Search strategy.

The search strategy utilised a range of systematic and serendipitous methods to identify relevant studies [25]. First, we searched two electronic database platforms–SCOPUS and Web of Knowledge–from inception until April 2015. These were searched in the title, abstract and keywords using key terms for translational research combined with thesaurus and free text terms for the subject areas of health and medicine (see Table 1). The titles and abstracts of returned articles were scrutinised and full text articles obtained for papers likely to meet the inclusion criteria. Each potentially relevant article was retrieved and read in full by two authors to determine whether it met the inclusion criteria. After identifying relevant papers through database searching, the reference lists of retrieved articles were searched for other studies which might meet the inclusion criteria. We searched personal biographies of experts in the field. Finally, we undertook citation tracking in Google Scholar of all included papers, to identify any additional studies. Fig 1 illustrates the flow of studies through the stages according to PRISMA [26].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flow diagram of the different phases of the systematic search and review based on PRISMA [26].

https://doi.org/10.1371/journal.pone.0160475.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Search strategy.

https://doi.org/10.1371/journal.pone.0160475.t001

Quality appraisal.

Two authors (NF, ES) independently appraised the quality of each study using a five-point checklist comparing quality scores given to each paper to reach a consensus [27]. This checklist assesses the methodological quality of studies and can be applied to empirical papers regardless of study design. Quality appraisal involved scoring each paper out of five according to how well the following criteria were met:

1. Are the aims and objectives of the research clearly stated?
2. Is the research design clearly specified and appropriate for the aims and objectives of the research?
3. Do the researchers provide a clear account of the process by which their findings were produced?
4. Do the researchers display enough data to support their interpretations and conclusions?
5. Is the method of analysis appropriate and adequately explicated?

Papers scoring four or five were considered to be of high quality. Papers scoring 1 to 3 only partially met the five criteria and were judged to be of low quality. Given the lack of consensus surrounding the use of quality appraisal in qualitative research, with suggestions that appraisal scores can be more reflective of the written report rather than the actual study and that there is a risk of discounting important studies for the sake of ‘surface mistakes’ [28, 29], we did not exclude studies with low quality scores. Instead we used this quality appraisal to assess the robustness of the synthesis.

Theme 1: Concepts of translational research.

Only one study in our sample explicitly looked at how scientists defined and understood the concept of translational research [44]. Morgan et al. conducted interviews with basic and clinician scientists in 2008 at a time when the ‘requirements of translational research were only beginning to emerge’ [44, p.948]. Both types of scientists reported awareness of the concept and its current policy emphasis, describing it as the ‘mantra of the moment’, but were unclear as to its meaning, and were able to give only a minimal definition such as ‘to try to move stuff from the lab to the clinic’ [44, p.948]. The other included studies, particularly those which utilised survey methods for data collection, did not report explicitly asking respondents to define translational research.

The type of translational research model adopted by an organisation or institution was perceived by scientists in a number of studies in our sample to influence the practice of translational research, with a linear model seen as an impediment to successful practice [33, 34, 40, 45, 53]. Institutional representation of translational research as a ‘pipeline’, and a process which requires acceleration, was articulated by basic and clinician scientists as problematic for a number of reasons: that science was seen as being reduced to solving problems and producing cures rather than discovering new knowledge [45, 53]; a lack of recognition that good science takes time [40, 45, 53]; and rapid selling of biomedical innovations to the public due to pressures to translate findings quickly for patient benefit without thinking through the implications of these developments or making better use of existing policies and interventions [34, 40, 45, 53]. For example, scientists working in the field of addiction research acknowledged that: ‘translation takes time, that bodies of knowledge are built slowly over many years, and that basic science has value even in the absence of swift translation’ [45, p.4].

These views contrast with institutional and policy interpretations of translational research which assume a linear model, impeded by ‘blocks’ that act as barriers to translating laboratory discoveries into improvements in human health [for example see 7]. The methods adopted by researchers in our sample allowed scientists to present a more subtle and nuanced view of the problems attributed to translational research and how they should be addressed [40, 51]. For example, basic and clinician scientists interviewed in one US study described challenges of translational research as ‘dilemmas’ as opposed to the oversimplified notion of ‘barriers’ and ‘blocks’ typically perceived by external agencies, including policy makers and research funders [40]. The preferred term ‘dilemmas’ indicates that there are sometimes very valid and important reasons why things do not unfold according to the expectations that policy makers and research funders may have.

Viewing translational research as a circular or iterative process, was seen to facilitate translational research practices by encouraging reciprocal interactions between the lab and the clinic. This was thought to promote a collaborative research environment, which in turn attracted basic and clinician scientists who wanted to collaborate (this is elaborated upon further in the theme ‘interdisciplinary collaboration’) [34, 41, 52, 57]. A case study of eight translational research organisations (TRO) in China elaborates on this point. In this study, TROs were established primarily by biomedical organisations. This limited the potential for TROs to address intractable problems because key disciplines required for successful translation, such as public health, health policy, social sciences, community engagement, had not been included [57].

Theme 2: Research processes.

Research processes were perceived by basic and clinician scientists across a number of studies to enable or hinder translational research practices. Research processes included: regulatory and ethics processes [32, 33, 38, 49, 52, 54, 56]; patient recruitment to research [32, 35, 38, 52]; and informatics and information technology [38–40, 46, 47].

Several studies identified complex regulatory processes, such as ethics and research governance, as barriers to translational research, in effect slowing it down [32, 38, 49, 52, 54, 56]. Two surveys of senior researchers working in US medical schools and academic health science centres, found that 38% and 54% of those surveyed identified complex regulatory requirements as particularly challenging for translational research, thus limiting the success of biomedical innovation being translated into benefits for patients [32, 54]. Four qualitative studies, from the field of stem cell research, explain why regulatory processes are particularly challenging for translational research [33, 49, 52, 56]. The rudimentary nature of stem cell research, yet to determine whether stem cell therapy should be defined as a drug or medical technology, complicated regulation in two studies in our sample [33, 49]. Scientists in both these studies reported that this led to confusion over which authority should oversee regulation–a human tissue authority or a food and drug authority. Scientific and technological developments and their accompanying ‘imagined futures’ therefore created uncertainty on the part of both scientists and regulators as both parties sought to develop and refine interpretations of rules and regulations, as this extract from a field note illustrates:

[The representative from the regulator] noted that the views upon this were different across the EU [European Union]. He reiterated that it’s hard to know until people make medicinal products what the regulations and guidance should be. But the [regulators] are asked to give guidelines anyway even though this is not known, and he said ‘it’s circular, it goes around and around and around’ [49, p.351].

In another Chinese study, also investigating stem cell research, regulation was complicated by numerous, overlapping regulatory jurisdictions inadvertently promoting inconsistency and minimal conformity with the law, resulting in scientists feeling powerless to change the system [56]. Ethical and social implications of scientific breakthroughs were perceived to add an additional layer of complexity to translational research. Scientists working on stem cells as a potential therapy for leukaemia [33] and diabetes [52] reported making a deliberate effort to follow strict regulatory processes to ensure acceptance and legitimacy of their research [33]. They argued that public attention to the ethical and social implications of stem cell research breakthroughs would limit the move from bench to bedside:

That then requires a wholesale change of ethical thought as to whether you can put some genetically modified cell back into a human, and that’s going to take years and years of legislation [52, p. 2061]

Basic and clinician scientists articulated mixed views about the role of patient recruitment and collection of patient samples and the implications this has for translational research [32, 35, 38, 52]. In one US study of medical school research leaders, 37% of respondents identified recruitment of research subjects as challenging. When asked whether formal institutional processes and adaptations to counter these challenges had been initiated, 34% identified processes to aid recruitment of research participants, with more than half (54%) reporting that this innovation had a moderate to large effect on the amount of clinical research conducted [32].

Inherent within translational research is the assumption that animal studies will lead to human studies. However, scientists from two studies in our sample [33, 52] commented on a tension between the ‘relevance of ‘human studies” and the ‘rigour of ‘animal experiments” [52, p.2061]. In one study, based in the UK, basic scientists reported that experiments on animals, due to their availability, were seen as preferable and more likely to lead to outputs in the form of publications. In comparison, experiments on donated human cells were considered to carry a greater risk in terms of output due to difficulties obtaining samples from patients and their families:

To plan a proper research programme you need regular access to the tissue and you are never going to get that with [human] donor material, particularly if the primary objectives are research, because the relatives just don’t see it as important to give permission for all this to be used for research. Quite rightly I think, they don’t feel it’s going to somebody else’s benefit and research is going to be a lower priority. [52, p.2061]

In the other study, based in China, drugs that had proved effective in animal studies were not effective in human clinical trials because patients’ health conditions were far more complicated than the animal models suggested, requiring scientists to constantly adjust their clinical trial protocols [33].

Solutions proposed to address the problems of recruitment to clinical trials and promote interactions between clinical researchers and potential recruits included community outreach and engagement projects and community advisory boards [38]. Results from one Canadian study suggested that state funded systems of care, such as Canadian Academic Health Science Systems, facilitated translational research due to the availability of a population of patients readily accessible to the researcher [35]. Scientists argued that it is the publicly funded system of care that has made access to patients and their data possible, contrasting this system with that of their neighbours in the United States, whereby the fragmented, privatised systems of care limits the potential for such a valuable resource:

[The population] is very special and to not take advantage of it would be a huge loss because we can do things here that other people can’t do. [35, p.721]

Basic and clinician scientists in five studies identified a lack of infrastructure to develop skills in translational research as a further organisational barrier. These included acquiring research skills and access to equipment and effective information technology systems. [38–40, 46, 47]. Effectiveness of clinical and translational science, particularly with regards to accessing information technology systems, was perceived by a number of these scientists to be determined by organisational and leadership factors [46, 47]. For example, a translational research grant awarded to a medical centre was praised by scientists for enabling the widespread provision of infrastructure such as bioinformatics, which was previously unavailable to individual scientists [40].

Theme 3: Research versus clinical care.

Basic and clinician scientists in four studies identified a perceived cultural divide between basic and clinical science as a key barrier to conducting translational research [44, 52, 54, 55]. In one of these studies, clinician scientists engaged in early stage clinical trials of stem cells understood that such a divide was due to differing scientific worldviews, language and needs among scientists. For example, one clinician scientist said:

[It’s] still very apparent that we do face, this massive void that exists between scientists and clinicians, that for most of the part, certainly in our area, seems to exist, with no great understanding of the needs of both. [55, p.504]

For other scientists, it was no great surprise that basic and clinician scientists were divided given they were educated within different faculties, with differing foundations, management approaches, and hierarchical systems but were then expected to come together to undertake translational research [52, 57]. In contrast, a UK study which interviewed basic and clinician scientists working in cancer genetics found no clear division between clinical practice and research [36]. While, in theory, research and clinical care may be seen as highly differentiated, in practice, as results from this study bear out, the situation may be more complex. The boundary between clinical care and research was characterised as ambiguous, fluid and flexible. This was attributed to the exploratory nature of medical practice sharing similar motivations and procedures to that of clinical research, with both seeking to further knowledge. The fluidity of the relationship was considered to benefit patients (who through research participation gained access to DNA tests not available within clinical services), as well as clinical staff, improving their clinical skills through engagement with research processes [36].

Organisational and structural factors were perceived by clinician scientists in several studies to influence the apparent separation between research and clinical care, which consequently acted as a barrier to translational research [32, 39, 40, 42, 51, 55]. These included: a lack of training in relevant research skills and training being too time intensive [32, 42]; a lack of time to undertake research among clinician scientists due to demanding clinical roles [32, 40, 55]; and the pressure of combining clinical service and research roles [32, 40]. For clinicians working in a community setting, a perceived lower value attributed to research compared to clinical care, with its associated lack of recognition, status and career progression, deterred their participation in research [39]. Clinician scientists reported that having to compete with full-time, non-clinical researchers, perceived as having more time and being better embedded in infrastructures to secure research funding, was a barrier to conducting translational research [42, 51]. This led to the perception among clinician scientists that research was not always considered to be an economically viable activity due to a lack of remuneration for clinical staff’s time and effort to conduct research [39]. The value of conducting translational research for clinician scientists was strongly linked to perceived patient benefit of the research; research considered likely to have low patient benefit was viewed as disruptive to clinical care [36, 39].

In a more positive light, clinician scientists interpreted their role as enabling translational research because they played a key role in bridging the perceived cultural divide between basic and clinical science [33, 39, 41, 44, 52], or by acting as ‘agents of change’ due to the strength and breadth of their knowledge [55, p. 503]. Clinician scientists viewed themselves as ‘boundary-spanners’ or ‘important bridges’ between the laboratory and clinic, acting as mediators and translators in collaboration with ‘pure’ clinicians and scientists, whom they relied upon to provide the in-depth clinical and scientific knowledge required for successful translational research [41, p. 542]. In a UK study one clinical-scientist clearly articulated the benefits of adding laboratory work to clinical responsibilities as part of a dual role:

Mechanistic studies are what I do. To do the clinical work [without the lab work] you wouldn’t have learned anything about how it worked and that seemed a bit of a shame to me…you might as well do the study properly. [44, p.950].

Intellectual differences between basic and clinician scientists in terms of what was considered to constitute ‘legitimate science’ shaped both positive and negative views of translational research among scientists [44, 52, 54]. Basic scientists articulated a greater emphasis on scientific discovery rather than translation research [44, 54], whereas clinician scientists placed more focus on patient-related outcomes to enable translational research [52]. For example, in a US survey of research investigators at a Medical School the most common reason given by basic scientists for not pursuing translational research was because it was not considered central to their research agenda; respondents with PhD degrees were significantly less likely to report they were conducting translational research compared to those with MD or MD/PhD degrees [54].

Basic scientists in five studies tended to hold negative views about translational research as a form of legitimate science [44, 45, 51, 54, 55]. They articulated the view that translational research was not central to their research agenda [45, 54], and perceived clinician scientists as having greater authority to conduct clinical trials to enable translational research practices [55]. Basic scientists perceived reward and career progression as more difficult to achieve in the field of translational research and shared significant concerns about how translational research as a form of science was viewed by their peers and promotion committees [44, 54]. They questioned how they would be able to retain standing in the field if they were publishing in translational research journals, instead of their key disciplinary journals, the former not ‘sufficiently valued by their peers to form ‘authentic’ knowledge’ [44, p.949]. The downsides of engaging in translational research for non-clinician scientists’ career progression were further emphasised by the tendency for translational research policy to focus on the needs of clinician scientists rather than providing new career structures for basic scientists to work within the requirements of translational research [44, 51]. For example, a basic scientist reported how exasperating she found it that so much attention was paid to ‘clinician scientists when other professional trajectories might also lead to the establishment of a class of translational investigators’ [51, p.8]. In comparison, most clinician scientists interviewed were more positive about the translational research drive. Translational research was viewed as aligning closer to their own research interests in answering particular health-related questions.

Theme 4: Interdisciplinary collaboration.

Interdisciplinary collaboration between basic and clinician scientists, but also with other professional groups, was perceived by scientists to facilitate translational research practices [37, 38, 40, 48, 52, 56, 57], by providing opportunities for knowledge exchange [37, 52], offering distinct forms of expertise [48, 56], and creating a working environment which encouraged communication and co-operation between different scientists [38]. Collaboration was seen as being best achieved through multi-disciplinary teams, working throughout the entire research process [40, 52], or through ‘team science’–a field of inquiry to understand and enhance processes which facilitate or inhibit collaboration of researchers across different fields and organisations [37, 48]. For example, one basic scientist working in human embryonic stem cell research in the UK said:

This is a good place to be because you have got people who make hES cells and this group is a recognised area or centre for expertise in beta cell biology. If you put the two together then you progress quite quickly [52, p.2058].

In contrast, scientists working on stem cell research in China in another study reported that the typical research team structure consisting of ‘one professor and many students’, with few, if any, middle ranking researchers was a hindrance to research efficiency and productivity:

In China, everybody is a professor; everybody works on their own project; there is no connection between groups. Everybody is their own team-leader. Thus it’s hard to make progress [56, p. 197].

Perceived facilitators of collaborative working in a translational research network included geographical proximity of different professionals to enable sharing of values and knowledge exchange [34, 43, 52], cohesive bioinformatics and clinical-informatics teams [46] and the appropriate institutional and structural arrangements to accommodate the range of professional expertise required to meet research objectives [56]. Conversely a lack of collaboration between industry and clinical science [50] and poor coordination between translational research and biomedical informatics teams [46] inhibited interdisciplinary collaboration.

The nature of relationships within interdisciplinary collaboration teams was perceived among scientists as another factor shaping collaborative working, which could either enable or hinder translational research. Specifically, translational research was considered to be facilitated by past working relationship practices [43], prior experience of collaboration among scientists across departments [48], and teams with key members who could act as knowledge brokers [43]. Institutions which portrayed themselves as places of collaboration were considered to enable translational research as this collaborative attitude would in turn attract other scientists who believed in collaborative ways of working as an integral part of conducting research [52].

In contrast, scientists identified a number of characteristics that hindered effective collaborative working relationships and practices. Such factors reflected previous professional groupings that encouraged silo-working (where departments or groups do not want to share information or knowledge beyond their group) rather than collaboration in a translational research network [43, 56, 57], and poor leadership or weak mentorship skills of leaders of teams [40, 57]. Other perceived barriers pointed to a lack of experience among scientists collaborating across different departments to form effective teams [48], as well as institutional arrangements which not only inhibit collaboration between research groups within the same institution but also make collaboration between institutions harder [56]. Finally, traditional relationships between academia and industry, with universities as providers of basic research knowledge and industry as translating this knowledge into applications and profits, were viewed by scientists to discourage collaborative working and therefore hinder translational research practices [34, 52].

Theme 5: Entrepreneurial science.

From a policy perspective, translational research endeavours to improve the health of the nation through the development of effective drugs and treatments, while simultaneously increasing wealth by generating income from a nation’s research capabilities. Translational research therefore comes with health and wealth-based performance measures for those working within translational research environments. A number of papers included in our synthesis suggest that scientists’ research practices are being shaped by this policy agenda [33–35, 41, 44, 51, 53, 57]. They reported a perceived shift in attitudes among scientists and organisations to become ‘entrepreneurial’, thereby facilitating translational research [33–35, 53]. Scientists who considered that both the intellectual and commercial benefits of their research would facilitate translational research, and those with perceived entrepreneurial skills, saw themselves at an advantage to secure funding for translational research.

In addition to the concept of the ‘entrepreneurial scientist’, the concept of the ‘entrepreneurial hospital’ was proposed as a resource for enabling translational research [35]. Scientists in this study considered accessible patient populations, attending publicly funded hospital-university collaborations for their care, as a resource to attract partnerships with industry:

So the drug company comes in for example to develop a drug with us on a phase I study. If we were to extract information at the molecular level for individual patients both of their DNA as well as changes to their tumour, we would get a very good understanding as to who’s responding to that treatment. We would then be able to advise the drug company as to where we see the best outcomes in terms of patient populations that respond well. This, presumably would be passed on when it comes to developing their Phase II and Phase III studies and actually improve the likelihood of success [35, p.721].

Accessible patient populations as a resource had value in terms of rationalising medicine and saving money as well as meeting the commercial aims of translational research. The challenge, as one scientist commented, was ensuring that the mandates of the company and the hospital are both met:

no matter how they pose the issue of what their mandate is, the company is there to safeguard shareholder value, right? And that usually means sales, how much money have you made? [Our organisation] requires access to those companies’ drugs in many cases to take care of our shareholders, right? But if you look at it as shareholder, what our shareholders are interested in is not making money but being treated appropriately [35, p.722].

Results from several papers suggest the need for a broader view of translational research whereby not-for-profit institutions have an entrepreneurial role to play [34, 35, 41, 57]. One study identified the need for organisations such as universities to develop the technology to enable translational research opportunities, particularly when industry had lost interest due to patent problems [41].

In two studies, scientists identified the lack of knowledge and interest concerning commercialising research as a barrier to translational research [34, 44]. In a UK study, basic scientists highlighted organisational and structural barriers as limiting their ability to exploit the commercial potential of their research. These were: a lack of awareness of how to patent scientific discoveries or even that this is a necessary step in translational research; and academic funding systems which employ scientists on closed contracts, with institutional measures of performance (and implications for renewal of contracts) based on publication record rather than patents. For example, one basic scientist observed:

my instinct is I’m wasting my grant time doing that sort of work, I’m rewarded for publications not patents [44, p.949].

However, respondents in a US study highlighted the value of technology transfer offices within a university for taking on industrial collaborations in situations where scientists were less interested in the commercial exploitation of their discoveries [34].