items to consider
•Health insurance coverage
•Subjective - health seeking behaviors
•Accessibility of health professional (rural vs. urban)
•Emergency department visits
SOURCE: CDC/NCHS
Severity Threshold
SOURCE: CDC/NCHS
Canada Limitation of normal activities
New Zealand Received medical treatment
Scotland Saw a doctor, nurse, or other health professional
South Africa Treated by a doctor or a nurse
United States Got medical advice or treatment
Survey Methodology
Items to consider
•Non-response rates
•Representation
•Sensitive questions and non-reporting
•Cost and resources
SOURCE: CDC/NCHS
Survey Methodology
SOURCE: CDC/NCHS
Canada NPHS -Telephone
CHHS – Telephone and face-to-face
New Zealand Face-to-face
Scotland Face-to-face
South Africa Face-to-face
United States Face-to-face
Unit of Analysis
Items to consider
•Comparability of numbers•Events (e.g., car crash)•Persons•Most severe•Most recent•Conditions (e.g., fractured leg)
SOURCE: CDC/NCHS
Unit of Analysis
SOURCE: CDC/NCHS
Canada Injury events
New Zealand Injured persons
Scotland Person based – most recent injury-related event
South Africa Injured persons
United States Injury events
Injury Definitions
Items to consider
•Self report vs. coding of self reports (ICD coding)
•Comparability
SOURCE: CDC/NCHS
Injury Definitions
SOURCE: CDC/NCHS
Canada ICD coded injuries
New Zealand ICD coded injuries
Scotland Self reported accidents
South Africa Self reported injuries
United States ICD coded injuries
Survey Frequency
Items to consider
•Trends over time
•Assess data
•Sample size
SOURCE: CDC/NCHS
Survey Frequency
SOURCE: CDC/NCHS
Canada Conducts surveys annually and reports results every two years
New Zealand Conducts survey every three years
Scotland Conducts survey every three years
South Africa Conducts survey every five years
United States Conducts survey and reports results annually
Screener Question
Items to consider
•Length
•Complexity
•Embedded examples
SOURCE: CDC/NCHS
Screener Question
SOURCE: CDC/NCHS
Canada “(Not counting repetitive strain injuries), in the past 12 months, that is from ## to yesterday, were you injured?”
New Zealand “In the last 12 months, have you had an injury which you received medical treatment?”
Scotland “How many accidents – about which you saw a doctor, nurse, or other health professional – have you had in the last 12 months?”
Screener Question
SOURCE: CDC/NCHS
South Africa “Did you have any injury that was treated by a doctor or a nurse in the last 30 days?”
United States “During the past 3 months, that is since ###, were you or anyone in the family injured or poisoned seriously enough that you got medical advice or treatment?”
US Survey
•Decrease injury rate
•Cognitive testing -screening question
•Length
•Definitions
•Severity
Screener QuestionProposed changes for US
• During the past 3 months, that is since ##, did you have an injury where any part of your body was hurt, for example, with a broken bone, sprain, burn, wound, cut, bruise, or animal or insect bite?
• During the past 3 months, how many times were you injured?
• Did you talk to or see a medical professional about any of these injuries?
• Of the # times that you were injured, how many of those times was the injury serious enough that you consulted a medical professional?
SOURCE: CDC/NCHS
Screener QuestionProposed changes for US
• During the past 3 months, that is since ##, were you poisoned by swallowing or breathing in a harmful substance such as bleach, carbon monoxide, or too many pills or drugs? (Do not include food poisoning, sun poisoning, or poison ivy rashes)
• During the past 3 months, how many different times were poisoned?
• Did you talk to or see a medical professional about any of these poisonings?
• Of the # times that you were poisoned, how many of those times was the poisoning serious enough that you consulted a medical professional?
SOURCE: CDC/NCHS
National Household Injury Surveys
Additional items to consider
• Placement of injury section within a larger survey
• Length of survey
• Sample selection
• Definitions of injuries and poisonings
• Narrative introduction to the section
• Weighting of data
• Respondent vs. Proxy
SOURCE: CDC/NCHS
National Household Injury Surveys
Generic considerations
•Continuous funding
•Strong lobbying
•Demonstrating usefulness of data
SOURCE: CDC/NCHS
Centers for Disease Control and Prevention
Website: http://www.cdc.gov/nchs/about/otheract/ice/housesur.htm
Considerations
• Should our goal be to compare our selves to each other to have better questions and survey methodology?
• Is it a worthwhile goal to standardize questions?
•Cause compatible with ICD?
•Should health behaviors be a priority?
• Is it even worth comparing numbers?