Barriers to research in palliative care in France
- Wadih Rhondali1,2,
- Julien Berthiller3,
- David Hui2,
- Sriram Yennu2,
- Veronique Lafumas1,
- Mathilde Ledoux1,
- Florian Strasser4 and
- Marilène Filbet1
- 1Department of Palliative Care, Centre Hospitalier de Lyon-Sud, Hospices Civils de Lyon, Lyon, France
- 2Department of Palliative Care and Rehabilitation Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
- 3Pole Information Médicale Evaluation Recherche, Hospices Civils de Lyon, Lyon, France
- 4Oncological Palliative Medicine, Department of Internal Medicine and Palliative Care Center, Cantonal Hospital, St Gallen, Switzerland
- Correspondence to Dr Wadih Rhondali, Department of Palliative Care, Centre de soins palliatifs Pavillon 1K, Centre Hospitalier Lyon-Sud, Hospices civils de Lyon, 165, chemin du Grand-Revoyet, Pierre-Bénite cedex 69495, France;
Purpose Palliative care (PC) needs expansion of its research capacity to improve the quality of care. This is particularly true for France that contributed less than 2% of all European PC research publications. We conducted a survey to assess the barriers French healthcare professionals (HCPs) involved in PC research had to face.
Methods We surveyed all 420 PC departments registered in the French National Association for Palliative Care (SFAP) database using a previously used questionnaire. We documented the ethical limitations, time constraints, financial resources, patient issues and methodological issues related to PC research.
Results We obtained 382 responses. Ninety-two (24.1%) HCPs were involved in a research project during the last 5 years. In univariate results, predictors of being involved in PC research were men (p=0.004), physician (p=0.022), working at a university hospital (p<0.001). There was a trend towards working in a PC unit (p=0.052). The main barriers to participating in PC research were lack of time (80.1%) and patient issues (47.4%). Lack of methodological support (33.0%) and financial limitations (30.4%) were also reported as major barriers.
Conclusions There is a consensus that PC research and publication in the English language for peer-reviewed journals must be expanded in France but at this stage, clinical teams still lack specific funding and appropriate support. More research is needed to confirm our results and to determine the best ways to develop PC research capacity in France.
It is much more difficult to develop evidence-based recommendations in palliative care (PC) than in other specialities such as oncology. However, these recommendations are necessary for PC to maintain credibility in academic medicine and to ensure that we deliver optimal evidence-based care to patients and their families.1 ,2 Over the past 15 years, PC has witnessed considerable growth in research, as indicated by an increased number of scientific publications.3 For example, among the 40 327 articles on PC indexed in the MEDLINE database (using the keyword ‘palliative care’), 6268 were published between 1995 and 2000, 8910 between 2000 and 2005 and 11 079 between 2005 and 2010.
However, when we searched for published randomised controlled trials (RCTs), which represent the ‘gold standard’ for biomedical research, we found a limited increase (221 articles between 1995 and 2000, 245 between 2000 and 2005 and 267 between 2005 and 2010) and even a decrease in the proportion of total PC articles (from 221 out of 6268 articles (3.5%) to 267 out of 11 079 (2.4%)).These results have been confirmed by a recent review of PC publications that also identified deficiencies in the quantity and quality of RCTs.4 ,5
In France, research development in PC has been organised in part by the French Society of Palliative Care, which created a research commission in 1998. However, during the 2011 congress of the European Association of Palliative Care (EAPC) in Lisbon, only 12 French abstracts (out of a total of 1106) were presented. Similarly, Hui et al4 identified only nine French publications out of 471 from Europe (2%) over a 12-month period.
To increase PC research in France, it is important to first understand the barriers to conducting research. Therefore, we conducted a survey to identify the potential barriers to conducting PC research in France.
We adopted the same methodology as a survey conducted in Switzerland by Binswanger et al.6 The original aim of that survey was to identify the potential barriers to conducting PC research in Switzerland. To explore those barriers, the researchers created a questionnaire that was based on the data collected previously from the individual (n=14) and group (n=14) interviews among PC healthcare professionals (HCPs) in Switzerland.
This questionnaire consisted of four sections: (1) demographics, (2) employment status, (3) research activities, (4) a question asking respondents to rank the top three barriers among a list of nine obstacles (ethical limitations, time constraints, financial resources, patient issues (eg, fatigue, vulnerability, inclusion criteria), methodological and support issues, interdisciplinary collaboration constraints, other limitations (eg, clinical or educational obligations), language limitation and political issues). The last section was documented only by participants involved in a research project in the last 5 years. Participants were allowed to select more than one answer from the list of obstacles, but they were requested to rank only the three main barriers.
For the current study, two PC physicians (WR, MF) independently translated this questionnaire to French. They discussed all the discrepancies to obtain a final questionnaire.
We sent the questionnaire to all French PC organisations (mobile teams and units) registered with the French Society of Palliative Care (n=420) on 1 September, 2010 by email along with a stamped envelope for return. It included an explanatory note for the survey. We asked the respondents to answer within a month. We included in our analysis all the questionnaires received during this time frame as well as the questionnaires received in the following month. We did not receive any questionnaire after this time point. We did not send a reminder because the results from the original Swiss study reported that most of the participants involved in that research answered within the requested time frame.
This study was submitted to our local institutional review board and we got a waiver as the survey was done for HCP without any private questions.
Categorical variables were expressed as numbers and percentages. We determined whether the distributions of values for continuous variables were normal or non-normal using the Kolmogorov–Smirnov method. Quantitative variables were expressed as means±SDs when the distribution was normal or medians and minimums and maximums when the distribution was not normal. Categorical variables were compared using the χ2 test or Fisher's exact test when the conditions of the χ2 test were not met. Quantitative variables were compared between the groups using Student's t test after a verification of the equality of variances when the data were normally distributed and using the Mann–Whitney non-parametric test for independent series when the data were not normally distributed.
A logistic regression was conducted to quantify the association between explanatory variables and the participants’ involvement in a research project. We first conducted univariate analysis, and then included the variables with a threshold significance of p<0.1 to build a multivariate model.
For the categorical variables, we reported the individual p values rather than the general p value for the full category. All tests were two-sided and we considered the threshold p<0.05 as statistically significant in multivariate analyses.
Statistical analyses were conducted using SAS V.9.1.3 (SAS Institute Inc, Cary, North Carolina, USA).
Responses were received from 134 (35%) physicians, 152 (40%) nurses and 96 (25%) psychologists (table 1). Participants were predominantly women (310/382; 81.2%) and most respondents were working in public institutions (70.4%) and/or general hospitals (58.6%). Ninety-two HCPs (24.1%) had participated in a research project during the last 5 years.
Participants involved in a research project in the last 5 years
In univariate analysis, HCPs involved in research in the last 5 years were significantly associated with male sex (OR=2.62, p=0.004) and physicians (vs nurses (OR=0.66) or psychologists (OR=0.41), p=0.022), working at a university hospital (vs a general hospital (OR=0.32, p<0.001)) and spent more time dedicated to PC (p<0.001). There was also a trend for respondents to be HCPs working at a PC unit (vs another type of unit, OR=0.41, p=0.052; table 2).
Multivariate analyses confirmed that sex, time spent on working in PC and locations of work were independent factors explaining the involvement in research in the last 5 years. Men were 2.1 times more likely to have done research than women in the last 5 years and each additional 1% of the time spent in PC significantly increased the chance of having done research by 1.5%. Similarly, staffs working at general and regional hospitals were 3.0 and 6.5 times less likely to be involved in research than staffs working at a university, respectively (table 2).
Reported barriers for palliative care research
More than 75% of the HCPs involved in a research project reported some difficulties in conducting research because of the vulnerability of the study population (tables 3 and 4). The HCP reported a high rate of dropouts or missing data (many patients are not able to complete the forms because of weakness or delirium).
They also reported a lack of protected time (unexpected clinical tasks often encroach on the time dedicated to research) that led to difficulties in both participating in and leading research projects.
Needs for specific information or support (several research areas represent an unknown area for HCP, such as biostatistics were reported by participants suggesting that HCPs involved in research should benefit from adequate training and appropriate support infrastructure.
When we assessed the three main barriers to conducting research, we found similar results between the entire sample and participants involved in a research project during the last 5 years: time constraints were reported by 306 (80.1%) HCPs, patient issues by 181 (47.4%), methodological and support issues by 126 (33.0%) and financial resources by 116 (30.4%). Interdisciplinary collaboration and ethical limitations were only reported by 63 (16.5%) and 60 (15.7%) of the HCPs, respectively (table 5).
The aim of this survey was to identify the potential barriers to conducting PC research in France. Our main findings suggest that French HCPs are significantly more likely to be involved in a PC research project in the last 5 years if they were physician, men and working at a university hospital. This is consistent with the fact that women represent less than 35% of the government sector researchers and less than 20% of the researchers in the business enterprise sector.7–11 However, our results should be interpreted with caution because 81.2% of our sample were women, which may limit the generalisability of our findings.
Furthermore, physicians in PC units and in the teaching hospitals were more likely to be involved in a research project. This result stresses the importance of PC units as places of training and teaching excellence.12 In France, there is a very low level of research training for nurses and psychologists, which results in almost no academic careers for nurses and mostly teaching careers for psychologists.
The main perceived barriers to participating in French PC research projects were lack of time (80.1%), patient issues (47.4%), lack of methodological support (33.0%) and financial resources (30.4%).
Time constraint has been previously reported as a barrier to participating in research projects. However, on the basis of previous research,13–16 we were surprised to find time constraint as the most important factor. HCPs involved in simultaneous clinical and research activities often have difficulties preserving their dedicated research time. This phenomenon might be partially explained by a heavy clinical load, which reduces the time available for research.
PC patients are generally frail, heterogeneous (both demographically and medically) and usually suffer from multiple physical and psychosocial symptoms. This vulnerability makes prospective studies with multiple assessments more challenging in PC patients than others. Moreover, patients are generally referred to PC late in the disease trajectory,17 which significantly limits enrolment of patients with higher function.
Cognitive assessment should be conducted carefully before including patients because confusion and delirium are very frequent in the PC setting.18–22 Delirium is often under-recognised, which is unfortunate because approximately 50% of the delirium episodes are potentially reversible.18 ,23 ,24 Undiagnosed delirium can lead to specific difficulties in the implementation of protocols (eg, patients will give wrong answers and have difficulties understanding the instructions), introduce potential bias and reduce the statistical power of a study (eg, attrition and missing data).
Furthermore, the PC population is often reluctant to participate in research projects such as RCTs. This reluctance is in contrast to patients in earlier stages of disease.13 This is because patients generally value the benefits of symptom control trials less than disease-modifying trials (ie, potential survival benefit).25 ,26
Staff and methodological issues
Methodology development and data analysis are also common research barriers in PC. Researchers must select or design assessment tools that do not overwhelm patients.3 ,27 Consequently, existing assessment tools may not be appropriate for PC patients because they have been developed for and/or validated in other populations.28 Another important issue for research in PC is the very high rate of attrition caused by patient death or increased symptom burden.29 These effects should be anticipated and correctly included in the study design and sample size calculations.28–30 Pilot studies should also be conducted when appropriate to specifically test the instruments used.
Financial resources have been discussed by several authors showing that funding for PC research is insufficient.14 ,15 ,31 However, financial burdens are also related to time constraints; grant writing and fund raising require time. Without funding, it is very difficult to start a research programme.
Another important consideration is that French PC investigators involved in cancer care need to compete against other oncology projects because there is no PC-specific funding from the French government. Funding agencies need to be convinced that investing money in PC can improve patient care and reduce healthcare costs. Only three of the 58 projects (5%) funded by the French Hospital Clinical Research Program for cancer in 2010 were related to PC or supportive care, two of which were interventional studies. To improve funding for PC, we recommend that all funding agencies have PC-specific panels. Such a change could ensure that PC specialists have more equitable access to research funds.
PC patients are particularly vulnerable and frail, and their level of dependence on caregivers and HCP is important. Thus, HCP should be very cautious during the informed consent process to minimise coercion. This unbalanced relationship between patients and HCP could partially explain why HCPs are often reluctant to include patients in PC research protocols. Such ethical barriers are not specific to PC and can be found in several settings. It is important to provide detailed study information to participants and their caregivers, who may require greater explanation of PC trials than participants in other trials.32
Limitations of this study
Our survey was declarative and anonymous. We were unable to check the publication record of HCP's self-reporting. Future studies should collect objective measures of research involvement such as the number of publications and grants. Indeed, the main barriers reported by our sample are similarly experienced by several other countries in Europe and around the world.33 Our list of barriers was not developed specifically for French investigators. This may result in underdetection of various barriers such as difficulties with the English language. In fact, the vast majority of the French research in PC is published in the National journal in French (Medecine Palliative. Soins de support – Accompagnement—Éthique). Unfortunately, this journal is not included in the bibliographic databases such as MEDLINE. As a result, few French PC publications are circulated internationally. Another barrier not captured in this study that might explain the small number of French publications, is the lack of specific training regarding research for undergraduate and postgraduate students.
Finally, there are many other factors that cannot be assessed with a single questionnaire and will probably require further qualitative study. For instance, physicians working in PC may do so to escape the impersonal aspects of research medicine and thus they may place less emphasis on palliative medicine.
There is an important consensus towards the development of PC research capacity in France. We propose several solutions to overcome the aforementioned methodological barriers. First, we need to minimise study burden and design eligibility criteria to account for frailty among PC patients.3 ,27 Second, we need to develop and standardise the measurement tools specifically for PC. Third, we need to maximise the enrolment of PC clinical trials through multicentre multinational involvement.
As Cecily Saunders recently stated, palliative medicine will only grow by conducting high-quality research that leads to evidence-based recommendations instead of recommendations based on expert opinion, which is currently the case.36 Therefore, it is necessary not only to increase the quantity but also the quality of studies.
Contributors WR was the principal researcher and was responsible for supervision and management of the research and publication (including drafting and submission). WR is also the guarantor of the paper. JB was the research statistician and he contributed not only to the research conception and design but also in the analysis of the data and the drafting of the manuscript. DH and SY were associate investigators and had contributed to the analysis of the data, the drafting of the manuscript. VL, ML, FS and MF were co-investigators and contributed to the research design, the data collection, the analysis and interpretation and to drafting and finalisation of the manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
- Received 30 August 2012.
- Revision received 20 February 2013.
- Accepted 13 March 2013.
- Published Online First 18 April 2013
- Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://group.bmj.com/group/rights-licensing/permissions