Misclassification Of Cause Research Articles

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

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. 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. 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%. 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.

Population Variation in Cause of Death: Level, Gender, and Period Effects

The trends in 13 cause of death categories are examined with respect to expectation of life, sex differences, and period effects while misclassification of cause of death is controlled. The results suggest that as mortality declines, 1) the increasingly U-shaped age pattern of mortality is a period effect associated with the infectious diseases, 2) the risks of both overall infectious and degenerative causes of death decline, and 3) infectious disease mortality declines more in males, while degenerative disease mortality declines more in females. Finally, the model shows that some contemporary populations are approaching the limits of reduction in mortality during infancy, childhood, and young adulthood. Past declines in the degenerative diseases, however, suggest that mortality may continue to decline.

A cohort study on mortality among wives of workers in the asbestos cement industry in Casale Monferrato, Italy.

The study investigates mortality from cancer and other diseases in a cohort of wives of asbestos cement workers in Casale Monferrato (northwest Italy). After the exclusion of women with an occupational record in the asbestos cement industry, the cohort comprised 1964 women. Their domestic exposure was estimated according to their husbands' periods of employment in the plant: 1740 had a period of domestic exposure whereas the remaining 224 married an asbestos cement worker only after he definitely stopped his activity in the asbestos cement plant; these have, therefore, been considered as unexposed. The cohort of wives was constructed entirely through official records in the town offices and is both exhaustive and unaffected by recall bias. At the end of follow up (1988) 1669 women were alive, 270 were dead and 25 (1.2%) were untraced. Main mortality analyses were only up to age 79 to reduce the misclassification of causes of death. Expected mortality was based on local rates. Mortality analyses were limited to the period 1965-88 due to the availability of local rates: in that period 210 deaths occurred among women with domestic exposure v 229.1 expected. There were four deaths from pleural tumours (one diagnosed as mesothelioma at necropsis) and six from lung cancer v. 0.5 and 4.0 expected respectively. Two further cases of mesothelioma were diagnosed by histological examination after the end of follow up. None of the three wives with histologically diagnosed mesothelioma had been engaged in industrial activities. Corresponding information for the other three cases could not be traced.

The decline of mortality in England and Wales 1861 to 1964: decomposition by cause of death and component of mortality.

The decline in mortality in England and Wales, between 1861 and 1964 is analysed, using competing-hazard models. These methods provide greater specificity concerning the potential biases due to misdiagnosis, and facilitate the examination of temporal shifts in the age-patterns of mortality. The results provide a number of insights into the historical decline of mortality that were not previously available. First, misclassification of causes of death as ill-defined conditions in historical data is not limited to the diseases of ageing, as has typically been assumed. The degenerative diseases have probably emerged as the most frequently misclassified causes of death only during recent times. Secondly, the unusual age distribution of deaths considered to be characteristic of the influenza epidemic of 1918–19 appeared more than ten years earlier and lasted for more than ten years after the pandemic. This suggests that this age-pattern of deaths was not a characteristic of the Swine influenza virus itself, and...

Accounting for Misclassification in the Cause-of-Death Test for Carcinogenicity

Abstract The cause-of-death test (Peto 1974; Peto et al. 1980) provides a test for carcinogenicity without requiring extreme lethality assumptions. The approach has been criticized, however, because cause-of-death determinations require subjective decisions by pathologists and may be unreliable. Using a missing data formulation, we derive a simple, intuitive modification of the test that allows for misclassification of cause of death. The modified statistic is a function of the misclassification probabilities. A sensitivity analysis may be performed using a plausible range of values for these probabilities or, in certain situations, the misclassification probabilities can be estimated from the data. Both approaches are applied to data from the ED01 experiment.

Accounting for Misclassification in the Cause-of-Death Test for Carcinogenicity

Abstract The cause-of-death test (Peto 1974; Peto et al. 1980) provides a test for carcinogenicity without requiring extreme lethality assumptions. The approach has been criticized, however, because cause-of-death determinations require subjective decisions by pathologists and may be unreliable. Using a missing data formulation, we derive a simple, intuitive modification of the test that allows for misclassification of cause of death. The modified statistic is a function of the misclassification probabilities. A sensitivity analysis may be performed using a plausible range of values for these probabilities or, in certain situations, the misclassification probabilities can be estimated from the data. Both approaches are applied to data from the ED01 experiment.

THE PROBLEM OF MISCLASSIFICATION OF MORTALITY STATISTICS

summaryThe problem of misclassification of cause of death in mortality tables is discussed through the introduction of a classification matrix. This is a stochastic matrix whose off‐diagonal entries are the probabilities of a misclassification. An example is given in Section 2 which shows how a classification matrix can be estimated. An unbiased linear estimate of the unknown true numbers of deaths in different categories is obtained in Section 3 and its asymptotic normality is established. It is shown that the usual classical estimate is biased. Section 4 deals with the special case of a classification matrix whose rows define independent multinomial distributions.

Postponing or Preventing Death? Trends in Infant Survival-Reply

In Reply.— Dujardin et al are correct in noting the limitations of vital records in studying underlying causes of death (determined in Georgia by nosologists, who review all diagnoses listed on death certificates), and their letter raises several issues regarding the question of postponement vs prevention of infant deaths. It is certainly plausible that some infants with chronic illness developing in the perinatal period may succumb late in infancy to infections, sudden infant death syndrome, or other diseases. We attempted to address this question in part by examining underlying causes of death in conjunction with other conditions listed on death certificates; however, multiple diagnoses were available on Georgia's computer records for only the most recent years of our study period. More importantly, we wish to emphasize that the rate of infant death for all causes combined declined for all birth-weight groups. Thus, misclassification of causes of death could not explain

MUMPS SURVEILLANCE IN ENGLAND AND WALES 1962-81

A study of routine data on mumps in England and Wales suggests that its epidemiological features are changing from those of an epidemic disease in young adults and older children to a more endemic disease in younger children. Infection now occurs at an earlier age, at which complications are less frequent and symptomless infection may be more common. The incidence of clinical disease may be falling. The high proportion of registered deaths in the middle aged and elderly may be an artifact due to misclassification of causes of death and to misdiagnosis. These changes lessen the need for routine immunisation. Indeed with low acceptance rates of measles vaccination of 50% at age 15 months, mumps/measles routine vaccination at the same age is contraindicated because similarly low acceptance rates might lead to an increase in the number of cases of mumps in older children and young adults, at which ages complications are more frequent.