Pharmacist-participated medication review in different practice settings: Service or intervention? An overview of systematic reviews

Introduction Medication review (MR) is a pharmacy practice conducted in different settings that has a positive impact on patient health outcomes. In this context, systematic reviews on MR have restricted the assessment of this practice using criteria such as methodological quality, practice settings, and patient outcomes. Therefore, expanding research on this subject is necessary to facilitate the understanding of the effectiveness of MR and the comparison of its results. Aim To examine the panorama of systematic reviews on pharmacist-participated MR in different practice settings. Methods A literature search was undertaken in Biblioteca Virtual em Saúde (BVS), Embase, PubMed, Scopus, The Cochrane Library, and Web of Science databases through January 2018 using keywords for "medication review", "systematic review", and "pharmacist". Two independents investigators screened titles, abstracts, full texts; assessed methodological quality; and, extracted data from the included reviews. Results Seventeen systematic reviews were included, of which sixteen presented low to moderate methodological quality. Most of reviews were conducted in Europe (n = 7), included controlled primary studies (n = 16), elderly patients (n = 9), and long-term care facilities (n = 8). Seven reviews addressed MR as an intervention and thirteen reviews cited collaboration between physicians and pharmacists in the practice of MR. In addition, thirteen terminologies for MR were used and the main objective was to identify and solve drug-related problems and/or optimize the drug use (n = 11). Conclusion There is considerable heterogeneity in practice settings, population, definitions, terminologies, and approach of MR as well as poor description of patient care process in the systematic reviews. These facts may limit the comparison, summarization and understanding of the results of MR. Furthermore, the methodological quality of most systematic reviews was below ideal. Thus, international agreement on the MR process is necessary to assess, compare and optimize the quality of care provided.


Introduction
Medication review (MR) is a pharmacy practice conducted in different settings that has a positive impact on patient health outcomes. In this context, systematic reviews on MR have restricted the assessment of this practice using criteria such as methodological quality, practice settings, and patient outcomes. Therefore, expanding research on this subject is necessary to facilitate the understanding of the effectiveness of MR and the comparison of its results.

Aim
To examine the panorama of systematic reviews on pharmacist-participated MR in different practice settings.

Methods
A literature search was undertaken in Biblioteca Virtual em Saú de (BVS), Embase, PubMed, Scopus, The Cochrane Library, and Web of Science databases through January 2018 using keywords for "medication review", "systematic review", and "pharmacist". Two independents investigators screened titles, abstracts, full texts; assessed methodological quality; and, extracted data from the included reviews.

Results
Seventeen systematic reviews were included, of which sixteen presented low to moderate methodological quality. Most of reviews were conducted in Europe (n = 7), included controlled primary studies (n = 16), elderly patients (n = 9), and long-term care facilities (n = 8). Seven reviews addressed MR as an intervention and thirteen reviews cited collaboration PLOS

Introduction
Medication Review (MR) has been defined as a structured assessment of patients' pharmacotherapy whose aim is to optimize the drug use and to improve health outcomes [1]. Despite that concept, MR is used as a generic term and its practice can be performed by some health professionals, such as physicians, nurses and pharmacists, causing confusion regarding its characterization and effectiveness. In the MR, pharmacists play an important role in the detection and resolution of drug-related problems (DRPs) through interaction with patients and/or healthcare professionals [2]. The MR conducted by pharmacists may be classified into three types: Prescription Review, Concordance and Compliance Review (Adherence Support Review), and Clinical Medication Review (with or without prescribing) [3,4]. This practice, the last type particularly, must be conducted in collaboration with health professionals [5]. There are models of MR performed by pharmacists with collaboration of other health professionals in countries where pharmaceutical education is clinic-oriented as well as in those where pharmacists have no formal clinical education [6][7][8][9][10][11].
In Australia, there are examples of MR in which after the pharmacist assesses the patient's pharmacotherapy, he sends a report with recommendations to the patient's physician. After agreement with the pharmacist, the physician discusses the proposed recommendations with the patient [8,9]. The implementation of these recommendations made by the pharmacist may be influenced by some factors such as: a good working relationship between the pharmacist and the health care team [2], the type of communication between the pharmacist and the team (verbal or written) [2,10,11] and the clinical relevance of the recommendations [10,11].
In this context, studies show that pharmacist-participated MR has a positive impact on patients in practice settings such as community pharmacies [12,13], long-term care facilities [14,15], outpatient clinic [16], home care [17,18] and hospitals [19,20]. Besides the identification and resolutions of DRPs [21][22][23], the studies highlight benefits such as increase in quality of life [24], decrease of hospitalizations and health costs [25,26]. In order to achieve such results, the implementation of MR demands changes in pharmacists' professional and social behaviour [18].
Although a previous overview of systematic reviews has corroborated the importance of MR for the improvement of patients' health outcomes, it restricted relevant aspects such as methodological quality, practice settings, and assessed outcomes [27]. Moreover, the mentioned overview didn't focus on concepts, terminologies and approach (as service or practice primary studies with or without collaboration of other health professionals. Systematic reviews that did not present the definition of MR were included only if the interventions described in primary studies accorded to the concept of MR adopted. In this overview, other systematic reviews were excluded because: i) full text was unavailable; (ii) MR was performed collaboratively by pharmacist and other health professionals but the pharmacist's role within the team was not clearly defined in the primary studies included.

Data extraction
Two investigators (R.O.S.S. and L.A.M.) extracted independently the following data: authors, publication year, main author's country, aim, country of primary studies, study design, practice setting, and population, number of primary studies included in the systematic review and meta-analysis, number of primary studies related to MR, assessed outcomes related to the drug use process, and economic, clinical, and human outcomes (ECHO model) [32], main results, concepts, terminologies, and approach (service or intervention) of MR, interprofessional collaboration, structure, processes, and outcomes variables [33,34] described in the systematic reviews as well as limitations or biases. Except for the number of primary studies included in the systematic reviews and meta-analysis, all data were extracted only from primary studies on MR. In case of data absence or clear pieces of information, it was considered that authors did not report such variable. Discrepancies among investigators were solved by consensus.
Study design, practice setting and population were determined according to the authors of systematic reviews. Regarding terminologies of MR, the words used in the search strategies of the included systematic reviews were not considered. If the review presented different MR definitions, the one presented in the methodology was adopted. Moreover, in the absence of a clear definition, components or objectives of MR were extracted. Interprofessional collaboration was considered present if at least one primary study reported it.
Degree of agreement between the two investigators (R.O.S.S. and L.A.M.) was excellent for title and abstract screenings (k1 = 0.942) and full-text screening (k2 = 0.805).
Number of primary studies in the systematic reviews, meta-analysis, and related to MR; assessed outcomes, main results; and structure, processes and outcomes variables Number of primary studies in the systematic reviews, meta-analysis and related to MR as well as assessed outcomes and main results are presented in Table 2. Number of primary studies included in the reviews varied from 5 [40] to 69 [52]. The minimum and maximum number of primary studies in the systematic reviews related to MR was 4 [44,[51][52][53] and 63, respectively [43]. Only seven of these reviews performed meta-analysis [37,40,41,44,45,47,49], in which the number of primary studies varied from 3 to 25 [37,40]. Pharmacist-participated medication review in different practice settings: An overview of systematic reviews Table 1. Description of the systematic reviews' aim; primary studies' countries, practice setting, study design and population; and AMSTAR score for quality assessment of systematic reviews followed or not by meta-analyzes.

Assessed outcomes
Main results [52] 69 N (0) 3 Medication-related falls RT 3-Most of pharmacist's recommendations were accepted by general practitioner. The mean number of medication changes per patient increased while the number of falls per patient decreased. There was no significant reduction in the rate of recurrent falls, injurious falls, or overall use of high-risk medications; the number of falls was reduced in the postintervention group resulting in future savings of US$7.74 per patient per day in one of the included studies. The use of the addressed drug classes decreased in the postintervention period [53] 21 N (0) 3 QALY RT 3-In Spain, the conSIGUE program was dominant with robust results (in terms of QALY gained, CEAC: 100%, with a WTP at 30,000€ per QALY gained), whereas in the two UK studies, the intervention was unlikely to appear cost-effective: HOMER program showed ICER = £82,678 per QALY gained, CEAC: 25%, with a WTP at £30,000 per QALY gained, only 1/5 scenarios led to an ICER< £30,000 per QALY gained while in MRpain program interventions were more costly and provide similar QALYs than usual care and results were uncertain due to the small sample size Pharmacist-participated medication review in different practice settings: An overview of systematic reviews Table 3. Definition, terminology, approach of medication review (service or intervention), and interprofessional collaboration.

Reference Definition of MR Main terminologies for MR MR approach
Was there interaction between other health professionals in the MR? (What professionals?) [37] DP 1-Structured evaluation of a patient's medicines, aimed at reaching agreement with the patient about drug therapy, optimizing the impact of medicines, and minimizing the number of medication related problems Pharmacist-participated medication review in different practice settings: An overview of systematic reviews

Discussion
Most reviews had main authors and primary studies from Australia, which includes research with elderly people in community pharmacies, long-term care facilities and hospitals. Australia is one of the first countries to incorporate MR in primary outpatient care and has remuneration programs to accredited pharmacists who offer such service [54][55][56][57]. Furthermore, elderly people are some of the priority patients of this practice according to international guidelines as well as children and pregnant women [55][56][57][58][59][60]. Regarding the practice settings, Bulajeva et al. (2014) [54] corroborate our results when reported that in Europe, MR has been performed in community environments, hospitals, and long-term care facilities. Abbreviations: DT 1 -Definition type 1 (concept); DT 2 -Definition type 2 (components); DT 3-Definition type 3 (objectives); NR-Not reported. https://doi.org/10.1371/journal.pone.0210312.t003 Pharmacist-participated medication review in different practice settings: An overview of systematic reviews Terminologies for MR are not standardized in literature. Among the most well-known are: Home Medicines Review, Medication Use Review and Residential Medication Management Review, in Australia [61][62][63]; Revisión de la medicación and Revisión sistemática de la medicación, in Spain [64,65]; Medication Review and Comprehensive Medication Review, in the United States [66,67]; Comprehensive Medication Review, in Finland [68,69]; and Medicines Use Review in the United Kingdom [70,71]. These terminologies result from the differences in patient complexity and characteristics of each country and practice setting where MR is performed [72][73][74].
In pharmacy, there is no consensus among concepts and terminologies of clinical practice [75][76][77][78][79]. Linguistic and cultural questions as well as the overlapping between "what we do" and "how we do" can be causes of these divergences. Consequently, lack of standardization of definitions and terminologies can confuse researchers and professionals who aim to compare results and to confirm the practice effectiveness [75,78]. Thus, definitions and terminologies internationally standardized can benefit the assessment of impact of MR [80]. Moreover, modelling of clinical pharmacy services should be used since it facilitates the standardization and comparison of MR and provides a holistic approach to the decision-making process and organizational change [81][82][83][84][85]. Therefore, establishing minimum quality standards for MR is important for comparison of the practice as well as for the optimization of the care provided and, consequently, of the patients' health outcomes.
Regardless of the terminology adopted in systematic reviews, the objective of MR, both as service as well as intervention, is to identify and solve DRPs, implement changes in patients' pharmacotherapy and improve health outcomes. This objective agreed with guidelines of the countries where this practice is more frequent. In this regard, interprofessional collaboration is necessary to reach this objective [55-60, 86, 87]. Collaborations among healthcare professionals declared in the systematic reviews can be seen as a positive factor to achieve better clinical, economic, and humanistic outcomes.
Interprofessional collaboration can be encouraged through specializations, since they motivate information sharing and communication between healthcare professionals. Furthermore, appropriate training of these professionals, such as pharmacists, is essential to develop abilities to the clinical practice, for example, critical thinking and collaborative interpersonal practice [88]. In Australia, only pharmacists who are trained and go through assessments are accredited and can provide MR [89]. In the United States, post-graduate pharmacist residency training, as well as physician residency training, has become a requirement for entry-level health-system pharmacy practice [90]. In this same country, the Accreditation Council for Pharmacy Education (ACPE) (2016) [91] established interprofessional collaboration as one of the accreditation standards for the professional program in Pharmacy leading to the Doctor of Pharmacy degree. According to this institution, the curriculum should prepare students to provide entrylevel, patient-centered care in different practice settings as member of an interprofessional team (with prescribers or other healthcare professional). Thus, literature supports our finding of "pharmacists qualified to provide MR" as a frequently described structure variable.
In this context, MR can be performed by physicians, nurses, and pharmacists. Despite using the same terminology, for physicians and nurses, MR is usually a component of clinical practice whose process has not been well described in literature [92,93]. In pharmacy, studies and guidelines of different countries described MR as a clinical pharmacy service or intervention [13,16,[94][95][96][97][98][99][100]. In this overview, the systematic reviews who presented practice components of MR (e.g. assessment of drug use history, health education, and review of patient's medications) addressed it as service or intervention. However, the discrepancy between the MR approach as an intervention and the concept of the intervention present in the literature is noticeable.
According to the Society of Hospital Pharmacists of Australia, "intervention" is "any action performed by a clinical pharmacist that directly results in the change of patient management or therapy" [101]. Suggett and Marriott (2016) [102], in turn, define "intervention" as a process in which the pharmacist identifies and makes recommendations in an attempt to prevent or resolve DRPs. The authors emphasize that the definition of "intervention" does not include MR without recommendations for changes in treatment.
In Brazil, "intervention" is a "professional action planned, documented and performed by the pharmacist for optimization of pharmacotherapy, promotion, protection and recovery of health, prevention of diseases and other health problems" [103]. Thus, we understand "pharmacist intervention" as an action whose goal is to improve patient health outcomes and that may result in changes in pharmacotherapy. In addition, "intervention" is a result of the situational analysis of the patient, and is part of the care plan, step of the patient care process.
Considering the patient care process, three stages are recommended in several health professions, including pharmacists: initial assessment, care plan, and assessment of outcomes. The first stage is a situational analysis in which the pharmacist gathers, analyzes and interprets information about patient's clinical conditions and pharmacotherapy, aiming to evaluate his or her drug-related needs. The second stage is the care plan whose purpose is to agree with the patient the actions necessary to manage his or her clinical conditions successfully with pharmacotherapy. The care plan includes goals of therapy, interventions (e.g. inclusion of new drug therapy, patient education, and referrals to other health professionals), and a schedule for assessment of outcomes. In the third stage, patient outcomes are assessed, documented, and compared to the goals of therapy [104].
From the presented patient care process, the most appropriate MR approaches are as "service" or "practice component". As a service, MR should include all steps of the patient care process. As a practice component, MR is part of other health services, such as medication reconciliation, and consists of the situational analysis of the patient's pharmacotherapy. Hence, future studies need to describe if MR is a clinical pharmacy service or a practice component. Only then, it will be possible to compare the impact of its results and assure the robustness of this practice.
Regardless of the MR approach, access to medical records is important for the clinical practice. Literature has reported that pharmacists should rely on medical records and technical drug information to make decisions based on evidences and provides the best possible patient care [105,106]. Guidelines highlight the need to obtain patient information from different sources, such as interviews with patient and caregiver, clinical laboratory tests, and medical records, considering that they are complementary [55][56][57][58][59][60]. Therefore, describing access to medical records as a structure variable is relevant, since the limitation of access to any sources of information can result in the restriction of clinical activities of pharmacists.
Although there is no standardization for terminologies and approaches of MR, there are variables related to the care process that are commonly used in this practice, such as: "number of drugs" and "number of interventions". Considering the objective of MR, literature confirms that the analysis of the number of drugs is necessary to the assessment of the impact of MR, especially because it can involve vulnerable patients which present polypharmacy, therapeutic duplication, drug interactions, and contraindications [55][56][57][58][59][60]. Thus, Cipolle et al. (2012) highlight that interventions related to the resolution of DRPs, usually in interprofessional collaboration, may result in the reduction of the number of drugs [104].
Despite Jokanovic et al. (2016) [27] mentioning the positive impact of MR in primary studies performed in community environments (e.g. blood pressure control, quality of life, and healthcare costs), some reviews included in our study show that MR results were contradictory, had little significant or were inconclusive. A systematic review conducted by Huiskes et al. (2017) [107] showed positive and negative effects for some outcomes variables described in our overview. According to the authors, the different results found may be a reflection of factors such as: 1) selection of patients, which may not fit the objective of MR; 2) how MR is performed in the clinical practice, since there is heterogeneity in the work processes and there is no gold standard on how it should be performed; and 3) outcomes and time of follow-up used to assess the impact of MR, variables that should be chosen according to the objective of MR and being more specific to diseases and drugs. Thus, heterogeneity of processes can affect the method of data analysis as well as the sensitivity and specificity of results, such as mortality, economy of drugs costs, hospital readmissions, quality of life and patients' satisfaction.
Another factor that influences the impact analysis of MR is the methodological quality of the reviews. Although systematic reviews are considered a key element used to the practices of patient care, low quality of reviews have limited processes of decision making and performance of healthcare systems. Our findings were corroborated by literature that has reported low methodological quality of systematic reviews on clinical pharmacy practice. Melchiors et al. (2012) [108] assessed the quality of 31 systematic reviews, in which 24 presented low and moderate quality. In overview of seven systematic reviews, Aguiar et al. (2014) [109] noticed that 71.4% of the reviews had low and moderate quality. Rotta et al. (2015) [110], in turn, found in overview of 49 systematic reviews that no review met all AMSTAR criteria. Therefore, future systematic reviews should value high methodological quality to result in more reliable evidences of real impact of clinical services.

Strengths and limitations
Strengths of this overview include: research in six different databases as well as manual search in the references of the included systematic reviews; use of MESH terms and text words for the literature search; use of 15 different terminologies to MR in the literature search; title, abstract, and full text screenings as well as quality assessment performed by two independent investigators. Moreover, systematic reviews were not excluded based on methodological quality, study design, practice setting, and population. Thus, our study presents a panorama of systematic reviews about pharmacist-participated MR. This overview comprised variables little explored in overviews of systematic reviews on MR, such as definitions and terminologies of MR; interprofessional collaboration; MR approach as well as structure, processes, and outcomes variables described in the systematic reviews.
This overview also presents limitations. Search in the grey literature was not performed. As most systematic reviews are found in databases, the inclusion of only indexed reviews may not have influenced the final sample. Data extraction and analysis of the variables studied were based on systematic reviews rather than primary studies, which may have resulted in overlapping of primary studies in the evaluation of results of MR. Many included reviews did not provide clear information or presented few details on primary studies' design, population, and practice setting; definitions of MR; and impact of MR, this might have compromised data extraction. Moreover, AMSTAR limitations, such as the subjectivity of items "no" and "cannot answer" and the dependence of quality of reports [111,112], could have influenced the assessment of methodological quality of systematic reviews included in this overview.

Toward a future research agenda
This overview of systematic reviews is a starting point to analyze the panorama of literature on pharmacist-participated MR in different practice settings regarding to the concepts, terminologies and approach of MR as well as interprofessional collaboration. From the findings of this overview, it is possible to identify the need for future systematic reviews and primary studies to clarify these variables. The lack of clarity about concepts, terminologies and MR approach as well as interprofessional collaboration extracted from the primary studies of the included systematic reviews may be due to the summarization of the results found in these primary studies and/or to the lack of clarity of the primary studies themselves. Thus, future systematic reviews should analyze these variables in the primary studies in order to reinforce the need to standardize concepts, terminologies and approach of MR in the literature. Moreover, the findings of this overview should also be addressed in future primary studies since any systematic review is only as good as the primary studies that compose it. That is, having in the literature high quality of systematic reviews is just as important as having high quality primary studies. Therefore, future studies, both systematic reviews and primary studies, should clearly present the variables studied in this manuscript in order to facilitate the understanding of effectiveness of MR and the comparison of its results.

Conclusion
In this overview, considerable heterogeneity of systematic reviews about MR was evidenced, especially regarding practice setting, population, MR approach and terminology. Description of patient care process of the primary studies is not clear in some reviews. These facts may limit the comparison, summarization and understanding of MR results. "Medication Review" was the most used terminology, whose main objective is the identification and resolution of DRPs to optimize the drug use. MR practice is mostly comprehended as "intervention", and its main collaborator is the physician. Moreover, methodological quality of most systematic reviews was below ideal. In the light of what was mentioned, it is necessary to come to an international agreement regarding the work process of MR, as a clinical service or practice component, improving, then, the assessment, comparison and optimization of care quality given to patients.  Pharmacist-participated medication review in different practice settings: An overview of systematic reviews