Related Experiment Videos
Querying the ill-defined stroke diagnoses on death certificates and their effects on type-specific mortality in
Tsung-Hsueh Lu1, Shiuh-Ming Huang
1Department of Public Health, Chung Shan Medical University, No. 110, Section 1, Chien Kuo North Road, Taichung 402, Taiwan. robertlu@ms1.hinet.net
Insights
Querying physicians significantly improved stroke subtype classification on death certificates in Taiwan. This method enhances cause of death statistics, revealing substantial changes in cerebral hemorrhage and infarction mortality rates.
Area of Science:
- Public Health
- Epidemiology
- Medical Statistics
Background:
- Ill-defined causes of death in Taiwan impede accurate stroke subtype mortality studies.
- Accurate classification of stroke subtypes is crucial for effective public health interventions.
Purpose of the Study:
- To evaluate the impact of physician queries on refining stroke subtype classification and mortality data.
- To assess the suitability of non-queried mortality data for type-specific stroke analysis.
Main Methods:
- Reviewed 2632 death certificates with ill-defined cerebrovascular disease codes (ICD-9 436, 437.9).
- Queried certifying physicians for 2035 hospital-issued certificates, obtaining responses for 1505 (74%).
- Analyzed changes in stroke subtype coding (cerebral hemorrhage, cerebral infarction) and age-adjusted death rates post-querying.
Main Results:
- Physician queries led to more specific coding for 79% of cases, primarily changing to cerebral hemorrhage (CH) or cerebral infarction (CI).
- Age-adjusted death rates increased significantly: CH by +16% (men) and +20% (women); CI by +100% (men) and +130% (women).
- Changes in CH/CI ratios were more pronounced in younger age groups (<65 years).
Conclusions:
- Physician querying is an effective method for improving the quality and specificity of stroke subtype data in mortality statistics.
- Unqueried mortality data in Taiwan are unsuitable for reliable type-specific stroke analysis due to significant under-classification.
- Enhanced data accuracy supports better epidemiological research and targeted public health strategies for stroke.
Abstract:
In Taiwan, the high proportion of stroke deaths of which the underlying causes of death (UCD) has been classified to ill-defined categories, has impeded valid type-specific mortality studies. The aim of this study was to assess the extent of changes in the distributions of stroke subtypes and mortality after querying the certifying physicians. There were 2632 death certificates issued from January through June 1998 in Taiwan of which the UCD was classified to ill-defined cerebrovascular disease categories (International Classification of Diseases, 9th Revision ICD-9 code 436 or 437.9). Staff of the Office of Statistics, Department of Health copied 2035 death certificates issued by hospitals and mailed them back to the certifying physicians to request more specific information on the stroke subtype. Responses were obtained for 1505 (74%) of the 2035 queries sent. Four fifths of these (1191/1505) resulted in a change to a more specific code. In most cases, the ICD-9 code was changed to 431 (cerebral hemorrhage, CH, 239/1505) or 433-4 (cerebral infarction, CI, 819/1505). After the querying, the percentage changes in the age-adjusted death rate (per 100,000 population) for CH were +16% for men and +20% for women; and for CI the percent changed were +100% for men and +130% for women. The changes increased as the age of the deceased increased. The changes in the CH/CI death ratios resulting from querying were greater in younger (less than 65 years old) age groups than in older age groups. Given the large gains of more specific information on stroke subtypes, querying is a useful method to improve the quality of cause of death statistics. However, the large changes in distributions and death rates of different stoke subtypes as a result of querying, show the non-querying mortality data are not suitable for type-specific analysis.