SciELO - Scientific Electronic Library Online

 
vol.22 suppl.3Evaluation study of the improvement of the quality of death information in hospitals of the states of Rio de Janeiro and São Paulo, Brazil, 2017Improving the usefulness of mortality data: reclassification of ill-defined causes based on medical records and home interviews in Brazil author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

Share


Revista Brasileira de Epidemiologia

Print version ISSN 1415-790XOn-line version ISSN 1980-5497

Rev. bras. epidemiol. vol.22  supl.3 Rio de Janeiro  2019  Epub Nov 28, 2019

http://dx.doi.org/10.1590/1980-549720190009.supl.3 

ORIGINAL ARTICLE

Change in the profile of causes of death after investigation of hospital deaths in Belo Horizonte, Brazil, 2017

Paulo Roberto Lopes CorrêaI 
http://orcid.org/0000-0002-2302-2179

Lenice Harumi IshitaniII 
http://orcid.org/0000-0002-7165-4736

Sônia LanskyI 
http://orcid.org/0000-0001-5533-4858

Mayara Rocha dos SantosII 
http://orcid.org/0000-0002-4985-4862

Renato Azeredo TeixeiraII 
http://orcid.org/0000-0002-1259-6812

Elisabeth Barboza FrançaIII 
http://orcid.org/0000-0001-6984-0233

IDiretoria de Promoção à Saúde e Vigilância Epidemiológica, Gerência de Vigilância Epidemiológica, Sistema de Informação sobre Mortalidade, Secretaria Municipal de Saúde - Belo Horizonte (MG), Brasil

IIGrupo de Pesquisas em Epidemiologia e Avaliação em Saúde, Universidade Federal de Minas Gerais - Belo Horizonte (MG), Brasil

IIIPrograma de Pós-Graduação em Saúde Pública, Universidade Federal de Minas Gerais - Belo Horizonte (MG), Brasil


ABSTRACT

Introduction:

Deaths certified with ill-defined causes or garbage codes (GC) compromise the analysis of mortality and its use for planning and evaluation of public health policies. The hospital investigation of these causes is one of the strategies qualifying the profile of mortality in the country.

Objective:

To evaluate the change in the hospital mortality profile after investigation of deaths certified with GC in 2017 in Belo Horizonte, Brazil.

Methods:

A sample of hospital deaths reported with GC in the Mortality Information System (SIM) of Belo Horizonte in 2017 was investigated and subsequently certified by a physician to compare the mortality profile before and after investigation.

Results:

After investigating 1,395 deaths out of 3,038 reported with GC, a reduction of 35.5% of these causes was observed. Groups of all ages presented decreases in GC occurrence. A higher proportional increase was observed for deaths due to ischemic heart diseases, Alzheimer’s disease, chronic obstructive pulmonary disease, ischemic and hemorrhagic stroke, and external causes of death (accidental falls, homicides and traffic/transport accidents).

Conclusion:

The investigation on reported hospital deaths is one of the strategies to improve mortality statistics, reducing the occurrence of GC among reported deaths and changing the mortality profile in these facilities. The importance of continuous physician training in cause-of-death certification is emphasized.

Keywords: Mortality; Information systems; Cause of death

RESUMO

Introdução:

Óbitos declarados com causas mal definidas, ou causas garbage (CG), comprometem a análise da mortalidade e sua utilização para planejamento e avaliação de políticas públicas de saúde. A investigação hospitalar destas causas é uma das estratégias para qualificação do perfil de mortalidade no país.

Objetivo:

Avaliar a mudança no perfil de mortalidade hospitalar após investigação de óbitos declarados com CG, em 2017, em Belo Horizonte.

Métodos:

Tomou-se uma amostra dos óbitos hospitalares notificados com CG no Sistema de Informação sobre Mortalidade de Belo Horizonte em 2017. Posteriormente, os dados foram investigados e analisados por médico, para comparação do perfil de mortalidade antes e depois do processo de investigação.

Resultados:

Após investigação de 1.395 óbitos, dentre 3.038 declarados com CG, houve redução de 35,5% para estas causas. Todas as faixas etárias apresentaram decremento na ocorrência de CG. Observou-se maior incremento proporcional para os óbitos causados por doenças isquêmicas do coração, doença de Alzheimer, doença pulmonar obstrutiva crônica, acidente vascular cerebral isquêmico e hemorrágico e as mortes por causas externas (quedas acidentais, homicídios e acidentes de trânsito/transporte).

Conclusão:

A investigação de óbitos hospitalares notificados é uma das estratégias para aprimorar as estatísticas de mortalidade, reduzindo a ocorrência de CG entre as mortes declaradas e alterando o perfil de mortalidade nestes estabelecimentos. Enfatiza-se a importância de educação permanente de médicos para qualificação das causas de morte.

Palavras-chave: Mortalidade; Sistemas de Informação; Causas de morte

INTRODUCTION

Mortality statistics are among the most important tools for analyzing the health situation of populations, subsiding the planning and evaluation of public policies1), (2. The Mortality Information System (SIM), implemented in Brazil in the 1970s, uses the death certificates (DC), an instrument of mandatory completion by physicians, as data source. However, in 2000, the quality of these sources is still precarious and unequal, with a high proportion of deaths due to ill-defined causes3) even with death records with above 90% coverage in the country.

We emphasize that the proportion of deaths whose underlying cause is declared by the physician as nonspecific, including the codes from chapter XVIII of the International Statistical Classification of Diseases and Health-related Problems - 10ª Review (ICD-10), and also codes of other chapters, among the problems involving the quality of information on mortality. These causes have been denominated incomplete, unhelpful causes or garbage codes (GC), as proposed by the Global Burden of Disease initiative (GBD); and compromise the analysis of mortality when occur in high proportion4), (5. GBD also proposes the grouping of GCs according to the severity of its impact on the mortality profile and, consequently, in the definition of assertive public policies6. In Brazil, 33% of about 1.3 million total deaths in 2016 were declared with GC, with a similar proportion among hospital deaths7.

Since 2005, the Brazilian Ministry of Health (MoH) has incremented actions aiming at improving the definition of causes of death in the country, implementing standardized instruments for investigation of deaths, training teams of death surveillance and establishing goals and quality indicators of SIM, having the transfer of financial incentives to municipalities as counterpart8)- (10. Recently, in 2018, the Protocolo de investigação de óbitos com causas classificadas como códigos garbage - sessenta cidades do Brasil (“Protocol for Investigating Deaths with Causes Classified as Garbage Codes - sixty cities of Brazil”) was implemented, relying on the support of partner institutions and including the investigation of hospital deaths reported with GC in 2017. Belo Horizonte was one of the cities included in this project11, with about 29% of these cases, which until then investigated deaths from ill-defined and external causes due to event of undetermined intent, in addition to all infant, fetal, maternal deaths and death of women of childbearing age.

This study analyzed the change in the profile of hospital mortality after investigation and correction of deaths declared with GC in Belo Horizonte in 2017, considering that most of deaths in hospital facilities should present a specific cause of death.

METHODS

This is an evaluative descriptive epidemiological study using the SIM database, available at the local base of the Municipal Health Secretariat of Belo Horizonte (SMS-BH), regarding the deaths of residents occurring in hospitals in this municipality in 2017. Belo Horizonte has a total of 2,375,151 inhabitants, according to the population census conducted by the Brazilian Institute of Geography and Statistics (IBGE) in 2010, and records approximate annual average of 15,000 deaths of residents, 70% of these in hospitals7.

The following GC were selected for investigation, considered priority by MoH in the protocol and referring to the following three-or four-character codes of the ICD-10: ill-defined causes (R00-R99, except R95); sequelae or cerebrovascular accident (stroke) unspecified as hemorrhagic or ischemic (I64, I67.4, I67.9, I69.4, I69.8); septicemia (A40-A41); heart failure and unspecified heart diseases (I50, I51); essential hypertension (I10); unspecified neoplasia (C26, C55, C76, C78, C79, C80); pulmonary embolism (I26); pneumonia (J15.9, J18); respiratory failure, not elsewhere classified (J96) and other respiratory disorders (J98); renal failure (N17, N19); external causes due to event of undetermined intent and unspecified accidents (Y10-Y34, X59); unspecified traffic accidents and homicides (V89, Y09). GC were grouped in severity levels (1 to 4) according to the list proposed by GBD/201712, considering its impact on public policies, which was defined by Anaconda13 as: very high (Level 1), for causes with serious implications; high (Level 2), GC with substantial implications; medium (Level 3), containing GC with important implications; and low (Level 4) in which GC have limited implications. According to GBD, levels 1 and 2 are the most important for their greater impact on mortality analysis.

The investigations were conducted according to the operational capacity of the epidemiological surveillance teams of the municipality and in three different stages of sample selection, presented below.

  • 1st stage: from the total of hospital deaths with priority GC between January and April 2017 (n = 475), approximately 80% (n = 374) were selected, which occurred in the 16 hospitals in the city with greater proportions of these events.

  • 2nd stage: We selected hospital deaths with priority GC (n = 289) identified by epidemiological surveillance in routine investigations of diseases and illnesses of compulsory notification, infant and maternal death, deaths of women of childbearing age and deaths by external causes due to event of undetermined intent.

  • 3rd stage: A probabilistic sample of all deaths with priority GC not investigated in the previous stages (n = 1,698) was selected. The prevalence of the three main causes of death among the uninvestigated cases (unspecified pneumonia, unspecified stroke and septicemia) was adopted as a parameter to estimate the sample of this stage, with 664 deaths. A proportional stratification was performed between hospitals with five or more deaths to be investigated to minimize the possibility of selection bias, and a collection of 20% or more cases was defined in each hospital to compensate for possible sample loss, resulting in a total sample of 797 deaths.

We used the form of Investigação de Óbitos por Causa Mal Definida Hospitalar (“investigation of hospital deaths due to ill-defined causes” - IOCMD-H), elaborated by the MoH in 2017 for hospital investigation of the deaths selected in the first and third stages. This form includes data from the clinical history and results of complementary exams recorded in the hospital records during the last hospitalization that culminated with the death investigated. Epidemiological surveillance teams (physicians, nurses and dentists) of the SMS-BH and the selected hospitals, including the teams of the Epidemiology Hospital Centers, were previously trained to investigate and use the IOCMD-H form. In the second stage, the hospital investigation reports of the epidemiological surveillance teams were evaluated for deaths related to compulsory notification diseases and illnesses and the standardized forms by the MoH for investigation of infant and maternal deaths. Regarding the deaths from external causes and initially coded with priority GC in chapter XX of ICD-10, the information from the technical reports elaborated by the forensic physicians of the Instituto Médico Legal (forensic institute - IML) of Belo Horizonte and research in electronic media to specify the circumstances involved in the deaths.

The investigations of the three stages were completed by a certifier physician to define a new underlying cause of death. In cases, in which the investigation did not allow the definition of another cause of death, the original underlying cause in the DC was maintained. All investigations conducted were subsequently forwarded to the same group of encoders of the previous DC for recoding of the post-investigation underlying cause. The reclassified cases were included in the SIM.

The proportional distributions of the selected deaths were analyzed in the three stages according to the following variables: sex, age range (0 to 19, 20 to 39, 40 to 59, 60 to 79, and 80 years or more), priority GC and the impact levels of the respective codes. We compared the proportion of deaths from GC and specific causes (according to the list of causes of GBD/2015), before and after the investigation, and stratified by age range to evaluate the change of causes of death comparing with the total number of hospital deaths.

The variables of interest were tabulated using the Microsoft Office Excel program; the Minitab-18 program was used for the analysis of frequency and measures of central tendency, establishing a 0.05 significance level.

This study was approved by the Ethics and Research Committee of the Universidade Federal de Minas Gerais (CAEE 75555317.0.0000.5149) and developed according to the ethical precepts established in Ordinance No. 466/2012 of the National Health Council.

RESULTS

In Belo Horizonte, of the 24,490 deaths reported in the SIM in 2017, 68.4% were of residents in the city, and of these, 58.6% occurred in hospitals in the capital. Among the 3,038 hospital deaths with GC, 2,361 (77.7%) were declared with priority GC. In total, 1,395 deaths were investigated: 374 and 289 in the first two stages, respectively, and 732 in the third, as shown in Figure 1.

Source: Mortality Information System of Belo Horizonte (SIM).

Figure 1 Flowchart of selected and investigated garbage code deaths. Belo Horizonte, 2017. 

The proportional distribution of men among total deaths with total GC and priority GC was similar (49.7 and 49.6%, respectively; p = 0.98) and the median of identical age (77 years; p = 0.56) (data not shown). Men predominated in the sample investigated, as shown in Table 1. Although most of the deaths investigated (76.1% of the total) were of adults older than 60 years of age, the proportion of deaths in the age group of 20 to 39 years in the second stage was high (15.2%), with a median age of 65 years. The median age among the deaths investigated in the three stages were different: 71 years in the first, 65 in the second, and 80 in the third (p < 0.00).

Table 1 Deaths investigated according to variables, level of garbage codes and stages of the investigation. Belo Horizonte, 2017. 

Variable First Second Third Total
n % n % n % n %
Sex Women 167 44.7 93 32.2 394 53.8 654 46.9
Man 207 55.3 196 67.8 338 46.2 741 53.1
Age group 0 to 19 11 2.9 22 7.6 3 0.4 36 2.6
20 to 39 24 6.4 44 15.2 15 2 83 5.9
40 to 59 76 20.3 52 18 87 11.9 215 15.4
60 to 79 149 39.8 90 31.1 253 34.6 492 35.3
80 and more 114 30.5 81 28 374 51.1 569 40.8
GC Priority External causes 124 33.2 223 77.2 36 4.9 383 27.5
Pneumonia 5 1.3 10 3.5 295 40.3 310 22.2
Unspecified Stroke 61 16.3 0 0 132 18 193 13.8
Septicemia 58 15.5 7 2.4 82 11.2 147 10.5
Ill-defined 33 8.8 46 15.9 23 3.1 102 7.3
Other causes 93 0 3 0 164 0 260 0
GC level Level 1 138 36.9 52 18 197 26.9 387 27.7
Level 2 139 37.2 225 77.9 53 7.2 417 29.9
Level 3 31 8.3 2 0.7 57 7.8 90 6.5
Level 4 66 17.6 10 3.5 425 58.1 501 35.9
Total 374 100 289 100 732 100 1,395 100

The main causes among the deaths investigated were the external causes due to event of undetermined intent, pneumonia and unspecified stroke, with 27.5% (n = 383), 22.2% (n = 310) and 13.8% (n = 193) of the deaths, respectively. While external causes had greater relative importance in the first (33.2%) and second (77.2%) stages, pneumonia accounted for 40.3% of the deaths investigated in the third stage. Unspecified stroke and septicemia showed higher proportions in the first (16.3%) and second (13.8%) stages. Table 1 shows that garbage codes classified in levels 1 and 2 (higher severity) combined corresponded to 74.1% and 95.9% of deaths in the first two stages, respectively, while level 4 presented a proportion of 58.1% of deaths in the third stage.

External causes due to event of undetermined intent, pneumonia and unspecified stroke stood out among adults older than 60 years, as shown in Table 2. In the age groups below 40 years, the external causes due to event of undetermined intent also predominated. It also draws attention to the predominance of level 2 attributed to GC in the age groups of 0 to 19 (77.8%), 20 to 39 (72.3%) and 40 to 59 years (35.3%), while level 4 prevailed in the age groups of 60 to 79 (33.7%) and 80 or more years (46.7%).

Table 2 Hospital deaths investigated by causes and levels of priority of garbage codes, by age group. Belo Horizonte, 2017. 

Priority garbage codes Age group (years) Total
0 to 19 20 to 39 40 to 59 60 to 79 > = 80
n % n % n % n % n % n %
External causes 28 77.8 60 72.3 70 32.6 120 24.4 105 18.5 383 27.5
Pneumonia 3 8.3 9 10.8 38 17.7 85 17.3 175 30.8 310 22.2
Unspecified Stroke 0 0 0 0 20 9.3 81 16.5 92 16.2 193 13.8
Septicemia 4 11.1 8 9.6 24 11.2 50 10.2 61 10.7 147 10.5
Ill-defined causes 1 2.8 2 2.4 21 9.8 42 8.5 36 6.3 102 7.3
Other causes 0 0 4 4.8 42 20 114 23 100 18 260 19
Levels of garbage codes
Level 1 5 13.9 12 14.5 55 25.6 145 29.5 170 29.9 387 27.7
Level 2 28 77.8 60 72.3 76 35.3 135 27.4 118 20.7 417 29.9
Level 3 0 0 2 2.4 27 12.6 46 9.3 15 2.6 90 6.5
Level 4 3 8.3 9 10.8 57 26.5 166 33.7 266 46.7 501 35.9
Total 36 100 83 100 215 100 492 100 569 100 1,395 100

Generally, deaths due to external causes increased in absolute values and in all age groups, drawing attention to deaths due to traffic accidents in the age group of 20 to 39 years, with over 1 death to 29 cases. From the deaths among children under 20 years of age, 30.6% (n = 11) remained with GC. Table 3 showed a relative reduction of 35.5% in total hospital deaths attributed to GC, from 3,038 to 1,959 deaths. This reduction was observed in all age groups, especially in the group between 20 and 39 years, which decreased from 35% to 16.4%, and the range under 20 years, with the lowest reduction, 26.6% to 20%. The relative reduction were 40.6%, 36.3% and 32.3% for the age groups of 40 to 59, 60 to 79 and 80 or more years respectively (data not shown).

Table 3 Main causes of hospital death according to age range, before and after investigation and variation. Belo Horizonte, 2017. 

Age group Main causes of death Before After Variation (%)
n % n %
0-19 years (n = 380) Homicides 18 4.7 30 7.9 66.7
Traffic accidents 1 0.3 5 1.3 400
20-39 years (n = 397) Accidental Falls 0 0 3 0.8 300
Traffic accidents 1 0.3 29 7.3 2,800
40-59 years (n = 1,571) Homicides 8 0.5 20 1.3 150
Hypertensive heart disease 0 0 7 0.4 700
Suicides 0 0 8 0.5 800
Traffic accidents 0 0 14 0.9 1,400
Accidental Falls 0 0 19 1.2 1,900
60-79 years (n = 3,686) Hypertensive heart disease 8 0.2 25 0.7 212.5
Ischemic stroke 6 0.2 36 1 500
Suicides 2 0.1 12 0.3 500
Accidental Falls 5 0.1 69 1.9 1,280
Traffic accidents 0 0 22 0.6 22,000
80+ (n = 3,394) Hypertensive heart disease 11 0.3 30 0.9 172.7
Traffic accidents 0 0 8 0.2 800
Ischemic stroke 3 0.1 37 1.1 1,133.3
Accidental Falls 10 0.3 124 3.7 1,140

Table 4 shows the distribution of the main causes of death before and after the investigation process, ordered from the highest to the lowest proportional value in comparison with the total number of deaths (n = 9,428). Ischemic heart diseases maintained the highest relative frequency (with an increase of 5 to 5.4%), remaining in the first place among the causes of death. An increase in the relative frequency of deaths attributed to chronic obstructive pulmonary disease (COPD) (from 3.9 to 4.5%), Alzheimer’s disease (3.9 to 4.5%) and ischemic stroke (2.2 to 3.5%, passing from 9º to 4º place) is observed. External causes of mortality present an important change, with accidental falls passing from 79º to 10º place; homicide from 20º to 15º place; and specified traffic/transport accidents, from 122º to 24º place.

*Comparing with total hospital deaths (n = 9,428); ** traffic accident specified; *** Excluded levels of garbage codes excluded in 46 (before) and 48 (after) deaths. D.: disease.

Table 4 Main causes of hospital deaths in order of occurrence, before and after investigation. Belo Horizonte, 2017. 

Regarding the levels attributed to each GC before investigation (n = 3,038), levels 1 and 4 presented a proportion of 33.2% and 33.8%, respectively, and level 3 had the lowest proportion (11.1%). Level 2 showed a higher reduction after the investigation, (55.3%), followed by levels 4 (37.5%) and 1 (29.7%); the lowest reduction occurred at level 3 (13.3%). The proportion of deaths remaining as GC after investigation (n = 1,959) was 36.5%, 15.4%, 15.1% and 33.1% for levels 1 to 4, according to severity levels, respectively. For levels 1 and 4 in the age groups of 60 to 79 years, the proportion reduced from 3.7% to 2.5% and from 3.5% to 2.4%, respectively; the highest decrease was observed over 80 years, from 4.8% to 3.3% and from 5.9% to 4%, respectively (data not shown).

Of the total 310 deaths investigated and reported as due to pneumonia (codes J15 to J18 of ICD-10), 11.3% were reclassified for Alzheimer’s disease, 10.3% for COPD, 6.5% for hypertensive heart disease and 14.5% maintained the same cause. The total number of hospital deaths reported with the codes R00 to R99 (chapter XVIII of ICD-10), reduced from 1.5% to 0.7% after the investigation (data not shown).

DISCUSSION

This study showed an important reduction of 35% in the proportion of GC after investigation of deaths and, therefore, reinforces the applicability of the process performed as a tool to qualify and improve the SIM. External causes (falls, homicides and traffic/transport accidents) and non-communicable chronic diseases (ischemic heart disease, Alzheimer’s and unspecified stroke) increased their importance as causes of death in hospitals after the investigation. However, 20.8% of the total hospital deaths in 2017 still remained with GC.

Few studies have analyzed the distribution of GC levels before and after investigating deaths. In this study, the highest proportion of GC with high level (level 2) observed in the second stage may be attributed to external causes due to event of undetermined intent in young adults, while level 4 in older adults stood out in the third stage, in which pneumonia and stroke not specified as ischemic or hemorrhagic are relatively more common. Among the deaths whose cause remained as garbage code after the investigation, a level change was observed for some of these causes, i.e., a Level 1 GC may have been reclassified to another level of lower impact due to improved information. An example of this situation is one death due to unknown cause (ICD-10 code R99: Level 1) in which a sequela of stroke unspecified if ischemic or hemorrhagic (ICD-10 code I 69.4: Level 4) was identified as a cause, after investigation.

In Brazil, GC still remains an important concern in the analysis of mortality, compromising the planning and evaluation of public policies. The quality of the physician’s completion of DC and the data coding process have been reported as relevant for the occurrence of CG14), (15. The lack of training and attention of the physician for the correct registration, the quality of information available in clinical records and the diagnostic resources available have been pointed out among the factors involved in the quality of the completion of the DC16. In this study, the identification of a well-defined cause of death or change in the GC level with lower impact corroborated the hypothesis that one of the main factors compromising the quality of the SIM refers to the completion of the DC by the physician since clinical data and diagnostic support to define a specific cause were available in hospital records.

Other studies have pointed out that the occurrence of deaths with GC may be related to their higher frequency in more advanced age groups, to sex, to the difficulty of physicians in defining the external causes due to event of undetermined intent and the existence of a death verification service (SVO) to investigate deaths from natural causes without medical care or epidemiological interest16), (18. In older adults, for example, the coexistence of several comorbidities in the same patient may hinder the physician’s identification of the underlying cause of death18. In Brazil, a study analyzing deaths in the age range of 15 to 59 years between 1998 and 2012 also showed a higher proportion of ill-defined causes in the most advanced ages19. Still on the proportion of deaths with GC, the analysis of 762,000 deaths occurring in 24 provinces in Ecuador between 2001 and 2013 found differences related to sex (higher in women) and age (higher in older adults), different from observed in this study, possibly due to the criteria used for sampling20. A higher proportion of GC in young men was observed among the deaths investigated in this study, which can be attributed to the preponderance of external causes due to event of undetermined intent verified in the first and, mainly, in the second stage. Furthermore, the highest proportion of deaths with GC investigated occurred in adults older than 60 years, except for the first two stages, in which deaths from external causes due to event of undetermined intent predominated and those investigated by the routine of the services of epidemiological surveillance.

In Belo Horizonte, the investigation of deaths with undetermined or unknown underlying cause (ICD-10 code R99) has been a priority for the city’s surveillance teams, even in those occurring in a hospital facility. This municipality does not have SVO to support the elucidation of these deaths and those without medical assistance. Although SIM information in the city is considered good regarding the ill-defined causes (chapter XVIII of ICD-10) - 5% of total deaths are reported with ill-defined causes - between 2011 and 2013, 68.4% of deaths occurred in hospitals, of which 28.7% have GC, compromising the analysis of mortality both for these establishments and for the municipality7. A study evaluating the investigation of deaths from ill-defined causes in Brazil in 2010 detected a reduction of 20% of these causes and around 65% of reclassification into defined causes21. In this study, total GC reduced 53%, reaching a value of 0.7% after the investigation. These findings indicate that a considerable proportion of deaths remains with ill-defined causes after investigation even in hospital deaths, for which a more specific cause is expected to be identified.

One limitation of this study was the investigation only 46% of the hospital deaths reported with GC selected in three stages with different methodologies, considering the operational difficulties for analyzing the total number of deaths. In addition, the study prioritized the investigation of priority GCs, as proposed by the MoH. In this sense, the change in the distribution of the causes of hospital death found in this study would possibly be even greater if all deaths with GC had been included in the investigation. However, our results indicate that there was a change in the hospital mortality profile after the investigation, with identification of several causes that had not previously been emphasized and became important.

CONCLUSION

The study showed that the investigation of hospital deaths with GC is a viable strategy that allows the qualification of the death certificate, improving the analysis of mortality profile in these establishments and enabling managers to direct more assertive care actions, with a view to coping with the main causes of death. A fundamental and timely strategy to minimize the occurrence of deaths reported with GC involves improving the medical records through training these professionals in qualified completing of causes of death in the DC. The improvement of health care services with qualified diagnostic support and the need to implant a SVO in the municipality is also noteworthy.

Acknowledgments

The authors thank the team of technicians of the epidemiological surveillance to the death of the central level of the Municipal Health Secretariat and the sanitary districts of Belo Horizonte, as well as the teams of the Epidemiology Hospital Centers for supporting the process of research and qualification of the mortality information system, enabling the achievement of the objectives of this study.

REFERENCES

1. Rede Interagencial de Informação para a Saúde. Indicadores básicos para a saúde no Brasil: conceitos e aplicações. 2a ed. Brasília, DF: Organização Pan-Americana da Saúde; 2008. [ Links ]

2. Frias PG, Szwarcwald CL, Morais Neto OL, Leal MC, Cortez-Escalante JJ, Souza Junior PRB, et al. Utilização das informações vitais para a estimação de indicadores de mortalidade no Brasil: da busca ativa de eventos ao desenvolvimento de métodos. Cad Saúde Pública. 2017;33(3):e00206015. http://dx.doi.org/10.1590/0102-311x00206015Links ]

3. Souza MFM, França EB, Cavalcante A. Carga da doença e análise da situação de saúde: resultados da rede de trabalho do Global Burden of Disease (GBD) Brasil. Rev Bras Epidemiol. 2017;20(Suppl 1):1-3. http://dx.doi.org/10.1590/1980-5497201700050001. [ Links ]

4. Murray CJL, Lopez AD. The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. Cambridge, MA: Harvard School of Public Health; 1996. [ Links ]

5. Naghavi M, Abajobir AA, Abbafati C, Abbas KM, Abd-Allah F, Abera SF, et al. Global, regional, and national age-sex specific mortality for 264 causes of death, 1980-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2017;390(10100):1151-210. https://doi.org/10.1016/S0140-6736(17)32152-9Links ]

6. Mikkelsen L, Richards N, Lopez AD. Redefining 'garbage codes' for public health policy: report on the expert group meeting, 27-28 February 2017. CRVS technical outcome series. Melbourne, Australia: Bloomberg Philanthropies Data for Health Initiative, Civil Registration and Vital Statistics Improvement, University of Melbourne; 2018. [ Links ]

7. Brasil. Ministério da Saúde. Sistema de Informação de Mortalidade [Internet]. Brasília, DF: Ministério da Saúde; 2017 [citado 15 maio 2019]. Disponível em: http://svs.aids.gov.br/dantps/cgiae/sim/tabulacao/Links ]

8. França EB, Cunha CC, Vasconcelos AMN, Escalante JJC, Abreu DX, Lima RB, et al. Avaliação da implantação do programa "Redução do percentual de óbitos por causas mal definidas" em um estado do Nordeste do Brasil. Rev Bras Epidemiol. 2014;17(1):119-34. http://dx.doi.org/10.1590/1415-790X201400010010ENGLinks ]

9. Brasil. Ministério da Saúde. Secretaria de Gestão Estratégica e Participativa. Departamento de Articulação Interfederativa. Caderno de diretrizes, objetivos, metas e indicadores: 2013-2015. Brasília, DF; 2013. [ Links ]

10. Brasil. Ministério da Saúde. A experiência brasileira em sistemas de informação em saúde. Brasília, DF: Organização Pan-Americana da Saúde, Fundação Oswaldo Cruz; 2009. 2 v. [ Links ]

11. Ishitani LH, Teixeira RA, Abreu DMX, Paixão LMMM, França EB. Qualidade da informação das estatísticas de mortalidade: códigos garbage declarados como causas de morte em Belo Horizonte, 2011-2013. Rev Bras Epidemiol. 2017;20(Suppl 1):34-45. http://dx.doi.org/10.1590/1980-5497201700050004Links ]

12. Roth GA, Abate D, Abate KH, Abay SM, Abbafati C, Abbasi N, et al. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018; 392(10159):1736-88. https://doi.org/10.1016/S0140-6736(18)32203-7Links ]

13. Mikkelsen L, Lopez AD. Improving cause of death information: guidance for assessing and interpreting the quality of mortality data using Anaconda. Melbourne: University of Melbourne, Civil Registration and Vital Statistics Improvement Group, Bloomberg Philanthropies Data for Health Initiative; 2017. [ Links ]

14. Hazard RH, Chowdhury HR, Adair T, Ansar A, Rahman AMQ, Alam S, et al. The quality of medical death certification of cause of death in hospitals in rural Bangladesh: impact of introducing the International Form of Medical Certificate of Cause of Death. BMC Health Services Research. 2017;17(1):688. doi: https://doi.org/10.1186/s12913-017-2628-y [ Links ]

15. Matos SG, Proietti FA, Barata RCB. Confiabilidade da informação sobre mortalidade por violência em Belo Horizonte, MG. Rev Saúde Pública. 2007;41(1):76-84. http://dx.doi.org/10.1590/S0034-89102007000100011Links ]

16. Messias KLM, Bispo Júnior JP, Pegado MFQ, Oliveira LC, Peixoto TG, Sales MAC, et al. Qualidade da informação dos óbitos por causas externas em Fortaleza, Ceará, Brasil. Ciênc Saúde Coletiva. 2016;21(4):1255-67. http://dx.doi.org/10.1590/1413-81232015214.07922015Links ]

17. Kanso S, Romero DE, Leite IC, Moraes EN. Diferenciais geográficos, socioeconômicos e demográficos da qualidade da informação da causa básica de morte dos idosos no Brasil. Cad Saúde Pública. 2011;27(7):1323-39. http://dx.doi.org/10.1590/S0102-311X2011000700008Links ]

18. Mello Jorge MHP, Laurenti R, Lima-Costa MF, Gotlieb SLD, Chiavegatto Filho ADP. A mortalidade de idosos no Brasil: a questão das causas mal definidas. Epidemiol Serv Saúde. 2008;17(4):271-81. http://dx.doi.org/10.5123/S1679-49742008000400004Links ]

19. Rodrigues NCP, Daumas RP, Almeida AS, O'Dwyer G, Andrade MKN, Flynn MB, et al. Risk factors for the ill-defined causes of death in the Brazilian states: a multilevel analysis. Ciênc Saúde Coletiva. 2018;23(11):3979-88. http://dx.doi.org/10.1590/1413-812320182311.27182016Links ]

20. Peralta A, Benach J, Borrell C, Espinel-Flores V, Cash-Gibson L, Queiroz BJ, et al. Evaluation of the mortality registry in Ecuador (2001-2013) - social and geographical inequalities in completeness and quality. Population Health Metrics. 2019;17(1). http://doi:10.1186/s12963-019-0183-yLinks ]

21. França E, Teixeira R, Ishitani L, Duncan BB, Cortez-Escalante JJ, Morais Neto OL, et al. Causas mal definidas de óbito no Brasil: método de redistribuição baseado na investigação do óbito. Rev Saúde Pública. 2014;48(4):671-81. http://dx.doi.org/10.1590/S0034-8910.2014048005146Links ]

Financial support: Funding from Vital Strategies as part of the Bloomberg Philanthropies Data for Health Initiative (Project 23998 Fundep/UFMG).

Received: June 12, 2019; Revised: July 18, 2019; Accepted: August 08, 2019

Corresponding author: Paulo Roberto Lopes Corrêa. Afonso Pena, 2333, 9th floor, Funcionários, CEP: 30.130-007, Belo Horizonte, MG, Brazil. Email: paulocorreabh@gmail.com

Authors’ contribution: Corrêa PRL, Ishitani LH, Lansky S, Santos MR and France EB conceived and delineated the study, organized the database and analyzed the results, wrote and revised the manuscript. Teixeira AR contributed to the conception of the study, sample selection and statistical analysis of the data. All authors approved the final version of the manuscript and declared themselves responsible for all aspects of the work, guaranteeing their accuracy and integrity.

Conflict of interests: nothing to declare

Creative Commons License Este é um artigo publicado em acesso aberto sob uma licença Creative Commons