A review of data-derived methods for assigning causes of death from verbal autopsy data

BACKGROUND: Verbal autopsy (VA) is an indirect method for estimating cause-specific mortality. In most previous studies, cause of death has been assigned from verbal autopsy data using expert algorithms or by physician review. Both of these methods may have poor validity. In addition, physician review is time consuming and has to be carried out by doctors. A range of methods exist for deriving classification rules from data. Such rules are quick and simple to apply and in many situations perform as well as experts. METHODS: This paper has two aims. First, it considers the advantages and disadvantages of the three main methods for deriving classification rules empirically; (a) linear and other discriminant techniques, (b) probability density estimation and (c) decision trees and rule-based methods. Second, it reviews the factors which need to be taken into account when choosing a classification method for assigning cause of death from VA data. RESULTS: Four main factors influence the choice of classification method: (a) the purpose for which a classifier is being developed, (b) the number of validated causes of death assigned to each case, (c) the characteristics of the VA data and (d) the need for a classifier to be comprehensible. When the objective is to estimate mortality from a single cause of death, logistic regression should be used. When the objective is to determine patterns of mortality, the choice of method will depend on the above factors in ways which are elaborated in the paper. CONCLUSION: Choice of classification method for assigning cause of death needs to be considered when designing a VA validation study. Comparison of the performance of classifiers derived using different methods requires a large VA dataset, which is not currently available.     

The validity of verbal reports in children's subtraction.

To assess the validity of verbal reports in children's subtraction, students in Grades 1, 3, and 5 were asked to solve a set of simple subtraction problems and were placed in the no-report, retrospective-report, or concurrent-report conditions. Two aspects of verbal report validity were assessed: reactivity, or whether providing a verbal report alters subsequent task performance, and veridicality, or whether the verbal reports are accurate reflections of solution strategies. Students in all grades and in both the retrospective-report and concurrent-report conditions were able to provide veridical strategy reports, and the instruction to verbally report had few effects on task performance. Informal findings indicate that students had less difficulty reporting retrospectively than concurrently. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On the validity of validity scales: The importance of defensive responding in the prediction of institutional misconduct.

This study examined the effects of defensive responding on the prediction of institutional misconduct among male inmates (N = 349) who completed the Personality Assessment Inventory (L. C. Morey, 1991). Hierarchical logistic regression analyses demonstrated significant main effects for the Antisocial Features (ANT) scale as well as main effects for the Positive Impression Management (PIM) scale in some instances. Significant ANT × PIM interactions also were evident, particularly when examining recommended cut scores on these scales. These results demonstrate the predictive validity of ANT with prisoners and reinforce the importance of context when considering whether validity scales are in fact valid. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Testing the validity of an instrument for assessing family of origin history

Family of origin history information is often gathered for research and clinical purposes in psychology, but there has been relatively little examination of the validity of the data obtained. This report describes several tests of the validity of the recently developed Family Background Questionnaire, a relatively behaviorally specific instrument designed to comprehensively assess characteristics of one's family of origin. Using a sample of 678 nonclinical and clinical participants, it was found that a history of incest, parental chemical dependency, clinical status, socioeconomic status, and birth order all predicted Family Background Questionnaire scores as hypothesized and that 38% of the variance in Total Scale scores was accounted for by these variables.     

Predictive validity of an implicit association test for assessing anxiety.

The Implicit Association Test (TAT) was adapted to measure anxiety by assessing associations of self (vs. other) with anxiety-related (vs. calmness-related) words. Study 1 showed that the TAT-Anxiety exhibited good internal consistency and adequate stability. Study 2 revealed that the IAT-Anxiety was unaffected by a faking instruction. Study 3 examined the predictive validity of implicit and explicit measures and showed that the TAT-Anxiety was related to changes in experimenter-rated anxiety and performance decrements after failure. Study 4 found that several behavioral indicators of anxiety during a stressful speech were predicted by the TAT. Taken together, these studies show that the TAT-Anxiety is a reliable measure that is able to predict criterion variables above questionnaire measures of anxiety and social desirability. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reliability and validity of scales for assessing behavior in epilepsy.

Behavior Rating Scales for Epilepsy (BRSE), developed to assess 8 characteristics that have been attributed to patients with epilepsy, were rated from a semistructured interview with 79 outpatients (aged 19–69 yrs) with epilepsy and from an independent interview with a friend or relative of each patient. Patients and observers also completed the Bear and Fedio Inventory (BFI) for assessing behavior change in epilepsy. Interrater reliability estimates exceeded .75 for 5 BRSE scales. Convergent and discriminant validity were demonstrated for all but 2 BRSE scales when comparing patient with observer-based rating. BRSE results showed correspondence with BFI-Observer scores for most scales but not with BFI-Patient scales. BFI-Observer scores also showed little association with BFI-Patient scores. The BRSE shows promise as a time-efficient, psychometrically sound assessment instrument for further research on behavioral aspects of epilepsy. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The reliability and validity of the family history method for assessing pathological gambling and gambling involvement.

The family history (FH) method, which involves the use of an informant to gather information about one or more family members, has been used in a number of previous gambling studies. However, no evaluation of the reliability and validity has been conducted on the use of the FH method for assessing pathological gambling (PG) and gambling involvement. The current study examined the test-retest and inter-rater reliability and the validity of the FH method for assessing PG and gambling involvement among a large community-based sample of adult twins (N = 4,764) reporting on their parents, co-twins, and spouses. The test-retest and inter-rater reliabilities of the FH reports of PG were high. Validity of the FH reports of PG was low, primarily because of substantial underestimation of pathology (low sensitivity). The test-retest and inter-rater reliabilities of the FH reports of gambling involvement (ever gambled, ever gambled monthly, and ever gambled weekly) were moderate and the sensitivities were quite high. The results of this study support the use of the FH method for studies of PG and gambling involvement. A number of potential explanations for the low sensitivity of FH reports of PG are elaborated. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The main causes of death in Spain, 1992; comment

BACKGROUND: To study the mortality from the leading causes of death in Spain in 1992 and trends since 1980. POPULATION AND METHOD: The number of deaths was obtained from mortality statistics. We included the 12 causes with the highest mortality rates in 1992 and calculated for each cause of death the age adjusted mortality rates for each year in the study period, the percent change from 1990 to 1992 and from 1980 to 1992, and the adjusted ratio of rates between men and women in 1992. RESULTS: The leading causes of death in 1992 were malignant neoplasms, with 24.3% of deaths and a mortality rate of 205.6 per 100,000 population; diseases of the heart, with 22.6% and a rate of 191.8 per 100,000; and cerebrovascular disease, with 12.7% and a rate of 107.6 per 100,000 population. Between 1980 and 1992 the adjusted mortality rate increased for four causes of death: malignant neoplasms; chronic obstructive pulmonary disease and similar diseases; nephritis, nephrotic syndrome and nephrosis; and suicide. From 1990 to 1992, the adjusted mortality rate declined for all other causes of death. From 1990 to 1992, the adjusted mortality rate declined for all causes of death except for malignant neoplasms and human immunodeficiency virus (HIV) infection, which rose 0.4% and 69%, respectively. The adjusted mortality rate was higher in men than in women for all causes of death except for diabetes mellitus and atherosclerosis. CONCLUSIONS: Except for malignant neoplasms and HIV infection, mortality from all other leading causes of death declined in 1992 with respect to 1990, independently of the trend experienced by each cause of death in the eighties.     

Choice reaction-time study assessing the verbal behavior of the minor hemisphere in normal adult humans.

Conducted 2 choice RT experiments in which a total of 12 right-handed college students scanned a memory set of 1 or 6 letters for the presence of a test letter. The memory set was presented binaurally and the test letter visually to either the right or left visual field. S indicated with his left hand whether or not the test letter was contained in the memory set. Results show the RT was shorter for the left visual field test letter when the set contained only 1 letter and for the right visual field test letter when the set contained 6 letters. The possibility that Ss matched the set and test letters visually when the set contained only 1 letter and verbally when the set contained 6 letters is discussed. Results suggest that the minor hemisphere displays little, if any, verbal abilities. (31 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Clarifying problems and offering solutions for correlated error when assessing the validity of selected-subtest short forms.

The correlation between a short-form (SF) test and its full-scale (FS) counterpart is a mainstay in the evaluation of SF validity. However, in correcting for overlapping error variance in this measure, investigators have overattenuated the validity coefficient through an intuitive misapplication of P. Levy's (1967) formula. The authors of the present article clarify that such corrections should be based on subtest-level versus FS-level data. Additionally, the authors propose a simple, modified equation incorporating FS-level scores that provides liberal and conservative validity measures for comparison across estimation methods, and they demonstrate its use in both a normative (N = 2,450) and clinical psychiatric (N = 216) sample. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On the causes of collapse and sudden death by Avicenna

Avicenna's views on the causes of collapse and sudden death are presented from a literal translation of the mediaeval Arabic text. Medical knowledge based on observation mixed with pure abstract reasoning forms the essence of Avicenna's medical writings.     

Evaluation of the validity of the MAND in assessing motor impairment in young children.

Objective: The identification of children with motor impairment is critical so that remedial intervention can be offered. This study examined the concurrent validity of the McCarron Assessment of Neuromuscular Development (MAND; McCarron, 1997), using the Movement Assessment Battery for Children (MABC; Henderson & Sugden, 1992) as the criterion measure. Method: One hundred eighteen children, 58 girls and 60 boys, 4, 5, and 6 years of age, formed the sample. Children were tested on both the MAND and MABC, which were then compared on their level of agreement in identifying children as having motor impairment. Results: Inconsistent support for the MAND’s discrimination accuracy was evidenced, and differences were found between the MAND and the MABC in the constructs defined for their development. With significant differences evident between the MAND and MABC, the MAND’s cross-loadings on the MABC were plotted against the canonical loadings on the MABC. Conclusions: A considerable degree of independence was found between the MAND and MABC in the classification of children with and without motor impairment. The MABC appeared to have broader variation, and hence greater versatility, than the MAND. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The effect of different sensitivity, specificity and cause-specific mortality fractions on the estimation of differences in cause-specific mortality rates in children from studies using verbal autopsies

BACKGROUND: Verbal autopsies (VA) are increasingly being used in developing countries to determine causes of death, but little attention is generally given to the misclassification effects of the VA. This paper considers the effect of misclassification on the estimation of differences in cause-specific mortality rates between two populations. METHODS: The bias in the percentage difference in cause-specific mortality between two populations has been explored under two different models: i) assuming that mortality from all other causes does not differ between the two populations; ii) allowing for a difference in mortality from all other causes. The bias is described in terms of the sensitivity and specificity of the VA diagnosis and the proportion of mortality due to the cause of interest. Methods for adjustment of sample size and adjusting the estimate of effect are also explored. RESULTS: The results are illustrated for a range of plausible values for these parameters. The bias is more extreme as both sensitivity and specificity fall, and is particularly affected even by a small loss of specificity. The bias also increases as the proportion of all deaths due to the cause of interest decreases, and is affected by the size of the true change in mortality due to the cause of interest relative to the change in mortality from other causes. Calculations from existing data suggest prohibitively large sample sizes may often be required to detect important differences in cause-specific mortality rates in studies using existing VA. CONCLUSIONS: Highly specific VA tools are needed before observed differences in cause-specific mortality can be interpreted. Loss of power due to misclassification may obscure real differences in cause-specific mortality.     

Ecological validity of role plays for assessing heterosocial anxiety and skill of male college students.

Examined the ecological validity of role-played dating interactions. 45 male undergraduates, categorized into low-, medium-, and high-frequency dating groups (Social Activity Questionnaire), participated in 4 heterosexual social (heterosocial) situations in a laboratory. Two situations were role plays simulating dating interactions, and 2 were more naturalistic waiting-period interactions. Controls for confederate familiarity and role-play order were used. Judges rated Ss as more skillful in role plays than waiting periods and rated low-frequency daters as more anxious than high-frequency daters. High-frequency daters rated themselves as more skillful than did low-frequency daters. Ss rated the waiting-period interactions as more like "real life" and their behavior in waiting periods as more representative of everyday heterosocial interactions. However ANOVAs did not result in major differences between role plays and waiting periods, and correlational analyses indicated that the relative ranking of Ss was similar in role plays and waiting periods. Implications for the use of role-play assessments in the heterosocial skill area are discussed. (21 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Requests for physician-assisted death: Guidelines for assessing mental capacity and impaired judgment.

This article fills a need in the theoretical and clinical literature by presenting a set of guidelines for assessing mental capacity and impaired judgment when a person makes a request to have physician-assisted death. Because Oregon is the only state to have a law allowing physician-assisted death, the guidelines are placed within the context of the Oregon Death with Dignity Act. A comprehensive overview of the legal and clinical issues associated with the assessment of capacity precedes the actual guidelines. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The effect of negative verbal cues upon verbal behavior.

Research which investigated the capacity to modify behavior, starting from the framework of Skinner's operant conditioning paradigm and utilizing verbal behavior, was found to have primarily explored the effect of positive secondary reinforcement. The present investigation explored the effect of a negative verbal cue presented under different schedules of reinforcement upon verbal behavior. Ss (male VA psychiatric patients) responded to projective techniques. The E varied the frequency of negative verbal reinforcement defined as "unh unh." The hypothesis that such a negative reinforcement would depress verbal behavior was substantiated, with a periodic schedule of reinforcement being more effective. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Usefulness of the codification of multiple causes of death in mortality statistics

BACKGROUND: The codification of multiple causes of death began in the US in 1917 and systematic publication of this data started in 1984. In Spain this began in 1988, and the data from this year have been taken as the basis for investigation. They have also been studied for regional differences. METHODS: A representative sample (595) of Spanish Standard Death Certificates (DC) was collected in Asturias for the year 1988. All were coded according to the International Classification of Diseases and a separate coding was made for each nosological entity included in the certificate (coding of multiple causes). The median, mode and the multiple cause/underlying cause ratio were also calculated. RESULTS: More than 80% of the certificates studied contained more than one cause of death. Chronic diseases are those which are accompanied by a greater number of causes and acute diseases those which appear alone. The highest ratios appear for diseases which are ill defined and also in those which are chronic. CONCLUSIONS: Our data show that information is lost in the production of the statistics of mortality and there are repercussions for the usefulness of these statistics.