Trends in mortality from ill-defined causes among the elderly in Brazil, 1979-2013: ecological study

ABSTRACT CONTEXT AND OBJECTIVE: Mortality measurements are traditionally used as health indicators and are useful in describing a population's health situation through reporting injuries that lead to death. The aim here was to analyze the temporal trend of proportional mortality from ill-defined causes (IDCs) among the elderly in Brazil from 1979 to 2013. DESIGN AND SETTING: Ecological study using data from the Mortality Information System of the Brazilian Ministry of Health. METHODS: The proportional mortality from IDCs among the elderly was calculated for each year of the study series (1979 to 2013) in Brazil, and the data were disaggregated according to sex and to the five geographical regions and states. To analyze time trends, simple linear regression coefficients were calculated. RESULTS: During the study period, there were 2,646,194 deaths from IDCs among the elderly, with a decreasing trend (ß -0.545; confidence interval, CI: -0.616 to -0.475; P < 0.000) for both males and females. This reduction was also observed in the macroregions and states, except for Amapá. The states in the northeastern region reported an average reduction of 80%. CONCLUSIONS: Mortality from IDCs among the elderly has decreased continuously since 1985, but at different rates among the different regions and states. Actions aimed at improving data records on death certificates need to be strengthened in order to continue the trend observed.


INTRODUCTION
Mortality measurements are traditionally used as health indicators 1 and are useful in describing a population's health situation through reporting injuries that lead to death. This allows authorities, among other things, to prioritize the allocation of resources in accordance with the mortality profile identified. 2 Mortality indicators assist in monitoring the trends of the most prevalent causes of death in a population and hence identify which segments are affected to a greater or lesser extent by certain diseases. The numerators of these indicators are obtained from the Mortality Information System (SIM) of the Ministry of Health, which also functions as a strategic tool for management of the healthcare system. 3 Mortality statistics are only infrequently used. This is partly because of lack of completeness of the data fields that comprise death certificates (DCs), particularly the field representing the underlying cause of death. The lack of information in this field limits the explanatory power of death records regarding mortality patterns in a population.
In situations in which it is not possible to identify the underlying cause of death, such as lack of medical care, failure of doctors to properly maintain assignations and records or missing information, the cause of death is classified as an ill-defined cause High proportions of reported deaths classified as ill-defined causes can significantly alter the mortality rates for specific diseases. This can distort a given community mortality profile and consequently reduce the potential use of these statistics for diagnosing the health of a given population and for planning and administering healthcare services for that population. 4 By international standards, Brazil was characterized as having high levels of ill-defined causes of death in the middle 1990s. 5 Over the last three decades, the Brazilian government has made significant investments that have improved vital registration systems over recent years. 6,7 The completeness of death counts increased from 80% in 1980-1991 to 95% in 2000-2010, while at the same time the percentage of ill-defined causes of deaths was reduced by about 53% in the country, but with large regional differences. The south and southeast have much better data quality than the rest of the country. 8 The distribution of ill-defined causes according to demographic characteristics, such as gender and age, is marked by higher incidence among men and the elderly (elderly is defined here as 60 years of age or older). Among the elderly, it is particularly difficult to identify the cause of death 9 because of the presence of comorbidities (hypertension, diabetes, cancer, arteriosclerosis, dyspnea upon exertion, osteoarthritis and reduced visual acuity, among others) that frequently occur among the elderly. Moreover, age can influence the clinical expression of signs and symptoms, 10 and it may be difficult to deal with the elderly, who may refuse to seek treatment and only do so in the later stages of the disease when there is greater impairment, which can hinder or even prevent establishment of diagnoses.
Therefore, it is essential to monitor the quality of the information relating to the underlying cause of death among the elderly and the information relating to the demand for healthcare and social services, so as to better develop care planning and health promotion in this age group. 11 Mortality trends can be identified from mortality rates in which the risk of death due to a specific cause is measured; or through proportional mortality, in which the relative importance of a disease or group of diseases is reported. In this study, we chose to work with proportional mortality to assess the weight of ill-defined causes of death among the elderly.

OBJECTIVE
The objective of this study was to analyze the evolution of proportional mortality as a result of ill-defined causes of death among the elderly in Brazil during the period 1979-2013.

METHODS
This ecological study used time series and exploratory analyses 2 in which secondary data were used. All deaths registered as illdefined causes among the elderly (detailed in Chapter XVI of ICD-9, for the period between 1979 and 1995; and in Chapter XVIII of ICD-10 from 1996 onwards) were included in this study.
This system was implemented between 1975 and 1976, and the computerized database became available for viewing/capture on the web pages of the Information Technology Department of the Brazilian National Health System (DATASUS), with data from 1979 on. 12 This study used data in the public domain. Thus, there was no need for approval from a research ethics committee. Two spatial scales were used for data analysis: Brazil and its macroregions (north, northeast, south, southeast and center-west). The proportional mortality due to IDCs among the elderly was calculated for each year of the study series (1979 to 2013) in Brazil, and the data were disaggregated according to sex and to the five geographical regions. The proportion of IDC deaths for each sex was calculated based on the total number of deaths for each sex.
Simple linear regression coefficients were calculated to examine the nature and significance of the temporal trend of proportional mortality from ill-defined causes among the elderly.
In this analysis, the variable of time, expressed in years, was entered into the model as the independent variable, and the variable proportions of deaths from IDCs, overall and separated according to gender and geographical region, functioned as the dependent variables. To perform the data analysis, we used Tabwin, 13 which is a public-domain spreadsheet provided by DATASUS, Ministry of Health; and the R software, which is a public-domain statistical package.

RESULTS
During the study period, there were 2,646,194 deaths from illdefined causes among the elderly, corresponding to an average of 75,606 deaths/year. The highest frequency of these deaths    (Figure 3).

DISCUSSION
The time series analysis on the proportional mortality from illdefined causes among the elderly revealed a marked and progressive decrease during the study period. This reduction was observed for the country as a whole, with regional variations that were accentuated in the north-south direction, ranging from a reduction of approximately 95.5% in the state of Espírito Santo,  1980  1981  1982  1983  1984  1985  1986  1987  1988  1989  1990  1991  1992  1993  1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008     The downward trend in the proportion of this type of death is widespread, but we believe that peculiarities are maintained in some age groups such as the elderly. The high occurrence of home deaths in this age group 9 also increases the chance of the death being classified as having an ill-defined cause. A study conducted in four state capitals that assessed mortality from illdefined causes among the elderly from 1996 to 2007 found that in Porto Alegre, 50% of deaths classified as IDCs occurred at home. 22 In the state of São Paulo in 2010, 48.2% of these events occurred at home, and only 44.2% occurred in hospitals or other healthcare facilities. Moreover, unattended deaths, which are one of the most common situations for ill-defined causes of death, represented 31% of all ill-defined causes of death. These deaths occur predominantly at home (42.6%). 14 Moreover, the classification of deaths did not differ according to sex. The downward trend in both sexes suggests that the factors that caused this reduction produced the same effects in both sexes.
The high proportions of IDC deaths observed in the 1980s may reflect a time when the need for services and healthcare professionals to assist the population was greatest. This need resulted from uneven spatial distribution of healthcare services and professionals in Brazil, which were primarily located in the southern and southeastern regions and in major urban centers.
The above mentioned problems added to the difficulty of the population's access to healthcare services and the organization of health surveillance services. 23 The disabilities of the organization of health surveillance services depicts a situation of serious neglect of a problem regarding information that was more marked that decade, thus reflecting a failure to comply with the mandatory registration of vital events, and specifically deaths, and the lack of importance given to this by those in charge of planning healthcare actions.
The household living situations of the elderly, whether in rural or urban areas, may explain the resistance to or greater difficulty in health service provision. The predominantly rural location pattern probably affects the demand for healthcare because rural populations have less access to and therefore make less use of healthcare services. 24,25 This may result partly from transportation difficulties, financial constraints and greater resistance to seeking medical care. We believe that seeking medical care was a more important issue in the mid-1980s. A significant proportion of the population over 60 years of age during that decade came from a cohort that was primarily born in rural areas, and a fraction of that population is still alive today. Brazil was historically a country with predominantly rural characteristics, but since the 1950s, Brazil has been undergoing a transformation into a more urbanized country. Only in the 1960s did the urban population exceed the rural population. 26 The low frequency of healthcare services use among people living in rural areas throughout their lives is a behavioral characteristic that may change as these people become older, given that the demand for healthcare services can be expected to increase. This increase in the frequency of healthcare services use among the elderly has been previously demonstrated. This stage of life is characterized by a greater biological vulnerability associated with higher prevalence of diseases and disabilities. 27 The cohorts born in the 1930s and subsequently differed from those that preceded them in relation to household status at birth. In 2000, the campaign began offering immunization to those over 60 years old. Campaigns towards the elderly may have contributed towards encouraging them to seek healthcare services, thus reducing the culturally constructed resistance.
The availability of healthcare services and professionals prepared to meet the needs of this contingent, which swells the population, is a challenge for both the state and society. Populations have the ability to extend their average lifespan and therefore age is an indicator of social evolution, which is influenced by the pattern of economic development and the technical/scientific attainment level of the society to which the population belongs. Such achievements are a source of concern for both society and the state, which need to adjust to new demands, and they have an impact on the economic and social structure. The challenge ahead for the twenty-first century is to provide quality-of-life support to a growing elderly population of primarily low socioeconomic and educational level and high prevalence of chronic diseases and disabilities. 28,29 Apart from these unwieldy problems, the reduction in the proportion of ill-defined causes of death points towards the possibility of achieving an even lower level if the ongoing actions are intensified. The Brazilian states with the worst indicators need to be prioritized, so as to identify the main causes of the poor quality of information and implement a series of actions to reverse this situation. Among the measures that could help reduce occurrences of deaths categorized as having ill-defined causes is continuous monitoring of what is causing death in this age group, with training and skill transfers for municipalities with greater difficulties, with the aim of reaching physicians in these regions and encouraging them to fill out the underlying cause of death on the death certificate.
This study was based on large spatial units, including macroregions and states, for the data analysis. Although this allowed approximation of occurrences and the spatial distribution of the event analyzed, the spatial unit size can be considered to be a limitation of this study. Use of smaller spatial units, such as regional health districts or even municipalities, would provide knowledge in greater detail, including identification of the localities with major problems in classifying deaths, and would allow interventions to be targeted to the most deprived locations.

CONCLUSIONS
Proportional mortality from ill-defined causes among the elderly was seen to present a marked progressive decrease during the study period. This reduction was observed for the country as a whole, with regional variations accentuated in the north-south direction. These variations require geographically differentiated interventions in order to reduce their occurrence. Thus, improving the quality of mortality statistics among the elderly is essential in order to provide valid and reliable data for producing information to support healthcare planning for this group of elderly individuals.