Characterization of primary health care for patients with diabetes based on the PMAQ-AB

This article aims to present a proposal for making the instruments used in the three cycles of the PMAQ-AB compatible and to analyze the information on access, coverage, structure, organization and provision of services in PHC related to care for DM in Brazil, according to regions, from the perspective of family health professionals and users. We performed an analysis of the degree of compatibility of the PMAQ-AB questions (2012, 2014 and 2017). To analyze the temporal evolution of the components, we performed a proportion difference test. We calculated the percentage difference between the perspective of professionals and users, per year analyzed, for Brazil. In general, there was an improvement in the quality of care and examinations, except for the diabetic foot. Worse results were found for the North region in relation to the other regions. Despite the structural improvement and the quality of care reported by professionals, there are significant gaps in the quality of care for patients with DM in the SUS. In the scenario of scarce investment added to the growing prevalence of DM, obstacles become progressively more challenging. Therefore, monitoring and evaluating the quality of services provided are essential tasks of the Brazilian Health System.


Introduction
The prevalence of Type 2 Diabetes Mellitus (DM) in Brazil was 9.2% in 2013, ranging from 6.3% in the North region to 12.8% in the Southeast 1 . Of these, it is estimated that half are unaware of the disease 1,2 . The progressive growth in the number of patients with type 2 DM is due to factors such as aging, nutritional transition, and urbanization. A study by the International Diabetes Federation estimated that Brazil had the fifth largest population with diabetes (16.8 million) in 2019, being only behind China, India, the United States, and Pakistan 3 . In addition, DM is among the diseases that most cause loss of healthy life years 4 .
With access to health services, adequate treatment, therapeutic adherence, and continuous monitoring, people with DM can have a superior quality of life. However, when neglected, this disease can cause severe transient and/or permanent complications, such as neuropathy, retinopathy, blindness, nephropathy, diabetic foot, and amputations 1 . Therefore, the Ministry of Health established in the year of 2013 a protocol for screening and care of people with DM in Primary Health Care (PHC) 5 . In this protocol, a list of clinical exams and care routines were listed that, when correctly performed, allow the early identification of the disease, and its treatment 5 .
It is important to emphasize that, in addition to the population's quality of life, the adequacy of health care for patients with DM can avoid unnecessary costs for the Unified Health System (SUS) 6,7 .
The monitoring and evaluation of the adequacy of care for patients with diabetes are essential to prevent the worsening of the disease and deaths. It is known that, as well as the prevalence, care for patients with diabetes varies widely across the national territory 1 . Knowing the gaps in the health system allows the evaluation and reformulation of public policies and strategies to face the DM problem, especially in a country of continental dimensions like Brazil. However, it is essential to maintain this monitoring, since it identifies points of optimization in the health care policy.
The National Program for the Improvement of Access and Quality of Primary Care (PMAQ-AB) was established in 2011 with the objective of expanding access and improving the quality of PHC, ensuring a quality standard in the country and greater transparency and effectiveness of governmental actions 8 . The three available cycles (2012, 2014, and 2017) present valuable infor-mation related to DM. Despite its relevance, it is noteworthy that over the cycles some issues were modified, making it difficult to keep on monitoring.
It is noteworthy that in the Brazilian context, the expression "primary care" is still used in initiatives adopted by the government, in the composition of widely used abbreviations, such as PMAQ-AB, PNAB, and SIAB, which would otherwise be incomprehensible. However, we adopted the term "Primary Health Care", which is currently in the international literature 9 .
This article aims to present a proposal for the compatibility of the instruments used in the three cycles of the PMAQ-AB and to analyze the information on access, coverage, structure, organization, and offer of services in PHC related to DM care in Brazil, according to regions, from the perspective of family health teams and users.

Materials and methods
The external evaluation instruments of the PMAQ-AB from the 2012, 2014, and 2017 cycles were used as data sources. According to the National Registry of Health Establishments (CNES), in 2012, 43,947 Basic Health Units (BHU) were accounted. However, the cycle 1 database outlines 38,812 BHU, 88.3% of the forecastrecast 10 . In cycles 2 and 3, 24,055 BHU (62% of the 2012 quantity) and 30,346 (78.2%) were measured, respectively. Modules II were used, aimed at family health teams, and Module III at BHU users.

Issues Compatibility
The PMAQ-AB instruments changed in the analyzed period. When comparing cycle 1 (2012) with cycles 2 (2014) and 3 (2017), it is noted that some questions were added, deleted, or modified, making it necessary to make them compatible, with a detailed description of the criteria adopted.
We used some compatibility features between the years to ensure the greatest possible comparability, with minimal loss of information, as described below: (a) Using synonyms or similar questions without changing the meaning, or with minor changes; (b) Use of questions with varying answer types (such as switching from a single answer to a multiple-choice answer or from a categorical answer to a continuous answer); (c) Use of complementary response categories (when the question is asked in one year in a positive sense and in another in a negative sense); and, (d) The use of proxy questions of the measured construct (when the questions refer to the same or a similar concept, even though the target population is different).
In the analysis of the compatible questions, a classification was carried out regarding the degree of comparability of the same, as follows: (1) Total: When the formulation of the question is the same, (2) High: When there are minor changes in the manner of asking but no change in the meaning (for example, when using synonyms or matching similar/complementary answer categories); and, (3) Medium: When there are significant changes in the way of asking but it is still possible to compare (when the terms used are not synonymous but refer to the same concept/ object of analysis, or when it refers to a different population group). Questions with low comparability, which require greater effort to approach concepts or investigate objects, were excluded from the analysis. Some questions were kept even though they were available for only one or two years, given their relevance to the topic.
The compatibilizations carried out in Module II (teams) were described in Chart 1, and those referring to Module III (users) are found in Chart 2. In these charts, you can see the identification of the question in this article, characterized by "TQ" (Team Question) or "UQ" (User Question), followed by the original variable used in the matching, the category or categories of response used, and the degree of comparability, according to previously defined criteria.

Analysis
From the compatibility, the respective response percentages were estimated for each question, both for Module II -Team (Table 1) and Module III -User (Table 2). For both, proportion difference tests were carried out between the years 2012, 2014, and 2017 for Brazil, with a significance level of 5%.
The questions that can be paired and compared between the Team and Users modules are presented in Table 3, to facilitate the analysis. Additionally, for this last table, the percentage difference between the teams' and users' perspectives was calculated for each analyzed year, only for Brazil.
The analyses were performed using the Statistical Package for the Social Sciences (SPSS), version 21.

results
In Module 2 (Teams), of the 28 selected questions, only four remained unchanged throughout the cycles. Most of the questions (17) had high compatibility, and for 6 problems, significant changes were found, but it was still possible to compare them using a conceptual or proxy approximation. Question TQ16 about offering educational and health promotion actions for the prevention and treatment of DM was discontinued after 2012. Other questions in this module, such as TQ05, UQ06, TQ11, TQ17, TQ19, TQ20, TQ27, and TQ28, were unavailable for one of the years (Chart 1).
In Module 3 (Users), of the 26 selected questions, half did not change between cycles, 11 had minor modifications, maintaining high comparability, and two questions were discontinued as of 2012 (UQ16 and UQ17). Several questions from this module were also excluded in one of the cycles, namely UQ04, UQ08, UQ09, UQ10, UQ11, UQ12, UQ14, UQ15, UQ24, and UQ26 (Chart 2). Table 1 presents the characterization of coverage, structure, organization, and service offering in PHC according to the teams. As for coverage (TQ01), it was found that the percentage of teams that reported having a population discovered by AB in the surroundings of the territory covered little changed between 2012 and 2017, reaching 34.4% in the last year. Regions such as the North (52%), and the Midwest (55%) had higher percentages. Great progress was observed in the implementation of electronic medical records (TQ02), especially in the North and Northeast, which went from 3.5% and 1.3% of units with medical records in 2012 to 12.0% and 8.5% in 2017.
Still in relation to the structure of the units, almost all the teams reported every year that the users who arrive at the BHU spontaneously seeking care have their needs heard and evaluated (TQ03). The proportion of teams that say they provide urgent and emergency care (TQ04) increased in the period in all regions.
In 2017, only the North region (86.1%) was still not close to 100%. Since 2012, almost all teams said that their respective units had an exam scheduling center (TQ05). Regarding the proportion of teams that report having reserved spaces when necessary (TQ06), scheduling appointments and actions for users of programs or priority groups that need continued care (TQ07), and for users with diabetes (TQ08), these were at around 90% every year.   Regarding attention, the proportion of teams that say they renew prescriptions for users of continued care/programs such as hypertension and diabetes without the need for an appointment (TQ09) increased in all regions in the analyzed period, reaching 94.3% in 2017. In 2012, between 32.5% (Midwest) and 56.2% (Southeast) of the teams reported that the units had protocols for prioritizing cases that needed referral (TQ10). In 2014, this percentage almost doubled in most regions, and in 2017, it practically reached the total.
Regarding the existence of specific protocols for patients with diabetes (TQ11), in 2012, approximately 70% of the units in the country had    As for the support and tracking of people with diabetes, the percentage of teams that offered educational actions, promotion, prevention, and treatment of DM (TQ16) was 89.5% in 2012 in the country. The proportion of support groups (TQ17) increased significantly between the years in all regions, reaching 71.5%. Despite this, even in the last year, the North region presented 60% of realization, below the national measure. The schedule of ACS visits (TQ18) was reported by the teams for most units in 2012 and progressively increased, reaching almost 100% in 2017; 88% of the teams, decreased to 73% in 2014, being less pronounced in the North and Midwest regions. Despite the growth, the proportion of teams that schedule appointments for people with DM on any day of the week and at any time (TQ20) remains low, at 60% in Brazil. In the North (49%) and Northeast (51.4%) regions, this value is even lower.
Regarding the tests indicated for people with diabetes, almost all teams reported that they requested and carried out tests for creatinine (TQ21), lipid profile (TQ22), electrocardiogram (TQ23), glycosylated hemoglobin (TQ24), fasting blood glucose (TQ25) and urine (TQ26), in all years investigated. The proportion of teams that reported performing a diabetic foot exam grew significantly between the years, in all regions, from an average of 57.8% to 80.1%. The proportion of eye fundus examinations (TQ28), despite remaining low, grew in the North and Northeast regions, reaching 29% and 38.6% in 2017, while the other regions showed a decrease in this percentage. In 2017, the performance of eye fundus examinations according to the teams was 32.9% in the country. Table 2 describes demographic and socioeconomic aspects of users and information on access, structure, and adequacy of services in PHC related to care for Diabetes Mellitus. It was found that about 80% of the responding users were female (UQ01). The elderly (UQ02) account for approximately 22% of respondents, ranging from 15% in the North and Northeast regions to almost 30% in the South and Southeast.
The proportion of respondents who can read and write (UQ03) gradually increased in the analyzed period, from 84.8% to 94.6% in Brazil. Until 2014, 1/3 of the respondents had paid work (UQ04) in the country, with this percentage being higher in the South (44.1%). In 2012, the proportion of respondents who were beneficiaries of the Bolsa Família Program (UQ05) dropped for all regions in 2014 (31.6% in Brazil) and remained unchanged in 2017.
When analyzing access to the health service, it was found that between 2012 and 2017, in Brazil, more than 80% of respondents reported living close to the health unit (UQ06). In all the years investigated, approximately half of the interviewees said that they are normally able to schedule an appointment for the same day (UQ07), except for the Southeast region (28.4% in 2017). The proportion of users who said they could be heard when they go to the health unit even without   making an appointment (UQ08) was 66.5% in 2012 and rose to 74.3% in 2017.
In 2012, the percentage of users who reported that the guidelines received by professionals always meet their needs (UQ09) was 52% in the country, lower in the North (42.5%) and Midwest (45.9%). This percentage increased in 2014, reaching over 80% in all regions except the North (75.3%). In 2014, over 87% of users thought that the team sought to solve their needs in the unit itself (UQ10). In Brazil, 63% sought care the last time they had a health problem that they considered urgent (UQ11). In the North region, this percentage was lower, at 56% in 2017, and in the South, it was higher (70%).
Regarding the structure and adequacy of care, almost all respondents thought that the place for care is a reserved place with privacy (UQ12) (94.6%). The proportion of users who reported that professionals performed a physical examination in consultations (UQ13) increased from 74% in 2012 to 69,2% nationwide. The biggest decrease occurred in the Southeast and South regions, which went from 78,9% and 76.4% in 2012 to 69.9% and 70% in 2017, respectively. Receiving guidance on care for their recovery (UQ14), was reported by 73% of respondents in 2014. The percentage of those who indicated that they had always received guidance from the team on the signs that indicate improvement or worsening (UQ15) was similar (67.7% in 2014).
Almost all respondents (95.3%) stated that during consultations, professionals make notes in their medical records or files (UQ16).  Although it remains low, the percentage that said it was easy to talk to the professionals who attended them (UQ18) increased, from 46.1% in 2012 to 54.3% in 2017. In 2012, only 25% of users said that the professionals looked for them to find out what happened and to resume care when they interrupted treatment or missed appointments (UQ19). In 2014, there was a slight drop in all regions, but in 2017 the value practically doubled, reaching 51.6% in the whole country.
It was found that the proportion of patients with DM (UQ20) ranged from 9.4% (North and Northeast) to 16.5% (Southeast) in 2017, without great variation from previous years. As for the specific aspects of care for patients with diabetes, in 2012, almost 90% of users said they had consulted with a doctor because of diabetes in the last six months (UQ21). In 2014, there was a small decrease in this proportion, going to 83% in Brazil and growing again in 2017 (86.4%).
The proportion of people with diabetes who had a blood test in the last six months (UQ22) was approximately 90% over the entire period. Despite having increased, the number of foot exams performed in the last six months (UQ23) was less than 35% of the country in 2017. The North region stands out negatively, with only 25.3% of this exam performed even in the last year. Less than half of users reported having received guidance from professionals about foot care in the last six months (UQ24). The percentage of users who reported leaving the service with the next appointment scheduled was 34.6% in 2012, falling to 28.3% in 2014 and rising to 39% in 2017. Among the regions, the South had the lowest percentage in the last year, with 31.2%, and the North and Northeast had the highest, with 39.6% and 46.7%, respectively. The use of medication for diabetes was around 90% for the entire period.
The comparative analysis between the perspectives of teams and users shows that the perception of users and teams about listening and providing guidance improved between 2012 and 2017, reducing the difference between them. The teams' perspective on the provision of urgent care improved between 2012 and 2017, but there is no variation in users' perception, which makes the difference between the increase (Table 3).
Regarding the scheduling of consultations and actions for users with DM, there was no significant variation between the groups, and the difference between the perspectives remain high, being greater than 50%. Regarding the active search of users when there is an interruption of treatment or lack of consultation, the greatest difference was found between perspectives, despite the decrease reported by the team from 2012 to 2014 and the users' reports having practically doubled between 2012 and 2017 in all regions.
The question that had the least variation between team and user and in all years was about the performance of the fasting blood glucose test. There is a variation of about 10% between teams and users in performing the blood glucose test, although all the values found are high (above 87%). As for the diabetic foot exam, the teams' perspective on its performance improved between 2012 and 2017, but there was no variation in the users' perception, which makes the difference between them increase.

Discussion
PHC is a relevant tool for organizing access to services in the health system. The monitoring and evaluation of the PHC implementation process have been the best for improving the health care provided around the world. Health prevention and promotion actions are the premises responsible for their effectiveness in guiding the flow of care 9 . In this study, considerable variations were not found for PHC coverage in the three PMAQ cycles. However, as pointed out by Kovacs et al. 11 , over the cycles, there is an improvement in the PMAQ score in geographic areas with lower economic levels, pointing to the importance of implementing the program to reduce health inequality. From the users' perspective, in general, there was an improvement in the structure of health units and in the adequacy of patient care.
The findings of this article are in line with what was observed in studies that indicate a greater number of teams in the Northeast and Southeast regions, and a smaller number in the North 12 . In general, the North region stood out negatively in relation to the others in terms of access, structure, and adequacy of health care.
Given the importance of PHC and regional differences, it is essential to prioritize health care in terms of promotion and prevention, especially for patients with DM, due to the high prevalence and consequent complications of the disease, to the detriment of the absence or inadequate care. Despite this, it was observed that only 50% of users said that professionals sought them out when they interrupted treatment or missed appointments.
Adequate guidance and well-defined care flows are important. Thus, the organization of the care flow through PHC allows to save finite resources in the health area, since the cost of prevention is much lower compared to the treatment and, occasionally, the complications caused by the disease and the consequent need for re-hospitalization 6 . It is noteworthy that the costs are not limited to the financial, they also reflect a considerable social impact, as well as denote a perception on the part of the user of low quality of service provided.
In the care of patients with diabetes, carrying out periodic examinations is essential in the prevention of diseases resulting from the dis-ease. Patient follow-up, as recommended in the DM protocol in the country, constitutes a more financially sustainable strategy, as a higher cost of treatment is expected for individuals who live longer with the disease or who have developed complications, due to the greater demand for medicines and frequency of examinations [13][14][15] . As observed by Muzy et al. 1 , for exams recommended for patients with diabetes, there was a higher demand than the supply.
The performance of diabetic foot examinations has grown considerably over the years, according to the teams. However, this result is not confirmed from the perspective of users, who reported that only 1/3 of the population had been tested in 2017. Less than half of users reported having at least received guidance from professionals on foot care. The low performance of foot examinations was also reported in other studies, which pointed to a possible association with a higher incidence of foot wounds and amputations 1,16 . Consequently, the lack of tools for continuous monitoring of the performance of the exam becomes even more serious, given its relevance and low cost of execution.
The performance of eye fundus exams by the teams remained low, reaching 32.9% in the country in 2017. The literature indicates that there is a greater demand than the offer of follow-up services for patients with diabetes, such as insufficient performance fundus examination (40%), with wide regional variation (North 25% -Southeast 52%), reflecting a high prevalence of retinopathy 1,17 . It is also noteworthy the low or no knowledge about this complication among SUS users 18 .
Although the glycated hemoglobin test is one of the essential tests for people with diabetes 19 , only 70% of subjects reported having done so in the last year. It is noted that more than half of the patients diagnosed with diabetes showed changes in the exams, indicating that the disease was not controlled 1 and highlighting problems in tracking and monitoring. Deficiencies in health care for patients with diabetes lead to higher morbidity, hospitalizations (15%), and visits to emergency rooms (27% -PMAQ). Carrying out care monitoring exams for patients with DM is a primary activity of PHC and is fundamental to reducing the burden of DM and its chronic complications.
The search for emergency and urgent care is an important indicator of service quality, as well as access to medication and adherence to medication therapy. The findings point to an increase in the search for urgent and emergency services, indicating potential problems in the care process. In this sense, the increase in the search for emergency care can be used as a sentinel event and enable the inference of other important indicators to be monitored, aiming to understand the obstacles in the flow of care that have an impact on the quality of PHC 20 .
A medical consultation is a moment of meeting with the health professional and encompasses the activities of measuring low-complexity clinical parameters, nutritional monitoring, guidance and verification of therapeutic adherence, and evaluation of drug treatment and prescription adjustments. Despite the increase in the investigated period, the percentage of users who reported leaving the service with the next appointment was still low in 2017 (39%). It is noteworthy that the good organization of care and continuity of care may reflect improved access to medication and treatment 21 and glycemic control 22 , particularly in the context of chronic diseases. It is noteworthy that the renewal of prescriptions does not necessarily reflect the optimization of care, but it may indicate a weakness in the health system, pointing to inadequate follow-up.
In general, this study highlights the mismatch between the increase in the prevalence of DM and the network of services needed to monitor patients with DM. It was not observed that there was such a considerable improvement in the indicators of users when compared to teams. When analyzing the compatible questions between the groups, the difference between the team's and the user's perspective is maintained; it reduces extraordinarily little or even increases. Despite an improvement in access and structure to health services, this was not reflected in specific care for patients with diabetes. In general, the indicators that have improved for this group are those that were already at acceptable levels. The lower access to services and structure in the North and Northeast regions can be reflected in the low proportion of patients with DM, which shows failures in screening and, consequently, in the adequacy of care.
In the scenario portrayed, the situation was already worrying, but a worsening is still expected with the imminent increase in the prevalence of people with diabetes and with the amplification of gaps in health care, potentiated by the underfunding of the SUS. There is still the aggravating factor of the COVID-19 pandemic of 2019 23 , which may have led to treatment abandonment, and to a lower frequency of examinations and medical consultations 24 and changes in lifestyle 25 among people with chronic diseases.
In the context of the COVID-19 pandemic, with the overwhelming number of cases and the crisis in the health system, PHC has a fundamental role in controlling the growth curve of the pandemic and in guaranteeing the sufficiency of ICU beds as well as an alternative, safer, and more efficient way to face the pandemic, to prevent the collapse of the health system 26 . The longer the person-to-person transmission takes, the more manageable the situation will be for SUS 27 . Considering patients with DM, this follow-up is even more relevant since there are biochemical characteristics and the additional risk that this disease can determine the progression of COVID-19 28 . COVID-19 patients who have diabetes are twice as likely to develop severe disease. Likewise, these patients are also at nearly twice the risk of mortality due to COVID-19 disease 29 .
The scope of this article does not cover supplementary health, or even individuals who are not being monitored by any health service. The data presented here reflect the care provided to people with diabetes who are already part of the public service. It is noteworthy that even when dealing only with people who have already had access to the diagnosis, the monitoring conditions are not ideal, especially when comparing the perspectives of the user and the PHC teams.
Considering the three cycles of the PMAQ-AB, it is important to note that in 2012, there may have been a bias toward voluntary participation as only the units that considered themselves best evaluated adhered to the external evaluation process, culminating in better results from the teams' perspective. In the second cycle, in 2014, there was an important reformulation in the construction of the instrument, with the exclusion of specific modules on chronic diseases, which caused a loss of continuity of research. In 2017, there was a rescue of the 2012 instrument referring to these specific modules, as well as an expansion of the coverage of the teams, approaching the totality of the register.
Finally, the loss of the PMAQ-AB is regrettable because until then it was the only SUS monitoring tool for some services recommended in the care protocol for patients with diabetes, such as the diabetic foot exam. It is also worth mentioning that the loss of essential questions in some years or even the incompatibility due to changes in the instruments reduces the longitudinal of the comparisons. Despite this, from the PMAQ-AB data, this study showed regional differences, both between the perspectives of teams and users and between the years, which reinforces the potential and relevance of the continuous production of this information.

final considerations
Despite the structural improvement and the quality of care reported by the teams, there is an important distance between this perspective and that of the users. It is possible that the user does not perceive or understand so clearly the service provided by the team, or that the team reports providing a better service than what was delivered. livered. Despite this, the differences presented here are important and deserve attention as they show significant gaps in the quality of care for patients with DM in SUS.
The current health scenario in Brazil is challenging. The event of the COVID-19 pandemic that began in 2019 presents itself as another complicating factor in this context of increasingly restricted investment in the health field. Faced with all the challenges already reported by the APS regarding the treatment of patients living with DM, the current health situation particularly faces us with increasingly challenging obstacles.
From this perspective of crisis and health constraints, it is important to emphasize that the current pandemic situation that the world is going through allows us to identify points of subjection in the health system and other support structures linked to it will have to face to effectively face it.
It is worth remembering that the problems that once existed do not stop manifesting themselves, nor the impediment of new ones to arising, so that the idea of recognizing weaknesses and actions to support the current system is fundamental, to try to respond to old and recent problems.
The scenario presented in 2012, although not ideal, took place in the context of strengthening the Unified Health System (SUS). With the increasing prevalence of DM and the reduction in investment in public health, it is essential to reflect on the care of people with diabetes in the coming years. Thus, the current pandemic scenario exacerbates health conditions that were already fragile and consequently causes greater discrepancies in problem-solving and guaranteeing access to and quality of health services.

Collaborations
J Muzy was responsible for the conception, study design, data analysis and interpretation, and the review and approval of the final version of the manuscript. M Campos was responsible for the conception, study design, analysis and data interpretation. I Emmerick was responsible for the analysis, data interpretation, review and approval of the final version of the manuscript. FG Avelar was responsible for the analysis, data interpretation and general review of the manuscript.