Discussion
The CATPCA results identified patterns (principal components - PC) for evaluating the adequacy of structure of PHUs and the work process of the family health teams in the primary care for NCDs in the state of Tocantins, Brazil. Three patterns were identified in the PHUs analysis and two patterns in the work process analysis of family health teams. These PC were composed of the variables that showed the greatest variability among the PHUs and among the teams evaluated, and those that better explained the variability in each PMAQ-AB cycle. However, it is important to highlight the need for further studies of this same methodology, which could vary for each Brazilian state.
The variables of the PC varied much more among the PHUs than among the teams, as observed from the high factor loadings found for the PHUs. In the analysis of PHUs, the three components obtained by the CATPCA (medications for the treat of NCDs, materials for cervical cancer screening, and provision of vaccinations and infrastructure) were similar for the two cycles, which favored the comparison among all of them in the assessment of adequacy. For the teams, two components were obtained by the CATPCA (health promotion and health education for people with diabetes and hypertension for Cycle 1 and health promotion and health site analysis and health education and user referral for Cycle 2), with different variables, if compared in one cycle and another. Variables related to prevention, integral health care and NCDs management, according to the logical model, did not remain in the final CATPCA model because they did not have a sufficient factorial load; therefore, they were not compared between Cycle 1 and Cycle 2.
Evaluation of the structure of PHUs indicated the low availability of medications for the treatment of NCDs. The only structure variables that increased significantly in the studied period were the percentages of captopril and seasonal influenza vaccinations and the number of sterilization, procedures, dressings and inhalation rooms. The work process of family health teams showed significant improvement regarding health promotion activities.
Tocantins has the highest national coverage of the FHS [
7], which is typical of underdeveloped regions such as the northern and northeastern regions of the country, where the Human Development Index (HDI), which is a proxy for socioeconomic conditions, is lower [
30]. Increasing FHS coverage in these poorer regions, with the aim of reducing health inequities and increasing access to Primary Health Care [
7], has not been sufficient to overcome the precarious conditions of the PHUs in these regions. They have the lowest scores of any region for infrastructure and lack consumables, medications and Information Technology (IT) equipment, such as computers and Internet access, with even worse findings for rural areas, indicating important social and structural inequalities in the municipalities [
31], which it is not the case for the municipalities with lower FHS coverage that have a higher prevalence of adequate PHUs [
32].
The availability of medications for the treatment of NCDs in the PHUs in Tocantins is below 50.0% for all types of medication studied and falls far short of the 80.0% recommended by the World Health Organization [
33], with insulin present in less than 40.0% of the PHUs. Studies evaluating the PHUs in the entire country have also demonstrated an availability of medications for the treatment of hypertension and diabetes below 50.0% in Cycle 1 of the PMAQ-AB [
34] and approximately 40.9% for the treatment of diabetes in Cycle 2 [
32]. The free provision of medications for the treatment of NCDs has increased in the country since 2011 through the Popular Pharmacy Program and expanded the access of users [
3]. However, several factors are related to adherence to treatment for hypertension and diabetes in Primary Health Care [
35], including the proximity of the primary care source to the users of the health service, a conditional factor for people with diabetes who use insulin [
36]. Furthermore, because Primary Health Care is well positioned to tackle NCDs as a gateway to the health system [
37], the PHUs are care points that are strategically located close to the ascribed population and may promote greater access to medications for users with NCDs who attend the same primary care source over time.
Similarly, screening for cervical cancer can be carried out adequately and efficiently in PHUs [
37]. The significant reduction of the screening materials identified in this study, despite having high availability in the PHUs of Tocantins, is worrying, since the maintenance of their adequacy over time is expected, especially considering that these materials are of low cost and low technological complexity. Similar studies have identified improvements in the quality of the screening due to the adequacy of the PHUs structure and the work process of the family health teams [
38]. The adequacy of the PHUs structure across the country was of 50.0%, with missed opportunities for screening occurring in the presence of persistent problems in the structure and work process [
39]. It is likely that the high prevalence in the provision of these materials was induced by PMAQ-AB [
11] because the Program guarantees greater financial resources to the PHUs that present improvements in the structure and work process of family health teams. In addition, cervical cancer screening is a priority for FHS in the primary care context in Brazil [
38,
39].
About PC3 (provision of vaccinations and infrastructure), the high rate of availability of the influenza vaccine in Tocantins is an important marker for the adequacy of the structure of the PHUs, since vaccinations against influenza and pneumonia have been shown to be effective in preventing complications and hospitalizations among users with chronic obstructive pulmonary disease [
40]. The observation that the PHUs have a sufficient amount of available vaccines can be an important factor for high vaccination coverage in Tocantins. The state has one of the highest vaccination coverage rates in the country (89.1%) according to findings from a similar study, which showed a national rate of 61.8% [
41] and influenza vaccination coverage in priority groups exceeding 70% between 2010 and 2012 [
42]. International studies have reported lower influenza vaccination rates in a Primary Health Care service of the United States of America [
43] and ones similar to those of Tocantins among users participating in a study performed in the city of Birmingham in the United Kingdom [
44].
An improvement in the infrastructure of the PHUs in Tocantins can be considered, which increased the percentage of sterilization, procedures, dressings and inhalation rooms in the PHUs in Cycle 2 of the PMAQ-AB. Some national studies, using data from the PMAQ-AB, related the improvement of the infrastructure of the PHUs to the increased proportion of some types of rooms to perform procedures and services for the users [
31,
32,
34]. The improved infrastructure of the PHUs was probably driven by the increase in financing for construction and PHUs renovation through the PHUs Requalify Program [
31], which began in 2010 throughout the country. In addition to the sterilization room, the findings of this study highlighted other important variables that did not present a significant difference in the period but were in the final CATPCA model, such as the vaccination room and provide vaccination, which are considered to be quality markers of the PHUs structure in the country [
31]. This important result found in the statistical approach used in this study that can be useful to evaluate the adequacy of structure in other contexts.
Despite an absence of significant differences in the period under study, lack of Internet access for just under 50.0% of the family health teams in the PHUs investigated is worrisome, as there is evidence confirming positive results from the use of IT in primary care for NCDs. Overall, this finding reflects the incipient process of IT incorporation into Primary Health Care in Brazil in terms of structure, with 41.9% of the family health teams working in contexts that present low levels of IT [
45]. The use of IT is fundamental in redesigning the primary care model for NCDs, especially considering its contribution to decision support and the implementation of clinical information systems [
46]. The implementation is mainly carried out through records and adherence to clinical protocols [
47] and it is fundamental for improving the performance of the health system. In a study carried out in 22 European countries, the need to improve the information infrastructure in the Primary Health Care of the countries was highlighted to achieve these objectives [
48]. Strategies such as the use of reminders regarding medication treatment [
35] and provision of the influenza vaccination [
49] through interactive voice response (IVR) have proven to be positive. Furthermore, there is an association between IT incorporation and quality of care by family health teams, with better quality certification results in the PMAQ-AB [
45].
Although the analysis of the work process of family health teams did not result in PCs with equal variables between cycles, as it happened in the PHUs analysis, it is worth highlighting those variables that showed significant differences in adequacy between Cycle 1 and Cycle 2 of PMAQ-AB. Analysis of the work process indicated health promotion as a key point in the work of these teams. The variables that significantly increased beetwen cycles were health education for men, registration of schoolchildren with health needs, evaluation of user satisfaction, health education schedule and physical activities and user receives referral form to seek scheduling.
These findings assume that family health teams are seeking to create greater links with the community through activities performed outside the PHUs environment, especially as schoolchildren were the target of the actions of one third of the teams in Cycle 2. This finding may be related to the advances of the School Health Program in 2008 [
50], which encouraged health teams to visit schools and carry out health promotion activities among schoolchildren.
In a similar study using the PMAQ-AB, the data did not differ significantly from the findings of the present study. The majority of the family health teams reported carrying out health promotion actions, such as school activities and physical activity [
51]. The findings of this study regarding health promotion are consistent with results from other studies in which health promotion was also identified as an important strategy in chronic care model evaluation contexts, with improvements in health education processes [
52], physical activity [
53] and satisfaction of the users [
54]. Brazilian studies evaluating the chronic care model in the FHS identified improvements in links with the community [
55] and positive changes in the care performed by the teams [
56].
Regarding health education, the findings of this study do not allow an evaluation of the nature of the health education carried out by the teams. It is not known whether these actions involved potent educational processes to encourage the change of behavior of the users or merely informed them about the damage of certain habits that are harmful to health. It is known that the health education and health promotion programs offered in PHUs are directed more toward adults and elderly individuals, usually users with hypertension and diabetes [
57], which justifies the high percentage of health teams developing activities for these groups in Tocantins in the two cycles of the PMAQ-AB. Furthermore, based on the National Health Survey (
Pesquisa Nacional de Saúde), more than 80.0% of users with hypertension and diabetes who used public or private health services were given advice regarding adequate diet, body weight and the practice of physical activity. However, adherence to beneficial practices such as healthy eating and physical activity was not greater among these users, although they prioritized avoiding certain harmful habits, such as the use of tobacco products. This finding showed that these specific groups preferred to avoid harmful habits than to adhere to healthy practices [
58].
The variables of health education for women, for older adults and for addressing healthy were those that decreased in the period. It is likely that family health teams have performed fewer health education actions for those ones in order to have more time to target these actions to men users. Historically, Primary Health Care in Brazil has been little directed to men’s health, with greater coverage for children, women and the elderly [
59,
60]. The increase in health education for men may be an indicator of improved access for this population to the FHS.
The variable user receives referral form to seek scheduling is an important variable related to care coordination, which can become more effective when access to scheduling is guaranteed, even if scheduled by the user himself or by the PHUs. It is known that referral of users can be more effective if requested by family health teams and if the waiting time of the user to schedule appointments or specialized exams is shorter as possible [
9].
Even though there was no significant difference between the PMAQ-AB cycles, self-management support is an important variable for evaluating the adequacy of the primary care for NCDs. However, this variable was not performed by almost 40.0% of the family health teams in the two PMAQ-AB cycles. Self-management support is one of the key components of the chronic care model that aims to improve chronic care in the Primary Health Care and to contribute to the redesign of health systems aimed at NCDs [
60].
This study did not intend to investigate the possible interactions between structure and process to achieve expected outcomes in relation to the primary care model for NCDs. However, the adequacy of the structure is an element that contributes in favor of the work process of the family health teams in their various work contexts. According to the Donabedian quality evaluation model, a good structure increases the likelihood of a good process, which in turn increases the likelihood of a good result [
14]. A study in South Africa found that, regardless of structure, a good process mediated the relationship between a good structure and a good outcome in the Primary Health Care model in South Africa [
61]. In fact, PHUs with significant structural deficiencies make it difficult to retain professionals in the work process of the family health teams [
62,
63]. Furthermore, the final analysis showed that the deficiencies can make it impossible to implement the necessary advances in the primary care model, with an emphasis on NCDs.
There are a number of limitations to this study. Of particular concern was that the database consisted of secondary data and there were too many variables to choose for the study. The variable human resource was excluded because it presented a high percentual of missing data. This study prioritized those variables according to the theoretical framework for primary care for NCDs, and systematized them in the logical model. In addition, it was not possible to collect outcome variables directly related to the user because they are not available in database. It was not possible to evaluate the adequacy of the FHS as a whole and make inferences about it due to lack of outcome variables directly related to the user. However, it was possible to evaluate in this study the adequacy of the structure and work process of family health teams to primary care for NCDs in the high coverage context of the FHS. Nevertheless, the PMAQ-AB database is currently the largest national FHS database and it has great potential for use in primary care evaluation research. Furthermore, PMAQ-AB is a program with voluntary adherence linked to financial incentive, which makes it possible that the teams that adhered were those with the best performance. There was a change in the external evaluation instrument, reducing the data analysis to the variables belonging to the two PMAQ-AB cycles. Moreover, the comparison between Cycles 1 and Cycle 2 in this study caused information losses related to those that did not participate in the two cycles. But this was minimized by the fact that Tocantins presented one of the highest rates of adherence to the Program in the country, encompassing almost all the PHUs and family health teams of the state.
Although some studies have used data from PMAQ-AB [
31,
32,
34,
38,
39,
41,
45,
51,
53,
59], this is the first one that evaluated the structure of PHUs and the work process of family health teams for NCDs primary care using the CATPCA method. This approach was chosen because the PMAQ-AB data is categorical and a method was needed to reduce a large number of variables and identify the most powerful indicators explaining the variability of variables between PMAQ-AB cycles.
The PMAQ-AB may have led to improvements in the overall organization of the work process of the family health teams included in this study, especially regarding health promotion. However, more significant process changes may require additional time for evaluation, and further studies are needed incorporating data from the new PMAQ-AB cycles.