INTRODUCTION

According to the World Health Organization, Chronic Non-Communicable Diseases (NCDs) are responsible for 71% of all deaths worldwide, mainly affecting low- and middle-income countries and posing a threat to public health^1,2.

In Brazil, the Ministry of Health estimates that 57.4 million people have NCDs – cardiovascular diseases, malignant neoplasms, chronic respiratory diseases, and diabetes mellitus –, which are responsible for 72% of deaths in the country³. However, its distribution and impact are not homogeneous across all social strata: vulnerable groups, such as the elderly and individuals with low income and lower educational levels tend to be the most affected⁴.

Taking into account that more than 70% of the national population depends exclusively on the Unified Health System (UHS)⁵, as well as the differences regarding the allocation of resources and health outcomes between the regions of Brazil, it is worth questioning the impact of the number of doctors working in the UHS on the number of hospitalizations and mortality rates due to NCDs in the country, along with the influence of income inequality in such outcomes.

Thus, the objective was to evaluate the relationship between the number of doctors in the Unified Health System and the number of hospitalizations and mortality rates due to NCDs, and how income inequality might influence these outcomes.

METHODS

RESULTS

Between 2016 and 2018 there were 2,423,251 deaths in adults over 20 years of age due to Chronic Non-Communicable Diseases (NCDs) in Brazil, with a slight predominance among men, who corresponded to 51% of total deaths. Among NCDs, diseases of the circulatory system reached first place with almost 50% of deaths, followed by malignant neoplasms with 656,696 deaths, respiratory diseases and, finally, endocrine-metabolic diseases. All NCDs showed a predominance of deaths in men, with the exception of endocrine-metabolic diseases. In the same period, the country spent US$3,198,354,081.93 on NCDs, with 52% of spending going to men (table 1).

Regarding variations over time, it is noted that, for cancer, there was an increase in the absolute number of deaths (APC 4.44; AAPC 2,501), but a reduction in mortality rates (APC -2.71; AAPC -2.51) for men, while there was an increase in both the absolute number of deaths (APC 7.92; AAPC 3.951) and mortality rates (APC 0.90; AAPC 0.59) for women. For the disease, there was an increase in spending and hospital admissions for both sexes both in absolute numbers and calculated rates. As for metabolic diseases, an increase in the absolute number of deaths was found (APC 6.32; AAPC 1,111.5), but reduced mortality rates in men (APC -1.36; AAPC -0.40). The same phenomenon was observed for deaths in women, for which there was a tendency for the absolute number to increase (APC 2.69; AAPC 562.5), but with a reduction in mortality rates (APC -4.76; AAPC -1.25). Overall, there was a trend towards an increase in spending, but a reduction in hospital admissions for both sexes. Regarding circulatory diseases, there was a reduction in the absolute number and mortality rates in men and women, with increased expenses and fewer hospitalizations. And, finally, for diseases of the respiratory system there was a tendency towards a fall/stability in hospitalizations, with a reduction in the absolute number and mortality rates in both sexes, but with an increase in absolute expenditure for men and women, at the same time in that there was a reduction in the spending rate for men (APC -2.31; AAPC -1,496.51) (table 1).

Regarding the multivariate models, it is noted that there is a negative correlation between the Gini index and the hospital admission rate for all NCDs (p value < 0.001), except for metabolic diseases for which there was a negative correlation with average monthly income (β = -0.14; CI95% -0.20 – -0.08; p value < 0,001). Still in this model, it is noted that the Gini index showed a negative correlation with the total amount spent on all NCDs (p value < 0.05) and, furthermore, there was a positive correlation between the average monthly income and the amount spent on circulatory and respiratory diseases (β = 622.98; CI95% 440.99 – 804.97; p value < 0,001; β = 80.38; CI95% 28.75 – 132.02; p value = 0.002, respectively) (table 2).

However, except for respiratory diseases (β = 35.758; CI95% 30.255 – 41.561; p value < 0,001), there was no correlation between the Gini index and mortality from NCDs, which is why it was not included in the multivariate model. On the contrary, there was a divergent correlation between mortality from NCDs and average monthly income, with the latter being negatively associated with mortality from metabolic and circulatory diseases (β = -0.109; CI95% -0.140 – -0.078; p value < 0,001; β = -0.184; CI95% -0.250 – -0.118; p value < 0,001, respectively), and positively associated with cancer mortality (β = 0.056; CI95% 0.034 – 0.079; p value < 0,001). Furthermore, there is a positive correlation between mortality due to any NCD and sex, where, for statistical reasons, this variable was considered as binary, with “0” linked to the female sex and the number “1” to the male sex. (p value < 0.001). It is interesting to note that, for cancer, the only medical specialty measured associated with lower mortality was surgical oncology (β = -95.236; CI95% -163.616 – -26.856; p value 0.007), while for metabolic and circulatory diseases, several medical specialties were negatively associated with mortality (p < 0.05). Similarly, bronchoscopy was the only specialty negatively associated with mortality from respiratory diseases (β = -72.770; CI95% -99.073 – -46.467; p value < 0,001), for which a negative association was also noted over time (β = -5.132; CI95% -8.568 – -1.695; p value 0.004) (table 3).

DISCUSSION

When analyzing the behavior of NCDs in Brazil between 2016 and 2018, we found that: there was a predominance of deaths and expenses involving men; there was a negative association between the Gini index and the hospital admission rate and expenditure on NCDs; there was no correlation between the Gini index and mortality from NCDs, but there was a divergent correlation between this and the average monthly income depending on the disease analyzed.

To understand these results, some points must be taken into consideration. First, measuring inequality at an individual and interpersonal level can prove to be extremely difficult. Brazil has large and historic income inequality that, despite falling across the country, persists among its regions^13,14. Thus, although the UHS and the Family Health Strategy (FHS) play a crucial role in combating NCDs at national level, they also serve as a parameter for monitoring differences between macro-regional outcomes, since the UHS replicates inequalities between regions through the unequal distribution of its services¹⁵.

In this sense, we found that average monthly income was negatively associated with mortality from NCDs, with the exception of cancer. Here, it is worth considering that, probably, in regions where there is a greater concentration of income, there is also the possibility of greater out-of-pocket expenses, complementing or even replacing the therapy offered by the UHS and, thus, reducing mortality rates¹⁶. Furthermore, it is known that there is an unequal distribution of high and medium complexity equipment, which remains concentrated in the country’s richest centers¹⁵. In this context, there is evidence of regional influence on life expectancy and mortality from NCDs, as well as the different local availability of doctors ^17-21. Thus, adding the aforementioned, it can be hypothesized that the positive association found between average income and cancer mortality is precisely due to the increase in the number of diagnoses secondary to the greater arsenal of high complexity available in greater quantity in richer areas of the country, which does not necessarily translate into better treatments or results, but might be, sometimes, a waste of public resources²².

Regarding the finding on the greater expenditure and number of deaths from NCDs in men, it must be taken into account that the burden of NCDs still remains greater in men around the world and, in Brazil, their mortality is, in general, higher^23,24. Regarding expenses, it is possible to hypothesize that, since men tend to seek medical assistance later – compared to women –, their disease is at a more advanced stage and, therefore, requires more intensive treatment – which leads to higher costs. Another equally valid explanation could be attributed to individual data – such as age – or the number of deaths itself, which can distort real values, falsely increasing expenditure on men²⁵.

Furthermore, it is worth commenting on another findings probably reflecting a mathematical variation: for both cancer and endocrine-metabolic diseases, an increase in the absolute number of deaths was observed, but with a reduction in mortality rates –probably due to population increase, which, albeit small, led to the dilution of the number of cases²⁶. Noteworthy, for endocrine-metabolic diseases, there was an increase in general expenses, but with a reduction in hospital admissions – which is consistent with the most recent data in the literature on admission rates for this group of diseases²⁷. This finding, in addition to reinforcing the high cost associated with the management of these diseases²⁸, could, indirectly, indicate better outpatient management and, thus, contributes to the hypothesis that there was a correct allocation of public financial resources.

Finally, this article has limitations. First, like any ecological study, data extrapolation must be done with caution: its main purpose must be to generate hypotheses for a better global understanding of the Brazilian medical situation. Furthermore, it should be noted that the values described as “ignored” in the databases were not collected for any variable studied, which may underestimate the effect found. Still in this regard, the specific doctor/inhabitant rate may overestimate the real medical density, however it can be useful in certain cases – after all, there is no 100% accurate method of using the rate – this is another practical approach to better understand the current medical landscape.

CONCLUSION

Chronic non-communicable diseases represent a major burden on the Brazilian public health system, with more than 2 million deaths in adults during the period studied. There was a predominance of deaths and expenditure on men, as well as a negative association between the level of inequality and NCD outcomes in the country, with no specific relationship with the number of doctors working in the Public Health System.