Home care in the Federal District : factors associated with the first occurrence of acute lower respiratory infection and death

Introduction: We aimed to evaluate the frequency and associated factors of acute respiratory infection and death among home care patients in Sobradinho/DF, Brazil. Methods: Data were obtained from patients’ medical records. Bivariate and multivariate analyses were performed. Odds ratio and 95% confidence interval were estimated. Results: Factors associated with respiratory infection were ages <40 and >80 years and accommodation. Female sex was a protective factor. Age >30 years and dependency for daily activities were associated with death. Conclusions: These factors can improve clinical results in this specific home care and open new opportunities and questions for future research.

Home care encompasses actions ranging from promotion of health; disease prevention, treatment, and rehabilitation; until palliative care 1 .Health actions are carried out in the patient's home by a multi-professional team trained for this purpose as well as the patient and his/her family 2 , with doctors and nurses being obligatory members of the team 3 .
In the 1980s, the discussion about caring for elderly patients with non-communicable chronic diseases within the public health system of the Federal District [Distrito Federal (DF)], Brazil, was intensified.The situation was overburdening hospital services, not only due to the complexity of assistance needed, but also due to the longer hospital stay, posing a risk for hospital infections 4 .In 1994, the Regional Health Authority (NRAD/RSS) in Sobradinho/DF, established a Home Care Public Program.Patients with non-communicable diseases with comorbidities who needed services beyond those provided by primary care centers attended the program.For example, patients with a disease that has gone beyond the curative possibilities or those with a moderate to serious level of functional dependency in activities of daily living (ADL) 5 .
From 2003 to 2010, 973 patients attended the program, of whom 750 (77%) were >60 years old.Among the elderly patients, 306 (41%) presented with a secondary diagnosis of infectious disease (ID) during their assistance in the program 6 .The accurate diagnosis of IDs is of great clinical importance, especially in elderly patients receiving home care as well as in institutions providing home care.IDs account for 1/3 of the mortality in individuals aged >65 years 7,8 .
An analytical cross-sectional study was conducted to estimate the frequency of the first occurrence of acute lower respiratory infection and death -independently of the its basic cause -and its associated factors in patients who attended the Home Care Public Program of the Regional Health Authority in Sobradinho (NRAD/RSS) in the DF, Brazil, to subsidize health authorities and improve clinical results.
The definition of first occurrence of acute lower respiratory infection case was of an infection not acquired during hospitalization.This infection was diagnosed based on the patients' clinical or laboratory records and appeared at least 30 days after the patient attended the program.In the program, the medical professional makes a diagnosis considering the patient's clinical state: presence of fever, changes in patient's general condition, reduced appetite, appearance or worsening of findings upon pulmonary auscultation, mental confusion without neurological disease, and presence of alterations in the leucogram.
Data were obtained from the patients' medical records.The study was conducted in the administrative region of Sobradinho.The study involved 973 patients who were admitted to Home de Barros LN and de Oliveira MRF -Home care: infection and death Care from June 1, 2003 to June 30, 2010.As an assumption of the study, the first admission of each patient was considered.This information was not used for patients whose records were not found in the files or in patients whose medical records contained unclear data.Data regarding patient's admission to the program was obtained by a single observer using a standardized questionnaire.
For the respiratory infection outcome the total number of patients (973) was studied.Regarding factors associated with death, patients who died or were discharged from the program after a clinical improvement (control group) were studied, a total of 730 patients, after excluding.patients who were still being actively monitored by the program at the time of the study.
The frequencies of the following variables were obtained and calculated: 1) from the individual: age, sex, and level of functional incapacity for activities of daily living (ADL), measured during the admission on the program and using the scale validated by the Spanish Red Cross 9 ; 2) from the environment: water treatment (water supplied by the public network or by other informal sources) and household waste and sewage disposal methods (both public collection and other informal destinations); 3) from the residence: presence of domestic animals and patient's type of accommodation (individual or shared room).The Spanish Red Cross scale was used to classify the patient's level of dependency, ranging from Level 0 (independence for ADL) to Level 5 (total dependency for ADL).To classify the dependency for ADL in the analyses, the following categories were adopted: non-dependent patient (Levels 0, 1, and 2) and dependent patient (Levels 3, 4, and 5).
A bivariate analysis was conducted to evaluate the association between studied variables and patient outcomes.A multivariate binary logistic regression model was constructed.Odds ratios (ORs) and 95% confidence intervals (95% CIs) were used to measure the association between variables.The binary logistic regression model adopted the method of sequential entry controlled by the researcher using the following criteria: a) category of the variable presented in three levels in the following order: characteristics of individuals, residence, and environmental control services for water, sewage, and household waste; and b) probability of the association -the variables were applied in the multivariate model when they presented a statistical association with patient outcomes (p < 0.15) in the bivariate analysis.After analyzing each level, variables with a p-value ≥0.20 were removed after fitting to continue the analysis.The variables that had a statistical association with the outcome remained in the model, having a p-value ≤ 0.05.Stratified analyses were performed to estimate the existence of interactions between the variables that, in the multivariate analysis, continued to be associated with the outcome at a p-value ≤ 0.05.The goodness of fit in the multivariate analyses was analyzed using the Hosmer-Lemeshow test.Statistical analyses were performed at a 5% significance level.

Ethical considerations
The study was approved by the Committee for Ethics in Research of the Foundation for Teaching and Research in Health Sciences, State Health Secretariat (Nº 0040, November 8, 2010).
Of the 973 patients, 111 experienced the first occurrence of acute lower respiratory infection.To estimate the association between ages and presence of the first occurrence of acute lower respiratory infection, patients were categorized into the following three age groups: ≤40 years old, 41-80 years old, and ≥81 years old.
The 41-80-year age group was considered as the reference category, with the lowest frequency of the first occurrence of acute lower respiratory infection based on the data description.The two remaining age groups were significantly associated with the presence of the first occurrence of acute lower respiratory infection, with ORs of 3.15 (95% CI=1.83-5.45 and p<0.001) and 1.87 (95% CI=1.20-2.83 and p=0.005) for age groups ≤40 years and ≥80 years, respectively (Table 1).
In the final model, the same variables remained associated with the first occurrence of acute lower respiratory infection (Table 2).
In the bivariate analysis of death, patients were categorized into three age groups: up to 30 years, 31-59 years, and ≥60 years.The up to 30 year age group was considered as the reference category.The 31-59 year and ≥60 year age groups were significantly associated with death.Being male and having a greater degree of functional dependence (ADL= 3,4, and 5) were also associated with death.Protective factors were shared accommodation and sewage disposal system outside the public network (Table 3).
In the multivariate analysis, the factors significantly associated with death were as follows: age >30 years, male sex, and greater degree of functional dependency.Shared accommodation was a protective factor (Table 3).In the final multivariable model, the factors significantly associated with death were age >30 years and greater degree of functional dependency (Table 2).
The products of interactions between the variables present in the final models were not introduced in the logistic regression analysis (homogeneity test non-significant at a p-value of 0.15).
This study describes some factors associated with the first occurrence of acute lower respiratory infection within the context of this health assistance modality.Notably, it is uncommon in Brazil to incorporate practices in home care services to prevent and control infections.There were only a few national studies that used this modality to determine the risk factors for infections 10 .The results of the study showed that being female was a protective factor for acute lower respiratory infection in patients under home care.In 2007, 733,209 patients in Brazil were hospitalized due to pneumonia, according to the Hospital Information System of SUS Hospital Information System of SUS.Majority of these patients were men 11 .This appears to be a worldwide tendency.
The fact that 11.41% of patients from this study presented a diagnosis of acute lower respiratory infection while they were on the program shows the importance of early identification of symptoms.Infectious diseases that affect the respiratory system are particularly worrying in elderly patients (aged >60 years) and are among the leading cause of admission.This group is the most susceptible to influenza and pneumonia that can lead to death 12 .It should be emphasized that a shared accommodation  constitutes a risk factor due to the agglomeration of people in the same environment, in a low socio-economic groups.Acute respiratory diseases are potentially preventable by home care, for example, an individual room must be provided; the environment should be kept clean and well ventilated; the patient should not remain completely immobile in bed to prevent the accumulation of secretions in the upper respiratory tract; and attention should be paid to factors associated with infection 13 .A greater degree of functional dependence was a factor that was independently associated with death, indicating a possible relationship between basic non-communicable diseases, loss of autonomy, and death.Regarding dying at home, according to a report from the World Health Organization, home is a place of emotional and physical associations, memories, and comfort 14 .Dying at home is sometimes recommended by professionals who work in home care facilities.Whenever feasible, with a guarantee that the patient will receive a good assistance at home, this alternative is suggested to family members, because this is a more dignified death, even if suffering is involved in the process of becoming ill.
This study had some limitations.Precluding the generalizability of the results are the absence of diagnosis by imaging to confirm the presence of a respiratory infection and the collection of secondary data from patients' medical records; furthermore, it was impossible to analyze the association between respiratory infection and vaccination against influenza, as this information was not available in the patients' records.One also has to consider the absence of data from other episodes of acute lower respiratory infection in individuals who had the first diagnosis of the disease and the lack of a description of the time elapsed between the patient's admission and the diagnosis of infection.Moreover, it should be emphasized that the study of comorbidities, which can act as risk factors for acute lower respiratory infection and death, was not the objective of this study, which also constitutes a limitation on the data.It is hypothesized that some of the factors associated with acute lower respiratory infection and death, in the multivariate models, may have been affected by the confounding effect of unstudied variables, such as socioeconomics, nutritional factors, use of medical equipment and procedures, presence of comorbidities, and other characteristics related to the caregiver.

TABLE 1 :
Results of bivariate and multivariate analyses on the outcome of the first occurrence of acute lower respiratory infection, based on the characteristics of the individual, residence and environmental control services (Sobradinho-DF, June 2003 to June 2010).

TABLE 2 :
Final logistic regression model of the multivariate analysis of studied outcomes based on the characteristics of the individual and residence (Sobradinho, DF, June 2003 to June 2010).