Recognising inaccuracies in Australian suicide and 'hidden suicide' data.

OBJECTIVE: To examine the effects of revision of Australian mortality statistics every year since 2007 on numbers and rates of suicide and 'hidden suicide'. METHOD: Nine months after the end of each year, the Australian Bureau of Statistics releases preliminary statistics concerning deaths registered in that year, together with revised and finalised data regarding previous years. Numbers and rates of suicide and of deaths coded to selected categories of accidental, undetermined and unknown cause deaths were tabled. RESULTS: Upward revision of suicide and accidental drug poisoning death numbers, three years after first release, show that true rates are substantially higher than initially released data suggested. Concomitant downward revision of rates of undetermined and unknown cause deaths supports evidence that at first release some suicides are coded to these categories. CONCLUSIONS: Australia's finalised suicide data are likely to be more accurate than equivalent data from nations that do not revise mortality data. More comprehensive investigation (including verbal or psychological autopsy) in doubtful cases in Australia and elsewhere would probably lead to reported suicide rates being higher.

[The accuracy of the causes of death with judicial intervention in Catalonia in the period 2015-2020. Effect of declaration through IML-Web application]. / La precisión de las causas de defunción con intervención judicial en Cataluña en el periodo 2015-2020. Efecto de la declaración a través de la aplicación IML-Web.

OBJECTIVE: On January first, 2020, the Institutes of Legal Medicine and Forensic Sciences (IMLCF) began to inform the causes of death directly to the National Statistics Institute (INE) through a web application (IML-Web). The objective of this study was to evaluate the impact of the implementation of this application on the quality of the data collected. METHODS: A descriptive study using deaths data with judicial intervention that occurred in Catalonia was carried out. The data of the period 2015-2018 and 2019 was compared with 2020. The percentages, with confidence intervals, of the causes of death that were not specific, according to different classifications, were calculated on the total of cases by period and territory. RESULTS: The total percentage of non-specific deaths had decreased, not significantly, by 1.6 points between the period 2015-2018 and 2020. The same indicator between 2019 and 2020 had decreased by 13.4 points. The percentage of non-specific deaths from external causes showed significant drops between both periods and 2020. In general, the indicators displayed territorial differences. CONCLUSIONS: The roll-out of the IML-Web implies, compared to 2019, an improvement in the quality of the data. On the other hand, compared to the period 2015-2018, the data show a similar level of quality. Generally, it is assessed that the information provided by IMLCF of Catalonia through the IML-Web is accurate, but still has room for improvement.

OBJECTIVE: A partir del 1 enero de 2020, los Institutos de Medicina Legal y Ciencias Forenses (IMLCF) empezaron a declarar las causas de muerte directamente al Instituto Nacional de Estadística (INE) mediante una aplicación web (IML-Web). El objetivo de este trabajo fue evaluar el impacto de la implementación de esta aplicación en la calidad de los datos recogidos. METHODS: Se realizó un estudio descriptivo utilizando datos de las defunciones con intervención judicial ocurridas en Cataluña. Se comparó la información del período 2015-2018 y de 2019 con la de 2020. Se calcularon los porcentajes, con intervalo de confianza, de las causas de defunción poco específicas, según diferentes clasificaciones, sobre el total de causas por período y división judicial. RESULTS: El porcentaje total de causas de defunción poco específicas se redujo, de forma no significativa, 1,6 puntos entre el período 2015-2018 y el año 2020. El mismo indicador entre el año 2019 y 2020 se redujo 13,4 puntos. El porcentaje de defunciones poco específicas de causas externas mostró reducciones estadísticamente significativas entre ambos períodos. En general los indicadores mostraron diferencias territoriales. CONCLUSIONS: La implementación del IML-Web en el año 2020 supone, en comparación con 2019, una mejora en la calidad de la información notificada. En cambio, si se compara con el período 2015-2018, los datos muestran una calidad similar. A nivel general se valora que la información proporcionada por el IMLCF de Cataluña a través del IML-Web es precisa, pero todavía tiene margen de mejora.

La precisión de las causas de defunción con intervención judicial en Cataluña en el periodo 2015-2020. Efecto de la declaración a través de la aplicación IML-Web / The accuracy of the causes of death with judicial intervention in Catalonia in the period 2015-2020. Effect of declaration through IML-Web application

Fundamentos: a partir del 1 enero de 2020, los institutos de medicina legal y ciencias forenses (imlcf) empezaron a declarar las causas de muerte directamente al instituto nacional de estadística (ine) mediante una aplicación web (iml-web). El objetivo de este trabajo fue evaluar el impacto de la implementación de esta aplicación en la calidad de los datos recogidos. Métodos: se realizó un estudio descriptivo utilizando datos de las defunciones con intervención judicial ocurridas en cataluña. Se comparó la información del período 2015-2018 y de 2019 con la de 2020. Se calcularon los porcentajes, con intervalo de confianza, de las causas de defunción poco específicas, según diferentes clasificaciones, sobre el total de causas por período y división judicial. Resultados: el porcentaje total de causas de defunción poco específicas se redujo, de forma no significativa, 1,6 puntos entre el período 2015-2018 y el año 2020. El mismo indicador entre el año 2019 y 2020 se redujo 13,4 puntos. El porcentaje de defunciones poco específicas de causas externas mostró reducciones estadísticamente significativas entre ambos períodos. En general los indicadores mostraron diferencias territoriales. Conclusiones: la implementación del iml-web en el año 2020 supone, en comparación con 2019, una mejora en la calidad de la información notificada. En cambio, si se compara con el período 2015-2018, los datos muestran una calidad similar. A nivel general se valora que la información proporcionada por el imlcf de cataluña a través del iml-web es precisa, pero todavía tiene margen de mejora.(AU)

Background: on january first, 2020, the institutes of legal medicine and forensic sciences (imlcf) began to inform the causes of death directly to the national statistics institute (ine) through a web application (iml-web). The objective of this study was to evaluate the impact of the implementation of this application on the quality of the data collected.methods: a descriptive study using deaths data with judicial intervention that occurred in catalonia was carried out. The data of the period 2015-2018 and 2019 was compared with 2020. The percentages, with confidence intervals, of the causes of death that were not specific, according to different classifications, were calculated on the total of cases by period and territory.results: the total percentage of non-specific deaths had decreased, not significantly, by 1.6 points between the period 2015-2018 and 2020. The same indicator between 2019 and 2020 had decreased by 13.4 points. The percentage of non-specific deaths from external causes showed significant drops between both periods and 2020. In general, the indicators displayed territorial differences.conclusions: the roll-out of the iml-web implies, compared to 2019, an improvement in the quality of the data. On the other hand, compared to the period 2015-2018, the data show a similar level of quality. Generally, it is assessed that the information provided by imlcf of catalonia through the iml-web is accurate, but still has room for improvement.(AU)

Beyond the virus: Examining the impact of COVID-19 on violent fatalities.

INTRODUCTION: The COVID-19 pandemic has had a significant impact on various aspects of society, including crime rates. In Serbia, it is important to examine the changes in violent deaths before and during the pandemic to inform public health and safety policies. MATERIALS AND METHODS: We conducted a retrospective, epidemiological, cross-sectional analytical study of medico-legal autopsies in the Department of Forensic Medicine and Toxicology at the University Clinical Center of Kragujevac. Our study sample comprises all forensic autopsy cases examined from January 2017 to December 2019 (151 cases), labeled as "Before," and from January 2020 to December 2022 (192 cases), labeled as "During" the pandemic period. Natural deaths, skeletal remains, and undetermined cases were excluded from our sample. RESULTS: The data show an increase in the total number of incidents reported from 152 in the "Before" period to 191 in the "During" period. The proportion of incidents involving males remained relatively stable at around 70%, while the proportion of incidents involving females increased. There was no statistically significant change in the proportion of incidents classified as accidental, while the proportion of incidents classified as homicide and suicide increased. The results show a statistically significant association between gender and incident type for both the "Before" and "During" periods. Deaths due to domestic violence have increased by 22.2% during the pandemic, which is cause for concern. In terms of demographic characteristics, males and younger individuals were more likely to be victims of violent deaths both before and during the pandemic. CONCLUSIONS: The COVID-19 pandemic had a significant impact on violent deaths in the Sumadija region (Central Serbia), with an overall increase in the number of violent deaths and a major impact on deaths due to domestic violence. Policies to address domestic violence should be prioritized during the pandemic and beyond, and strategies should be developed to mitigate the effects of future pandemics or lockdowns.

Reporting system on mortality statistics in Lao People's Democratic Republic.

The system to collect information on mortality statistics in Lao PDR is not well established, accurate and timely death information is therefore not available. This article reports the system and process to make the mortality statistical data of Lao PDR. The country has a paper-based resident registration system, using a death notification document, a death certificate, and a family census book. The death notification document is important as it provides the cause of death, which is issued from a health facility and the village office. In the event of a death occurring at home, the family representative needs to report to the village office verbally to obtain a death notification document. On the other hand, if the death occurred in a medical facility, a death notification document from a health facility is provided. The family representative should bring the death notification document to the district Home Affairs office to register the death and obtain a death certificate. After that, the family representative needs to bring the death certificate to the district Public Security office for an amendment in the family census book. ICD-10 is under development regarding death notification from health facilities under the Ministry of Health. However, it is unclear how death notification from village offices can adopt ICD-10 as the majority of deaths occur outside health facilities. A comprehensive and integrated mortality reporting system is necessary in order to create a holistic health policy and welfare for the country.

Análisis estadístico implicativo para identificar factores pronósticos de mortalidad por linfomas en niños y adolescentes

Fundamento: el análisis estadístico implicativo surgió en los años 80 para resolver problemas de la didáctica de las matemáticas. Recientemente se fundamentó su empleo en las Ciencias Médicas para identificar factores de riesgo y pronósticos. Objetivo: evaluar la utilidad del análisis estadístico implicativo en la identificación de los factores pronósticos que más inciden en la mortalidad por linfomas en niños y adolescentes. Método: se realizó un estudio de casos y controles en niños y adolescentes con diagnóstico de linfoma Hodgkin y no Hodgkin atendidos en el Hospital Docente Pediátrico Sur Dr. Antonio María Béguez César de Santiago de Cuba en el período de enero 2008 a enero 2021. Se analizó como variable dependiente el estado del paciente fallecido o vivo al momento del estudio y como covariables se tomaron: el estadio de mal pronóstico, la presencia de síntomas B, el subtipo histológico, la presencia de tres o más sitios extraganglionares, la metástasis, edad y presencia de masa tumoral. Se aplicaron dos técnicas estadísticas, la regresión logística binaria y el análisis estadístico implicativo. Resultados en los casos fue más frecuente el linfoma no Hodgkin mientras que en los controles predominó el Hodgkin. Ambas técnicas reconocieron el subtipo histológico y la afectación extraganglionar como factores pronósticos desfavorables. El análisis estadístico implicativo reconoció además el estadio y la presencia de metástasis. Conclusión: el análisis estadístico implicativo es una técnica que complementa la regresión logística binaria en la identificación de factores pronósticos, lo que permite mejor comprensión de la causalidad.

Background: the implicative statistical analysis arose in the 80s to solve problems in the didactics of mathematics. Its use in the Medical Sciences to identify risk factors and prognoses was recently founded. Objective: to evaluate the usefulness of the implicative statistical analysis in the identification of the prognostic factors that most affect mortality from lymphomas in children and adolescents. Method: a case-control study was carried out in children and adolescents diagnosed with Hodgkin and non-Hodgkin lymphoma treated at the Dr. Antonio María Béguez César Sur Pediatric Teaching Hospital in Santiago de Cuba from January 2008 to January 2021. The state of the deceased or alive patient at the time of the study was analyzed as the dependent variable and the following were taken as covariates: poor prognosis stage, presence of B symptoms, histological subtype, presence of three or more extranodal sites, metastasis, age and presence of tumor mass. Two statistical techniques were applied: binary logistic regression and implicative statistical analysis. Results: non-Hodgkin's lymphoma was more frequent in the cases, while Hodgkin's lymphoma predominated in the controls. Both techniques recognized the histological subtype and extranodal involvement as unfavorable prognostic factors. The implicative statistical analysis also recognized the stage and the presence of metastases. Conclusion: the implicative statistical analysis is a technique that complements the binary logistic regression in the identification of prognostic factors, which allows a better understanding of causality.

Real-Time Classification of Causes of Death Using AI: Sensitivity Analysis.

BACKGROUND: In 2021, the European Union reported >270,000 excess deaths, including >16,000 in Portugal. The Portuguese Directorate-General of Health developed a deep neural network, AUTOCOD, which determines the primary causes of death by analyzing the free text of physicians' death certificates (DCs). Although AUTOCOD's performance has been established, it remains unclear whether its performance remains consistent over time, particularly during periods of excess mortality. OBJECTIVE: This study aims to assess the sensitivity and other performance metrics of AUTOCOD in classifying underlying causes of death compared with manual coding to identify specific causes of death during periods of excess mortality. METHODS: We included all DCs between 2016 and 2019. AUTOCOD's performance was evaluated by calculating various performance metrics, such as sensitivity, specificity, positive predictive value (PPV), and F1-score, using a confusion matrix. This compared International Statistical Classification of Diseases and Health-Related Problems, 10th Revision (ICD-10), classifications of DCs by AUTOCOD with those by human coders at the Directorate-General of Health (gold standard). Subsequently, we compared periods without excess mortality with periods of excess, severe, and extreme excess mortality. We defined excess mortality as 2 consecutive days with a Z score above the 95% baseline limit, severe excess mortality as 2 consecutive days with a Z score >4 SDs, and extreme excess mortality as 2 consecutive days with a Z score >6 SDs. Finally, we repeated the analyses for the 3 most common ICD-10 chapters focusing on block-level classification. RESULTS: We analyzed a large data set comprising 330,098 DCs classified by both human coders and AUTOCOD. AUTOCOD demonstrated high sensitivity (≥0.75) for 10 ICD-10 chapters examined, with values surpassing 0.90 for the more prevalent chapters (chapter II-"Neoplasms," chapter IX-"Diseases of the circulatory system," and chapter X-"Diseases of the respiratory system"), accounting for 67.69% (223,459/330,098) of all human-coded causes of death. No substantial differences were observed in these high-sensitivity values when comparing periods without excess mortality with periods of excess, severe, and extreme excess mortality. The same holds for specificity, which exceeded 0.96 for all chapters examined, and for PPV, which surpassed 0.75 in 9 chapters, including the more prevalent ones. When considering block classification within the 3 most common ICD-10 chapters, AUTOCOD maintained a high performance, demonstrating high sensitivity (≥0.75) for 13 ICD-10 blocks, high PPV for 9 blocks, and specificity of >0.98 in all blocks, with no significant differences between periods without excess mortality and those with excess mortality. CONCLUSIONS: Our findings indicate that, during periods of excess and extreme excess mortality, AUTOCOD's performance remains unaffected by potential text quality degradation because of pressure on health services. Consequently, AUTOCOD can be dependably used for real-time cause-specific mortality surveillance even in extreme excess mortality situations.

Neural Translation and Automated Recognition of ICD-10 Medical Entities From Natural Language: Model Development and Performance Assessment.

BACKGROUND: The recognition of medical entities from natural language is a ubiquitous problem in the medical field, with applications ranging from medical coding to the analysis of electronic health data for public health. It is, however, a complex task usually requiring human expert intervention, thus making it expansive and time-consuming. Recent advances in artificial intelligence, specifically the rise of deep learning methods, have enabled computers to make efficient decisions on a number of complex problems, with the notable example of neural sequence models and their powerful applications in natural language processing. However, they require a considerable amount of data to learn from, which is typically their main limiting factor. The Centre for Epidemiology on Medical Causes of Death (CépiDc) stores an exhaustive database of death certificates at the French national scale, amounting to several millions of natural language examples provided with their associated human-coded medical entities available to the machine learning practitioner. OBJECTIVE: The aim of this paper was to investigate the application of deep neural sequence models to the problem of medical entity recognition from natural language. METHODS: The investigated data set included every French death certificate from 2011 to 2016. These certificates contain information such as the subject's age, the subject's gender, and the chain of events leading to his or her death, both in French and encoded as International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) medical entities, for a total of around 3 million observations in the data set. The task of automatically recognizing ICD-10 medical entities from the French natural language-based chain of events leading to death was then formulated as a type of predictive modeling problem known as a sequence-to-sequence modeling problem. A deep neural network-based model, known as the Transformer, was then slightly adapted and fit to the data set. Its performance was then assessed on an external data set and compared to the current state-of-the-art approach. CIs for derived measurements were estimated via bootstrapping. RESULTS: The proposed approach resulted in an F-measure value of 0.952 (95% CI 0.946-0.957), which constitutes a significant improvement over the current state-of-the-art approach and its previously reported F-measure value of 0.825 as assessed on a comparable data set. Such an improvement makes possible a whole field of new applications, from nosologist-level automated coding to temporal harmonization of death statistics. CONCLUSIONS: This paper shows that a deep artificial neural network can directly learn from voluminous data sets in order to identify complex relationships between natural language and medical entities, without any explicit prior knowledge. Although not entirely free from mistakes, the derived model constitutes a powerful tool for automated coding of medical entities from medical language with promising potential applications.

The Need to Address Fragmentation and Silos in Mortality Information Systems: The Case of Ghana and Peru.

Objectives: We aimed to understand the information architecture and degree of integration of mortality surveillance systems in Ghana and Peru. Methods: We conducted a cross-sectional study using a combination of document review and unstructured interviews to describe and analyse the sub-systems collecting mortality data. Results: We identified 18 and 16 information subsystems with independent databases capturing death events in Peru and Ghana respectively. The mortality information architecture was highly fragmented with a multiplicity of unconnected data silos and with formal and informal data collection systems. Conclusion: Reliable and timely information about who dies where and from what underlying cause is essential to reporting progress on Sustainable Development Goals, ensuring policies are responding to population health dynamics, and understanding the impact of threats and events like the COVID-19 pandemic. Integrating systems hosted in different parts of government remains a challenge for countries and limits the ability of statistics systems to produce accurate and timely information. Our study exposes multiple opportunities to improve the design of mortality surveillance systems by integrating existing subsystems currently operating in silos.

Legal Frameworks: A Starting Point for Strengthening Medicolegal Death Investigation Systems and Improving Cause and Manner of Death Statistics in Civil Registration and Vital Statistics Systems.

Medicolegal death investigation systems, which generally fall within one of three types-medical examiner, coroner, or law-enforcement-led systems-investigate deaths that are unnatural or suspicious. The current quality of cause of death statistics on deaths investigated within medicolegal death investigation systems globally limits effective public health response. A starting point to strengthening global medicolegal death investigation systems and improving the quality of cause and manner of death reported to civil registration systems is through a strong legal framework. Two resources, the United Nations Statistics Division Guidelines on the Legislative Framework for Civil Registration, Vital Statistics and Identity Management and the Global Health Advocacy Incubator Legal and Regulatory Toolkit for Civil Registration, Vital Statistics and Identity Management, present recommendations and provide guidance to country stakeholders in reviewing and revising their medicolegal death investigation legal frameworks. Physician determination of cause and manner of death, defined criteria for case referral to the medicolegal death investigation system, an amendment process, and investigation collaboration are four core considerations for medicolegal death investigation system legal frameworks. A strong medicolegal death investigation legal framework is a necessary starting point, but it is not sufficient for ensuring the timely, accurate, and complete reporting of cause and manner of death in national vital statistics.

Evaluating Quality in Adult Cardiac Surgery.

National and institutional quality initiatives provide benchmarks for evaluating the effectiveness of medical care. However, the dramatic growth in the number and type of medical and organizational quality-improvement standards creates a challenge to identify and understand those that most accurately determine quality in cardiac surgery. It is important that surgeons have knowledge and insight into valid, useful indicators for comparison and improvement. We therefore reviewed the medical literature and have identified improvement initiatives focused on cardiac surgery. We discuss the benefits and drawbacks of existing methodologies, such as comprehensive regional and national databases that aid self-evaluation and feedback, volume-based standards as structural indicators, process measurements arising from evidence-based research, and risk-adjusted outcomes. In addition, we discuss the potential of newer methods, such as patient-reported outcomes and composite measurements that combine data from multiple sources.

Routine mortality surveillance to identify the cause of death pattern for out-of-hospital adult (aged 12+ years) deaths in Bangladesh: introduction of automated verbal autopsy.

BACKGROUND: In Bangladesh, a poorly functioning national system of registering deaths and determining their causes leaves the country without important information on which to inform health programming, particularly for the 85% of deaths that occur in the community. In 2017, an improved death registration system and automated verbal autopsy (VA) were introduced to 13 upazilas to assess the utility of VA as a routine source of policy-relevant information and to identify leading causes of deaths (COD) in rural Bangladesh. METHODS: Data from 22,535 VAs, collected in 12 upazilas between October 2017 and August 2019, were assigned a COD using the SmartVA Analyze 2.0 computer algorithm. The plausibility of the VA results was assessed using a series of demographic and epidemiological checks in the Verbal Autopsy Interpretation, Performance and Evaluation Resource (VIPER) software tool. RESULTS: Completeness of community death reporting was 65%. The vast majority (85%) of adult deaths were due to non-communicable diseases, with ischemic heart disease, stroke and chronic respiratory disease comprising about 60% alone. Leading COD were broadly consistent with Global Burden of Disease study estimates. CONCLUSIONS: Routine VA collection using automated methods is feasible, can produce plausible results and provides critical information on community COD in Bangladesh. Routine VA and VIPER have potential application to countries with weak death registration systems.

Sensitivity and Specificity of the National Death Index for Multiple Causes of Death in People With HIV.

OBJECTIVES: Inaccuracies in cause-of-death information in death certificates can reduce the validity of national death statistics and result in poor targeting of resources to reduce morbidity and mortality in people with HIV. Our objective was to measure the sensitivity, specificity, and agreement between multiple causes of deaths from death certificates obtained from the National Death Index (NDI) and causes determined by expert physician review. METHODS: Physician specialists determined the cause of death using information collected from the medical records of 50 randomly selected HIV-infected people who died in San Francisco from July 1, 2016, through May 31, 2017. Using expert review as the gold standard, we measured sensitivity, specificity, and agreement. RESULTS: The NDI had a sensitivity of 53.9% and a specificity of 66.7% for HIV deaths. The NDI had a moderate sensitivity for non-AIDS-related infectious diseases and non-AIDS-related cancers (70.6% and 75.0%, respectively) and high specificity for these causes (100.0% and 94.7%, respectively). The NDI had low sensitivity and high specificity for substance abuse (27.3% and 100.0%, respectively), heart disease (58.3% and 86.8%, respectively), hepatitis B/C (33.3% and 97.7%, respectively), and mental illness (50.0% and 97.8%, respectively). The measure of agreement between expert review and the NDI was lowest for HIV (κ = 0.20); moderate for heart disease (κ = 0.45) and hepatitis B/C (κ = 0.40); high for non-AIDS-related infectious diseases (κ = 0.76) and non-AIDS-related cancers (κ = 0.72); and low for all other causes of death (κ < 0.35). CONCLUSIONS: Our findings support education and training of health care providers to improve the accuracy of cause-of-death information on death certificates.

Is the role of forensic medicine in the covid-19 pandemic underestimated?

We believe that forensic medicine should play a significant role in the COVID-19 pandemic. Forensic pathologists should ask and answer various questions through autopsy cases during the COVID-19 period, thus providing a significant contribution to science. Some of the potential roles of forensic medicine in this issue include: determining the exact cause of death among the deceased who were SARS-CoV-2 positive, contribution to the accuracy of mortality statistics, understanding pathological mechanisms of COVID-19, tracking the presence of the virus over time, survival of the virus after death as well as dealing with medicolegal issues. A detailed multidisciplinary analysis of autopsy samples would undoubtedly help understand this new illness and its clinical management. Therefore, autopsies during the COVID-19 pandemic should not be an exception, but certainly a rule.

Cause of Death Certification in COVID-19 Deaths.

INTRODUCTION: Appropriate cause of death reporting is vital in the pandemic circumstance for effective planning of the control measures. Accurate reporting and registration of the reason for death are crucial in planning of health programs in turn contributing for the national development. BACKGROUND: All births and deaths occurring across India should be mandatorily registered per the Registration of Births and Deaths Act passed in the year 1969. The act also requires the issuance of cause of death certificate by the doctor attending the departed during his last illness. Data obtained from the cause of death certificate provides cause-specific mortality profile, which is required to analyze the health trends of the population. REVIEW RESULTS: This article discusses the available guidelines on the appropriate documentation of cause of death in the confirmed or suspected coronavirus disease-2019 (COVID-19) infection resulting into death. CONCLUSION: Proper certification of the cause of death leads to better epidemic surveillance. Scrutiny of the clinical sequences from the cause of death certificate is useful to prioritize the allocation of resources for critical care management and to augment our knowledge about underlying causes resulting in mortality from COVID-19. CLINICAL SIGNIFICANCE: Dissemination of available guidelines on proper documentation of the cause of death in confirmed/suspected COVID-19 cases will reduce the errors in cause of death reporting. HOW TO CITE THIS ARTICLE: Veeranna CH, Rani S. Cause of Death Certification in COVID-19 Deaths. Indian J Crit Care Med 2020;24(9):863-867.

Erroneous Reporting of Deaths Attributed to Pneumonia and Influenza at 2 New York City Teaching Hospitals, 2013-2014.

OBJECTIVES: Cause-of-death information, reported by frontline clinicians after a patient's death, is an irreplaceable source of public health data. However, systematic bias in cause-of-death reporting can lead to over- or underestimation of deaths attributable to different causes. New York City consistently reports higher rates of deaths attributable to pneumonia and influenza than many other US cities and the country. We investigated systematic erroneous reporting as a possible explanation for this phenomenon. METHODS: We reviewed all deaths from 2 New York City hospitals during 2013-2014 in which pneumonia or influenza was reported as the underlying cause of death (n = 188), and we examined the association between erroneous reporting and multiple extrinsic factors that may influence cause-of-death reporting (patient demographic characteristics and medical comorbidities, time and hospital location of death, type of medical provider reporting the death, and availability of certain diagnostic information). RESULTS: Pneumonia was erroneously reported as the underlying cause of death in 163 (86.7%) reports. We identified heart disease and dementia as the more likely underlying cause of death in 21% and 17% of erroneously reported deaths attributable to pneumonia, respectively. We found no significant association between erroneous reporting and the multiple extrinsic factors examined. CONCLUSIONS: Our results underscore how erroneous reporting of 1 condition can lead to underreporting of other causes of death. Misapplication or misunderstanding of procedures by medical providers, rather than extrinsic factors influencing the reporting process, are key drivers of erroneous cause-of-death reporting.

[On the issue of morbidity and mortality encoding].

The encoding of cases of disease and death is the translation of clinical diagnoses into alphanumeric code with observance of certain rules. Properly selected ICD-10 codes ensure reliability of statistical data that undoubtedly affects quality of managerial decisions. The article considers results of analysis of information from primary medical documentation (in-patient medical record, discharged patient statistical record) and its comparing with the ICD-10 codes. The detailed analysis of encoding errors is presented. The structure of hospital morbidity and mortality of patients of palliative care department of multidisciplinary hospital is analyzed.

A Deep Artificial Neural Network-Based Model for Prediction of Underlying Cause of Death From Death Certificates: Algorithm Development and Validation.

BACKGROUND: Coding of underlying causes of death from death certificates is a process that is nowadays undertaken mostly by humans with potential assistance from expert systems, such as the Iris software. It is, consequently, an expensive process that can, in addition, suffer from geospatial discrepancies, thus severely impairing the comparability of death statistics at the international level. The recent advances in artificial intelligence, specifically the rise of deep learning methods, has enabled computers to make efficient decisions on a number of complex problems that were typically considered out of reach without human assistance; they require a considerable amount of data to learn from, which is typically their main limiting factor. However, the CépiDc (Centre d'épidémiologie sur les causes médicales de Décès) stores an exhaustive database of death certificates at the French national scale, amounting to several millions of training examples available for the machine learning practitioner. OBJECTIVE: This article investigates the application of deep neural network methods to coding underlying causes of death. METHODS: The investigated dataset was based on data contained from every French death certificate from 2000 to 2015, containing information such as the subject's age and gender, as well as the chain of events leading to his or her death, for a total of around 8 million observations. The task of automatically coding the subject's underlying cause of death was then formulated as a predictive modelling problem. A deep neural network-based model was then designed and fit to the dataset. Its error rate was then assessed on an exterior test dataset and compared to the current state-of-the-art (ie, the Iris software). Statistical significance of the proposed approach's superiority was assessed via bootstrap. RESULTS: The proposed approach resulted in a test accuracy of 97.8% (95% CI 97.7-97.9), which constitutes a significant improvement over the current state-of-the-art and its accuracy of 74.5% (95% CI 74.0-75.0) assessed on the same test example. Such an improvement opens up a whole field of new applications, from nosologist-level batch-automated coding to international and temporal harmonization of cause of death statistics. A typical example of such an application is demonstrated by recoding French overdose-related deaths from 2000 to 2010. CONCLUSIONS: This article shows that deep artificial neural networks are perfectly suited to the analysis of electronic health records and can learn a complex set of medical rules directly from voluminous datasets, without any explicit prior knowledge. Although not entirely free from mistakes, the derived algorithm constitutes a powerful decision-making tool that is able to handle structured medical data with an unprecedented performance. We strongly believe that the methods developed in this article are highly reusable in a variety of settings related to epidemiology, biostatistics, and the medical sciences in general.

[Spatial autocorrelation of breast cancer mortality in the Metropolitan Region, Chile: an ecological study]. / Autocorrelación espacial de mortalidad por cáncer de mama en la Región Metropolitana, Chile: estudio ecológico.

INTRODUCTION: Breast cancer is the most common malignancy in women worldwide and Chile, being the leading cause of female cancer death. A wide variation in mortality has been reported, with geographic clusters of higher risk. OBJECTIVE: To spatially analyze mortality from breast cancer in women in the Metropolitan Region in 2015. METHODS: Ecological study of location. We used death records in 2015 (C50 according to ICD10) and population projections of the Statistics Institute to estimate mortality rates. We calculated crude breast cancer mortality rates and standardized mortality ratios and performed a spatial epidemiological analysis of breast cancer mortality in women, estimating the global and local Moran I index to assess spatial autocorrelation. We present the results in maps according to the 2016 pre-census cartography. RESULTS: There were 622 deaths from breast cancer in the Metropolitan Region in 2015. The mean age was 66 years (SD: 15.5). 92.4% of deaths were registered in urban or central areas. However, the highest mortality rates were observed in peripherical districts. No global spatial autocorrelation was observed in the region (Morans I 0.007 p = 0.134). However, at the local level, four districts differ significantly from their neighbors. CONCLUSIONS: The risk of dying from breast cancer in the Metropolitan Region of Chile is concentrated in women from peripherical communes. Four districts in the region present different risks from their neighboring districts. It is necessary to investigate local realities to prevent deaths from this pathology.

INTRODUCCIÓN: El cáncer de mama es la neoplasia maligna más común en las mujeres en todo el mundo y en Chile, siendo la primera causa de muerte oncológica femenina. Se ha reportado amplia variación en la mortalidad, con focos geográficos de mayor riesgo. OBJETIVO: Analizar espacialmente la mortalidad por cáncer de mama en mujeres de la Región Metropolitana en 2015. MÉTODOS: Estudio ecológico. Se utilizaron los datos de los registros de defunciones del año 2015 (C50 según CIE10), y las proyecciones poblacionales del Instituto Nacional de Estadísticas. Se calcularon tasas de mortalidad por cáncer de mama brutas y razones de mortalidad estandarizadas. Se realizó un análisis epidemiológico espacial estimando el índice I de Moran Global y Local para evaluar autocorrelación espacial. Los resultados se presentan en mapas (cartografía pre-censo 2016). RESULTADOS: Se registraron 622 defunciones por cáncer de mama en la Región Metropolitana en 2015. La edad promedio de las mujeres fallecidas fue de 66 años (desviación estándar: 15,5). El 92,4% de las muertes se registró en zonas centrales o urbanas. Sin embargo, las mayores tasas de mortalidad se observaron en comunas periféricas. No se observó autocorrelación espacial global en la región (I de Moran de 0,007; p = 0,134). A nivel local, cuatro comunas se diferencian de forma significativa de sus vecinas. CONCLUSIONES: El riesgo de morir por cáncer de mama en la Región Metropolitana de Chile se concentra en comunas periféricas. Cuatro comunas de la región presentan riesgos diferentes de sus comunas vecinas, por lo que es necesario explorar factores que explican la desigual distribución de las muertes.