Risk strata and quality of care for the elderly in Primary Health Care

Objective to identify patterns of associations between the degree of compliance to laboratory test requests by risk strata and the parameters of quality of care outcomes in primary health care (PHC). Method a cross-sectional study involving 108 elderly patients with hypertension and/or diabetes treated in PHC. A semi-structured questionnaire and electronic medical record data were used. To evaluate the quality of care, the Patient Assessment of Chronic Illness Care (PACIC) questionnaire was used. Descriptive analysis, multiple correspondence analysis and k-means grouping were performed. Results it was observed low compliance of the care practice, standing out as the worst parameter the evaluation of the diabetic foot (2.2%). Three clusters were identified, with cluster 1 having the highest number of individuals (37.0%), with better indicators of quality of care, evidenced by above 50% of compliance with laboratory tests (75.0%), high PACIC score (47.2%), control of blood pressure (70.0%) and metabolic levels (95.0%), and satisfaction with health (92.5%) and health access (90.0%). In contrast, cluster 3 (29.6%) was made up of individuals with worse outcomes of care. Conclusion low compliance of care practice and asymmetries among health actions and users’ needs were observed, indicating failures in the care process in PHC.


Introduction
Diabetes mellitus (DM) and hypertension are currently among the most prevalent chronic conditions. Together, they are considered as the main primary causes of renocardiovascular diseases in the world scenario (1) and responsible for 13.0 million deaths worldwide in 2015, of which 7.8 million were caused by complications of high blood pressure levels and 5.2 million due to metabolic uncontrol (2) . This scenario is alarming and suggests the need to invest in improvements in the quality of management of these morbidities, especially in Primary Health Care The gateway to the Brazilian public health system (Sistema Único de Saúde -SUS) consists of the Family Health Strategy, focused on primary health care and centered on the family and the community (4)(5)(6)(7) . Its main function is to provide person-centered care, with priority in actions of health promotion, disease prevention, health surveillance, assistance, and follow-up of the enrolled population, in an attempt to impact the quality of care provided to users with chronic conditions in PHC (3) .
Some of the strategies recommended in this regard are related to the implementation of evidence-based clinical practices, as well as the monitoring of these interventions. However, although there is a high degree of agreement on the best practices of prevention, diagnosis and treatment of DM and hypertension, as described in several clinical guidelines and protocols, the use of these practices is still incipient, especially when it refers to the implementation of actions according to risk stratification of the population addressed. This becomes a barrier to achieve better care results based on current scientific knowledge (8)(9)(10)(11) .
A Brazilian study on the quality of care for the person with DM at a basic health unit in the city of Ribeirão Preto, São Paulo, based on 138 medical records, revealed that the actions implemented in the care are fragmented. No body mass index record was found; feet and ophthalmological assessment was recorded in only 15.2% and 4.3% of medical records, respectively; and only 1.4% of the results of laboratory tests had registered microalbuminuria levels (12) .
These results would probably be more alarming if service provision according to the needs of the people addressed was taken into account. This was a concern already expressed by a scholar (13) who alerted to the possibility that users with greater clinical risk might not be prioritized in the offer of actions to their needs, since health systems still chose planning of supply to the detriment of needs that is differentiated according to risk stratification. This aspect deserves attention, since, as a consequence, efforts and resources can be put in unnecessary, misguided, and inefficient interventions; reason why this issue has to be better explored.
Based on the above, this research hypothesizes that there are asymmetries between health care provision by the Family Health teams and the needs of the elderly with DM and/or hypertension, which implies a less favorable scenario for the segment with worse care results.
Currently there are not previous studies that have been conducted on the association between quality of care outcomes and providers' compliance with care protocols for the elderly with DM and/or hypertension, considering the request for laboratory tests by risk stratification of individuals.
Therefore, the objective of this study was to identify patterns of associations between the degree of compliance to requests for laboratory tests by strata of cardiovascular risk and metabolic control, as well as the parameters of results of quality of care provided to the elderly with chronic conditions in PHC.

Method
This cross-sectional study is part of the second phase of the population-based study on "Aging and Kidney Disease" (en-DoRen) at the regional level, conducted from August 2014 to January 2017 in one of the nine health districts of Belo Horizonte, Minas Gerais, Brazil (14) .

The baseline sample from the "Aging and Kidney
Disease" study is comprised of 300 randomized individuals. This analysis considered the data of the participants who met the following inclusion criteria: age ≥ 60 years; self-report DM and/or hypertension or having a confirmed diagnosis in an electronic medical record; having been followed for at least one year by the Family Health team of that district; having attended some type of health service of the basic health unit (BHU) in the last three years; having the electronic medical record located. The elderly with severe cognitive impairment (Mini-Mental State Examination -MMSE ≤ 9) were excluded from the analysis. www.eerp.usp.br/rlae 3 Silva LB, Silva PAB, Santos JFG, Silqueira SMF, Borges EL, Soares SM.
The sample size was based on conservative criteria, adopting a prevalence of 50% of the different outcomes (care parameters per cardiovascular risk stratum and metabolic control), which delimits the maximum variability of the sample size. Using the calculation method proposed by Lwanga and Lemeshow (15) and considering absolute accuracy of 10%, with a significance level of 5%, a sample size of 96 individuals was found.
Adopting 10% for possible losses, the total number was 106 people.
At the baseline of the "Aging and Kidney Disease" study, 143 PHC users were identified, of whom 118 were diagnosed with DM and/or hypertension and 10 participants were excluded because the electronic medical record was not found (n = 1) and because the time since the last visit to the BHU was superior to three years (n = 9). After compliance with the inclusion and exclusion criteria, the sample was estimate for 108 elderly people.
The data collection involved a household survey conducted by two of the authors and six previously trained scholarship students. A semi-structured was used, as well as a pre-tested questionnaire containing sociodemographic data (sex, age, schooling, income), behavioral data (smoking), clinical data (self-referenced morbidities, MMSE), satisfaction levels (satisfaction with health and satisfaction with access to health services) and quality of care in the PHC (Patient Assessment of Chronic Illness Care -PACIC scores), in its version translated and adapted into the Portuguese language (16) .
Anthropometric data were obtained at participants' homes and, subsequently, the body mass index was calculated. Blood pressure levels were also measured and biological material was collected to measure the glycemic, lipid profile and renal function parameters.
Further details are described in a previous publication (14) .
After home survey, secondary data from electronic medical records were collected in January 2017. A standard form was filled with information on morbidities; care parameters, including number of medical and nursing consultations and tests (electrocardiogram, fundoscopy, chest X-ray, and diabetic foot evaluation); date of request of the laboratory tests of interest, as well as those examinations requested, but that for some reason were not performed. The overall cardiovascular risk classification was categorized as low, moderate, and high risk, according to the protocol of the Municipal Health Department of Belo Horizonte, Minas Gerais (18) , for its simplicity and ease of adaptation to the municipal resources.
In particular, the degree of compliance to laboratory test requests was calculated by dividing the sum of the laboratory tests in accordance with the above mentioned protocol (17) by the total number of recommended tests, and expressing it as a percentage.
For people with hypertension, the total number of tests recommended was nine and for people with DM, ten tests. Subsequently, they were categorized into: 0% (T1) -(no tests completed); 1-50% (T2) (1 to 4 tests met for hypertension and 1 to 5 tests met for DM); 51-100% (T3) (5 to 9 tests met for hypertension and 6 to 10 tests met for DM).
The quality of care was conceptually supported in two Donabedian dimensions (19)   to say that the proximity between the points reflects similarity or association, whereas the distancing of the points is considered non-similar (21) .
The implementation of the MCA was based on the presence-absence matrix structure of the data, in which the n individuals (matrix rows), characterized by m attributes, that is, the variables of interest (matrix column) are arranged. Each axis in the graph explains a percentage of the total data variability (inertia) (21) . For the interpretative process of the MCA, the steps described by Carvalho (21) were followed. The The graphical representation of the fusion coefficients of both methods indicated a marked fall in the distances between the coefficients up to the third cluster, being considered as an optimal cut-off point (21) .

Results
The study sample consisted of 108 elderly people,

Discussion
This study, including the elderly with DM and/or hypertension followed in PHC, showed low conformity of the care practice according to stratification of cardiovascular risk and metabolic control, which affects the quality of care achieved. Also, three clusters of individuals with distinct profiles were identified, being the first and third cluster formed by the elderly that concentrated the best and the worst quality indicators of the care, respectively.
Even with different methods, which limit the comparison between the data obtained, other national and international studies also emphasized the incipience of the use of the clinical protocol of DM and hypertension recommended for the improvement of the assistance by physicians and nurses working in PHC (10)(11)(22)(23) .
In this sense, challenges still need to be overcome to ensure improved care for people with chronic health conditions. Mainly regarding the periodicity of laboratory tests, the present study identified percentages that were to the health needs of the user (24) .
In general, there was more compliance to hypertension recommendations when compared to DM recommendations, whose highest proportion was found  (25) .
In the literature, there have been successful experiences on monitoring the absenteeism rate, such as those using information systems to identify the user in detail and the organization of the services. They also include monitoring of individual care targets and rate of service utilization. However, for such a strategy to be translated into practical results, it is necessary for the information system to promote the aggregation of the clinical data of the users, in order to assist the planning process, the recognition of risk groups, with special needs, and the management of comprehensive care, including the issuance of warnings and alerts for the maintenance of bond with the user (23) .
Regarding the nursing consultation, there was a low systematic record about this care practice in most of the medical records, being below that defined by the protocol (minimally performed semiannually) (17) . Thus, the aim is to guarantee management based on the health needs of the population (28) , taking into account one of the most important principles of the SUS, namely, equity (13) .  (24) .
In addition, it is important to highlight the role of specialized ambulatory care units that are part of the network, whose performance is not limited only to their assistential role, but also in the supervision and permanent education of PHC professionals. From this perspective, it is necessary to identify which professionals have these skills developed and, once identified, provide the necessary means to organize the professionals' agenda in order to guarantee time and space for these activities. Previous experiences show that specialized professionals are crucial in the permanent education of general practitioners (13) .
In addition, this study may support the nurse's role, motivating them to play the role of articulator of the work process in the Family Health Strategy, helping to rethink the provision of care to the elderly who experience a chronic health condition, encouraging the performance of a multidisciplinary team truly committed to an equitable, comprehensive, and resolutive practice. In this context, not only nursing, but all professionals should fully accomplish their duties and competencies, recording them in medical records, valuing the information contained in these documents, thus improving the quality of the information generated.
Finally, this study reinforces the need to institutionalize health assessment policies as a process of transforming PHC practices and strengthening those already existing in Brazil, such as Administrative Rule no. 483, of April 1, 2014 (24) , that in redefining the Health dissemination of reports with the data and indicators monitored is a strategy that can effectively systematize these policies (29) .
The results of the study should be interpreted with caution because of its transversal design, which makes it impossible to establish a temporal and causal relationship between care parameters and the quality of care in PHC. Another limitation refers to the secondary data obtained from electronic medical records, which depends on the quality of health professionals' records (information bias) and the type of analysis applied, without any control of the confounding factor. However, these limitations do not make the findings unviable, rather, they may help managers and health professionals in strengthening policies aimed at assessing the quality of care in PHC.

Conclusion
The study showed low conformity of care practice, with emphasis on the evaluation of diabetic foot and the request of specialized tests, indicating failures in the process of care in PHC. It was also verified that the clustering technique proved to be interesting as a clinical management tool, allowing the identification of distinct groups within the same health service, consequently directing specific interventions.
It was also possible to confirm the existence of asymmetries between health care provision by the Family Health teams and the needs presented by the elderly with DM and/or hypertension, presenting a less favorable scenario for the segment with worse care results.