Use of health services and medicines by hypertensive and diabetic patients in the municipality of Rio de Janeiro , Brazil

This study aimed to describe and analyze factors associated with emergency care and hospitalization of hypertensive and diabetic patients in the municipality of Rio de Janeiro, especially those related to the use of medicines. This is across-sectional study using secondary database from a household survey that approached hypertensive and diabetic patients. The outcome variables were: 1) seek for emergency careasa resultof complications related to hypertension and diabetes in the 12months preceding the interviews; 2) hospitalizationin the same period andfor the same reasons. Uni and bivariate analysis between exposure variables and each of the outcomes were performed using chi-square test at a significance level of 10%, which originated multivariate logistic regression analysis. Negative self-evaluation of health status was associated with both outcomes in the multivariate analysis. Having stopped taking the medications was associated with hospitalization and having missed a medical appointment in the last six months was associated with search for emergency care.


Introduction
Brazil has undergone a process of epidemiological transition, with Chronic Noncommunicable Diseases (CNCD) becoming the main cause of morbidity and mortality and leading to a significant economic impact, although in coexistence with infectious diseases of major magnitude 1 .
Arterial Hypertension (AH) and Diabetes Mellitus (DM), both CNCDs, are included in this context.The self-reported prevalence of these diseases was, respectively, 21.4% and 6.2% of the Brazilian population over 18 years of age in 2013 2 .The complications resulting from the lack of control of these diseases generate negative impacts both on the quality of life of the affected individuals and on the health system.Such complications were responsible, in 2012 alone, for 228,323 hospitalizations in the Unified Health System (SUS), at the cost of approximately 114 million reais 3 .This is not to mention hospitalizations in private hospitals and numerous other services performed in emergency services, currently not computed by the available information systems.
Among the factors that may contribute to the lack of control of AH and DM are those related to inappropriate use of medicines.For reasons such as the need for treatment for long periods and therapeutic regimens that include polypharmacy and high prevalence in the more advanced age groups, the carriers of these diseases have greater difficulty to adhere to the drug therapy 4 , in addition to being at risk of adverse drug events 5 .
Adherence to AH and DM treatment includes both continuous maintenance by individuals of the prescribed drug treatment and lifestyle change (LSC), from conscious obedience to guided behaviors regarding diet and physical activity.Low adherence to treatment is one of the main causes of reduction of clinical benefit and control of AH and DM, leading to health and psychosocial complications and reduction of quality of life 6,7 .Such complications may culminate with the worsening of the situation, leading to the search for emergency services and hospitalizations 8,9 .
A systematic review 10 found that different aspects of four major domains (characteristics of the health system, access, characteristics of the practices and care organization) were associated with avoidable hospitalizations in patients with chronic conditions.A study that investigated factors associated with hospitalizations among diabetic patients in the SUS 11 found that in 28% of the cases the patients reported that it was impossible to schedule a consultation at the primary health care unit, about one quarter had not attended any consultation in the last 12 months in the PHC units, and 18% reported failure in medication use.Luiza et al. 12 indicate that access to medicines may have repercussions on the use thereof, which will ultimately be expressed both in the use of health services (hospitalizations, outpatient or emergency care), and in the health outcomes per se.
In order to contribute to reduce the rate of abandonment of chronic treatments and reduce the overload on health units in the city of Rio de Janeiro, the program Medicine at Home (RECA-SA in Portuguese) was implemented in January 2002 in Rio de Janeiro and expanded gradually, starting to function in 100% of the units participating in the Hypertension and Diabetes Mellitus Programs in October 2004.
With RECASA, patients started receiving anti-AH and anti-DM drugs at a registered address.The main inclusion criterion of the program was the confirmation by a physician that the patient was in the control phase of the disease, with defined drugs and stable doses, but it is not known whether these criteria were strictly followed.RE-CASA was interrupted in 2008 due to operational problems in the delivery of medicines, and is still not working.
Primary Health Care (PHC) has been considered the best strategy for coping with most CNCDs 13 .Through its attributes, it becomes possible to employ technologies considered more effective, capable of interfering not only in healing and rehabilitation, but also in the determinants and conditioning factors of individual and collective health, aiming at the comprehensiveness of care 14 .
In this sense, PHC has gained a central role in the health system around the world, having in Brazil the role of coordinator of the health care network.
Several efforts have been made in the last decades in the search for indicators to measure the effectiveness of primary care.In the 1990s, the concept of "Primary Care-Sensitive Conditions -PCSC" was developed to denote the set of health problems for which the effective action of PHC would reduce the risk of hospitalizations and is therefore a direct indicator of the effectiveness of this level of health care.Alfradique et al. 15 have adapted the use of PCSC to the Brazilian reality and identified that in 2006, the number of hospitalizations due to PHC-sensitive conditions in the SUS as a result of AH and DM was 146,426 and 120,878 respectively, representing 1.5% and 1, 2% of all admissions in the country, excluding childbirth.These data may be due to problems in access and resolution of PHC of the country.
Given the importance of access and use of medications in the evolution of patients with diabetes and hypertension, either postponing or preventing hard outcomes or injuries of target organs, it is important to investigate this association.
The objective of this article is to describe and analyze the factors associated with the search for emergency care and hospitalization by patients with AH and DM, especially those related to the use of medications, taking as a cut hypertensive and diabetic patients attended at primary health care units of the municipality of Rio de Janeiro in the year 2009.

Method
This is a cross-sectional study of a secondary database, from a population survey carried out in the context of the research "The Medicine at Home Program (RECASA) as a model of public provision of medicines -Analysis of implantation in the city of Rio de Janeiro", conducted from January to February 2010 by the Nucleus of Pharmaceutical Assistance/ENSP/Fiocruz in partnership with the Municipal Secretary of Health and Civil Defense of the municipality of Rio de Janeiro (SMSDC/RJ).This study had two objects of analysis: 1) Basic Health Unit (BHU) level, in which interviews were made with managers and health professionals, direct observation and consultation of documents; 2) Household level, in which interviews were carried out directly with AH/DM patients enrolled in the RECA-SA program, in addition to direct observation of medicines used and stored at home for the studied diseases.In this article, we used data from the household level.
The source study sample was based on the enrollment in the RECASA program, in which the target population was 139,222 patients distributed in 87 BHUs.We used proportional sampling and, considering the multiplicity of outcomes of interest, we have established a conservative prevalence of 50%, considering a 95% confidence interval, a maximum variance of 0.25 and a sampling error of 4%, resulting in out of 598 users.Throughout the field process, replacements were required.To do so, a reserve bank based on the sample was created, following the parameters mentioned above.The final sample resulted in 547 households, with one person per household being interviewed, representing a loss of 8.5% in relation to the sample plan.The applied questionnaire contained mostly closed questions, distributed in 5 blocks (general data, data on health conditions, use of health services, experience with the RECASA program, patient's treatment), with some aspects collected through observation (verification of medicines and medical prescription at home).The source study method is presented in more detail by Mendes et al. 16 .
In the present study, outcome variables were: 1) search for emergency care because of complications related to AH/DM in the 12 months prior to the interview, as referred by the patients; 2) Admission referred to in the 12 months prior to the interview for the same reason.
The exposure variables, measured in a self-reported interview, included the following aspects: sociodemographic (age; sex; marital status; living alone; ethnicity; occupation and economy class) 17 , health conditions (disease, time of diagnosis of AH/DM -in cases in which the interviewees had both diseases, we adopted the oldest diagnosis: self-evaluation of health status (very good or good, fair or poor/very poor); degree of difficulty to perform work tasks or household chores (none, mild, fair, high, extreme)); aspects related to the use of health services (regular attendance at scheduled consultations; scheduling the next consultations; degree of satisfaction with the pharmacy service; indication of physical exercise and indication of diet by a BHU physician) and aspects related to the use of medicines (we measured the reference to problems related to adherence to the drug treatment expressed by the questions -"usually fails to take medications", "usually having medicines left over", being considered non-adherent those patients who answered always or almost always); not practicing self-medication -"all medicines used for AH/DM are prescribed by prescribers", "usually asks for information on medicines only for health professionals"); problems in medicine care at home (proportion of households with "expired drugs/ damaged packaging", measured by direct observation at home); "time without change of medical prescription; time of drug treatment".
Regarding the statistical treatment of the data, a descriptive and bivariate univariate analysis was performed between the exposure variables and each of the outcomes using a chi-square test (χ 2 ) at the level of significance of 90%, using the program Statistical Package for the Social Sciences (SPSS®) V 20.0.
The variables that showed statistically significant associations at the 90% level in the bivariate analysis gave rise to logistic regression models for each of the outcomes separately.The method used was the stepwise backward at the significance level of 10%, with an entry probability of 0.20 and an exit probability of 0.10.Their respective proportions of total logistic model classification (PCT -Overall), adjusted odds ratios (ADOR) and respective confidence intervals of 95% were presented.
The source project was approved by the Committee of Ethics in Research of the Escola Nacional de Saúde Pública Sérgio Arouca(CEP/ ENSP).The authors do not identify any conflict of interest.

Results
There was a prevalence of elderly people, with a mean age of 62.7 years (ranging from 24 to 95 years), female predominance (68%) and slightly more than half of the interviewees were married, followed by widows/widowers.Approximately 90% of individuals did not live alone, that is, they had some type of family support and 64.9% described themselves as belonging to some nonwhite ethnicity.One-third of the sample studied did not work, among whom the majority were retirees or dedicated themselves to household care.Approximately 90% of the individuals belonged to economic classes C, D or E (Table 1).
About 96% had hypertension, 20.8% diabetes and 17.2% had both diseases.Of the set of interviewees, 87.4% knew the diagnosis and 86.4% had been using drugs for these diseases for more than 5 years.Approximately 73% had not had the content of their prescription for AH and DM changed for more than 1 year (Table 1).
Sixty-seven percent self-rated their health status as fair, poor, and very poor; however, when this same question was asked considering the comparison with people of the same age, this percentage fell to 58.2%.Approximately 14% reported experiencing a high degree of difficulty when performing work tasks or household chores.Nineteen percent reported having missed a scheduled consultation at the BHU in the last 6 months and 38.2% had not scheduled the date of the next consultation.About seventy-nine per-cent said they were satisfied or very satisfied with the service performed at the BHU pharmacy.Regarding the solutions offered for treatment, 65% and 56%, respectively, reported the physicians, in addition to prescribing medications, used to recommend diet and exercise (Table 1).
About 22% of patients reported they used to stop taking their medications and 23.2% reported that they had left over medication; in addition, 38% of patients reported asking for information about medications for others who are not health professionals.Almost all said that all medications used for AH/DM were prescribed by authorized prescribers.Direct observation of these medications showed that about half of the patients had expired medications or had damaged packaging at home (Table 1).
One hundred and thirteen individuals (20.7%) reported having sought emergency services and 33 (6.0%) had been hospitalized for complications of hypertension and diabetes in the previous year.
These were significantly associated with the use of emergency service by AH and DM carriers, both in bivariate analysis and in logistic regression: fair, poor or very poor health status; difficulty in carrying out routine activities; medicine prescription changed in the last year; and having missed a scheduled consultation in the six previous months (Table 2 and Table 3).Evaluation of the health status as fair, poor or very poor in relation to people of the same age was associated with the search for emergency care only in the bivariate analysis.
These were significantly associated with hospitalization in the logistic regression: female sex, non-white ethnicity, diabetes, evaluation of health status as fair, poor or very poor, having set the date of the next consultation, having stopped taking medications, having medication left over and having not requested information about medicines only from health professionals.
In the multivariate analysis, through the logistic regression, for the outcome "admission", the factor associated with greater magnitude was "self-assessment of the health status in relation to other people of the same age as fair, poor or very poor," presenting adjusted OR of 5.3; whereas for the outcome "seeking emergency care" the factor was " high or extreme degree of difficulty in performing work tasks/household chores", with adjusted OR of 2.4.

Discussion
The sociodemographic profile found in this study is close to that identified by other studies carried out with patients with chronic diseases in Brazil regarding age distribution, with a higher concentration of elderly people 18,19 , women [19][20][21] , married people 19 or living together with a family member, belonging to economic classes C, D and E, and not included in the formal or informal labor market 21 .Likewise, the majority of the popu-lation with AH and DM in Brazil makes continuous use of medication 22 .Hospitalization due to AH/DM complications was higher among women.These are more attentive to the symptoms, seek a better knowledge of the diseases, express better what they feel and seek more frequently the outpatient health services 23 .Likewise, the literature reports higher rates of hospitalization due to preventable causes among women compared to men 24 .Probably, by seeking more health services, women are ear- lier diagnosed and referred for hospitalization, although this may result in iatrogenic effect because they are also more subject to avoidable hospitalizations.
Negative self-evaluation of health status and the difficulty in performing work tasks or household chores were statistically significant both with hospitalization and with the search for emergency services.Studies indicate that the way the individual perceives the illness or feels the disease adequately represents their health status and influences the search for care services [25][26][27] .These data reinforce the importance of medical anamnesis and patient listening.
In the present study, the data were obtained in a cross-sectional survey.Thus, the experience of hospitalization and use of emergency services had occurred in the last 12 months, while the perception of health status was related to the moment of the interview.Our findings suggest that the patients in a more advanced stage of the disease and, therefore, with a worse perception of the health status, did not have, for the most part, such a situation reduced by the experience of hospitalization or emergency care.
Individuals whose medical prescription was unchanged for less than 1 year at the time of the survey also showed a greater search for emergen-cy services, with a statistically significant difference in relation to those with a stable prescription for more than one year.The instability of the therapeutic approach suggests a lack of control of the symptoms related to AH/DM, which may have influenced the search for emergency services.Thus, the variability between prescriptions of hypertensive and diabetic patients may be an easily traceable trait event of disease severity.
The proportion of hospitalization among the individuals with DM was higher, and this difference was statistically significant (p = 0.024).Although in the present study we did not explore the reasons for admissions in the interviews, according to Suzuki et al. 28 the highest rate of hospitalization for diabetic patients arises from ineffective glycemic control, associated with the patient and family unpreparedness in relation to the disease.This lack of control contributes to the development of incapacitating complications such as ulcer in the lower limbs, diabetic foot, amputations, diabetic retinopathy, blindness and chronic renal failure leading to prolonged and recurrent hospitalizations, resulting in high bed occupancy 29 .
In our study, the mean age was 62.7 years, characterizing an elderly population, thus more susceptible to injuries due to falls, a fact that may have contributed to the greater proportion of hospitalization of DM patients.Yau et al. 30 have identified that older adults with diabetes have a higher risk of hospitalization for fall-related injuries than the elderly without diabetes.Compared to hypertension, the treatment of diabetes may be more complex due to aspects such as the great variety of complications resulting from the disease, the use of insulin with the consequent need to handle the glucometer and its inputs and the emotional imbalance 31 .
Having missed a scheduled consultation at the BHU in the 6 months prior to the interview was statistically significant both with hospitalization and with the search for emergency services.Systematic review 32 pointed out that the greater continuity of care as well as the number of consultations in primary care is associated with the lower risk of hospitalization due to PCSC, reinforcing the potential effectiveness of this level of care, a finding corroborated by a national study 33 .However, Oliveira 34 found that 57% of individuals with AH and DM who received emergency care in a Brazilian municipality reported receiving follow-up at a basic health unit as well as a visit from the Community Health Worker and the Family Health Strategy Programs.Thus, it is important to stress that the potential effectiveness of care received in primary health care is certainly linked to different factors for its guarantee.Among these, the organization and the accessibility to the services influence the continuity of the care received.Therefore, they may be related to the proportion of hospitalizations due to the primary care-sensitive cause.In a review,van Loenen et al. 35 have found that having care accessibility in the primary care system on a regular and continuous basis seems to be more important to reduce avoidable hospitalizations than how these primary care services are organized.This study also points out that the availability of primary care medical supplies and long-term relationships between physicians and patients reduce hospitalizations for chronic primary care-sensitive conditions.
There was a greater chance of hospitalization among those who reported having scheduled the next medical consultation at the BHU.It is quite plausible that the previous experience of hospitalization, probably caused by a poor evolution of the disease, increases the awareness about it and about the importance of regular care.
The consultation on the use of medicines to unqualified persons and the lower adherence to the drug treatment were associated with the greater hospitalization due to complications of AH and DM.
Problems in adherence to drug therapy are pointed out by several studies as one of the main causes for the lack of control and aggravation of CNCDs 36 .A significant proportion of outpatients have suffered adverse drug events, with several being treatable or preventable.Thus, adequate attention to problems with the use of medicines can prevent both the use of emergency services and hospitalizations 36,37 .In fact, a systematic review showed that the median prevalence of hospital admissions related to these events is 4.3%, of which 59% are preventable 38 .
Regarding the limitations, we can mention that this was a cross-sectional study, which has as one of its weaknesses the impossibility of establishing relations of causality.There were questions related to previous experiences and others to situations present at the time of the interview.The reference to previous experiences brings the problem of memory bias.However, we consider that hospitalization and emergency care use in a 12-month period are experiences that are striking enough to be remembered.Our study only addresses patients enrolled in a specific primary care program in the city of Rio de Janeiro at the time, and there was no comparison group.The initial intention of the study was to evaluate the RECASA program.However, the long period of interruption of the program, which turned out to be permanent, made it impossible to verify its specific effects.

Final thoughts
Both questions about the patient's relationship with health professionals and about the rational use of medications were shown to be associated with the search for emergency services/hospitalization by individuals with AH and DM among those enrolled in a basic health care program in Rio de Janeiro.
Therefore, there is the need for actions by both health service professionals and government institutions to prioritize health prevention activities and to address the various care needs of chronic patients and their families, reinforcing the attributes of primary health care.The changes in prescriptions are easily traceable events, so that they show good potential to be used as sentinel indicators of patient evolution, and can subsidize multiprofessional interventions.Likewise, the data of this study point to the need to greater investment in the promotion of rational use of medicines and in the strengthening of primary health care.

Table 1 .
Sociodemographic data, health status and use of health services by users of the Medicine at Home program.Municipality of Rio de Janeiro.2010.

Table 2 .
Association between the search for emergency care/hospitalization and different exposure variables in the bivariate analysis among users of the Medicine at Home program.Municipality of Rio de Janeiro.2010.

Table 3 .
Logistic regression between outcome (search for emergency/hospitalization) and exposure variables for users of the Medicine at Home program.Municipality of Rio de Janeiro.2010.