Evaluation Of The Completeness And Data Quality Of Mortality Information System Used In Iran

Background Measures of mortality, such as age-specific death rates, life expectancy, causespecific death rates and years of life lost (YLLs) are commonly used to measure the health status of a population. In Iran, the Iranian National Organization for Civil Registration (NOCR) and the Iranian Ministry of Health and Medical Education (MOHaME) have both operated death registration systems. Other mortality data sources include censuses and surveys conducted by the Statistical Centre of Iran (SCI) and MOHaME. Due to incomplete data, late registration, missing data and errors in reporting or classifying the cause of death, the information from the NOCRrs Death Registration System has not been widely used for health policy. Independently, the MOHaME (Deputy of Research and Technology) collected data on causes of death from cemeteries over the period 1965-2001. However, the quality of cause of death data from this system was judged to be poor, and hence it was discontinued in 2002, and replaced by a new comprehensive death registration system operated by the MOHaME (Deputy of Health). Substantial progress has been made in integrating cause of death data from various systems into a single coordinated death registration and certification system, although the quality and utility of the data have never been evaluated. In 2006, 29 out of 30 provinces in Iran were covered by this death registration system. It is expected that all 30 provinces (including Tehran province) will be covered by the end of 2007. The broad goal of this thesis is to critically evaluate Iranian mortality data now becoming available for the entire population and, using the Burden of Disease Framework, develop estimates of age-sex-cause-specific mortality rates in Iran, and derive summary measures of mortality from these estimates. Specifically, the aims of the research conducted for this thesis are to: . b Evaluate the completeness and quality of the different data sources on mortality in Iran. . b Estimate national trends and provincial differentials in levels of mortality in Iran according to various mortality indices, including child mortality (5q0), adult mortality (45q15), and life expectancy at birth. . b Assess the extent and pattern of misclassification of death from seven leading causes of ill-defined diagnoses1 emanating from the MOHaMErs Death Registration System, particularly with regard to cardiovascular disease. . b Estimate the leading causes of death and Years of Life Lost (YLLs) for Iran based on corrected mortality and cause of death data for Iran for 2004. Methods A systematic review of all available studies on infant mortality in Iran was carried out and the most probable trend in child mortality over the period 1964-2004 was estimated. Death registration data between 1992 and 2004 were assessed for completeness to estimate the level of adult mortality. Life tables for 2004 were constructed for Iran based on these data, corrected for underegistration of deaths. To estimate provincial differences in levels of mortality, data from the death registration system (MOHaME) for 2004 for each province were separately evaluated for data completeness and life tables were constructed for provinces after correction for under-enumeration of death registration using the Brass Growth Balance method. For those provinces where it was not possible to adjust the data on adult deaths by using the Brass Growth Balance method, adult mortality was predicted based on adult literacy using information from provinces with reliable data. To assess the extent and patterns of misclassification of causes of death, a fieldwork study involving detailed medical records for 1,426 hospital deaths classified to seven ill-defined or vague causes of death in the death registration system were reviewed by trained physicians. These physicians had no prior knowledge of the cause of death recorded on the routine death certificate. Causes of death were coded using automated software (ACME), where possible, and compared with the cause assigned in the routine death registration system. Leading causes of premature death in Iran were then estimated on the basis of corrected values of Years of Life Lost (YLLs) by major causes. 1 The seven ill-defined and/or vague diagnoses were: lsenility without mention of psychosisr, lunknownr, lother cardiac diseasesr, lother and unspecified disorders of the circulatory systemr, lother respiratory diseasesr, lheart failurer and lhypertensive diseaser. Results Infant mortality decreased from an estimated 154 deaths per 1000 live births in 1964 to 26 in 2004. Based on an estimated 65% registration of female deaths and 89% for males, the risk of adult mortality between 15 to 60 years (45q15) in 2004 was estimated to be 0.124 and 0.175 for females and males, respectively. Life expectancy at birth in 2004 was estimated at 71.2 years for females and 68.7 for males. The average completeness of death registration for ages 5 years and older across all systems was 76% for the period 2001n2004. Important provincial differentials in mortality exist in Iran. Child mortality (5q0) in 2004 varied between 25 per 1000 live births in Tehran and Gilan provinces (both sexes) and 47 per 1000 live births in Sistan and Baluchistan province. For adults, provincial differences in mortality were significantly greater for males than females. Adult mortality (45q15) for females varied between 0.117 in Tehran province and 0.133 in Kerman province; for males 45q15 ranged from 0.149 in Tehran province to 0.218 in Kerman provinces. Life expectancy at birth for females was highest in Tehran province (73.8 years) and lowest in Sistan and Baluchistan (70.9 years). For males, life expectancy ranged from 70.9 years in Tehran province to 65.7 years in Sistan and Baluchistan province. Socioeconomic status in Iran (as assessed by literacy) is lowest in Sistan and Bluchistan, Kordistan and West Azerbaijan provinces and highest in Tehran, Semnan and Esfahan provinces. The probable underlying pattern of causes of death in Iran is substantially different to that suggested by the death registration system. Of the 582 cases of ill-defined or vague diagnoses at ages 15-69 years included in this study, less than 12% were still classified as such after detailed medical records review. The remainder were reassigned to various specific causes including ischaemic heart disease (33%), cerebrovascular disease (13%) and injuries (10%). A similar pattern of misclassification is apparent for death at older ages (70 years and above). Nearly half of the 738 cases of ill-defined deaths at these ages were reassigned to ischaemic heart disease and stroke (46%). Chronic obstructive respiratory diseases (7%) and digestive diseases (5%) were also misdiagnosed at the older ages.After adjustment, ischaemic heart disease is estimated to be the leading cause of death in Iran for both females (26.2% of all deaths) and males (23.2%), followed by cerebrovascular disease in females (17.3%) and road traffic accidents (13.1%) for males. The third leading cause of death for females is perinatal conditions (5.2%) and for males, cerebrovascular disease (9.6%). There are considerably different patterns of YLLs for males and females. Ischaemic heart disease is the leading cause of YLLs for females, accounting for 14.5 % of all YLLs, followed by perinatal conditions (13.4%), road traffic accidents (8.5%) and cerebrovascular disease (8.2%). For males, road traffic accidents are by for the leading cause of YLLs (21.6%), followed by ischaemic heart disease (12.5%), perinatal conditions (12.4%), and congenital anomalies (5.6%). Conclusion While the completeness of the newly-launched death registration system operated by the Iranian MOHaME appears to be acceptable in the majority of provinces, further efforts are needed to improve the quality of data on mortality in Iran. A significant proportion of deaths in the Iranian MOHaMErs Death Registration System are being classified to cause groups that are of little relevance to either epidemiological research or health policy. Substantial changes to procedures for diagnosing causes of death are urgently required if the data are to effectively guide health policies and programs in Iran. This suggests that policies to reduce road traffic accidents, perinatal conditions (such as birth trauma and birth asphyxia) and major vascular diseases are urgently required in Iran.