Internações hospitalares por causas sensíveis à atenção primária no Distrito Federal , 2008 Hospitalizations due to ambulatory care-sensitive conditions in the Federal District , Brazil , 2008

RESULTADOS: As causas sensíveis à atenção primária (CSAP) representaram cerca de 20% das internações no Sistema Único de Saúde. As causas mais frequentes foram: gastroenterites (2,4%), insufi ciência cardíaca (2,3%) e infecção do rim e trato urinário (2,1%). Constataram-se coefi cientes de internações por causas sensíveis à atenção primária relevantes no grupo infantil (< 1 ano), redução importante nos grupos etários seguintes (um a 29 anos) e aumento gradativo até as idades mais avançadas. Comparados aos dos homens, os coefi cientes de internações foram discretamente maiores em mulheres jovens (20 a 29 anos) e menores em mulheres com mais de 49 anos.


INTRODUCTION
Primary care in health promotion and the prevention of disease and hospitalizations, disability and early deaths, is greatly important.a Structuring the primary health care services based on the Family Health Strategy (FHS), and covering the health needs of the population, is the main strategy in the reorientation of the health care model in Brazil.Good results are obtained in many countries, where health systems are driven by FHS.b Even so, there are few evaluative tools that support performance monitoring in the primary health care, thus limiting the implementation of any necessary adjustments.
A list of hospitalizations due to Ambulatory Care-Sensitive Conditions (ACSC), published in Ordinance nº.c ACSC comprise a number of health problems that can be prevented by timely and effective ambulatory care, control of acute events and managing of chronic diseases or conditions. 2,4,7It refers to a set of events that would hardly progress to the point of requiring hospitalization, if properly handled through health promotion, prevention, early treatment and monitoring, at ambulatory level. 3spitalizations for infectious diseases preventable by immunization (measles, tetanus, diphtheria and others) can be avoided. 7It is also possible to avoid diseases, such as gastroenteritis, whose complications can be alleviated by early diagnosis and treatment.If there is coverage and quality in primary health care, there will be a reduction in hospitalizations for acute complications caused by non-transmissible diseases (such as the diabetic coma), as well as a decrease in readmissions and length of hospitalization period due to different diseases. 3spitalizations due to ACSC has become a valuable tool for monitoring access to services and quality assessment of primary health care in the world, over the last decade.a With the adaptation and validation of a Brazilian list of hospitalizations due to ACSC, the study of these events is an important tool for health researchers and managers to assess the performance of primary health care services in Brazil.

METHODS
A descriptive cross-sectional study on hospitalizations for ACSC and the total number of hospitalizations in the Federal District (DF), between January and December 2008, was carried out.
Hospital admission orders from the Hospital Database of the Brazilian National Health System (SIH-SUS), which contains information identifying patients and related treatments, was the source of information on hospitalizations for ACSC.The variables included: patient individual characteristics, clinical information and primary diagnosis for admission.
The tables containing the ACSC distribution were based on the variable "main diagnosis" from the code of the International Classifi cation of Diseases − 10 th Revision (ICD-10) and from the primary diagnosis for hospital admission, available in the SIH-SUS database.
In 2008, the number of hospitalizations through the SUS was 188,106.The hospitalizations related to childbirth (31,103) were excluded because they represent the natural outcome of pregnancy and are infl uenced by the fertilization rate, according to method adopted by other authors. 1 Thus, the proportions and the coeffi cients of 157,003 hospital admissions were analyzed.
The Microsoft Office Excel and IBM-SPSS 16.0 softwares were used for encoding and classifying the hospitalizations due to ACSC.
The study was approved by the Research Ethics Committee of the Faculty of Medicine, Universidade de Brasília (CEP-FM 001/2010).

RESULTS
ACSC were responsible for 19.5% of hospital admissions through the SUS (coeffi cient of 121 per 10,000 inhabitants -Table ).
The DF has an estimated population of 2,557,158 inhabitants d and is surrounded by 42 small and midsized towns belonging to the states of Goiás and Minas Gerais, which are also known as the entorno region.The DF has low coverage regarding the Family Health Strategy service (FHS, 5.6% in 2008) and Community Health Agents (CHA, 39 teams). 10The supply and demand for health services in the DF are affected, among other factors, by patients living in the entorno region: 11% of emergency treatments and 20% of hospitalizations in the DF are related to non-residents.e Thus, hospitalizations due to ACSC refl ect not only the fl aws of primary care in the DF, but also part of the unmet needs in health care in the entorno region.
The analysis of the main diagnoses of hospitalization due to ACSC are shown in the  Signifi cant hospitalization rates due to ACSC were observed in young children (< 1 year), with a significant decrease in the successive age groups, up to 30 to 39 years.A gradual increase in the age groups above 40 years was also observed.There were no signifi cant differences in the hospitalization rates related to gender up to 39 years of age, except for a slight increase in relation to female patients in the 20 to 29 years age group.Higher hospitalization rates due to ACSC were observed in males when compared to females, at 40 years of age or older (Figure 1).
Lower hospitalization rates due to ACSC were identifi ed among women from the 10 to 14 years to the 50 to 59 years age group.In contrast, among males, lower hospitalization rates due to ACSC is persistent only up to 40 to 49 years age group.
Hospitalizations due to gastroenteritis, heart failure, cerebrovascular diseases, bacterial pneumonia and skin and subcutaneous tissue infections occurred more frequently in males (Figure 2).
The hospitalization rate due to kidney and urinary tract infections was almost double among women, whereas the coeffi cient related to diabetes mellitus was only slightly higher in females (Figure 3).
The hospitalization rate due to gastroenteritis was more relevant in patients younger than 9 years and slightly more common in males (Figures 2 and 3).The hospitalization rate due to heart failure and cerebrovascular diseases were higher after 60 years of age, and more relevant among men.The hospitalization rate due to diabetes mellitus showed an increase in patients older than 60 years, especially among those older than 70 years, being more relevant in women.
The hospitalization rate due to bacterial pneumonia showed a slight relevance, preferably reaching the over 5 years age group and the elderly over 70 years of age to a lesser extent, and with greater relevance among men.The hospitalization rate due to kidney and urinary tract infections was relevant in all female age groups, including the 20 to 29 years age group.This cause of hospitalization could be identifi ed with some relevance in elderly men.

DISCUSSION
The results showed that the hospitalization rate due to ACSC was 121 per 10,000 inhabitants, representing 19.5% of total hospitalizations.The three leading causes of hospitalization due to ACSC were: gastroenteritis, heart failure and kidney and urinary tract infections.This set of causes for hospitalizations (35.4% of total hospitalizations due to ACSC and 6.9% of the total hospitalizations through the SUS) is considered sensitive to primary care, because the effective attention at this level of care could prevent such events.These hospitalizations represent a warning sign that should be considered by managers and health professionals in the SUS, in order to identify fl aws in primary care in the DF and entorno region.
Some authors present plausible evidence for these assumptions; however, controversial results are discussed in the literature, possibly due to diffi culties in the methodological approach. 5,8,9Through a literature review on hospitalizations for ACSC and the exposure to at least one characteristic of primary health care (18 articles in the past 13 years) it has been found that most studies pointed to a protective association of variables of structure, process and performance of primary care services, in relation to the risk of hospitalizations due to ACSC. 8 Preventive consultations (sporadic or not), vaccines and ongoing child health care were associated, in various studies, with lower risk of hospitalization due to ACSC. 8 The authors found no consistent results for associations between the care model (closer to or further away from the principles of primary health care) and the risk of hospitalization for ACSC. 8ias & Magajewski 5 analyzed the association between the adequacy of primary care and rates of hospitalization due to ACSC in the state of Santa Catarina.The authors classifi ed the primary health care of the selected municipalities as adequate or inadequate, according to coverage and quality criteria, and described the trends in rates of hospitalization due to ACSC through the SUS in these municipalities.More favorable results, but with no statistical signifi cance, were observed in the trends and magnitudes of these rates of hospitalization in the municipalities with adequate primary care when compared to those classifi ed as inadequate. 5A study carried out in Bagé, Southern Brazil, 9 showed a protective statistically signifi cant association between hospitalizations due to ACSC and residency in areas where there was "organization of the primary health care (according to the manager's assessment)", when there was a traditional primary care model. 9However, the authors show divergent results to those previously described.
In this study, 19.5% of hospital admissions through the SUS were classifi ed as ACSC.In a study performed in Brazil, 1 this proportion was 28.5%, slightly higher than the data found for the DF.In Bagé, 9  Similarly to the DF, the most frequent cause of hospitalization due to ACSC in Brazil 1 was gastroenteritis, followed by heart failure and bacterial pneumonia.In a study conducted in Minas Gerais a , the three major ACSC linked to hospitalizations were (in this order): bacterial pneumonia, heart failure and gastroenteritis.The hospitalization rate due to gastroenteritis was the highest within the ACSC identifi ed in the DF in 2008, particularly affecting children under 5 years of age.This fi nding is indicative of fl aws in preventive and curative services, in the primary care sphere, which should be timely and decisive at the fi rst manifestations of such event in this age group, thus preventing hospitalization.The oral rehydration solutionan intervention of low technological demand and costshowed great effectiveness in preventing deaths from gastroenteritis in Brazil, according to previous studies. 13he hospitalization rate due to gastroenteritis in the DF (14.9 per 10,000) represents half of the estimated rate for Brazil (34.7 per 10,000), and is slightly lower than that observed in Minas Gerais (19.5 per 10,000).a The mean estimated value of this indicator for Brazil is strongly infl uenced by regions with poor living conditions, sanitation and health care, such as some remote areas in the North and Northeast regions, which offer different living conditions when compared to the DF.Moreover, the paradox of observing such high rate of hospitalizations for gastroenteritis, the most frequent cause of hospitalization due to ACSC, in a Federative Unit with a high HDI deserves refl ection.
The high rates of hospitalization due to heart failure in the DF (14.35 per 10,000), especially among the elderly, confi rm data related to Brazil 1 (16.8 per 10,000), where this is the second most frequent cause of hospitalization due to ACSC.However, this indicator is markedly lower than the hospitalization rate for this cause in Minas Gerais (30 per 10,000).a The main risk factors known for heart failure are hypertension, smoking, obesity, sedentary lifestyle and familial history.Being the DF a young Federative Unit, the proportion of elderly (7.16% were 60 years or older in 2009) is below the national average level (10.15%), which reinforces the need to investigate the prevalence of risk factors associated with these events, as well as the quality of care received in the primary health care network.h The major hospitalization rates in the DF were due to kidney and urinary tract infections (13.2 per 10,000), especially in females in all age groups, but particularly among young women from 20 to 29 years.The magnitude of this indicator in the DF was not different from those found in other studies.a On the other hand, the hospitalization rate due to infection of the ear, nose and throat is four times higher in the DF (2.72 per 10,000) than the national average (0.6 per 10,000).This should be investigated through more detailed studies to elucidate the causes of such high rates in the DF.
The elderly showed the most relevant coeffi cients of hospitalization due to ACSC in the DF, similar to those described by Perpetuo &Wong b (2006) for Minas Gerais.The most frequent causes were heart failure, cerebrovascular diseases and diabetes mellitus.In Santa Catarina 5 between 1999 and 2004, one of the ACSC analyzed was diabetes mellitus and its association with the adequacy of primary care estimated by the coverage and quality of service.Hospitalizations for diabetes mellitus tended to have a steeper decline in the population with adequate ambulatory care than among municipalities without this service, suggesting vulnerability of this event to the good quality primary health actions.
The hospitalization rate due to ACSC among children under 1 year of age were relevant in the DF, as described in other studies. 3The most frequent causes were gastroenteritis and bacterial pneumonia, events that require low complexity attention for their diagnosis and management, when timely handled.
In general, in the absence of effective prevention of hospitalization due to ACSC, it is possible to anticipate increased hospital costs and use of hospital beds, which could otherwise be available to patients in more serious conditions.Further studies are needed to examine the factors associated with hospitalizations due to ACSC, including evaluative research on the quality and coverage of primary care available in the DF.It will also be necessary to expand the understanding on the process of health care use by the population residing in the different areas of this Federative Unit.The analysis of the "therapeutic itinerary" covered by the SUS patient in a given territory is an approach proposed to understand the complex web of events involved in the search for health care. 6To know the itinerary of the patients hospitalized due to ACSC in the DF and the fl aws in this process would be very useful.The present study has limitations.The main purpose of the data recorded in the hospital admission orders is reimbursement to the hospital for the services provided.There may be fl aws in this record (intentional or not) that may compromise the accuracy of the variables, especially the diagnosis used to identify hospitalizations due to ACSC.

Table .
Absolute frequency, proportion (%) and hospitalization rates (per 10,000 inhabitants) due to ambulatory care-sensitive conditions (ACSC), according to diagnostic groups registered through the hospital admission order (HAO) a , performed in the hospital network of the Brazilian National Health System (SUS).Brazil and the Federal District (DF), 2008.
a HAO: hospital admission order 11me authors analyze the quality of the data found in the SIH/SUS forms.Veras & Martins11analyzed the correlation of variables contained in SIH/SUS forms with data from medical records and found that, for the primary diagnosis (high frequency), this concordance h Instituto Brasileiro de Geografi a e Estatística (BR).DATASUS.Brasília; 2008 [citado 1 nov 2010].Disponível em: www.datasus.gov.br