Analysis of antimicrobial prescriptions in Primary Healthcare *

* Extracted from the dissertation: “Prescrição de antimicrobianos na Atenção Básica”. Programa de Pós-Graduação em Enfermagem, Universidade Federal do Piauí, 2018. 1 Universidade Federal do Piauí, Departamento de Enfermagem, Teresina, PI, Brazil. 2 Universidade de São Paulo, Escola de Enfermagem de Ribeirão Preto, Ribeirão Preto, SP, Brazil. 3 Universidade Nova de Lisboa, Instituto de Medicina e Higiene Tropical, Lisboa, Portugal. 4 Centro Universitário Uninovafapi, Teresina, PI, Brazil. ABSTRACT Objective: To evaluate prescription receipts for antimicrobial prescriptions prescribed in primary healthcare units in a capital city in the Northeast region of Brazil. Method: An evaluative, analytical study. Information from the central distribution of essential medicines and the receipts of antimicrobial prescriptions were used. Results: There were 2,232 prescription receipts analyzed, in which metronidazole (250 mg) was prescribed in 28% of the evaluated prescription receipts, the “pill” pharmaceutical form in 30.7%, and the “oral” administration form in 78.2%. In the prescriptions prescribed by nurses, 80.7% were intended for users with sexually transmitted infections. With the exception of the pharmaceutical form, only 34.7% of the prescriptions were in accordance with the Nursing protocol recommendations. There is still no information on the concentration (43.7%), the dosage (39.9%) and the treatment time (36.8%). Conclusion: The evaluated receipts of antimicrobial prescriptions do not accurately follow the guidelines of Resolution No. 20/2011, nor of the instituted Nursing protocol.


INTRODUCTION
The irrational prescription of antimicrobials in Primary Healthcare (PHC) has significantly contributed to the phenomenon of antimicrobial resistance (AMR), since the consumption of these drugs has considerably increased in the last decades in the community (1) . Professionals responsible for the prescription do not always do it with all the necessary clinical rigor and routinely prescribe antimicrobials for infections which are not caused by bacteria (1)(2) , thereby intensifying this process. PHC prescriptions still involve other errors, such as: duration of the wrong treatment; prescription omissions; and non-compliance by the user with the guidelines contained in the prescriptions, among others (2) .
The seriousness of this situation in Brazil culminated in a stance by the National Health Surveillance Agency (ANVISA -Agência Nacional de Vigilância Sanitária) to issue Collegiate Board Resolution (RDC) no. 20/2011 with the aim of promoting actions to control antimicrobials in the context of PHC, which links the sale and dispensing of these medications to retaining special control prescription receipts in two copies (from the user and the health unit that dispensed the medication) (3) .
Following this trend, the Ministry of Health (MoH), supported by Law No. 7.498, of June 25, 1986 (Law of Professional Nursing Practice), issued Ordinance No. 2.488/2011 in the same year, which reaffirms prescription of specific medications within the PHC as a specific attribution of nurses, provided that it occurs according to "protocols or other technical regulations established by the Federal, State, Municipal or Federal District manager" (4) .
This publication allows state governments to establish protocols within the scope of PHC, a movement followed by the Municipality of Teresina, Piauí state, which created the Nursing Protocol in Primary Care and Outpatient Services in 2012, being composed of 10 subprotocols in several areas and constituting the Ministry of Health Manuals and Portals as reference (5) . This protocol issues local guidelines for nurses to prescribe antimicrobials within two subprotocols: Sexually Transmitted Infections (STI) and Child Health, being pointed out as an important strategy to streamline the care provided to these diseases in PHC.
Although in effect for some years, the impact and compliance of resolution 20/2011 have been little analyzed in the scientific literature produced in Brazil, in addition to an evaluation of protocol conformity in PHC, for which no studies have been found in Northeast Brazil. In view of the importance of the correct prescription of antimicrobials to reduce the impact of bacterial resistance, the objective of this study was to assess the adequacy of receipts for antimicrobial prescriptions retained in PHC health units in the city of Teresina-PI, with the guidelines of RDC no. 20/2011 for all prescribing professionals (doctors, nurses and dental surgeon) and the Nursing protocol in primary healthcare used by the Municipality.

Study deSign
This is a cross-sectional analytical study with retrospective data collection.

Scenario
The study was conducted in the city of Teresina-PI, Brazil. One used the central distribution of essential medicines database linked to the Pharmaceutical Care Management of the Teresina Municipality, together with the receipts for antimicrobial prescriptions retained in the municipal health units which are composed of the Basic Health Units (UBS -Unidades Básicas de Saúde) and the Psychosocial Care Centers (CAPS -Centros de Atenção Psicossocial).
The UBS are subdivided into zones, namely: the east/ southeast area, which comprises 36 UBS; the northern zone, with 25 UBS; and the south zone, with 27 UBS. CAPS are referral and treatment centers for people with psychological distress and are divided into three models, each with its own particularities: 4 (four) CAPS II; 1 (one) CAPS III; 1 (one) CAPS AD (alcohol and drugs); and 1 (one) CAPS i (6) . Data were collected from January to March 2017.
In view of the lack of records on the number of antimicrobial prescription receipts retained in the health units of the Municipality under study, it was necessary to operate with a viable record, and therefore the data referring to the units of the distributed medicine were used for the sample calculation. Based on this, the inclusion criterion would be based on the health units which retained the second receipt copy of the prescription with at least one prescribed antimicrobial drug unit to consolidate the sample of prescription receipts to be evaluated in the study.
The study population consisted of data referring to the distribution of antimicrobial drug units between June 2015 and July 2016 (14 months); and to receipts of antimicrobial prescriptions retained in health units within the same time interval. The time frame for data collection ( January 2017) is based on Ordinance GAB/SMS No. 07/2015 of August 18, 2015 (7) , which determines that the minimum time for filing a prescription for control purposes is two years.
The quantity referring to the distribution of antimicrobial drug units dispensed monthly from June 2015 to July 2016 was used for calculating the sample size (1,651,516 units distributed), the distribution in the UBS according to the zone -east/southeast (617,810), south (581,483) and north (438,716), and by the CAPS models presented in the city which received antimicrobial medication units in the indicated period, namely: II (6,825), alcohol and drugs (5,090) and III (1,592). It is noteworthy that only the CAPS for children and adolescents of the four CAPS models did not receive antimicrobial medication units during the studied period. Therefore, CAPS II, III and AD models were part of the study.

Selection criteria
The study included UBS and CAPS which had retained the second receipt copy of prescriptions with at least one antimicrobial medication unit prescribed in the period from June 2015 to July 2016, and those which received more units of this medication according to each zone; as well as CAPS, according to service modality. UBS and CAPS which started retaining prescriptions less than two years ago were excluded, as well as prescriptions corresponding to health protocols for leprosy and tuberculosis, as they present different formalities for requesting antimicrobials.

Sample definition
The sample was obtained from the formula based on the estimate of the population average for infinite populations: Multistage sampling was applied based on Hulley (8) , in which stratified sampling proportional to the demand for antimicrobial medication units distributed in UBS (east/ southeast, south and north) and for each type of CAPS (II, III and AD) was initially used. The sample consisted of 2,232 units of antimicrobial drugs. This sample was obtained by establishing a minimum of 1 (one) antimicrobial medication unit for the receipts retained in the health units.

data collection
An instrument was used to collect the characteristics contained in the antimicrobial prescription receipts which presents categories based on Chapter III of RDC No. 20/2011, which discusses the prescription of antimicrobial drugs. Thus, its variables were subjected to the process of semantic analysis content validation to guarantee information reliability. The variables of interest were: identification data of the health units, elements characterizing the prescriptions and the drugs prescribed in the prescription.
A group of evaluators was selected in the content validation process, being composed of expert judges trained to analyze the content, presentation, clarity and understanding of the instrument. Seven judges participated to validate the research instrument. From the 23 items of the instrument, 5 (five) were not considered validated by the judges, as they obtained a Content Validity Index less than or equal to 80%; 3 were excluded; and 2 (full name of the user and prescribed antimicrobial) remained, as they are not mandatory items in the RDC.
Regarding the receipts of the antimicrobial prescriptions performed by nurses, characteristics included in the PHC protocol of the municipality were verified for the nurse's performance (prescribed antimicrobials, section of the Primary Healthcare protocol, dose, concentration, pharmaceutical form, posology, interval between administrations, treatment time and type of treatment).
After validating the instrument based on RDC 20/2011, a pilot study was carried out in a UBS in the east/southeast area with 112 prescription receipts, corresponding to 5% of the sample. These receipts were not included in the study sample. The inclusion and exclusion criteria of UBS and CAPS were followed after conducting the pilot study.
Three UBS participated in the study, one from each zone, and three CAPS models. A total of 835 prescription receipts were collected in the UBS in the east/southeast zone, 786 in the UBS in the south zone, 593 in the UBS in the north zone, 9 in the Southeast CAPS, 2 in the CAPS III and 7 in the CAPS AD, respectively, totaling the 2,232 prescription receipts. The systematic sampling technique was used in each health unit for the receipt selection.
The prescription receipts were initially organized within the medication dispensing units of each UBS according to the months ( June 2015 to July 2016), arranged side by side. The 1 st (first) prescription was removed from each month, then the 11 th (eleventh), then the 21 st (twenty-first), and so on, until the number of prescriptions to be evaluated in each health unit ends, respecting the cut-off used so that the prescription receipts were collected from all months.

data analySiS and proceSSing
Two databases were produced, one for the information collected by the instrument that evaluates the prescription receipts according to Chapter III of the RDC no. 20/2011, and another database for the data which analyzes the antimicrobial prescription receipts according to the PHC protocol of the municipality the nurses work in. The data were processed in SPSS version 22.0, and descriptive statistics were calculated. The Kolmogorov-Smirnov test was performed to verify the data normality for the inferential analysis, and the Pearson's Chi-Squared test was used to associate qualitative variables.

ethical aSpectS
The study followed the principles of Resolution 466/12 of the National Health Council, approved on November 4, 2016, under

RESULTS
There were 2,232 prescription drug receipts analyzed based on substances classified as antimicrobials dispensed in UBS (99.2%) and CAPS (0.8%) units with regard to the patient, the prescriber and the medication. Thirteen (13) antimicrobials were described in the evaluated prescription receipts, in which metronidazole (250mg) was the most prescribed antimicrobial (Table 1).   The association test between the mandatory data for antimicrobial prescription contained in RDC No. 20/2011 (3) and the professional revealed a statistical difference between the presence of complete information about antimicrobials and the prescribing professional (p<0.001) ( Table 4). The association between the mandatory data and the evaluated municipality area revealed a statistical difference between the prescriptions by the medical professional and the regional east/southeast (28.5%) and south (26.9%); as well as the prescriptions of the professional nurse and the northern region (10.0) ( Table 5).

DISCUSSION
The official recommendations (RDC no. 20/2011 (3) and the Nursing protocol (4) in PHC) are not strictly followed in antimicrobial prescription receipts. The private prescriptions of the health establishment presented a lot of missing data, constituting essential information for the purpose of creating the RDC and the Nursing protocol, namely the control of drugs based on substances classified as antimicrobials and legally qualified prescriptions by nurses as a PHC professional, respectively.
Only one prescription receipt within the total of 2,232 followed all the recommendations proposed by the RDC; and more than 50% of the prescriptions prescribed by the professional nurses were not in accordance with the protocol guidelines. This inadequacy to local recommendations is also pointed out by a multicenter study carried out in 53 countries with these same professionals (9) .
Regarding the elements which characterize the patient's identification, the percentage of prescription receipts which presented the full name (74.7%) and age (0.2%) were much less than ideal (100%). In the context of patient safety, correct identification is essential in handling medication, since the chances of other errors occurring (e.g., dose, administration route, medication and wrong patient) are maximized in the absence of the patient's name in a legible manner (10)(11) . The percentage (98.3%) with regard to the issue date of the prescription receipt was close to the ideal (100%), although still lower than that recorded (99.3%) in a study conducted in Minas Gerais state, Brazil (12) . On the other hand, information on the pharmaceutical form was absent on only one prescription receipt, and the treatment time was absent in five, with this finding being common in the national literature (13)(14) . Still, the prescriber identification was high (93.7%), of which the majority was prescribed by doctors (60.8%), followed by nurses (19.8%), corroborating information of another national study (12) .
Metronidazole was the most prescribed antimicrobial on the prescription receipts (63.4%), which goes against the national and international literature, which cite amoxicillin as the most prescribed antimicrobial (14)(15)(16) . Regarding the prescriptions prescribed by professional nurses based on the Nursing protocol in Primary Healthcare and Outpatient clinics in the studied municipality (4) , it is clear that the antimicrobials were mostly destined for the STI protocol (80.7%), which is an important area of Nursing practice, with 94.8% of these being related to the treatment of vaginal discharge, thus meeting the protocol competency requirements of the studied municipality.
Bacterial vaginitis, trichomoniasis and vulvovaginal candidiasis (17)(18) are among the causes of vaginal discharge in women of childbearing age. The first two can be treated in PHC with metronidazole in the form of a pill or vaginal cream. This drug needs to be monitored, as it induces the selection of resistant Trichomonas vaginalis strains, and this resistance occurs in 4% to 10% of cases (19)(20) . Therefore, the importance of the correct prescription is once again signaled, especially by nurses since they present a considerable amount of prescriptions for STIs, as observed in this study.
Regarding the characteristics of the prescribed antimicrobials and taking into account compliance with the items recommended in each Nursing subprotocol (STI and Child health), it was shown that only 45.6% were in agreement with the protocoled dose and 41.4% in accordance with the indicated concentration. A considerable number of prescription receipts (43.7%) still lacked the "concentration" data, which is extremely worrying.
Regarding the pharmaceutical form and protocol posology, 93.9% and 33.6% of antimicrobials were in agreement, respectively. However, high absence of posology in the prescription was found in 39.9% of the receipts. The time interval in hours was longer than recommended by 47.3%, and the treatment time was not included in 36.8% of the prescriptions.
With regard to the association between the mandatory elements of the medication and the prescribing professional, it was found that being a "doctor" increased the likelihood of a prescription with the presence of data on the medication, as reported by the international literature (21) . Doctors are the professionals who prescribe the most both in Brazil (12)(13)(14) and in other countries (15)(16) . The prescription of medications is intrinsic to the medical professional, so that their training represents a consubstantial basis which justifies the greater number of prescriptions of these medications. Therefore, mistakes made by these professionals have an important bearing on patient safety due to the volume and importance of prescriptions.
The medication prescriptions by professional nurses are limited to public health programs and routines which have been previously approved in health institutions (public or private), supported by the Professional Exercise Law and are additionally in all Protocols or Booklets of PHC which reaffirm the medication prescriptions as the nurse's assignment. Medication prescription by nurses occurs in other countries such as the United Kingdom, Sweden, the United States of America, Canada, Australia and Ireland (22) .
A recent documentary study on rules, guidelines, policies and institutional support for nurses to prescribe medications also highlights the existence of few studies which have analyzed medication prescriptions by nursing professionals. However, they emphasize that nurses' autonomy has been widely reaffirmed by the Nursing class entities and in the PHC protocols adopted by the Ministry of Health, with arguments about the prescription being based on three categories: autonomy and competence for prescribing medications; corporate policies which compromise the full exercise of Nursing; and transformation of health and nursing care in PHC (23) .
Corroborating the study above and reflecting on prescription expansion by nurses in Brazil, in fact expanding prescriptions increases the access of medicines to users. However, this expansion must be seen from the perspective of the dimension of comprehensive user care. That said, one supports implementation of the Systematization of Nursing Care (SNC) to the protocols published by the Ministry of Health as guiding evidence-based practice and as a guide for nurses' actions in the context of PHC.
The findings of this research provide evidence that Nursing prescription, as well as that of the other health professionals evaluated in this study, need in-depth and continuous analysis, especially in the case of specific prescription for antimicrobials, which has a potential negative impact, especially in the increase of AMR.
This study has some limitations. With regard to the adopted method, cross-sectional studies do not enable a scope for conducting a follow-up of the investigation object, which would provide identification of the antimicrobial therapy used according to the patient's medical diagnosis. However, the method used in this study achieves the listed objectives. In addition, the protocol's conformity assessment was only possible for Nursing because only this category had its specific protocol in the municipality under study.

DESCRIPTORES
Antiinfecciosos; Medicamentos bajo Prescripción; Control de Infecciones; Enfermería de Atención Primaria. demonstrates a failure in the system in providing quality services and patient safety as a whole.
Non-systematized prescription of antimicrobials can lead to incorrect, exacerbated or inferior use to that recommended for the drug, causing consequences to the health system. The problem is alarming, and this study confirms the need for interventions aimed at PHC, in which the contingent of prescription receipts is greater and there is difficulty in ensuring the quality and correct use by the community as there is no control system.
It is recommended to create a single prescription format for prescription which follows the guidelines established for that purpose; the use of a prescription control system integrated with SUS; the use of prescriptions in electronic format; and the qualification of prescribing professionals.