Dataset Information

Performance of physician-certified verbal autopsies: multisite validation study using clinical diagnostic gold standards.

ABSTRACT:

Background

Physician review of a verbal autopsy (VA) and completion of a death certificate remains the most widely used approach for VA analysis. This study provides new evidence about the performance of physician-certified verbal autopsy (PCVA) using defined clinical diagnostic criteria as a gold standard for a multisite sample of 12,542 VAs. The study was also designed to analyze issues related to PCVA, such as the impact of a second physician reader on the cause of death assigned, the variation in performance with and without household recall of health care experience (HCE), and the importance of local information for physicians reading VAs.

Methods

The certification was performed by 24 physicians. The assignment of VA was random and blinded. Each VA was certified by one physician. Half of the VAs were reviewed by a different physician with household recall of health care experience included. The completed death certificate was processed for automated ICD-10 coding of the underlying cause of death. PCVA was compared to gold standard cause of death assignment based on strictly defined clinical diagnostic criteria that are part of the Population Health Metrics Research Consortium (PHMRC) gold standard verbal autopsy study.

Results

For individual cause assignment, the overall chance-corrected concordance for PCVA against the gold standard cause of death is less than 50%, with substantial variability by cause and physician. Physicians assign the correct cause around 30% of the time without HCE, and addition of HCE improves performance in adults to 45% and slightly higher in children to 48%. Physicians estimate cause-specific mortality fractions (CSMFs) with considerable error for adults, children, and neonates. Only for neonates for a cause list of six causes with HCE is accuracy above 0.7. In all three age groups, CSMF accuracy improves when household recall of health care experience is available.

Conclusions

Results show that physician coding for cause of death assignment may not be as robust as previously thought. The time and cost required to initially collect the verbal autopsies must be considered in addition to the analysis, as well as the impact of diverting physicians from servicing immediate health needs in a population to review VAs. All of these considerations highlight the importance and urgency of developing better methods to more reliably analyze past and future verbal autopsies to obtain the highest quality mortality data from populations without reliable death certification.

SUBMITTER: Lozano R

PROVIDER: S-EPMC3160925 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundInterVA is a widely disseminated tool for cause of death attribution using information from verbal autopsies. Several studies have attempted to validate the concordance and accuracy of the tool, but the main limitation of these studies is that they compare cause of death as ascertained through hospital record review or hospital discharge diagnosis with the results of InterVA. This study provides a unique opportunity to assess the performance of InterVA compared to physician-certified verbal autopsies (PCVA) and alternative automated methods for analysis.MethodsUsing clinical diagnostic gold standards to select 12,542 verbal autopsy cases, we assessed the performance of InterVA on both an individual and population level and compared the results to PCVA, conducting analyses separately for adults, children, and neonates. Following the recommendation of Murray et al., we randomly varied the cause composition over 500 test datasets to understand the performance of the tool in different settings. We also contrasted InterVA with an alternative Bayesian method, Simplified Symptom Pattern (SSP), to understand the strengths and weaknesses of the tool.ResultsAcross all age groups, InterVA performs worse than PCVA, both on an individual and population level. On an individual level, InterVA achieved a chance-corrected concordance of 24.2% for adults, 24.9% for children, and 6.3% for neonates (excluding free text, considering one cause selection). On a population level, InterVA achieved a cause-specific mortality fraction accuracy of 0.546 for adults, 0.504 for children, and 0.404 for neonates. The comparison to SSP revealed four specific characteristics that lead to superior performance of SSP. Increases in chance-corrected concordance are attained by developing cause-by-cause models (2%), using all items as opposed to only the ones that mapped to InterVA items (7%), assigning probabilities to clusters of symptoms (6%), and using empirical as opposed to expert probabilities (up to 8%).ConclusionsGiven the widespread use of verbal autopsy for understanding the burden of disease and for setting health intervention priorities in areas that lack reliable vital registrations systems, accurate analysis of verbal autopsies is essential. While InterVA is an affordable and available mechanism for assigning causes of death using verbal autopsies, users should be aware of its suboptimal performance relative to other methods.

Project description:BackgroundVerbal autopsy (VA) is used to estimate the causes of death in areas with incomplete vital registration systems. The King and Lu method (KL) for direct estimation of cause-specific mortality fractions (CSMFs) from VA studies is an analysis technique that estimates CSMFs in a population without predicting individual-level cause of death as an intermediate step. In previous studies, KL has shown promise as an alternative to physician-certified verbal autopsy (PCVA). However, it has previously been impossible to validate KL with a large dataset of VAs for which the underlying cause of death is known to meet rigorous clinical diagnostic criteria.MethodsWe applied the KL method to adult, child, and neonatal VA datasets from the Population Health Metrics Research Consortium gold standard verbal autopsy validation study, a multisite sample of 12,542 VAs where gold standard cause of death was established using strict clinical diagnostic criteria. To emulate real-world populations with varying CSMFs, we evaluated the KL estimations for 500 different test datasets of varying cause distribution. We assessed the quality of these estimates in terms of CSMF accuracy as well as linear regression and compared this with the results of PCVA.ResultsKL performance is similar to PCVA in terms of CSMF accuracy, attaining values of 0.669, 0.698, and 0.795 for adult, child, and neonatal age groups, respectively, when health care experience (HCE) items were included. We found that the length of the cause list has a dramatic effect on KL estimation quality, with CSMF accuracy decreasing substantially as the length of the cause list increases. We found that KL is not reliant on HCE the way PCVA is, and without HCE, KL outperforms PCVA for all age groups.ConclusionsLike all computer methods for VA analysis, KL is faster and cheaper than PCVA. Since it is a direct estimation technique, though, it does not produce individual-level predictions. KL estimates are of similar quality to PCVA and slightly better in most cases. Compared to other recently developed methods, however, KL would only be the preferred technique when the cause list is short and individual-level predictions are not needed.

Project description:BackgroundComputer-coded verbal autopsy (CCVA) is a promising alternative to the standard approach of physician-certified verbal autopsy (PCVA), because of its high speed, low cost, and reliability. This study introduces a new CCVA technique and validates its performance using defined clinical diagnostic criteria as a gold standard for a multisite sample of 12,542 verbal autopsies (VAs).MethodsThe Random Forest (RF) Method from machine learning (ML) was adapted to predict cause of death by training random forests to distinguish between each pair of causes, and then combining the results through a novel ranking technique. We assessed quality of the new method at the individual level using chance-corrected concordance and at the population level using cause-specific mortality fraction (CSMF) accuracy as well as linear regression. We also compared the quality of RF to PCVA for all of these metrics. We performed this analysis separately for adult, child, and neonatal VAs. We also assessed the variation in performance with and without household recall of health care experience (HCE).ResultsFor all metrics, for all settings, RF was as good as or better than PCVA, with the exception of a nonsignificantly lower CSMF accuracy for neonates with HCE information. With HCE, the chance-corrected concordance of RF was 3.4 percentage points higher for adults, 3.2 percentage points higher for children, and 1.6 percentage points higher for neonates. The CSMF accuracy was 0.097 higher for adults, 0.097 higher for children, and 0.007 lower for neonates. Without HCE, the chance-corrected concordance of RF was 8.1 percentage points higher than PCVA for adults, 10.2 percentage points higher for children, and 5.9 percentage points higher for neonates. The CSMF accuracy was higher for RF by 0.102 for adults, 0.131 for children, and 0.025 for neonates.ConclusionsWe found that our RF Method outperformed the PCVA method in terms of chance-corrected concordance and CSMF accuracy for adult and child VA with and without HCE and for neonatal VA without HCE. It is also preferable to PCVA in terms of time and cost. Therefore, we recommend it as the technique of choice for analyzing past and current verbal autopsies.

Project description:BackgroundThe most common method for determining cause of death is certification by physicians based either on available medical records, or where such data are not available, through verbal autopsy (VA). The physician-certification approach is costly and inconvenient; however, recent work shows the potential of a computer-based probabilistic model (InterVA) to interpret verbal autopsy data in a more convenient, consistent, and rapid way. In this study we validate separately both physician-certified verbal autopsy (PCVA) and the InterVA probabilistic model against hospital cause of death (HCOD) in adults dying in a district hospital on the coast of Kenya.MethodsBetween March 2007 and June 2010, VA interviews were conducted for 145 adult deaths that occurred at Kilifi District Hospital. The VA data were reviewed by a physician and the cause of death established. A range of indicators (including age, gender, physical signs and symptoms, pregnancy status, medical history, and the circumstances of death) from the VA forms were included in the InterVA for interpretation. Cause-specific mortality fractions (CSMF), Cohen's kappa (κ) statistic, receiver operating characteristic (ROC) curves, sensitivity, specificity, and positive predictive values were applied to compare agreement between PCVA, InterVA, and HCOD.ResultsHCOD, InterVA, and PCVA yielded the same top five underlying causes of adult deaths. The InterVA overestimated tuberculosis as a cause of death compared to the HCOD. On the other hand, PCVA overestimated diabetes. Overall, CSMF for the five major cause groups by the InterVA, PCVA, and HCOD were 70%, 65%, and 60%, respectively. PCVA versus HCOD yielded a higher kappa value (κ = 0.52, 95% confidence interval [CI]: 0.48, 0.54) than the InterVA versus HCOD which yielded a kappa (κ) value of 0.32 (95% CI: 0.30, 0.38). Overall, (κ) agreement across the three methods was 0.41 (95% CI: 0.37, 0.48). The areas under the ROC curves were 0.82 for InterVA and 0.88 for PCVA. The observed sensitivities and specificities across the five major causes of death varied from 43% to 100% and 87% to 99%, respectively, for the InterVA/PCVA against the HCOD.ConclusionBoth the InterVA and PCVA compared well with the HCOD at a population level and determined the top five underlying causes of death in the rural community of Kilifi. We hope that our study, albeit small, provides new and useful data that will stimulate further definitive work on methods of interpreting VA data.

Project description:BACKGROUND:Verbal autopsies with physician assignment of cause of death (COD) are commonly used in settings where medical certification of deaths is uncommon. It remains unanswered if automated algorithms can replace physician assignment. METHODS:We randomized verbal autopsy interviews for deaths in 117 villages in rural India to either physician or automated COD assignment. Twenty-four trained lay (non-medical) surveyors applied the allocated method using a laptop-based electronic system. Two of 25 physicians were allocated randomly to independently code the deaths in the physician assignment arm. Six algorithms (Naïve Bayes Classifier (NBC), King-Lu, InSilicoVA, InSilicoVA-NT, InterVA-4, and SmartVA) coded each death in the automated arm. The primary outcome was concordance with the COD distribution in the standard physician-assigned arm. Four thousand six hundred fifty-one (4651) deaths were allocated to physician (standard), and 4723 to automated arms. RESULTS:The two arms were nearly identical in demographics and key symptom patterns. The average concordances of automated algorithms with the standard were 62%, 56%, and 59% for adult, child, and neonatal deaths, respectively. Automated algorithms showed inconsistent results, even for causes that are relatively easy to identify such as road traffic injuries. Automated algorithms underestimated the number of cancer and suicide deaths in adults and overestimated other injuries in adults and children. Across all ages, average weighted concordance with the standard was 62% (range 79-45%) with the best to worst ranking automated algorithms being InterVA-4, InSilicoVA-NT, InSilicoVA, SmartVA, NBC, and King-Lu. Individual-level sensitivity for causes of adult deaths in the automated arm was low between the algorithms but high between two independent physicians in the physician arm. CONCLUSIONS:While desirable, automated algorithms require further development and rigorous evaluation. Lay reporting of deaths paired with physician COD assignment of verbal autopsies, despite some limitations, remains a practicable method to document the patterns of mortality reliably for unattended deaths. TRIAL REGISTRATION:ClinicalTrials.gov , NCT02810366. Submitted on 11 April 2016.

Project description:BackgroundThe Indian Sample Registration System (SRS) with verbal autopsy methods provides estimations of cause specific mortality for maternal deaths, where the majority of deaths occur at home, unregistered. We aim to examine factors that influence physician agreement and coding choices in assigning causes of death from verbal autopsies.Methodology/principal findingsAmong adult deaths identified in the SRS, pregnancy-related deaths recorded in 2001-2003 were assigned ICD-10 codes by two independent physicians. Inter-rater reliability was estimated using Landis Koch Kappa classification ?0.4--poor to fair agreement; >0.4 ?0.6--moderate agreement; >0.6 ?0.8--substantial agreement; >8--high agreement. We identified factors associated with physician agreement using multivariate logistic regression. A central consensus panel reviewed cases for errors and reclassified as needed based on 2011 ICD-10 coding guidelines. Of 1130 pregnancy-related deaths, 1040 were assigned ICD-10 codes by two physicians. We found substantial agreement regardless of the woman's residence, whether the death was registered, religion, respondent's or deceased's education, age, hospital admission or gestational age. Physician agreement was not influenced by the above variables, with the exception of greater agreement in cases where the respondent did not live with the deceased, or early gestational age at the time of death. A central consensus panel reviewed all cases and recoded 10% of cases due to insufficient use of information in the verbal autopsy by the coding physicians and rationale for this reclassification are discussed.ConclusionIn the absence of complete vital registration and universal healthcare services, physician coded verbal autopsies continues to be heavily relied upon to ascertain pregnancy-related death. From this study, two independent physicians had good inter-rater reliability for assigning pregnancy-related causes of death in a nationally-represented sample, and physician coding does not appear to be heavily influenced by case characteristics or demographics.