Objectives The integration of community pharmacy services (CPSs) into primary care practice can be enhanced by assessing (and further addressing) the elements that enable (ie, facilitators) or hinder (ie, barriers) the implementation of such CPSs. These elements have been widely researched from the perspective of pharmacists but not from the perspectives of other stakeholders who can interact with and influence the implementation of CPSs. The aim of this study was to synthesise the literature on patients’, general practitioners’ (GPs) and nurses’ perspectives of CPSs to identify barriers and facilitators to their implementation in Australia.
Methods A meta-synthesis of qualitative studies was performed. A systematic search in PubMed, Scopus and Informit was conducted to identify studies that explored patients’, GPs’ or nurses’ views about CPSs in Australia. Thematic synthesis was performed to identify elements influencing CPS implementation, which were further classified using an ecological approach.
Results Twenty-nine articles were included in the review, addressing 63 elements influencing CPS implementation. Elements were identified as a barrier, facilitator or both and were related to four ecological levels: individual patient (n=14), interpersonal (n=24), organisational (n=16) and community and healthcare system (n=9). It was found that patients, nurses and GPs identified elements reported in previous pharmacist-informed studies, such as pharmacist’s training/education or financial remuneration, but also new elements, such as patients’ capability to follow service's procedures, the relationships between GP and pharmacy professional bodies or the availability of multidisciplinary training/education.
Conclusions Patients, GPs and nurses can describe a large number of elements influencing CPS implementation. These elements can be combined with previous findings in pharmacists-informed studies to produce a comprehensive framework to assess barriers and facilitators to CPS implementation. This framework can be used by pharmacy service planners and policy makers to improve the analysis of the contexts in which CPSs are implemented.
- Community pharmacy services
- health service research
- qualitative meta-synthesis
- determinants of practice
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 and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Statistics from Altmetric.com
- Community pharmacy services
- health service research
- qualitative meta-synthesis
- determinants of practice
Strengths and limitations of this study
The particular method chosen for this review (ie, qualitative meta-synthesis) is aimed at synthesising qualitative literature and so enabled a rich description of the barriers and facilitators perceived by GPs, patients and nurses who can influence the implementation of CPSs in Australia.
A systematic search was conducted in three comprehensive electronic databases (ie, PubMed, Scopus and Informit), one of which (ie, Informit) is particularly relevant to the specific context where the results will be applied.
A set of quality appraisal criteria was used to appraise all the studies included in this review to ensure minimal quality.
Qualitative meta-synthesis was conducted by one researcher according to a three-stage method for thematic synthesis.
This review was restricted to a specific implementation context (ie, Australia), to which its results are directly relevant and will be immediately applied and actions will be taken.
The implementation of new health interventions and services into established healthcare practices and systems has been found to be challenging.1–4 The inherent complexity of both health services and healthcare systems may be fundamental to the implementation problem.5 6 According to current health planning approaches, the implementation of health services can be enhanced by comprehensively assessing the context in which they will be delivered. Analysis of the context should consider the stakeholders who can influence or be affected by the health service, as well as the social, physical, economic and policy environments that can enable or hinder the normalisation of the service.2 7 Early identification of these elements (including how they relate to or interact with each other) is a key step for developing suitable strategies and interventions to enhance health service implementation.
In the implementation science literature, several terms are used to refer to the elements that can influence service implementation and practice change. Some generally known examples, which are commonly used interchangeably in the literature,8 are: barriers and facilitators,9 determinants of practice,7 implementation factors10 or constructs.2 The current use of these terms encloses different concepts. For the purpose of this review and to avoid the terminological debate, we have used the term ‘influential element’ as a neutral term.
Amid increasing awareness of the uniqueness of the community pharmacy setting and the positive contribution pharmacists can make to healthcare,11 there has been a shift towards pharmacists providing more professional, patient-centred services. However, the implementation and sustainability of community pharmacy services (CPSs) and the integration of community pharmacists into primary healthcare teams remain a challenge worldwide.12 13 In consistence with this international trend, Australian community pharmacies are eager to provide CPSs and receive remuneration from the government for its provision but are experiencing challenges in the implementation, uptake and sustainability of CPSs.14 Extensive research has been conducted to identify the elements that from the perspective of community pharmacists (ie, service provider) can influence the implementation of CPSs.14–16 However, considering the view of a single stakeholder group is insufficient to comprehensively analyse the complexity of a particular implementation context. These limited analyses can lead to the development of inadequate implementation strategies and interventions. Patients, general practitioners (GPs) and primary care nurses are key stakeholders who interact with or are affected by CPSs and may be able to strongly influence the implementation of such services. These stakeholders may have their own particular views about CPSs and so can complement the findings from previous pharmacy-informed research.14 15 Patients’, nurses’ and GPs’ views and experiences regarding CPSs have been explored in several qualitative studies,17–21 but no review that collates and analyses such information exists. Qualitative meta-synthesis aims to synthesise qualitative literature to provide a new, more comprehensive interpretation of the findings that goes beyond the depth and breadth of the original studies and to broaden the range of concepts identified.22 23 Thus, the aim of this study was to synthesise such qualitative literature to describe the broad range of elements that, from the patients’, GPs’ and nurses’ perspectives, can hinder or enable the implementation of CPSs in Australia.
Search strategy, screening and eligibility criteria
A systematic search was conducted in May 2015 in three electronic databases (ie, PubMed, Scopus and Informit), without time limits, to identify qualitative studies addressing patients’, nurses’ or GPs’ views about CPSs in Australia. A CPS was assumed to refer to an action or set of actions delivered in or organised by a community pharmacy to optimise the process of care, with the aim of improving health outcomes and the value of healthcare.24 For the purpose of this review, CPSs are specific health programmes that are implemented in addition to routine professional activities performed by community pharmacists, which do not require any specific or extra implementation effort (ie, they are part of normal community pharmacy practice). Since medicine dispensing is the main routine activity in the community pharmacy, it was not considered as a CPS so it was excluded. Articles that did not address a specific CPS but interprofessional collaboration (ie, between community pharmacists and other healthcare professionals) were included as they can also provide insight into the elements influencing the implementation of CPSs. Full search strategies are available on online supplementary appendix 1. In addition, the references from the included papers were searched manually for additional relevant studies. A two-step process was performed by one researcher to select studies for the analysis. As a first step, titles and abstracts were screened to identify and exclude non-relevant literature. In the second step, full texts of the remaining articles were reviewed to exclude those that: (1) were not related to CPSs; (2) did not address patient, nurse and/or GP perspective; (3) did not use qualitative research methodology25; (4) did not clearly identify the stakeholder (ie, patient, nurse or GP) as the source of the information; and (5) were not accessible in any of the research team university libraries or unattainable following contact with the authors.
Supplementary file 1
All the included articles were checked by the same researcher for ‘elementary quality assessment’ using the first three criteria delineated by Dixon-Woods et al 26 to appraise qualitative research: (1) was the research question clear?; (2) was the research questions suited to qualitative inquiry?; and (3) were (A) sampling, (B) data collection and (C) analysis clearly described? Articles were excluded when no answer, or an unclear answer, was given to at least one of the three questions.
Qualitative meta-synthesis was conducted by one researcher according to the three-stage method for thematic synthesis described by Thomas et al. 27 The first stage of the analysis involved free line-by-line coding of the original data (study participants’ quotes) and the study authors’ interpretation of the original data. The process of coding involves summarising text from the results and discussion sections of each article into one or more descriptive issues (ie, codes) to capture meaning. The second stage of the process involved grouping codes into one or more descriptive themes. Subsequent articles were coded into pre-existing themes, and new themes were created when considered necessary. To simplify the terminology throughout this article, themes were interpreted as elements (ie, influential elements) that could positively (ie, facilitators) or negatively (ie, barriers) influence CPS implementation or practice change. A barrier was defined as ‘any type of obstacle (material or immaterial) which can impede the dissemination, implementation and/or sustainability of a CPS’, while a facilitator was defined as ‘any type of element (material or immaterial) which can help to overcome barriers and/or accelerate the dissemination or implementation’ of a CPS.16 Themes that were related to similar issues were further grouped to create one broad barrier or facilitator. The identified influential elements were reviewed by a second researcher to assess clarity, consistency and understanding. At the third stage, barriers and facilitators were organised using an adapted version of the Ecological Model (table 1),28 which classified them into four different levels: patient, interpersonal, organisational and community/system. The four levels defined in table 1 were used as an overarching structure, with further subheadings created during analysis, for appropriate allocation and organisation of the influential elements into the levels. The ecological model has been widely and successfully used for planning services in a variety of settings, targeting different populations and problems.29 30 Coding of papers that were identified manually was conducted last. NVivo V.10 software (QSR International Pty; Doncaster, Victoria, Australia) was used to help manage and analyse the data. Once all the influential elements were identified, a second round of analysis was conducted to identify where a connection or relationship was mentioned between two or more elements. Again, both study participants’ quotes and study authors’ data interpretation were reviewed for this purpose. A network representing the identified relationships was generated using a ForceAtlas2 layout31 with Gephi V.0.8. This article has been written following existing guidelines for reporting the synthesis of qualitative research (the Enhancing transparency in reporting the synthesis of qualitative research: ENTREQ Statement).32
The systematic and manual search identified 243 articles once duplicates were removed. After title and abstract screening, 124 full-text articles were assessed for eligibility of which 29 articles were included in the qualitative meta-synthesis (all of them fulfilled the appraisal criteria) (figure 1). A description of the papers included in the review can be found in table 2. Of the 29 included papers, 15 addressed patients’ perspectives only, 2 addressed nurses’ perspectives only, 6 addressed GPs’ perspectives only, 2 addressed nurses’ and GPs’ perspectives together, 3 addressed patients’ and GPs’ perspectives together and 1 addressed the views of all three participants. Twenty-three articles were related to a specific CPS, two were related specifically to interprofessional collaboration, three were related to both CPSs and interprofessional collaboration and one addressed concordance-based healthcare. The articles employed semistructured interviews (n=23) and/or focus groups (n=11) as methods of data collection.
During the first stage of data extraction, 181 patient, 30 nurse and 91 GP codes were created. At the completion of the coding process, 63 influential elements were identified (table 3). These elements were found to exist as a barrier, facilitator or both. In several studies patients, nurses and GPs were able to describe approaches or strategies to overcome specific barriers.17–20 33–43 These strategies have been reported in table 3 as additional facilitators (marked with an asterisk). During coding of the manually identified papers, it seemed that conceptual saturation may have been reached, since no new barriers or facilitators were identified.
Individual patient level
All the 16 elements at the patient level were identified by patients. GPs and nurses did not identify any additional patient-related barriers and facilitators. Influential elements at this level were related to the patients’ needs, preferences, perceptions and expectations, capabilities or previous experiences with community pharmacists and services. Patients’ health-related concerns, understanding or perception of their health problems are important elements that influence patients’ need for healthcare and so their decisions to use CPSs. Most patients held positive views about CPSs and the role of the pharmacist in providing such services.38 40 44 Some articles highlighted that positive experiences were related to the patient feeling comfortable and welcomed in the pharmacy.44–46 When CPSs required a formal referral from the GP, some patients deterred from requesting the services. These patients perceived that by requesting a CPS they would be bothering the GP36 or offending and compromising their relationship with the GP.18 40 47Patients also reported that having a negative experience with a CPS also deterred them from accessing and using such CPSs in the future.46
Influential elements at the interpersonal level were related to two categories or sublevels: (1) individual healthcare professionals (which also includes professional pharmacy staff) and (2) relationships (or interactions) between individuals (which includes both the relationships between healthcare professionals and between those professionals and patients).
Individual healthcare professionals
Seven elements were identified and related to characteristics of the community pharmacists (n=4), nurses (n=4) and GPs (n=4) and characteristics of non-provider personnel (ie, other community pharmacy staff members, eg, pharmacy assistant) (n=5). Articles reported that GPs’ and nurses’ service support varied depending on their perceptions or understanding of CPSs and the role of pharmacists. Home medicine review services had a great deal of approval and support from the GP perspective.40 42 On the other side, pharmacists providing immunisations raised some conflicting views among GPs since they believed this was the role of the GP or nurse practitioner.42 Some studies highlighted that GPs had a limited understanding of the capabilities of the pharmacist as service providers with pharmacists perceived as drug sellers in a retail environment.34–36 48 49 Both patients and GPs implied the need for pharmacists to undergo upskilling and training to be qualified to provide some CPSs.34 37 47
Relationships (or interactions) between individuals
Articles reported that well-established relationships between the pharmacist and the nurse or the GP, including collaborative relationships, were essential for the success of a CPS.17 19 20 35 41 50 Multidisciplinary education and training for healthcare professionals was suggested as a way to improve healthcare professional competence.49 Similarly, characteristics of the relationship between the patient and the pharmacist (eg, trust) was a key element that influenced pharmacy choice, contributed to the patient adhering to the CPS, and accepting the intervention.18 33 36 38 44–46 51 Some articles reported the influence of family and friends on patient utilisation of CPSs (eg, providing support and influencing motivation),35 49 and others commented on the integration of partners into the CPS (eg, provision of group sessions with partners).35 45
Also at the organisational level, influential elements were divided into two sublevels: (1) the community pharmacy setting (n=8) and (2) the service itself (n=8).
The community pharmacy setting
Some articles identified the accessibility of the pharmacy facilitated interprofessional relationships between GPs and pharmacists42 48 and influenced patient17 38 45 and nurse41 participation in CPS. In some articles, non-English speaking patients reported that the lack of multilingual staff limited their awareness and access to CPSs.47 52 Other articles noted GP and nurse concerns regarding the lack of pharmacies that provide CPSs41 and insufficient accredited pharmacists to perform CPSs.40 43
The community pharmacy service
Concerns regarding the validity and accuracy of the tools and instruments used (eg, medical devices and medication charts) were raised by GPs and nurses.19 42Patients and nurses commented that having the same service provider at each encounter facilitated rapport building between the patient and the pharmacist38 45 51 and caused fewer errors when it came to preparing dose administration aids.19 Furthermore, patients, nurses and GPs reported on the involvement/participation of healthcare professionals other than pharmacists in the provision of CPSs,38 or to act as a point of liaison,20 to improve the quality and efficiency of the service. The cost of the service was a key element, mentioned by all stakeholders, that could either discourage41 49 or motivate45 patients to use services. In particular, it was mentioned that smaller, manageable cost payments for patients could facilitate CPS use.41
Community and healthcare system level
Nine influential elements were identified at this level. Several articles identified the need for adequate remuneration for GPs and pharmacists for participating in and providing CPSs17 42 50 52 as well as the implementation of an electronic system of information sharing between these two healthcare professionals.19 20 36 43 GPs also cited the availability of competing, government-funded health programmes and their high level of workload and lack of time as contributing to their low participation in CPSs.40 Where services were available, remunerated and widely supported by GPs and patients, such as home medicine reviews (ie, a medication review service), GPs mentioned complex bureaucratic procedures (eg, completing tedious documents) may discourage their use.17 20 40 43 48 Despite this, the home medicine review service was generally considered successful by GPs and a frequently reported reason for this was the presence of a clear protocol guiding service delivery.20 42 48 GPs also suggested increased and improved collaboration between pharmacy and GP professional representative bodies may improve awareness of the services and encourage participation. The media was perceived to have an important role in improving awareness of and promoting CPSs. Finally, some broad comments suggesting some additional issues at the higher levels of the healthcare system were mentioned, such as ‘better and more responsible organisation of the healthcare system’.43
With regards to the interactions between the identified influential elements, 12 articles out of 29 mentioned some form of a relationship between certain elements.20 33 41 42 44 46–48 50 51 53 54 As shown in online supplementary appendix 2, a total of 27 relationships between 25 elements were found, with 10 elements presenting two or more relationships with others (two elements showed five or more interactions). As a result of the limited, unsystematic information reported in the articles, a sparse network disclosing the recognised relationships between elements was obtained (see online supplementary appendix 2).
To the best of our knowledge, this is the first review that summarises comprehensive information on the elements that, according to patients, nurses and GPs, can enable or hinder the implementation of CPSs. Patients, GPs and nurses are key members of the primary healthcare team and their support and expectations for CPSs can highly influence their implementation.1 19 42 54–57 Thus, by synthesising and organising the influential elements identified by these key stakeholders, this review can optimise future analyses of barriers and facilitators to the implementation of CPSs and so potentially enhance their integration into primary practice. Importantly, this work was intentionally restricted to a specific implementation context (ie, Australia), to which its results are directly relevant and will be immediately applied. Focusing only on Australia is not considered a limitation of the study, rather it is a sensible decision that allows knowledge about a particular context of interest to be gained. Including studies conducted in contexts or healthcare systems other than Australia (eg, UK, USA and so on), where barriers and facilitators to CPS implementation can be dissimilar in nature and expressed differently, may have brought irrelevant or inappropriate information to this analysis, and so hinder the understanding of the context of interest. However, it should be noted that Australia is a country with a large experience in CPS implementation and where significant research has been conducted in this regard compared with other countries worldwide. Therefore, it is expected that the comprehensive list of influential elements identified in this context may be relevant to start investigating barriers and facilitators to CPS implementation in countries with less experience. Furthermore, the elements identified in this review can provide insight to pharmacy service planners in other countries to guess and avoid some problems in the implementation of CPSs beforehand.
Barriers and facilitators to the implementation of CPSs in Australia have been well researched and reported from the perspective of community pharmacists.14 15 56 58 In this regard, the results of this review confirms that patients, nurses and GPs also recognise some of the influential elements reported in previous pharmacist-informed studies, such as the pharmacist’s education and training, collaboration between the pharmacist and the GP, accessibility of the pharmacy setting and financial remuneration. However, this study provides additional insight into further barriers and facilitators, across different ecological levels, that are relevant to other key stakeholders and so are less likely to be reported by pharmacists, for example, patients’ capability to follow the procedures of the service, GPs’ workload, nurses’ attitudes towards other healthcare professionals/services, the actual relationships between GP and pharmacy professional bodies or the availability of multidisciplinary training and education. These results highlight the importance of engaging key stakeholders other than pharmacists to better understand the contexts in which CPSs are implemented. In other words, disregarding the input of these stakeholders (or considering only the views of pharmacists) may lead to an incomplete and biased understanding of the implementation context which, in turn, can result in service underutilisation, unsuccessful implementation and limited service impact.59 Generally, involving relevant stakeholders throughout the development, implementation and evaluation of health programmes is crucial to increase the chances of any of those initiatives being effective and successfully implemented.6 29 30 60 Indeed, this is equally relevant to CPS planning.61 62
Semistructured interviews and/or focus group with healthcare professionals and patients appear to be appropriate methods to identify a large number of unique influential elements.63 Thus, pharmacy service planners can continue to use these methods to identify determinants of pharmacy practice in their own context. Although, the type of qualitative method used may affect the type of barriers/facilitators identified, it is more likely that the aims of the studies included in this review, their target population and/or the specific service/topic addressed by the study may have had a stronger influence in the type of barriers or facilitator identified.
The results of this review can assist pharmacy service planners and researchers to better identify the elements that may be enabling or hindering the implementation of existing CPSs. By combining the list of influential elements generated in this review with previous findings in pharmacists-informed studies, a comprehensive framework to assess barriers and facilitators to CPS implementation can be produced. Assessing and understanding the elements influencing pharmacy practice and service implementation must be a key early step in developing appropriate, multilevel programmes (ie, including interventions targeting elements at different levels) aimed at enhancing the integration of CPSs into the healthcare system.29 30 62 64 Also, influential elements should be prompted and assessed when designing new CPSs. Identifying elements prior to designing a new CPS may guide both the early adaptation of the service to the context, as well as the early development of tailored implementation programmes to better fit (or change) the implementation context. As an analysis of influential elements is likely to yield a large number of items, it would not be feasible to address each and every one of those elements. Thus, once elements have been identified for a specific context, further efforts are required to prioritise those elements that are most relevant and can be practically addressed.8 65 In this regard, McMillan et al 66 provide a summary of methods used to determine priorities and how they have been used in pharmacy practice research, which can guide pharmacy service planners in this regard.
The analysis conducted in this review revealed three concerns that must be considered to improve future studies aimed at identifying influential elements. On the one hand, some influential elements at the community and healthcare system level were too broadly described (ie, ‘organisation of the health system’) and further exploration is needed to clearly understand the specific ‘items’ that they encompass. Presumably, the list of determinants of practice described by Flottorp et al 7 (ie, Tailored Implementation in Chronic Disease checklist) can provide more detail regarding influential elements at the higher community and healthcare system level and so can initially assist to better frame future analysis of barriers and facilitators to CPS implementation. Particularly, the determinants under the domains ‘Incentives and resources’, ‘Capacity for organisational change’ and ‘Social political and legal factors’ seem particularly relevant for this purpose. Importantly, to bring further insight on the elements at the community and healthcare system level, it would be important to include and explore the perspectives of other potential key stakeholders, such as other healthcare providers (eg, specialists), caregivers, representatives of healthcare organisations and professional bodies, policy makers and so on. Furthermore, future studies aimed at identifying barriers and facilitators to CPS implementation must better describe and understand the relationships between elements.2 7 This may help to understand how elements influence each other and which elements are more suitable to be addressed (based on the overall effect that they can produce on other elements) when designing implementation efforts.
The network analysis intended in this study was strongly constrained by the limited and unsystematically reported information about the relationships between influential elements. As a result, it was decided not to report further results of the network analysis beyond its pictorial representation. The potential of a full network analysis should be considered in future studies aimed at analysing elements that influence the implementation of CPSs. A suitable network analysis can help to better understand the complex relationships between these elements, detect the core elements that may primarily explain the implementation challenge and provide insight on the key leverage points that should be targeted within the network to enhance service implementation. Ideally, accurate information on relevant attributes of the influential elements (and the interactions between them) should be collected by the authors of the primary studies to increase the potential of a network analysis, for example, the frequency of occurrence, the direction of the relationships, the domain or level where the element is located (ie, patients, healthcare professionals, professional interactions and so on), the relative relevance of each element or the effect on implementation outcomes (ie, performance as barrier or facilitator).
Following the particular method chosen for this review (ie, qualitative meta-synthesis),22 23 only primary research articles that used qualitative methods were included. Meta-synthesis enabled a rich description of elements perceived by GPs, patients and nurses to influence implementation of CPSs in Australia. Future reviews that synthesise the quantitative literature on this topic are encouraged. Appraising qualitative research is controversial because of the difficulty of using information about quality to inform syntheses (eg, even studies with flaws in methodology can provide valuable information).26 Furthermore, there is no gold standard on appraising qualitative studies.32 The elementary quality assessment conducted in the current review was aimed at ensuring minimal quality while identifying a broad range of elements that might influence CPS implementation. Lastly, the papers included in this review were not restricted by the time at which they were published, since the aim of the study was to include all relevant papers that can inform about any influential element that has been noted in practice. It is important to acknowledge that as contexts can change over time, the effect of influential elements can also change, cease to exist or new elements can emerge. It is therefore important to regularly monitor elements and prioritise those that must be addressed.
This qualitative meta-synthesis identified a broad range of elements that, according to patients, GPs and nurses, can enable (ie, facilitators) or hinder (ie, barriers) the implementation of CPSs. These influential elements are located at different ecological levels and should be considered together with those previously identified in pharmacy-informed studies to comprehensively analyse the barriers and facilitators to the implementation of CPSs. Future studies aimed at that purpose must involve multiple stakeholder groups (ie, others than only pharmacists) and better understand the relationships between influential elements to increase the usefulness and interest of their findings. Further to the identification of the influential elements, key stakeholders should keep involved in developing suitable, multilevel programmes aimed at enhancing CPS implementation.
We would like to acknowledge Antonio E Mendes (Universidade Federal do Parana, Brazil) for his collaboration in the network analysis.
75th International Pharmaceutical Federation(FIP) World Congress of Pharmacy and Pharmaceutical Sciences; 29 September-3October, 2015; Dusseldorf, Germany
Contributors Conception or design of the work: LNH, FF-L, TL and DS-H. Data collection: LNH, DD and LF-T. Data analysis and interpretation: LNH, JCM, CB and DS-H. Drafting the article: LNH, FF-L, TL and DS-H. Critical revision of the article: LNH, JCM, FF-L, TL and DS-H. Final approval of the version to be published: all authors.
Funding LNH was awarded a University of Technology Sydney (UTS) President’s 424 Scholarship and a UTS Chancellors Research Scholarship. This work is part of a larger UTS Chancellor’s Postdoctoral Research Fellowship awarded to Dr DSH (UTS ID number: 2013001605).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement No additional data are available.
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.