Diabetes Reporting on Death Certificates in Kentucky, 2002 -2003 Melonie Heron, Ph. D. & Robert N. Anderson, Ph. D. Mortality Statistics Branch Division of Vital Statistics Centers for Disease Control and Prevention National Center for Health Statistics
Diabetes Checkbox Questions Implemented in July of 2002 1. Did the deceased have Diabetes? 2. Was Diabetes an immediate, underlying, contributing cause of or condition leading to death?
Cause of Death Coding and Statistics • All information reported in Parts I and II is coded (up to 20 conditions) • Underlying cause of death is selected from among the conditions reported • Mortality statistics usually based on underlying cause of death
Issues in the reporting of diabetes on the death certificate • Certifier may not know about a diagnosis of diabetes or its relation to other comorbid conditions • Causal sequence not always obvious in the presence of other co-morbid conditions • Diabetes most often reported in Part II
Upside of checkbox questions • 1 st question - Ability to identify diabetic decedents • 2 nd question indicates whether diabetes should have been reported in Part I or II • Reminder to certifier to report diabetes
Downside • Could result in underreporting in cause of death (COD) section of the death certificate • Cause of death coding standards and rules • Causes not reported in COD section are not coded • Causes not coded do not make it into official statistics • Loss of detail – no info about type
Responses • 12, 320 death certificates had a checked response of “yes” to at least one of the questions • 12, 231 had a checked response of “yes” to question 1 • 84 had a checked response of “no” to question 1, but checked “yes” to question 2 • 5 had no response to question 1, but checked “yes” to question 2
Diabetic Decedents (“Yes” to Question 1), KY: 7/2002 – 12/2003 47. 3% 52. 2% N=12, 231 Question 2 – Was diabetes an immediate, underlying or contributing cause of death?
Response in Checkbox 2 vs. Reported Cause of Death • 52% (3, 048) of deaths with response “yes” to checkbox 2 had diabetes reported in the cause of death section • 27% (1, 560) of deaths with response “yes” to checkbox 2 had diabetes reported as the underlying cause of death • 17% (1, 064) of deaths with response “no” to checkbox 2 had diabetes reported in the cause of death section
What does Checkbox 2 tell us? • Diabetes is not reported in the cause of death section in about 40% of cases in which diabetes is identified as an immediate, underlying or contributing cause • But…is this checkbox question an indicator or a cause of underreporting?
Do the Checkboxes Have an Effect on the Trend in Diabetes Reporting?
Diabetes Deaths: US, 1990 -2003
Diabetes Deaths: Kentucky, 1990 -2003
Diabetes Deaths: US, 1990 -2003
Diabetes Deaths: Kentucky, 1990 -2003
Conclusions • The diabetes checkboxes clearly have a marked effect on where (Part I or II) diabetes is reported • Increases in Part I and decreases in Part II • Slight increase in diabetes as the underlying cause of death • Decrease in overall reporting of diabetes
Alternative Methodology • Can use national health surveys linked to mortality data to: • Explore the extent to which diabetes is reported as a cause of death for those who self-reported diabetes. • Analyze the impact of diabetes on risk of mortality, generally and by cause of death
NHIS-NDI Linked Data • National Health Interview Survey (NHIS) • National survey conducted annually by NCHS since 1957 • Monitors the health of the U. S. population • National Death Index (NDI) • National file of death record information compiled by NCHS from state registries • Deaths added annually, 1979 -present • Helps researchers determine the mortality status of study participants
Study Approach • Linked NHIS and NDI files for 1997 to 2000 • Includes mortality follow up of NHIS participants until 2002 • Restricted analyses to survey participants who were eligible for follow up and determined to be dead • Compared self-reports of diabetes on the survey to subsequent recording of diabetes as a cause of death on the death certificate
Percent of deceased diabetics for whom diabetes was reported as a cause of death, NHIS-NDI Linked Data, 1997 -2002
Future Linked Data Research • Trends in self-reports of diabetes vs. death certificate recording of diabetes • Extend analyses past 2002 • Any changes in the Kentucky trends? (assuming sufficient number of observed deaths) • Trends in mortality from other causes of death for diabetic decedents • Analyze risk of death from diabetes and other causes using multivariate modeling
THE END Contacts: mheron@cdc. gov, RNAnderson@cdc. gov Centers for Disease Control and Prevention
Скачать презентацию Diabetes Reporting on Death Certificates in Kentucky 2002
af7bcb66456ee59aa681ae70411e73e5.ppt