Background Evidence on the effect of pay-for-performance (P4P) schemes on provider performance is mixed in low-income and middle-income countries. Brazil introduced its first national-level P4P scheme in 2011 (PMAQ-Brazilian National Programme for Improving Primary Care Access and Quality). PMAQ is likely one of the largest P4P schemes in the world. We estimate the association between PMAQ and hospitalisations for ambulatory care sensitive conditions (ACSCs) based on a panel of 5564 municipalities.
Methods We conducted a fixed effect panel data analysis over the period of 2009–2018, controlling for coverage of primary healthcare, hospital beds per 10 000 population, education, real gross domestic product per capita and population density. The outcome is the hospitalisation rate for ACSCs among people aged 64 years and under per 10 000 population. Our exposure variable is defined as the percentage of family health teams participating in PMAQ, which captures the roll-out of PMAQ over time. We also provided several sensitivity analyses, by using alternative measures of the exposure and outcome variables, and a placebo test using transport accident hospitalisations instead of ACSCs.
Results The results show a negative and statistically significant association between the rollout of PMAQ and ACSC rates for all age groups. An increase in PMAQ participating of one percentage point decreased the hospitalisation rate for ACSC by 0.0356 (SE 0.0123, p=0.004) per 10 000 population (aged 0–64 years). This corresponds to a reduction of approximately 60 829 hospitalisations in 2018. The impact is stronger for children under 5 years (−0.0940, SE 0.0375, p=0.012), representing a reduction of around 11 936 hospitalisations. Our placebo test shows that the association of PMAQ on the hospitalisation rate for transport accidents is not statistically significant, as expected.
Conclusion We find that PMAQ was associated with a modest reduction in hospitalisation for ACSCs.
- public health
- health economics
- health policy
Data availability statement
Data are available in a public, open access repository. All data are publicly available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is already known?
Despite the rapid expansion over the last decade, it is unclear the contribution of pay-for-performance (P4P) schemes in low-income and middle-income countries, particularly when related to health outcomes.
Brazilian National Programme for Improving Primary Care Access and Quality (PMAQ) is likely one of the largest P4P schemes in the world, reaching around 39 000 Family Health Teams.
What are the new findings?
We show that expansion of the PMAQ was associated with a modest reduction in avoidable hospitalisations for all age groups, with stronger impact for children under 5 years.
When stratifying by age group and diseases that accounted for a high proportion of ambulatory care sensitive conditions, we found that PMAQ expansion was associated with a reduction in hospitalisations for gastroenteritis for 0–4 years, asthma for 5–19 years and renal disease for 20–64 years.
What do the new findings imply?
The findings provide evidence that P4P can contribute to improve quality of care at primary healthcare setting, particularly by reducing avoidable hospitalisations.
Recent reviews have challenged low-income and middle-income countries (LMICs) that have adopted pay-for-performance (P4P) schemes in primary healthcare (PHC), by showing very low-certainty and evidence about the effect of P4P on provider performance, utilisation of services, patient outcomes and resource use.1–3 Moreover, evidence from LMICs is skewed towards donor-oriented approaches, since most P4P schemes have been driven and funded by international organisations,4 often implemented disassociated from existing health system institutions.5 6 Additionally, the effects of P4P schemes have typically been assessed at only one point in time.4
After nearly two decades of continuing reforms and strengthening of PHC,7 Brazil introduced its first national-level P4P scheme, called the Brazilian National Programme for Improving Primary Care Access and Quality (PMAQ, acronym in Portuguese) in 2011. PMAQ was rolled out over three rounds of implementation: round 1 (November 2011–Mar 2013), round 2 (April 2013–September 2015) and round 3 (October 2015–December 2019). During three rounds, the Brazilian Ministry of Health provided financial incentives for a wide variety of structure, process and outcome indicators,8 aiming to improve access and quality of PHC. To achieve these goals, some challenges should be overcome, such as poor health facility infrastructure, inadequate working conditions, high turnover of health professionals, low integration of PHC facilities with healthcare networks (secondary and tertiary levels), rare evaluative culture at facility level, and underinvestment in information technologies to support decision-making process.9
Participation in PMAQ has increased rapidly over the three rounds of implementation, reaching 38 864 family health teams (FHTs) in the third round, representing 89.5% of the total number of teams across the country in 2018. PMAQ has received RS 13.5 billion (US$ 2.7 billion) from 2011 to 2019. Despite 9 years of implementation, PMAQ is still underevaluated in terms of its effects on endpoint outcomes. Studies relied mainly on descriptive analyses,10 based on a subset of data on structure, access, service organisation and management indicators, generally for a specific health condition such as diabetes,11 antenatal care,12 cervical cancer screening,13 user’s satisfaction14 and work processes.15
Hospitalisation rate for ambulatory care sensitive conditions (ACSCs) have been used as an indirect indicator of the effectiveness and quality of PHC.16 ACSCs are a set of diseases and health problems for which timely and high-quality PHC reduces the risk of inpatient admission. Previous studies have investigated the relationship between P4P scheme and ACSCs showing mixed results. The Portuguese experience showed no significant impact of the P4P scheme on the hospitalisation rate for ACSCs17 ; while the English experience found a reduction on the ACSCs incentivised by the Quality Outcome Framework (QOF).18 19 In the USA, two out of three states that implemented P4P schemes presented reduction in ACSC hospitalisations.20 In Brazil, we identified a study that found a decline in hospitalisation for ACSCs due to PMAQ in the first 4 years of PMAQ implementation.21 However, the latter study covered the first two rounds of PMAQ (2010–2014) and used a dichotomous measure of exposure (=1 if municipality participated in PMAQ), which may not reflect the implementation intensity of PMAQ over time.
By improving access and quality of primary care, it is expected that PMAQ would deliver more prevention and better treatment at PHC level, which would result in fewer hospitalisations for ACSCs. Based on that, the objective of our study is to estimate the association of PMAQ on hospitalisations for ACSCs, using a fixed effect (FE) panel data approach of 5564 municipalities from 2009 to 2018, controlling for observed demographic, socioeconomic and health supply-side covariates.
Brazil created the Unified Health System (Sistema Único de Saúde - SUS, in Portuguese) in 1990, financed by general taxation and since then it has made consistent progress towards achieving universal health coverage by providing formally under the law, universal, equitable and comprehensive healthcare free of charge at the point of service.22 Around 75% of the Brazilian population receives healthcare only through the SUS, while 25% have private insurance coverage.23 PHC was nationally implemented in 1994, focusing on families and communities and integrating medical care with health promotion and public health actions, including epidemiological surveillance.24 PHC is provided mainly through three channels: (1) the FHT, composed of at least one physician, a nurse, a nurse assistant and a community health agent; (2) the oral health team (OHT), composed of at least one dentist and a dentist assistant and (3) the Family Health Support Unit (NASF), composed of mental health, rehabilitation, nutrition, maternal and childcare staff, pharmacy and social assistance workers.
PMAQ was one of the strategies aimed at improving PHC that occurred simultaneously and as part of a broader set of changes introduced at the federal level. In 2011, when PMAQ was implemented, the National Primary Healthcare Policy was revised,25 resulting in26–30: (1) increasing of financial transfers from the Ministry of Health to PHC and adoption of equity criteria in health resource allocation; (2) expansion of the Family Health Strategy (FHS) modalities that could receive federal funding, including fluvial health facilities that delivery care to populations living in deprived riverside communities; (3) implementation of Requalifica UBS, a programme that provided financial resources to improve the structure of PHC facilities, such as construction and refurbishments; (4) investments in information systems, aiming at implementing a new PHC information system (e-SUS) and telehealth (Telessaúde Brasil Rede); (5) expansion of NASF, including other healthcare workers; (6) expansion of the programmes for health promotion and prevention in schools (Programa Saúde na Escola) and homecare delivery (Programa Melhor em Casa); (7) implementation of the Programme to Value Primary Healthcare Professionals (PROVAB), which offered training and scholarships for doctors who work in highly vulnerable and deprived areas; and (8) implementation of More Doctor Programme (Programa Mais Médicos) in 2013, involving an emergency expansion of PHC doctors (allowing the hiring of foreign doctors), an increasing number of PHC residency positions, and increasing investments in the infrastructure of PHC facilities.
The PMAQ scheme
The PMAQ scheme was intended to strengthen primary care by increasing resources allocated from the Ministry of Health to municipalities, with the level of resources determined by the performance of primary care teams within the municipality. Municipalities decide whether to disburse financial incentives directly to healthcare providers or teams or to use the financial resources for other purposes related to PHC.8 Participation in PMAQ is voluntary, and municipalities have autonomy to indicate which of their teams would be engaged in the programme. Only the FHT and the OHT were eligible to participate in the first round of the PMAQ (November 2011–March 2013). NASF was included in the second (April 2013–September 2015) and third rounds (October 2015–December 2019). Only 50% of FHTs within a municipality could participate in the first round of PMAQ. In the following rounds, this restriction did not apply.
Teams are evaluated through self-assessment, routine monitoring and external evaluation, resulting in hundreds of structure, process and outcome indicators in each round (598 in round 1; 914 in round 2; and 660 in round 3).31 Structural indicators include availability of drugs and equipment, patient’s privacy during consultations and procedures, and users’ accessibility. Process indicators include content of antenatal care, treatment completion rates, consultation for routine monitoring (ie, diabetes, hypertension, asthma), number of consultations for selected diseases, and proportion of appointments that are scheduled. Outcome indicators include patients’ satisfaction, birth weight of children and prevalence of chronic disease.
We conducted a FE panel data analysis to measure the association of PMAQ on the hospitalisation rate for ACSCs over the period 2009–2018. FE models provide a method for assessing exposure/outcome associations (PMAQ/hospitalisation for ACSCs, in our study) adjusting for time-invariant confounders (municipal characteristics) and measuring time-varying confounders (socioeconomic, demographic, and supply-side variables).32 The unit of analysis is the municipality because the final decision about the PMAQ engagement is taken at this administrative level. We introduced novel aspect to the model by considering a measure of PMAQ implementation intensity. Instead of having a dichotomous measure of exposure (yes/no), we used the percentage of PMAQ participating teams related to the total number of FHTs for each municipality, which may better reflect the expansion of PMAQ over time (three rounds of PMAQ).
Variables and sources
The primary outcome is the age-adjusted hospitalisation rate for ACSCs, defined as the number of hospitalisations for ACSCs among people under 64 years of age per 10 000 population. Hospital admissions were available at municipal level. We excluded patients aged over 65 years because such individuals are likely to have a high prevalence of other diseases than those related to ACSCs, which may not be affected by timely and high-quality PHC. We used the Brazilian list for ACSCs, which was developed by the Ministry of Health based on the International Classification of Diseases (ICD-10).33 A complete description of the diseases and health conditions included in the Brazilian list of ACSCs can be found in online supplemental table S1. The list included 19 diseases or health conditions and all of them were incentivised by PMAQ. Additionally, we also included alternative outcomes, such as hospitalisation rate for ACSCs stratified by age group (0–4 years, 5–19 years and 20–64 years) and the most frequent causes of ACSCs (causes that account for more than 10% of hospitalisations for ACSCs). We used age-adjusted hospitalisation rates for adults (aged 20–64 years) and the overall population (aged 0–64 years). Age-adjusted rates were calculated using the direct method of standardisation.34 Data were extracted from the Hospital Information System (SIH/DATASUS).35
Our exposure variable is defined as the percentage of the PMAQ teams in terms of total number of FHT. FHTs include teams of the FHS and primary health teams. The PMAQ participating teams that were assigned as ‘unsatisfactory’ or ‘declassified’ by the Ministry of Health were excluded from the database, since these teams have not been exposed to PMAQ scheme during the whole round. This assumption was relaxed in the sensitivity analysis. ‘Unsatisfactory’ means that teams have not complied with the rules of PMAQ and ‘declassified’ encompasses three situations: (1) teams have asked to be formally removed from the PMAQ; or (2) teams have refused to be evaluated by the external evaluation; or (3) teams did not have a dental chair. Data were obtained from the Primary Healthcare System (SAPS/Ministry of Health).36 Although PMAQ started in November 2011, municipalities and FHT had been aware of its existence since very early 2011. On this basis, we considered 2011 as initial year of PMAQ programme.
We included a set of covariates usually associated with hospitalisation rate for ACSCs that could be regarded as confounding variables such as: coverage of PHC, hospital beds per 10 000 population, education index, gross domestic product (GDP) per capita (adjusted by General Price Index) and population density. For education index, we used the FIRJAN Index of Municipal Development for education (IFDM), which includes six indicators: (1) enrolment in early childhood education, primary school leaver, age-grade distortion in primary school, percentage of teachers with a degree qualification in primary school, average daily lesson hours in primary school and Basic Education Development Index result in primary school.37 The education index ranges from 0 to 1, where higher values indicate greater development of the municipal units. For the education index and GDP per capita, data were not available for 2018 and 2017–2018, respectively, in which case we estimated annual values by linear extrapolation. PHC coverage was obtained from the Primary Healthcare System (SAPS/Ministry of Health),36 hospital beds from the National Registry of Health Facilities (CNES/DATASUS),38 education index from the Federation of Industries of the State of Rio de Janeiro (FIRJAN),37 and GDP per capita and population were from the Brazilian Institute of Geography and Statistics (IBGE).39 All covariates were available at municipal level.
Descriptive analyses were undertaken, including national trends of overall and most frequent ACSC hospitalisation rates stratified by age group. FE panel data regressions were used to assess the association between municipal-level hospitalisation rates for ACSCs and the percentage of FTHs that participated in PMAQ.
The equation for the FE model can be written as:
where is the hospitalisation rate for ACSCs in the municipality i at year t. is the percentage of FHTs participating in PMAQ. is a vector of covariates that capture time-varying characteristics of municipality i at year t. is the municipality FE, which account for time-invariant characteristics that could affect the hospitalisation rates. is the year FE and captures any time-specific shock. is the error term. Standard errors were clustered at the level of municipality.
FE estimates account for time invariant unobserved heterogeneity and omitted variable bias, and hence are more appropriate for programme evaluation than random effects (RE) estimates.40 Hausman specification test41 was also performed to test whether there is a systematic difference between FE and RE estimates. The null hypothesis was rejected, indicating that FE is preferable to RE.
We performed several sensitivity analyses. First, we estimated equation (1) by age group and most frequent causes of ACSCs. Second, we also included ‘unsatisfactory’ or ‘declassified’ teams in PMAQ participating teams, that is, the numerator of exposure variable (PMAQ) increased. Although these teams were not fully exposed to the incentive scheme, they adhered at the beginning of the PMAQ round. Third, we used a dummy to define PMAQ exposure (=1 if at least one team have participated in PMAQ). Fourth, we used the crude hospitalisation rate for ACSCs instead of the adjusted rate, conducting the same analysis for overall and adult hospitalisations. Fifth, we estimated our main model, equation (1), using the negative binomial distribution. Finally, we performed a placebo test to support the validity of our empirical strategy. Placebo test was conducted using the hospitalisation rate for transport accident (ICD-10 codes: V01–V99).
Patient and public involvement statement
Patients and public were not directly involved in this research. We used publicly available data.
Since its implementation, PMAQ expanded rapidly, reaching on average 50.1% of the FHTs in the first round (2011–2012), 75.7% in the second round (2013–2015) and 86.9% in the third round (2016–2018), not considering teams classified as unsatisfactory or declassified. In 2018, 95.6% of the Brazilian municipalities had at least one FHT enrolled in PMAQ (table 1).
The average adjusted hospitalisation rate for ACSCs for population 0–64 years old decreased by 26.8% (from 127.9 per 10 000 population in 2009 to 93.6 per 10 000 in 2018). Considering the most frequent causes of hospitalisations for ACSCs, there was a large reduction in hospitalisation rates due to gastroenteritis for all age groups (ranging from 44.3% to 52.9%). Hospitalisation rates for asthma also showed a large decline over the period (around 54% for both younger groups of age) (figure 1).
Regarding the covariates used in the FE model, only hospital beds has decreased over the period under investigation, a reduction of 14.6% from 2009 to 2018. Although there is some oscillation within the period, the other covariates increased from the base line to 2018 (table 2).
Table 3 presents the association between the percentage of teams enrolled in PMAQ and ACSC rates. All estimates included year dummies and used robust standard errors clustered by municipality. An increase of PMAQ of one percentage point decreased the hospitalisation rate for ACSC by 0.0356 per 10 000 population (aged 0–64 years). Taking the participation in PMAQ of 89.5% of the total teams, this corresponds to a reduction of 3% of the mean hospitalisation rate for ACSC, or approximately 60 829 (95%CI 19 705 to 1 01 879) hospitalisations in 2018. The impact is stronger for children under five (−0.0940 per 10 000 children under 5 years old), representing a reduction of around 11 936 hospitalisations in this age group in 2018.
The estimated coefficients of PMAQ for the most frequent causes of hospitalisations for ACSCs by age group are presented in figure 2. Higher levels of team enrolment in PMAQ decreased the hospitalisation rate for gastroenteritis only for 0–4 years. Higher levels of team enrolment in PMAQ was also associated with fewer hospitalisation rate for asthma (0–4 years and 5–19 years) and renal disease (20–64 years). For other diseases, PMAQ coefficient was not statistically significant (online supplemental appendix 1).
Our sensitivity analyses included a placebo test to support the validity of our empirical strategy. As a placebo variable, we use hospitalisation rates for transport accidents (0–64 years), which is not expected to be related to access and quality of primary care, and therefore with exposure to PMAQ. As expected, the association between PMAQ and the hospitalisation rate for transport accidents is not statistically significant (online supplemental appendix 2).
Additionally, equation (1) was estimated using an alternative definition of PMAQ. First, we included teams that were assigned as ‘unsatisfactory’ or ‘declassified’ by the Ministry of Health (column 2, online supplemental appendix 3), showing similar results as shown in table 2. The PMAQ coefficient was −0.0379 (p<0.001) for those aged 0–64 years. Second, we used a dummy to define PMAQ exposure (=1 if at least one team have participated in PMAQ), which showed no statistical significance (column 3, online supplemental appendix 3). Third, we used the crude hospitalisation rate for ACSCs instead of the adjusted rate (column 4, online supplemental appendix 3). PMAQ was associated with a reduction on ACSCs, with a coefficient of −0.0369 (p<0.001) for those aged 0–64 years. Fourth, we estimated equation (1) using the negative binomial distribution (column 4, online supplemental appendix 3). The effect of the variables is expressed as incidence rate ratio (IRR). We found that one unit increase in PMAQ leads to a reduction in the rate of 0.04%. Taking the participation in PMAQ of 89.5% of the total teams in 2018, it represents a reduction of 3.58% in the mean hospitalisation rate for ACSCs. This shows a modest reduction in the hospitalisation rate, as found in our main analyses (table 3).
This study found the roll out of PMAQ across municipalities was associated with a modest but significant reduction in avoidable hospitalisations (all ACSCs) for populations aged 0–64 years during the three rounds of the programme (2011–2018). PMAQ was associated with a reduction of around 60 829 (95% CI 19 705 to 101 879) ACSCs for populations aged 0–64 years in 2018. When stratifying by age group and diseases that accounted for a high proportion of ACSCs, PMAQ expansion was associated with a reduction in hospitalisations for gastroenteritis for 0–4 years (−0.048; 95% CI −0.088 to –0.007), asthma for 5–19 years (−0.007; 95% CI −0.013 to –0.002) and renal disease for 20–64 years (−0.010; 95% CI −0.016 to –0.003). Our results are in line with other study carried out in Brazil, which found that PMAQ was associated with a 9% reduction in hospitalisations for ACSCs.21 We found that an increase of 89.5% of PMAQ participating reduce by 3% of the mean hospitalisation rate for ACSCs. However, our approach differs from their study in some important ways. We use a longer time period, consider the implementation intensity of PMAQ (instead of a dichotomous variable) and control for year dummies. In our sensitivity analyses, the dichotomous measure of exposure became non-statistically significant after including year dummies.
Qualitative studies provide some insights into the potential mechanisms underpinning the effect of PMAQ on avoidable hospitalisations. Previous studies have reported improvements in work processes and planning at the team level. These managerial tools are important to achieve improvements in access and quality of care. For example, home visits, support from other professionals and referral to specialised care when needed have been associated with lower rates of avoidable hospitalisations.42 In the state of Paraná, workers engaged in PMAQ have reported improvements in health information registries (allowing better monitoring of target populations) and the indicators from the external evaluation provided a good overview of what should be improved at team level.43 On the other hand, qualitative studies conducted in state of Goiás and São Paulo showed a low perception of the PMAQ impact on clinical practices and work process,15 44 mainly due to a top-down approach adopted during the implementation of PMAQ, lack of knowledge about the programme and absence of feedback regarding the results of external evaluation. These problems tend to demotivate health worker and consequently jeopardise the access and quality delivered though PMAQ teams.
Other studies have investigated the effect of P4P schemes on ACSCs worldwide. In England, hospitalisation for ACSCs incentivised by QOF reached a reduction of 2.7% (95% CI 1.6% to 3.8%) and 8.0% (95% CI 6.9% to 9.1%) in the first and the seventh year after QOF was introduced, respectively, compared with ACSCs that were not incentivised by the programme.18 A more recent study conducted in England suggested that QOF was associated with a smaller reduction in incentivised ACSC hospitalisations (IRR 0.993; 95% CI 0.990 to 0.995), where a 1% increase in quality of care corresponds to a decrease of 187 hospitalisations for all incentivised conditions in the period of 2015–2016 at national level.19
In Portugal, a difference-in-difference study indicated that municipalities that implemented P4P scheme at FHTs had no statistically significant impact on hospitalisations for ACSCs compared with municipalities that did not implement the scheme.17 It is worth noting that our study included 19 diseases-related ACSCs, while the English and Portuguese studies included only 8 and 4 diseases-related ACSCs incentivised by its P4P scheme, respectively, and this might have led to a difference in interpretation, as the current study shows a much broader description of impact of P4P scheme. In the USA, using a propensity score matching of physicians exposed to P4P versus non-exposed to P4P and a difference-in-difference approach at patient level, a study found that two out of three states that implemented P4P schemes had a reduction on ASCS hospital admissions.20 These results are robust to several regression specifications and matching methods.
Strengths and weaknesses of this study
Although we have not claimed that our results reflect the impact of PMAQ on avoidable hospitalisations, we provided robust estimates of its association on ACSCs. We controlled for confounders at municipality level and we performed several sensitivity analyses. Moreover, we used a rich panel dataset covering all Brazilian municipalities over a 10-year period. Regardless, some limitations have remained. First, we have treated our exposure as a homogenous variable, but in fact PMAQ can be implemented in different ways at the municipal level. Municipalities decide whether to distribute the PMAQ rewards in kind (training activities, additional drugs and supplies, equipment) or in cash to teams. Second, we used the official Brazilian Ministry of Health list of ACSCs as outcome, which comprised 19 diseases or health conditions. However, not all of them were explicitly incentivised by PMAQ scheme. As a result, our estimates are likely to be conservative. Third, other programmes were implemented in the PHC in the period investigated in this study, such as the More Doctors Programme (Programa Mais Médicos (PMM)). PMM was designed to overcome the insufficiency and turnover of physicians across the country, especially in remote and deprived areas. Studies have shown that PMM was associated with a modest but significant reduction on ACSC hospitalisations45 and amenable mortality.46 We have not controlled for PMM because data were not publicly available for the whole period analysed in this study (particularly data from 2018). However, as PMM have increased the supply of physicians across the country, we partially captured its contribution by controlling the PHC coverage.
Implication for policy
Our results suggest that P4P can contribute to improve quality of care at PHC setting, particularly by reducing avoidable hospitalisations. However, caution is needed in any extrapolation of our results to other LMICs. PMAQ has beneficiated from a long period of investment in PHC in Brazil, although this investment was insufficient in view of the health needs of the Brazilian population. The role of FHT has been improved since it was implemented in the mid-1990s, as well as political arrangements for cooperation between municipalities and community participation on health decisions. This context probably provided institutional background to enhance better results of PMAQ.
Moreover, decision-makers must include monitoring and evaluation framework along to health policy implementation to identify expected and unexpected results. It would allow decision-makers to make timely adjustments, avoiding inefficiencies and health inequalities among population groups.
Finally, PMAQ was discontinued by the Brazilian Ministry of Health in December 2019, establishing new bases for PHC funding.47 Although P4P had remained, it was completely reformulated, replacing hundreds of indicators, external evaluation, and equity concerns by very few indicators (just seven), restricted to few health conditions. Several authors have criticised the new funding of PHC and its impact on access and quality of care.48 49 Our results can shed some light on the debate on P4P adjustments in the Brazilian context.
Data availability statement
Data are available in a public, open access repository. All data are publicly available.
Handling editor Valery Ridde
Contributors Conception and design of the work: LXR, JOMB and ENS. Analysis: LXR. Interpretation of data for the work: all authors. Drafting the work: LXR and ENS. Revising it critically: all authors. Final approval of the version to be published: all authors.
Funding This research was funded by the Medical Research Council, Newton Fund and the Brazilian National Council for the States Funding Agencies (CONFAP) under the UK- Brazil Joint Health Systems Research Call (grant MR/R022828/1). Funding from CONFAP came from Fundação de Amparo à Pesquisa do Distrito Federal (FAPDF), Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco (FACEPE) and Fundação de Apoio à Pesquisa do Estado da Paraíba (FAPESQ).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.