Unknown and ill-defined causes of death in the mortality of the populations of Serbia, Croatia, North Macedonia, and Slovenia, in the period between 2007 and 2016

Introduction: Data on the cause of death form the cornerstone for analyzing the health situation and disease in countries, and they make a major contribution to building evidence for health policies.

Aim: The aim of this study was to determine the extent to which diagnoses from the group – Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (R00-R99), International Classification of Diseases (ICD - Revision X) were used as the main cause of death in Serbia, Croatia, North Macedonia, and Slovenia in the ten-year period, 2007 – 2016.

Materials and methods: Methods of descriptive and analytical statistics were used in this research. An analysis of data on the causes of death (R00-R99 ICD - Revision X), by gender, during the ten-year period (2007 – 2016) was performed. Linear regression was also used as an analytical method to analyze the trend.

Results: During the ten-year period, in Serbia, the R00-R99 diagnoses were among the five most common groups of diagnoses of death, i.e., in each year they took third place, with a percentage of 4.7%. In the observed period, in the surrounding countries, there was an increase in the death rate in Slovenia, with the highest rate in 2016 (19.9), while in Croatia there was a decrease in the death rate related to the diagnoses from group XVIII ICD - X (R00-R99). In Macedonia, the rate had a linear trend, with a slight decline in 2012 (52.3) and 2013 (58.7). In the observed period, an increase in the death rate of the population of Serbia with an unknown cause of death was observed, with particularly high rates in 2009 and 2016. Comparative analysis has shown that R00-R99 diagnoses are represented more in the mortality statistics of Serbia than in Slovenia and Croatia, and less than in Northern Macedonia.

Conclusion: Urgent interventions are needed to improve the quality of mortality statistics and data on the causes of death in the described countries.

INTRODUCTION

Policies and programs for fighting diseases and injuries are based on timely information on the nature and scope of health problems. The most commonly used data for the purpose of health policies are statistical data on the number of deceased people, given by age and sex, as well as by cause of death. The data on cause of death represent the cornerstone of the analysis of the health situation and disease in countries, and they greatly contribute to the development of databases for health policies [1]. The World Health Organization (WHO) recommends avoiding the use of undetermined and unknown causes of death in the death certificate, as the belief is that this terminology does not offer information on the conditions that had led to death [2]. Also, there are suggestions that the death rate for deaths attributed to symptoms, signs and ill-defined causes of death may be a potential indicator of access to and use of health services [3]. The percentage of ill-defined deaths or deaths with unknown cause, is one of the indicators of the very quality of the data on the cause of death. The reliability of these data is necessary, not only for the assessment of trends and variations in the population’s health, but also for the assessment of inequalities in health among the population groups [4].

Mortality data provide the most important health indicators for the assessment and comparison of the health status, at the local, state, and international levels, since in each developed country, as well as in most developing countries, these data are regularly and comprehensively collected [5]. Also, reliable and valid data on mortality are important for epidemiological research, as well as for public health policy and priority goals [6]. Mortality statistics is one of the basic sources of health-related information, and, in many countries, it is the most reliable source of health-related data [7].

The International Classification of Diseases, Tenth Revision (ICD-X), is reference classification of the WHO. It is widely accepted and recommended as a guidebook for international reporting on health. The purpose of ICD is to enable systematic recording, analysis, interpreting, and comparison of collected data on mortality and morbidity, in different countries or regions, and for different time periods [8]. The ICD is divided into groups and has a total of 21 groups. Groups I to XVII relate to diseases and other states of illness; Group XIX relates to injuries, poisoning, and consequences of external factors. The remaining groups complete the framework of possible conditions covered by diagnostic data. Group XVIII comprises symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified [8]. From the point of view of death prevention, it is important to break the chain of events which leads to an adverse outcome, illness or disability. The most effective public health goal is to prevent the impact of the causative agent. This is why the underlying cause of death is defined as “the disease or injury that initiated the train of events leading directly to death, or the circumstances of the accident or violence which produced the fatal injury” [9, p. 35, ICD-X, Volume 2]. These rules provide for international comparability of data on mortality and help in the standardization of managing unclear medical documentation. The rule is that, when more than one condition is entered into the death certificate, only one underlying condition should be selected as the one leading to all the others. Although there are rulebooks in use, when it comes to mortality data, international comparability is still limited, due to the differences amongst health systems, as well as nationally modified rules [9],[10].

The following states are considered unknown and ill-defined: ICD-X, R00-R94 or R95-R99, Group XVIII – Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (Appendix 1). The categories from this group should not be used as codes for the “underlying cause”, unless the symptom, sign, or abnormal finding was the obvious underlying condition that was treated and investigated during the treatment episode and was not connected with any other condition marked by the doctor [11].

The aim of this study was to determine the extent of the use of the diagnoses from the group - Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (R00-R99), ICD, Tenth Revision, as the underlying cause of death. This analysis will enable the comparison of the quality of the data on cause of death in our country with that of the other countries in the region: Croatia, North Macedonia, and Slovenia, over a ten-year period (2007 – 2016).

METHODS

Units of observation and tested variables

The units of observation are the following: the overall population, the participation in total mortality (%), and the rate of persons deceased due to unknown and ill-defined causes of death (R00-R99 ICD – Revision X) in Serbia, Croatia, North Macedonia, and Slovenia, in the period 2007 – 2016. The results of the analysis of the data trend regarding the participation of persons deceased due to R00-R99 in overall mortality (%) is also presented.

Data sources

Data published by the Statistical Office of the Republic of Serbia on the mortality of the citizens of Serbia were used in the study [12-15]. Official mortality statistics data of Croatia [16-17], North Macedonia [18],[19], and Slovenia [20],[21] were used for the purpose of international comparison.

Statistical analysis

Methods of descriptive and analytical statistics have been used in this study. Analysis of data on the causes of death (R00-R99 ICD - Revision X), by sex, age and regions in Serbia, as well as the trend line, the basic index and the chain index of change, over a ten-year period (2007 – 2016) have been carried out. Linear regression with the support of the program of the American National Cancer Institute (Joinpoint Regression Trend Analysis Software - Version 4.9.0.0) was used for trend analysis. The most significant results of the study are represented in graphs and tables, through percentages and the rate of participation of the diagnoses from the R00-R99 ICD-X group in the mortality statistics of the citizens of Serbia and the countries form the region, and have been discussed in the context of relevant reference literature.

RESULTS

In the observed period, a rise in the rate of deceased persons per 100,000 citizens of Serbia with a diagnosis from the group ICD XVIII – Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (hereinafter: R00-R99) was recorded. The highest rate was recorded in 2009 (72.0/100,000) as well as in 2016 (70.6/100,000) (Figure 1).

Figure 1. The rate of deceased persons per 100,000 citizens with the diagnosis from the XVIII ICD-X (R00-R99) group, in the Republic of Serbia, between 2007 and 2016

During the observed period, in the Republic of Slovenia, a rise in the rate of persons deceased from R00-R99 was recorded, with the highest rate noted in 2016 – 19.9/100,000 (Figure 2).

Figure 2. The rate of deceased persons per 100,000 citizens with the diagnosis from the XVIII ICD-X (R00-R99) group, in the Republic of Slovenia, between 2008 and 2016

In the Republic of Croatia, during the observed period, between 2008 and 2015 (at the time of the study being carried out, year 2015 was the last year with available data), a drop in the rate of deceased persons with diagnoses from the R00-R99 group was recorded, thereby the values in 2006 and 2015 were the same (Figure 3).

Figure 3. The rate of deceased persons per 100,000 citizens with the diagnosis from the XVIII ICD-X (R00-R99) group, in the Republic of Croatia, between 2008 and 2015

Between 2007 and 2016, in the Republic of North Macedonia, the rate of deceased persons with cause of death diagnoses from the R00-R99 group had a linear trend, with a mild decrease in 2012 and 2103 (Figure 4).

Figure 4. The rate of deceased persons with the diagnosis from the XVIII ICD-X (R00-R99) group, in the Republic of North Macedonia, between 2007 and 2016

In the observed period, the participation of R00-R99 in the overall mortality was statistically significantly higher in Serbia than in Slovenia (p < 0.001) (Figure 5). In the observed period, in Serbia, there is no statistically significant trend of change in the proportion of R00-R99 (APC = 0.0; p = 0.995), rather, there is a linear trend of stagnation. In Slovenia, within the observed period, there is a statistically significant linear trend of increase in the proportion of R00-R99, with an average rate of annual change of 6.6%. (APC = 6.6; p < 0.001).

Figure 5. Trend Analysis for the diagnoses R00-R99 (Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified), comparison between the Republic of Slovenia and the republic of Serbia, between 2007 and 2016

In the observed period, the participation of R00- 99 in the overall mortality was statistically significantly smaller in Serbia than in North Macedonia (p < 0.001) (Figure 6), although there isn’t a statistically significant trend of change in the proportion of R00-R99, in the observed period, either in Serbia or in North Macedonia (APC = 0.0; p = 0.995 and APC = -0.5; p = 0.669, respectively). In both countries, a stable linear trend is present, indicating stagnation and minimal change.

Figure 6. Analysis of the trend for the diagnoses R00-R99 (Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified), comparison between the Republic of Serbia and the Republic of North Macedonia, between 2007 and 2016

At the time of the conducting of the research, there were only data from the period between 2008 and 2015 available for Croatia. In that period the values of the proportion of R00-R99 were statistically significantly greater in Serbia, as compared to Croatia (p < 0.001) (Figure 7). There is a linear trend of change in the proportion of R00-R99 (APC = -0.4; p = 0.464, and APC = -2.4; p = 0.068, respectively) both in Serbia and Croatia, i.e., there are no statistically significant changes.

Figure 7. Analysis of the trend for the diagnoses R00-R99 (Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified), comparison between the Republic of Serbia and the Republic of Croatia, between 2007 and 2015

DISCUSSION

Despite WHO recommendations to avoid the use of undetermined and unknown causes of death in the death certificate, due to the belief that this terminology does not offer information on the conditions that had led to death [3], as well as the existence of legislation regulating the way that the death certificate should be filled out both in Serbia, as well as in the countries from the region, in the observed countries, a large number of these diagnoses is prevalent as the underlying cause of death in the death certificate. In the period between 2007 and 2016, in Serbia, the diagnoses from the group XVIII ICD-X (R00-R99) were amongst the five most common groups of diagnoses stated as the underlying cause of death, i.e., in each year, they were ranked third; also, an increase in their numbers was registered in years 2009 and 2016. Additionally, in the observed period, in Slovenia, there was an increase in the death rate, with the highest rate observed in 2016, while in Croatia and North Macedonia, a mild decrease of death rates of persons diagnosed with R00-R99 as the cause of death was noted.

An increase or continuation of the level of ill-defined causes of death in the mortality statistics represents a great challenge for researchers all over Europe. Experience from other countries teaches us the way to improve data on mortality and cause of death, and, as an example, we can use the fact that the Government of Turkey has, as of 2009, conducted several reforms, at the Turkish Statistical Institute, to improve the system of reporting on the cause of death [22]. After analysis, it has been noted that there is a significant lack of data related to demographic and epidemiological variables, especially when it comes to the death of infants and to the detailed recording of the cause of death [22].

The occurrence of ill-defined and poorly coded underlying causes of death in the death certificate undermines the usefulness of this data, especially when the access to coding is not standardized [23]. Such is the example from research indicating that there is a high degree of inconsistency in the system related to the coding of mental disorders and behavioral disorders, diseases of the nervous system, endocrine disorders, certain cardiovascular diseases and ill-defined causes of death, as the underlying cause of death [23]. For the purpose of improving mortality statistics data, the introduction of a software system for the coding of data on the cause of death in Serbia would contribute to better reporting. As of January 2017, software versions of ICD-X have been in use in the EU countries. Upon the implementation of the new software for coding the cause of death, the National Records of Scotland registered the first increase, and then decrease in the number of deaths with a certain diagnosis as the underlying cause of death [24]. The National Center for Health Statistics of the American National Center for Disease Control and Prevention developed an automatic coding system of causes of death, and the developed computer programs are widely in use, which has reduced the risk of systemic errors caused by inconsistent interpretation and application of coding rules. The software is available for the automatization of the coding of medical data in the death certificate, in keeping with WHO guidelines. This is one of the rare sources of health-related data comparable for small geographic areas and available for a longer period of time [25].

Study limitations

For some causes of death, mortality statistics reflects the coding practice rather than the actual epidemiological situation, and, therefore, the reliability of these research results should be observed in that context. The coverage and quality of data on cause of death vary amongst countries, and even valid, reliable, and comparable assessments of the trends of the cause of death, in the best of systems, are limited by problems such as the changes in the ICD-X, the use of tabular lists where significant details on cause of death are lost, as well as by many deaths which have been assigned a diagnosis that cannot be considered an underlying cause of death. Namely, underregistration is particularly serious in rural areas with poor transport and a poor accessibility to health care centers [26].

It is necessary to research a longer period and a greater number of variables (e.g., by sex, by age, and individual diagnoses, as well as by the source of data in the death certificates), and then to perform a comparative analysis against the diagnoses of the medical examiners, in order to better understand the problem and define the target interventions. Health care institutions, especially hospitals, are a common source of population mortality data, by age, sex, and cause. These statistical data are important markers of the quality of hospital care and provide for necessary investments, for both national and local health care policies. Continued training is necessary, with respect to the appearance, content and proper coding of data on the cause of death, both for doctors in health care institutions and for medical examiners, who are involved in filling out death certificates. It is also very important that decision makers in the health care system should recognize the significance of precise and quality data on mortality, which is why investing into the strengthening of the vital registration system is necessary, in order to achieve improvement at all levels, in the sense of properly coded underlying causes of death and the incorporation of these data into final reports on mortality statistics.

CONCLUSION

In the observed period, an increase in the death rate of Serbian citizens with unknown cause of death has been noted, with exceptionally high rates in 2009 and 2016. Comparative analysis has shown that R00-R99 diagnoses are more present in the mortality statistics of Serbia, as compared to Slovenia and Croatia, and less, as compared to North Macedonia. Urgent interventions are necessary in order to improve the quality of data on the underlying cause of death in death certificates, with the aim of improving the trust in their accuracy and providing for these data to be the primary proof which epidemiological research and health policies are based on. This research has provided a broader view of the quality of coding data on the underlying cause of death in death certificates, at the regional level.

APPENDIX 1

Group XVIII in the ICD-X, Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified comprises the following 13 subgroups:

R00-R09 Symptoms and signs involving the circulatory and respiratory systems
R10-R19 Symptoms and signs involving the digestive system and abdomen
R20-R23 Symptoms and signs involving the skin and subcutaneous tissue
R25-R29 Symptoms and signs involving the nervous and musculoskeletal systems
R30-R39 Symptoms and signs involving the urinary system
R40-R46 Symptoms and signs involving cognition, perception, emotional state and behavior
R47-R49 Symptoms and signs involving speech and voice
R50-R69 General symptoms and signs
R70-R79 Abnormal findings on examination of blood, without diagnosis
R80-R82 Abnormal findings on examination of urine, without diagnosis
R83-R89 Abnormal findings on examination of other body fluids, substances and tissues, without diagnosis
R90-R94 Abnormal findings on diagnostic imaging and in function studies, without diagnosis
R95-R99 Ill-defined and unknown causes of mortality