Impact of misclassification on measures of cardiovascular disease mortality in the Islamic Republic of Iran: a cross-sectional study

OBJECTIVE: To assess the extent and pattern of misclassification of death from non-specific diagnoses emanating from the Iranian death registration system, and to correct the data for health policy and planning. METHODS: Detailed medical records for 1426 hospital deaths classified to seven ill-defined or vague causes of death were reviewed by trained physicians, who then completed standard death certificates. Underlying causes of death from the review were compared with the cause assigned in registration data. FINDINGS: The probable underlying pattern of causes of death in the Islamic Republic of Iran is substantially different to that suggested by the death registration system. About 88% of 582 cases with non-specific diagnoses at ages 15-69 years were reassigned to various specific causes including ischaemic heart disease (33%), stroke (13%) and injuries (10%). A similar pattern of misclassification is apparent for 738 deaths at older ages (70 years and over), with 46% being reassigned to ischaemic heart disease and stroke. CONCLUSION: A significant proportion of deaths in the Iranian death registration system are being classified to cause groups of little relevance to epidemiological research or health policy. Reassignment of these deaths would increase the proportion of deaths from ischaemic heart disease and cerebrovascular diseases each by 32%, diabetes mellitus by 68% and chronic lower respiratory diseases by 73%. Substantial changes to procedures for diagnosing causes of death are urgently required if registration data are to effectively guide health policies and programmes in the Islamic Republic of Iran.     

Validation of verbal autopsy procedures for adult deaths in China

BACKGROUND: Vital registration of causes of death in China is incomplete with poor coverage of medical certification. Information on the leading causes of mortality will continue to rely on verbal autopsy (VA) methods. A new international VA form is being considered for data collection in China, but it first needs to be validated to determine its operating characteristics. METHODS: Detailed medical records and clinical evidence for 3290 deaths (mostly adults) among residents of six cities representative of the urban Chinese population were reviewed by a panel of physicians and coded by experts to establish a reference underlying cause of death. Independently, families of the deceased were interviewed using a structured symptomatic questionnaire and a separate death certificate was prepared for each matching case (2102). Validity of the VA procedure was assessed using standard measurement criteria of sensitivity, specificity, and positive predictive value. RESULTS: VA methods perform reasonably well in identifying deaths from several leading causes of adult deaths including stroke, several major cancer sites (lung, liver, stomach, oesophagus, and colorectal), and transport accidents. Sensitivity was less satisfactory in detecting deaths from several causes of major public health concern in China including ischaemic heart disease, chronic obstructive pulmonary disease, diabetes, and tuberculosis, and was particularly poor in diagnosing deaths from viral hepatitis, hypertension, and kidney diseases. CONCLUSIONS: VA is an imprecise tool for detecting leading causes of death among adults. However, much of the misclassification generally occurs within broad cause groups (e.g. CVD, respiratory diseases, and liver diseases). Moreover, compensating patterns of misclassification would appear to suggest that, in urban China at least, the method yields population-level cause-specific estimates that are reasonably reliable. These results suggest the possible utility of these methods in rural China, to back up the low coverage of medical certification of cause of death owing to poor access to health facilities there.     

Evaluation of the quality of cause of death statistics in rural China using verbal autopsies

BACKGROUND: Death registration systems in rural China are in a developmental stage. The Disease Surveillance Points (DSP) system provides the only nationally representative information on causes of death. In this system, there are no standard procedures or instruments for ascertaining causes of death; hence available statistics require careful evaluation before use. AIM: To assess the reliability of data from the DSP. METHODS: 14 DSP sites were selected through stratified sampling, enrolling 2482 deaths registered during June-November 2002. Defined verbal autopsy (VA) procedures were used to derive underlying causes of death. kappa Measures of agreement between VA and registered diagnoses were computed. VA diagnoses were used as references to compute sensitivity and positive predictive values. Finally, patterns of cause-specific discordance between the two data sources were studied. RESULTS: kappa Scores indicate only moderate agreement for chronic obstructive pulmonary disease (COPD) and ischaemic heart disease (IHD). The results also suggest that COPD is undercounted in registration data. A significant degree of cross attribution of cause was found between the two data sources. Overall, the VA was successful in limiting deaths with ill-defined causes to <10%. CONCLUSIONS: Diagnoses based on structured symptom questionnaires in the VA approach seem plausible and reliable as compared with lay-reported diagnoses in registration data. Concerns with attribution of cause of death due to COPD, IHD and tuberculosis in registration data suggest caution in their use for research and health programme purposes. The VA methods tested in this study offer promise for implementation in the routine registration system.     

Estimating mortality and causes of death in Turkey: methods, results and policy implications

BACKGROUND: Cause-specific mortality statistics are primary evidence for health policy formulation, programme evaluation, and epidemiological research. In Turkey, a partially functioning vital registration system in urban areas yields fragmentary evidence on levels and causes of mortality. This article discusses the application of innovative methods to develop national mortality estimates in Turkey, and their implications for national health development policies. METHODS: Child mortality levels from the Demography and Health Survey (DHS) were applied to model life tables to estimate age-specific death rates. Reported causes of death from urban areas were adjusted using re-distribution algorithms from the Global Burden of Disease (GBD) Study. Rural cause structure was estimated from epidemiological models. Local epidemiological data was used to adjust model-based estimates. RESULTS: Life expectancy at birth in 2000 was estimated to be 67.7 years (males) and 71.9 years (females), about 8-10 years lower than in Western Europe. Leading causes of death include major vascular diseases (ischaemic heart disease, stroke) causing 35-38% of deaths, chronic obstructive lung disease and lung cancer in men, but also perinatal causes, lower respiratory infections and diarrhoeal diseases. Injuries cause about 6-8% of deaths, although this may be an underestimate. CONCLUSIONS: Mortality estimates are uncertain in Turkey, given the poor quality of death registration systems. Application of burden of disease methods suggests that there has been progress along the epidemiological transition. Key health development strategies for Turkey include improved access to communicable disease control technologies, and urgent attention to the development of a reliable, nationally representative health information system.     

Estimated causes of death in Thailand, 2005: implications for health policy

Background  

Almost 400,000 deaths are registered each year in Thailand. Their value for public health policy and planning is greatly diminished by incomplete registration of deaths and by concerns about the quality of cause-of-death information. This arises from misclassification of specified causes of death, particularly in hospitals, as well as from extensive use of ill-defined and vague codes to attribute the underlying cause of death. Detailed investigations of a sample of deaths in and out of hospital were carried out to identify misclassification of causes and thus derive a best estimate of national mortality patterns by age, sex, and cause of death.     

A critical assessment of mortality statistics in Thailand: potential for improvements

This study evaluates the collection and flow of mortality and cause-of-death (COD) data in Thailand, identifying areas of weakness and presenting potential approaches to improve these statistics. Methods include systems analysis, literature review, and the application of the Health Metrics Network (HMN) self-assessment tool by key stakeholders. We identified two weaknesses underlying incompleteness of death registration and inaccuracy of COD attribution: problems in recording events or certifying deaths, and problems in transferring information from death certificates to death registers. Deaths occurring outside health facilities, representing 65% of all deaths in Thailand, contribute to the inaccuracy of cause-of-death data because they must be certified by village heads with limited knowledge and expertise in cause-of-death attribution. However, problems also exist with in-hospital cause-of-death certification by physicians. Priority should be given to training medical personnel in death certification, review of medical records by health personnel in district hospitals, and use of verbal autopsy techniques for assessing internal consistency. This should be coupled with stronger collaboration with district registrars for the 65% of deaths that occur outside hospitals. Training of physicians and data coders and harmonization of death certificates and registries would improve COD data for the 35% of deaths that take place in hospital. Public awareness of the importance of registering all deaths and the application of registration requirements prior to funerals would also improve coverage, though enforcement would be difficult.     

Population Health Metrics Research Consortium gold standard verbal autopsy validation study: design, implementation, and development of analysis datasets

Background

Verbal autopsy methods are critically important for evaluating the leading causes of death in populations without adequate vital registration systems. With a myriad of analytical and data collection approaches, it is essential to create a high quality validation dataset from different populations to evaluate comparative method performance and make recommendations for future verbal autopsy implementation. This study was undertaken to compile a set of strictly defined gold standard deaths for which verbal autopsies were collected to validate the accuracy of different methods of verbal autopsy cause of death assignment.

Methods

Data collection was implemented in six sites in four countries: Andhra Pradesh, India; Bohol, Philippines; Dar es Salaam, Tanzania; Mexico City, Mexico; Pemba Island, Tanzania; and Uttar Pradesh, India. The Population Health Metrics Research Consortium (PHMRC) developed stringent diagnostic criteria including laboratory, pathology, and medical imaging findings to identify gold standard deaths in health facilities as well as an enhanced verbal autopsy instrument based on World Health Organization (WHO) standards. A cause list was constructed based on the WHO Global Burden of Disease estimates of the leading causes of death, potential to identify unique signs and symptoms, and the likely existence of sufficient medical technology to ascertain gold standard cases. Blinded verbal autopsies were collected on all gold standard deaths.

Results

Over 12,000 verbal autopsies on deaths with gold standard diagnoses were collected (7,836 adults, 2,075 children, 1,629 neonates, and 1,002 stillbirths). Difficulties in finding sufficient cases to meet gold standard criteria as well as problems with misclassification for certain causes meant that the target list of causes for analysis was reduced to 34 for adults, 21 for children, and 10 for neonates, excluding stillbirths. To ensure strict independence for the validation of methods and assessment of comparative performance, 500 test-train datasets were created from the universe of cases, covering a range of cause-specific compositions.

Conclusions

This unique, robust validation dataset will allow scholars to evaluate the performance of different verbal autopsy analytic methods as well as instrument design. This dataset can be used to inform the implementation of verbal autopsies to more reliably ascertain cause of death in national health information systems.     

Chronic diseases now a leading cause of death in rural India--mortality data from the Andhra Pradesh Rural Health Initiative

INTRODUCTION: India is undergoing rapid epidemiological transition as a consequence of economic and social change. The pattern of mortality is a key indicator of the consequent health effects but up-to-date, precise, and reliable statistics are few, particularly in rural areas. METHODS: Deaths occurring in 45 villages (population 180 162) were documented during a 12-month period in 2003-04 by multipurpose primary healthcare workers trained in the use of a verbal autopsy tool. Algorithms were used to define causes of death according to a limited list derived from the international classification of disease version 10. Causes were assigned by two independent physicians with disagreements resolved by a third. RESULTS: A total of 1354 deaths were recorded with verbal autopsies completed for 98%. A specific underlying cause of death was assigned for 82% of all verbal autopsies done. The crude death rate was 7.5/1000 (95% confidence interval, 7.1-7.9). Diseases of the circulatory system were the leading causes of mortality (32%), with similar proportions of deaths attributable to ischaemic heart disease and stroke. Second was injury and external causes of mortality (13%) with one-third of these deaths attributable to deliberate self harm. Third were infectious and parasitic diseases (12%). Tuberculosis and intestinal conditions each caused one-third of deaths within this category. HIV was assigned as the cause for 2% of all deaths. The fourth and fifth leading causes of death were neoplasms (7%) and diseases of the respiratory system (5%). CONCLUSION: Non-communicable and chronic diseases are the leading causes of death in this part of rural India. The observed pattern of death is unlikely to be unique to these villages and provides new insight into the rapid progression of epidemiological transition in rural India.     

Global and regional cause-of-death patterns in 1990.

Demographic estimation techniques suggest that worldwide about 50 million deaths occur each year, of which about 39 million are in the developing countries. In countries with adequate registration of vital statistics, the age at death and the cause can be reliably determined. Only about 30-35% of all deaths are captured by vital registration (excluding sample registration schemes); for the remainder, cause-of-death estimation procedures are required. Indirect methods which model the cause-of-death structure as a function of the level of mortality can provide reasonable estimates for broad cause-of-death groups. Such methods are generally unreliable for more specific causes. In this case, estimates can be constructed from community-level mortality surveillance systems or from epidemiological evidence on specific diseases. Some check on the plausibility of the estimates is possible in view of the hierarchical structure of cause-of-death lists and the well-known age-specific patterns of diseases and injuries. The results of applying these methods to estimate the cause of death for over 120 diseases or injuries, by age, sex and region, are described. The estimates have been derived in order to calculate the years of life lost due to premature death, one of the two components of overall disability-adjusted life years (DALYs) calculated for the 1993 World development report. Previous attempts at cause-of-death estimation have been limited to a few diseases only, with little age-specific detail. The estimates reported in detail here should serve as a useful reference for further public health research to support the determination of health sector priorities.     

A medical cause of death validation study of adult Aboriginal deaths in the Northern Territory of Australia in 1992

Judged on the criterion of equity, premature adult Aboriginal mortality is the most serious public health problem faced in Australia today. There have been a number of published epidemiological studies that have analysed Aboriginal cause of death data, but this is the first study to formally validate such data. The study sample included all adult Aboriginal people who lived and died in the Northern Territory in 1992, excluding residents of the Alice Springs region. The appropriateness of underlying cause of death codes was assessed by a single reviewer in light of death certificates, medical records, postmortem records and interviews with key health professional informants. Data were collected on 220 deaths. 8% (17 out of 220) of deaths were classified erroneously at the ICD-9 chapter level. Errors in death certification accounted for 64% (11 out of 17) of the chapter errors and diagnostic and coding errors for 18% (3 out of 17) each. The overall impact on mortality statistics was less severe because some cross-chapter classification errors cancelled each other out. Misclassification errors aggregated mainly in chapter VII (circulatory diseases) of the ICD-9 classification which was overcounted by 3.2%, and chapter VIII (respiratory diseases) which was overcounted by 1.3%. Before correction for misclassification error, circulatory diseases were judged to cause the highest proportion of deaths, whereas after correction, respiratory diseases accounted for the highest proportion. Despite this, the overall quality of the medical cause of death statistics was of a sufficiently good standard from a public health perspective to broadly inform health policy. Future attempts to improve the validity of medical cause of death statistics for Australian Aboriginal people should focus on the education of medical practitioners about the purpose and process of death certification.     

Effect of misclassification of causes of death in verbal autopsy: can it be adjusted?

BACKGROUND: Verbal autopsy (VA) is an indirect method of ascertaining cause of death from information about symptoms and signs obtained from bereaved relatives. This method has been used in several settings to assess cause-specific mortality. However, cause-specific mortality estimates obtained by VA are susceptible to bias due to misclassification of causes of death. One way of overcoming this limitation of VA is to adjust the crude VA estimate of cause-specific mortality fractions (CSMF) using the sensitivity and specificity of the VA tool. This paper explores the application of sensitivity and specificity of VA data obtained from a hospital-based validation study for adjusting the effect of misclassification error in VA data obtained from a demographic surveillance system. METHOD: Data from a multi-centre validation study of 796 adult VA, conducted in Tanzania, Ethiopia and Ghana, were used to explore the effect of distribution of causes of death in the validation study population and the pattern of misclassification on the sensitivity and specificity of VA. VA estimates of CSMF for six causes (acute febrile illness, diarrhoeal diseases, TB/AIDS, cardiovascular disorders, direct maternal causes and injures) were obtained from a demographic surveillance system in Morogoro Rural District in Tanzania. These were adjusted for misclassification error by using sensitivity and specificity values of VA obtained from the validation study in a model proposed for correcting the effect of misclassification error in morbidity prevalence surveys. RESULTS: Sensitivity and specificity of VA differed between the three validation study sites depending on the distribution of causes of death. These differences were explained by variations in the level and pattern of misclassification between sites. When these estimates of sensitivity and specificity were applied to data from the demographic surveillance system with a comparable structure of causes of death the difference between crude and adjusted VA estimates of CSMF ranged from 3 to 83%. CONCLUSION: Estimates of sensitivity and specificity obtained from hospital-based validation studies must be used cautiously as a de facto 'gold standard' for adjusting the misclassification error in CSMF derived from VA. It is not possible to use sensitivity and specificity estimates derived from a location-specific validation study to adjust for misclassification in VA data from populations with substantially different patterns of cause-specific mortality.     

Four million neonatal deaths: counting and attribution of cause of death

Each year there are an estimated four million neonatal deaths and at least 3.2 million stillbirths. Three-quarters of the world's neonatal deaths are counted only through five-yearly retrospective household surveys. Without these surveys we would have no data, but limitations remain particularly in detecting deaths on the first day of life. Comparable reliable neonatal cause of death data through vital registration are available for less than 5% of the world's neonatal deaths, necessitating modelled estimates for the majority of the world. Improving the quantity, quality and frequency of data for numbers and causes of neonatal deaths is essential to effectively guide the increasing investments to reduce these deaths. Advancing the data requires general investment in information systems and specific improvements of tools and methods for both household surveys and verbal autopsy, particularly the use of consistent case definitions and hierarchical attribution of cause of death. An important paradigm shift is from historical categories for cause of death ('perinatal causes') to programmatic categories which are consistent with the International Classification of Diseases. If neonatal deaths remain uncounted, they cannot count in policy and in programmes.     

Myanmar mortality registration: an assessment for system improvement

Background

The vital registration system in Myanmar has a long history and geographical coverage is currently high. However, a recent assessment of vital registration systems of 148 countries showed poor performance of the death registration system in Myanmar, suggesting the need for improvement. This study assessed the quality of mortality data generated from the vital registration system with regard to mortality levels and patterns, quality of cause of death data, and completeness of death registration in order to identify areas for improvement.

Methods

The study used registered deaths in 2013 from the vital registration system, data from the 2014 Myanmar Population and Housing Census, and mortality indicators and COD information for the country estimated by international organizations. The study applied the guidelines recommended by AbouZahr et al. 2010 to assess mortality levels and patterns and quality of cause of death data. The completeness of death registration was assessed by a simple calculation based on the estimated number of deaths.

Results

Findings suggested that the completeness of death registration was critically low (less than 60%). The under-registration was more severe in rural areas, in states and regions with difficult transportation and poor accessibility to health centers and for infant and child deaths. The quality of cause of death information was poor, with possible over-reporting of non-communicable disease codes and a high proportion of ill-defined causes of death (22.3% of total deaths).

Conclusion

The results indicated that the vital registration system in Myanmar does not produce reliable mortality statistics. In response to monitoring mortalities as mandated by the Sustainable Development Goals, a significant and sustained government commitment and investment in strengthening the vital registration system in Myanmar is recommended.

     

Preventing type 2 diabetes mellitus in Qatar by reducing obesity,smoking, and physical inactivity: mathematical modeling analyses

Cause of death statistics are an important tool for quality control of the health care system. Their reliability, however, is controversial. Comparing death certificates with their corresponding medical records is implemented only occasionally but may point to quality problems. We aimed at exploring the agreement between information in the cause of death statistics and hospital discharge diagnoses at death. Selection of disease categories was based on ICD-10 Tabulation List for Morbidity and ICD-10 Mortality Tabulation List 2. Index cases were defined as deaths having occurred among Swiss residents 2010–2012 in a hospital and successfully linked to the Swiss National Cohort. Rare, external and ill-defined causes were excluded from comparison, leaving 53,605 deaths from vital statistics and 47,311 deaths from hospital discharge statistics. For 95% of individuals, respective information from the 2000 census could be retrieved and used for multiple logistic regression. For 83% of individuals the underlying cause of death could be traced among hospital diagnoses and for 77% the principal hospital diagnosis among the cause of death information. Mirroring different evaluation of complex situations by individual physicians, rates of agreement varied widely depending on disease/cause of death, but were generally in line with similar studies. Multiple logistic regression revealed however significant variation in reporting that could not entirely be explained by age or cause of death of the deceased suggesting differential exploitation of available diagnosis information. Substantial regional variation and lower agreement rates among socially disadvantaged groups like single, less educated, or culturally less integrated persons suggest potential for improving reporting of diagnoses and causes of death by physicians in Switzerland. Studies of this kind should be regularly conducted as a quality monitoring.     

Comparing Swedish hospital discharge records with death certificates: implications for mortality statistics

BACKGROUND: The quality of mortality statistics is of crucial importance to epidemiological research. Traditional editing techniques used by statistical offices capture only obvious errors in death certification. In this study we match Swedish hospital discharge data to death certificates and discuss the implications for mortality statistics. METHODS: Swedish death certificates for 1995 were linked to the national hospital discharge register. The resulting database comprised 69 818 individuals (75% of all deaths), 39 872 (43%) of whom died in hospital. The diagnostic statements were compared at Basic Tabulation List level. RESULTS: The last main diagnosis and the underlying cause of death agreed in 46% of cases. Agreement decreased rapidly after discharge. For hospital deaths, the main diagnosis was reported on 83% of the certificates, but only on 46% of certificates for non-hospital deaths. Malignant neoplasms and other dramatic conditions showed the best agreement and were often reported as underlying causes. Conditions that might follow from some other disease were often reported as contributory causes, while symptomatic and some chronic conditions were often omitted. In 13% of cases, an ill-defined main condition was replaced by a more specific cause of death. CONCLUSIONS: There is no apparent reason to question the death certificate if the main diagnosis and underlying cause agree, or if the main diagnosis is a probable complication of the stated underlying cause. However, cases in which the main diagnosis cannot be considered a complication of the reported underlying cause should be investigated, and assessments made of the feasibility and cost-effectiveness of routinely linking hospital records to death certificates.     

Can regional variation in "avoidable" mortality be explained by deaths outside hospital? A study from Sweden, 1987-90.

STUDY OBJECTIVE: This study aimed to calculate the proportion of deaths outside hospital in Sweden for some conditions for which the acute medical management may be important to the outcome and to analyse whether the proportion of deaths outside hospital can explain regional variations in mortality from these causes of death. DESIGN: The place of death was registered on all death certificates in Sweden during the period 1987-90. The proportion of deaths outside hospital was calculated at the national level for selected causes of death. Variation in cause-specific mortality among the 26 administrative health areas in Sweden was analysed. Death rate ratios were calculated with standardisation for age and sex using the national rate as standard. The correlation between the proportion of deaths outside hospital in each health area and the cause specific mortality irrespective of place of death was calculated. For areas with a significantly high death rate the ratios for mortality outside hospital as well as in hospital were analysed in order to decide which component of mortality represented a high mortality risk. SETTING AND PARTICIPANTS: All death registration in Swedish citizens and other residents in Sweden aged under 70 years between 1987 and 1990 which gave diabetes, asthma, ischaemic heart disease, cerebrovascular diseases, or ulcer of the stomach or duodenum as the underlying cause of death. MAIN RESULTS: For asthma (58%) and ischaemic heart disease (54%), most deaths occurred outside hospital. For most causes of death, however, no correlation was found among the health areas between the proportion of deaths outside hospital and the SMR for mortality irrespective of the place of death. A high death rate was associated with a high proportion of deaths outside hospital, for diabetes in one area in the north of Sweden (Norrbotten) and for ulcer of the stomach and duodenum in one large municipality (Göteborg). CONCLUSIONS: The high proportion of deaths outside hospital at the national level for some of the conditions studied suggests that in-depth studies of the process preceding death and the functioning of medical care are needed. In most cases, however, no evidence was found that regional variation in mortality could be explained by death outside hospital. The results for diabetes in Norbotten and ulcer of stomach and duodenum in Göteborg indicate that in-depth studies on the quality of care are required.     

Counting the dead and what they died from: an assessment of the global status of cause of death data

OBJECTIVE: We sought to assess the current status of global data on death registration and to examine several indicators of data completeness and quality. METHODS: We summarized the availability of death registration data by year and country. Indicators of data quality were assessed for each country and included the timeliness, completeness and coverage of registration and the proportion of deaths assigned to ill-defined causes. FINDINGS: At the end of 2003 data on death registration were available from 115 countries, although they were essentially complete for only 64 countries. Coverage of death registration varies from close to 100% in the WHO European Region to less than 10% in the African Region. Only 23 countries have data that are more than 90% complete, where ill-defined causes account for less than 10% of total of causes of death, and where ICD-9 or ICD-10 codes are used. There are 28 countries where less than 70% of the data are complete or where ill-defined codes are assigned to more than 20% of deaths. Twelve high-income countries in western Europe are included among the 55 countries with intermediate-quality data. CONCLUSION: Few countries have good-quality data on mortality that can be used to adequately support policy development and implementation. There is an urgent need for countries to implement death registration systems, even if only through sample registration, or enhance their existing systems in order to rapidly improve knowledge about the most basic of health statistics: who dies from what?     

A comparison of two cause-of-death classification systems for deaths among women of reproductive age in Menoufia, Egypt

Mortality data ascertained from sources other than a death registration system can validate the accuracy of the system, but this information is rarely obtained. Data on 1979 deaths among reproductive age women were collected in the 1981-1983 Reproductive Age Mortality Survey (RAMOS) in the governorate of Menoufia, Egypt, and compared with data on these deaths as recorded by the Egyptian death registration system. Although the distribution of the causes of death were similar, there were substantial differences between classification systems for deaths due to particular causes. Over half of the deaths classified differently by the systems were those assigned to circulatory disease on the death certificate. In contrast, there was a high rate of agreement between systems in the classification of trauma deaths. About half (52.4%) of cancer deaths had the same site-specific cancer diagnosis assigned by RAMOS. The percentage of deaths assigned to maternal causes was three times higher in RAMOS (19.2%) than on death certificates (6.1%). Reported mortality rates for this often-preventable cause of death have been substantially underestimated in national death registration systems. Such underreporting masks the need for additional prenatal care and maternal health programmes.     

Estimating cause-specific mortality from community- and facility-based data sources in the United Republic of Tanzania: options and implications for mortality burden estimates

OBJECTIVE: To compare mortality burden estimates based on direct measurement of levels and causes in communities with indirect estimates based on combining health facility cause-specific mortality structures with community measurement of mortality levels. METHODS: Data from sentinel vital registration (SVR) with verbal autopsy (VA) were used to determine the cause-specific mortality burden at the community level in two areas of the United Republic of Tanzania. Proportional cause-specific mortality structures from health facilities were applied to counts of deaths obtained by SVR to produce modelled estimates. The burden was expressed in years of life lost. FINDINGS: A total of 2884 deaths were recorded from health facilities and 2167 recorded from SVR/VAs. In the perinatal and neonatal age group cause-specific mortality rates were dominated by perinatal conditions and stillbirths in both the community and the facility data. The modelled estimates for chronic causes were very similar to those from SVR/VA. Acute febrile illnesses were coded more specifically in the facility data than in the VA. Injuries were more prevalent in the SVR/VA data than in that from the facilities. CONCLUSION: In this setting, improved International classification of diseases and health related problems, tenth revision (ICD-10) coding practices and applying facility-based cause structures to counts of deaths from communities, derived from SVR, appears to produce reasonable estimates of the cause-specific mortality burden in those aged 5 years and older determined directly from VA. For the perinatal and neonatal age group, VA appears to be required. Use of this approach in a nationally representative sample of facilities may produce reliable national estimates of the cause-specific mortality burden for leading causes of death in adults.