Improving Validity of Cause of Death on Death Certificates

Advancing Underlying Cause of Death Inference Through Wide and Deep Model

Introduction

Accurately filling out death certificates is essential for death surveillance. However, manually determining the underlying cause of death is often imprecise. In this study, we investigate the Wide and Deep framework as a method to improve the accuracy and reliability of inferring the underlying cause of death.

Methods

Death report data from national-level cause of death surveillance sites in Fujian Province from 2016 to 2022, involving 403,547 deaths, were analyzed. The Wide and Deep embedded with Convolutional Neural Networks (CNN) was developed. Model performance was assessed using weighted accuracy, weighted precision, weighted recall, and weighted area under the curve (AUC). A comparison was made with XGBoost, CNN, Gated Recurrent Unit (GRU), Transformer, and GRU with Attention.

Results

The Wide and Deep achieved strong performance metrics on the test set: precision of 95.75%, recall of 92.08%, F1 Score of 93.78%, and an AUC of 95.99%. The model also displayed specific F1 Scores for different cause-of-death chain lengths: 97.13% for single causes, 95.08% for double causes, 91.24% for triple causes, and 79.50% for quadruple causes.

Conclusions

The Wide and Deep significantly enhances the ability to determine the root causes of death, providing a valuable tool for improving cause-of-death surveillance quality. Integrating artificial intelligence (AI) in this field is anticipated to streamline death registration and reporting procedures, thereby boosting the precision of public health data.

Incompleteness and misclassification of maternal death recording: a systematic review and meta-analysis

OBJECTIVE

To quantify the extent of incompleteness and misclassification of maternal and pregnancy related deaths, and to identify general and context-specific factors associated with incompleteness and/or misclassification of maternal death data.

METHODS

We conducted a systematic review of incompleteness and/or misclassification of maternal and pregnancy-related deaths. We conducted a narrative synthesis to identify methods used to capture and classify maternal deaths, as well as general and context specific factors affecting the completeness and misclassification of maternal death recording. We conducted a meta-analysis of proportions to obtain estimates of incompleteness and misclassification of maternal death recording, overall and disaggregated by income and surveillance system types.

FINDINGS

Of 2872 title-abstracts identified, 29 were eligible for inclusions in the qualitative synthesis, and 20 in the meta-analysis. Included studies relied principally on record linkage and review for identifying deaths, and on review of medical records and verbal autopsies to correctly classify cause of death. Deaths to women towards the extremes of the reproductive age range, those not classified by a medical examiner or a coroner, and those from minority ethnic groups in their setting were more likely misclassified or unrecorded. In the meta-analysis, we found maternal death recording to be incomplete by 34% (95% CI: 28-48), with 60% sensitivity (95% CI: 31-81.). Overall, we found maternal mortality was under-estimated by 39% (95% CI: 30-48) due to incompleteness and/or misclassification. Reporting of deaths away from the intrapartum, due to indirect causes or occurring at home were less complete than their counterparts. There was substantial between and within group variability across most results.

CONCLUSION

Maternal deaths were under-estimated in almost all contexts, but the extent varied across settings. Countries should aim towards establishing Civil Registration and Vital Statistics systems where they are not instituted. Efforts to improve the completeness and accuracy of maternal cause of death recording, such as Confidential Enquiries into Maternal Deaths, are needed even where CRVS is considered to be well-functioning.

Trimethylamine N-oxide is associated with long-term mortality risk: the multi-ethnic study of atherosclerosis

AIMS

Little is known about associations of trimethylamine N-oxide (TMAO), a novel gut microbiota-generated metabolite of dietary phosphatidylcholine and carnitine, and its changes over time with all-cause and cause-specific mortality in the general population or in different race/ethnicity groups. The study aimed to investigate associations of serially measured plasma TMAO levels and changes in TMAO over time with all-cause and cause-specific mortality in a multi-ethnic community-based cohort.

METHODS AND RESULTS

The study included 6,785 adults from the Multi-Ethnic Study of Atherosclerosis. TMAO was measured at baseline and year 5 using mass spectrometry. Primary outcomes were adjudicated all-cause mortality and cardiovascular disease (CVD) mortality. Secondary outcomes were deaths due to kidney failure, cancer, or dementia obtained from death certificates. Cox proportional hazards models with time-varying TMAO and covariates assessed the associations with adjustment for sociodemographics, lifestyles, diet, metabolic factors, and comorbidities. During a median follow-up of 16.9 years, 1704 participants died and 411 from CVD. Higher TMAO levels associated with higher risk of all-cause mortality [hazard ratio (HR): 1.12, 95% confidence interval (CI): 1.08-1.17], CVD mortality (HR: 1.09, 95% CI: 1.00-1.09), and death due to kidney failure (HR: 1.44, 95% CI: 1.25-1.66) per inter-quintile range, but not deaths due to cancer or dementia. Annualized changes in TMAO levels associated with higher risk of all-cause mortality (HR: 1.10, 95% CI: 1.05-1.14) and death due to kidney failure (HR: 1.54, 95% CI: 1.26-1.89) but not other deaths.

CONCLUSION

Plasma TMAO levels were positively associated with mortality, especially deaths due to cardiovascular and renal disease, in a multi-ethnic US cohort.

Systematic Review of Advanced AI Methods for Improving Healthcare Data Quality in Post COVID-19 Era

At the beginning of the COVID-19 pandemic, there was significant hype about the potential impact of artificial intelligence (AI) tools in combatting COVID-19 on diagnosis, prognosis, or surveillance. However, AI tools have not yet been widely successful. One of the key reason is the COVID-19 pandemic has demanded faster real-time development of AI-driven clinical and health support tools, including rapid data collection, algorithm development, validation, and deployment. However, there was not enough time for proper data quality control. Learning from the hard lessons in COVID-19, we summarize the important health data quality challenges during COVID-19 pandemic such as lack of data standardization, missing data, tabulation errors, and noise and artifact. Then we conduct a systematic investigation of computational methods that address these issues, including emerging novel advanced AI data quality control methods that achieve better data quality outcomes and, in some cases, simplify or automate the data cleaning process. We hope this article can assist healthcare community to improve health data quality going forward with novel AI development.

An Automated Method of Causal Inference of the Underlying Cause of Death of Citizens

It is of great significance to correctly infer the underlying cause of death for citizens, especially under the current worldwide situation. The medical resources of all countries are overwhelmed under the impact of coronavirus disease 2019 (COVID-19) and countries need to allocate limited resources to the most suitable place. Traditionally, the cause-of-death inference relies on manual methods, which require a large resource cost and are not so efficient. To address the challenges, in this work, we present a mixed inference method named Sink-CF. The Sink-CF algorithm is based on confidence measurement and is used to automatically infer the underlying cause of death of citizens. The method proposed in this paper combines a mathematical statistics method and a collaborative filtering and analysis algorithm in machine learning. Thus, our method can not only effectively achieve a certain accuracy, but also does not rely on a large quantity of manually labeled data to continuously optimize the model, which can save computer computing power and time, and has the characteristics of being simple, easy and efficient. The experimental results show that our method generates a reasonable precision (93.82%) and recall (90.11%) and outperforms other state-of-the-art machine learning algorithms.

Proposing Causal Sequence of Death by Neural Machine Translation in Public Health Informatics

Each year there are nearly 57 million deaths worldwide, with over 2.7 million in the United States. Timely, accurate and complete death reporting is critical for public health, especially during the COVID-19 pandemic, as institutions and government agencies rely on death reports to formulate responses to communicable diseases. Unfortunately, determining the causes of death is challenging even for experienced physicians. The novel coronavirus and its variants may further complicate the task, as physicians and experts are still investigating COVID-related complications. To assist physicians in accurately reporting causes of death, an advanced Artificial Intelligence (AI) approach is presented to determine a chronically ordered sequence of conditions that lead to death (named as the causal sequence of death), based on decedent's last hospital discharge record. The key design is to learn the causal relationship among clinical codes and to identify death-related conditions. There exist three challenges: different clinical coding systems, medical domain knowledge constraint, and data interoperability. First, we apply neural machine translation models with various attention mechanisms to generate sequences of causes of death. We use the BLEU (BiLingual Evaluation Understudy) score with three accuracy metrics to evaluate the quality of generated sequences. Second, we incorporate expert-verified medical domain knowledge as constraints when generating the causal sequences of death. Lastly, we develop a Fast Healthcare Interoperability Resources (FHIR) interface that demonstrates the usability of this work in clinical practice. Our results match the state-of-art reporting and can assist physicians and experts in public health crisis such as the COVID-19 pandemic.

Redistribution of garbage codes to underlying causes of death: a systematic analysis on Italy and a comparison with most populous Western European countries based on the Global Burden of Disease Study 2019

Background

The proportion of reported causes of death (CoDs) that are not underlying causes can be relevant even in high-income countries and seriously affect health planning. The Global Burden of Disease (GBD) study identifies these ‘garbage codes’ (GCs) and redistributes them to underlying causes using evidence-based algorithms. Planners relying on vital registration data will find discrepancies with GBD estimates. We analyse these discrepancies, through the analysis of GCs and their redistribution.

Methods

We explored the case of Italy, at national and regional level, and compared it to nine other Western European countries with similar population sizes. We analysed differences between official data and GBD 2019 estimates, for the period 1990–2017 for which we had vital registration data for most select countries.

Results

In Italy, in 2017, 33 000 deaths were attributed to unspecified type of stroke and 15 000 to unspecified type of diabetes, these making a fourth of the overall garbage. Significant heterogeneity exists on the overall proportion of GCs, type (unspecified or impossible underlying causes), and size of specific GCs among regions in Italy, and among the select countries. We found no pattern between level of garbage and relevance of specific GCs. Even locations performing below average show interesting lower levels for certain GCs if compared to better performing countries.

Conclusions

This systematic analysis suggests the heterogeneity in GC levels and causes, paired with a more detailed analysis of local practices, strengths and weaknesses, could be a positive element in a strategy for the reduction of GCs in Italy.

Not discussed: Inequalities in narrative text data for suicide deaths in the National Violent Death Reporting System

Background

The rate of suicide in the US has increased substantially in the past two decades, and new insights are needed to support prevention efforts. The National Violent Death Reporting System (NVDRS), the nation’s most comprehensive registry of suicide mortality, has qualitative text narratives that describe salient circumstances of these deaths. These texts have great potential for providing novel insights about suicide risk but may be subject to information bias.

Objective

To examine the relationship between decedent characteristics and the presence and length of NVDRS text narratives (separately for coroner/medical examiner (C/ME) and law enforcement (LE) reports) among 233,108 suicide and undetermined deaths from 2003–2017.

Methods

Generalized estimating equations (GEE) logistic and quasi-Poisson modeling was used to examine variation in the narratives (proportion of missing texts and character length of the non-missing texts, respectively) as a function of decedent age, sex, race/ethnicity, education, marital status, military history, and homeless status. Models adjusted for site, year, location of death, and autopsy status.

Results

The frequency of missing narratives was higher for LE vs. C/ME texts (19.8% vs. 5.2%). Decedent characteristics were not consistently associated with missing text across the two types of narratives (i.e., Black decedents were more likely to be missing the LE narrative but less likely to be missing the C/ME narrative relative to non-Hispanic whites). Conditional on having a narrative, C/ME were significantly longer than LE (822.44 vs. 780.68 characters). Decedents who were older, male, had less education and some racial/ethnic minority groups had shorter narratives (both C/ME and LE) than younger, female, more educated, and non-Hispanic white decedents.

Conclusion

Decedent characteristics are significantly related to the presence and length of narrative texts for suicide and undetermined deaths in the NVDRS. Findings can inform future research using these data to identify novel determinants of suicide mortality.

An Automatic Approach Designed for Inference of the Underlying Cause-of-Death of Citizens

It is very important to have a comprehensive understanding of the health status of a country’s population, which helps to develop corresponding public health policies. Correct inference of the underlying cause-of-death for citizens is essential to achieve a comprehensive understanding of the health status of a country’s population. Traditionally, this relies mainly on manual methods based on medical staff’s experiences, which require a lot of resources and is not very efficient. In this work, we present our efforts to construct an automatic method to perform inferences of the underlying causes-of-death for citizens. A sink algorithm is introduced, which could perform automatic inference of the underlying cause-of-death for citizens. The results show that our sink algorithm could generate a reasonable output and outperforms other stat-of-the-art algorithms. We believe it would be very useful to greatly enhance the efficiency of correct inferences of the underlying causes-of-death for citizens.

Improving the Use of Mortality Data in Public Health: A Comparison of Garbage Code Redistribution Models

Objectives. To describe and compare 3 garbage code (GC) redistribution models: naïve Bayes classifier (NB), coarsened exact matching (CEM), and multinomial logistic regression (MLR).Methods. We analyzed Taiwan Vital Registration data (2008-2016) using a 2-step approach. First, we used non-GC death records to evaluate 3 different prediction models (NB, CEM, and MLR), incorporating individual-level information on multiple causes of death (MCDs) and demographic characteristics. Second, we applied the best-performing model to GC death records to predict the underlying causes of death. We conducted additional simulation analyses for evaluating the predictive performance of models.Results. When we did not account for MCDs, all 3 models presented high average misclassification rates in GC assignment (NB, 81%; CEM, 86%; MLR, 81%). In the presence of MCD information, NB and MLR exhibited significant improvement in assignment accuracy (19% and 17% misclassification rate, respectively). Furthermore, CEM without a variable selection procedure resulted in a substantially higher misclassification rate (40%).Conclusions. Comparing potential GC redistribution approaches provides guidance for obtaining better estimates of cause-of-death distribution and highlights the significance of MCD information for vital registration system reform.