Health services research is vital in any medical specialty. In psychiatry, it has become more high-profile with the reconfiguration of different approaches to delivering care to patients. Research in this area often appeals to clinicians, who feel that the findings may be applicable to ‘real life’ clinical experience. However, many become disillusioned when faced with unexpected problems, not only with regulatory bodies such as ethics committees and funding organisations, but also by the practicalities of recruiting patients and involving fellow clinicians in their studies.
The Developing Evidence-Based Implementation Trials (DEBIT) project (Reference Thompson, Barley and SullivanThompson et al, 2007) was run in four mental healthcare trusts in the south west of England, between February and July 2004. Its aim was to find a better way of changing clinical practice than the usual method of disseminating evidence-based guidelines, as there is wide appreciation that this is not successful (Reference Bero, Grilli and GrimshawBero et al, 1998). The research team developed a complex intervention consisting of a workbook, an educational visit to senior doctors and a reminder. The study targeted the prescription of antipsychotic polypharmacy on general adult psychiatric wards. A cluster randomised controlled trial design was employed and psychiatric units (clusters) were randomly allocated to the dissemination of guidelines alone, or the guideline plus the complex intervention. The primary outcome was a cross-sectional audit of antipsychotic prescribing taken from medication charts. The secondary outcome was a self-report questionnaire, designed to assess a change in beliefs about polypharmacy.
The objective of this paper is to highlight some of the problems that were experienced during the trial and suggested solutions. Although our study was a relatively large cluster randomised controlled trial and used a reasonably sophisticated methodology, many of the problems encountered were generic and are also common to smaller, simpler and more local research. Each stage of the study has unique problems and issues and we try to reflect this.
A trust ‘champion’
Most health services research is conducted, or at least coordinated, by university departments, whereas the research participants (staff and/or patients) are recruited by the health service. This can create immense operational difficulties for the coordinating research team, particularly when the trust is not a local one.
A trust liaison person is needed. This person facilitates the operation of the study in the trust and acts as a contact point. The person needs to be enthusiastic, proactive and well known within the trust, as well as efficient and persuasive. Such people are difficult to find. Senior staff commonly fulfil these criteria, but they are frequently too busy to help and are not readily available to the research team. On the other hand, junior staff might not have enough influence among their colleagues. In spite of these difficulties, getting the right person on board is key to the smooth running of a study.
Any study requires a certain amount of telephone research, particularly in the planning stage, to make useful contacts and to raise the profile of the study. Reaching key people by telephone can be extremely time-consuming. Wards are busy and occasionally chaotic places and for clinicians patients’ needs are priority, not research. Consequently it can take around four phone calls to achieve one successful conversation with a busy ward manager. It can also be surprisingly difficult to find out who works where. Staff lists are often essential for a study, but can be extremely fluid, both for doctors and nursing staff. Owing to prevailing conditions, trusts sometimes rely heavily on locum and agency staff, making running a study highly problematic, since these people are often not informed about the research and are unlikely to have much interest in facilitating it. Face-to-face visits were the most effective method of raising the profile of our project, but they can be time-consuming, especially when long distances are involved.
Sufficient time should be allocated to the preparation stage for key contacts to be made. Personal visits by the liaison person to key staff can shorten this period. They must have comprehensive knowledge of the study in order to be able to impart information confidently.
It is useful to make sure that up-to-date staff lists are available and that any temporary staff are aware of the study. Telephone enquiries, as well as brief informative leaflets and large notices for ward offices, are a good idea.
It is often easier to attend local meetings (such as acute care forums) to get the message across rather than have individual meetings with busy staff (although this contact, if feasible, is invaluable).
In order to persuade a National Health Service (NHS) organisation to take on a project the research team needs to effectively ‘sell’ the benefits of the study both for the service and the patients.
This does require a degree of zeal on the part of the research team and it helps to have researchers with some knowledge of the stresses and time pressures experienced by the NHS staff. Basic marketing strategies such as designing a project logo or a slogan that can be used, for example on stationery, are useful. This makes the study immediately identifiable among the many project requests, letters and other mailings that trust employees receive. Pharmaceutical companies are expert in these techniques and may be able to share some ‘tools of the trade’ (e.g. by running a short seminar on ideas and techniques needed to ‘sell’ concepts to mental health trusts). We feel that this can be done without compromising the independence of the study.
It is usually necessary to obtain honorary contracts for researchers who are not employed by the trust where the research is to be carried out. This can be very time-consuming, especially when Criminal Records Bureau checks need to be made, and can hold up the launch of the study.
This must be completed at the preparation stage, since the study cannot legally commence until all research staff have an honorary contract.
Chiefs or Indians?
In order for the study to run smoothly, as many people as possible in the trust must be aware of its aims and objectives. It is often difficult to determine at which level this profile-raising exercise should be aimed. Should the information be targeted at those actually taking part in the study, or at senior management, with the hope that there will be efficient dissemination of information down the chain? The bottom-up approach is labour-intensive since there are far more people to speak to, and lower-grade staff often do not have protected time for such matters. A top-down approach can also be problematic since enthusiasm and consent to proceed from senior members of staff do not automatically translate to junior clinicians.
A detailed strategy of the timing and manner of making contacts within the trust must be drawn up. This is important so as to involve, rather than antagonise people. The ideal solution is a top-down and bottom-up approach. Medical and nursing directors as well as clinical governance leads need to be contacted to obtain consent for the involvement of their trust. Contacts also need to be built with junior staff since they will be more directly involved and their enthusiasm is as much, if not more important. Operational staff often also believe that senior managers have unrealistic ideas about what life on the ground is like and what is actually possible. Middle management staff such as ward managers, senior nursing officers or professional development nurses are a good source of contact with junior nursing staff. Consultants can be a useful lead-in to junior doctors. It should be stressed that there is no substitute for personal contact with key collaborators and that relationships should be carefully nurtured.
Owing to pressures associated with timing of baseline and post-intervention data collection, respondents are often given limited time to complete and return their questionnaires. This can result in a low response rate and therefore results that cannot be related to a wider population.
It is vital that enough time is allowed for participants to respond and for reminders to be sent out. Two reminders sent 2 weeks apart is usually sufficient.
Variations in trust structure and clinical governance
Proper delivery of the study is impossible without an intimate knowledge of the structure of each trust, particularly with regard to clinical governance mechanisms. This knowledge is also important at the planning stage, but trusts can be reluctant to provide this information until extensive details about the research have been provided.
Ethical and research governance issues may also hinder the flow of information from the trust. Data protection is often a factor here. Trusts vary in their willingness to provide identifiable data about staff and patients.
The study should be run pragmatically and a certain degree of heterogeneity in clinical governance mechanisms between centres should be accepted as inevitable. These differences should be absorbed into study procedures as quickly as possible, without causing difficulties for the trust or the participants. Participation must be made as easy as possible. Finding out about the structure of each trust will also help to decide which members of staff to target.
Most health services research requires NHS staff to perform some study-related tasks. No study can run smoothly unless they are prepared to take a certain degree of responsibility for the project. As it is, there is often confusion over who should be implementing the trial within the clinical governance group.
Root the study within the clinical governance structure. This often requires much ‘oiling of the wheels’, but it gives the researcher some control over the process. It also means that, in theory at least, the clinical governance committee can strongly engage and motivate staff to contribute to the research. This will depend, however, on the degree of the committee's involvement in the project. To this end, the study must be relevant to the local agenda.
To achieve that, it is useful to find out whether any similar projects are running locally, and if so, how they relate to your project. The degree of support your project will get from the trust can depend both on the number of projects that the trust is already running and the attitude of the staff involved.
If the clinical governance committee is not directly engaged, it can be effective to intervene at middle management level. For instance, suggest that consultants mention the project during teaching sessions with their juniors. Alternatively, all junior staff could be sent a brief letter explaining the trial, signed by the principal investigator.
Our intervention resulted in a significant change in the prescription of antipsychotic polypharmacy. These findings were in line with Grimshaw's review of guideline dissemination and implementation strategies, which showed a moderate effect for most strategies (Reference Grimshaw, Thomas and MaclennanGrimshaw et al, 2005).
The only valid measurements of our success in overcoming the operational hurdles detailed above are the evaluations carried out at pre-defined points of the study.
Each part of the intervention was checked for adherence. The workbook contained a tear-out feedback form, which staff were encouraged to complete and return to their ward manager; 50% of these forms were returned. The educational visit to senior doctors by a clinical pharmacist achieved 93% coverage. Of those originally visited, 44% also received a follow-up phone call. The pharmacy reminder system, which consisted of stickers placed on the charts of patients whose prescriptions indicated antipsychotic polypharmacy, was carried out with 61% accuracy. The research team felt confident that the solutions we developed to overcome the operational problems enabled an impressive response to our intervention.
Since the DEBIT trial there have been two important developments that should make it easier to plan and conduct research. The Mental Health Research Network (MHRN) has been formed to assist researchers with many of the difficulties detailed here. Clinical studies officers based within the network can provide invaluable assistance with local issues for multicentre research. The Research and Development Division in the Department of Health has also devised the ‘research passport’ to avoid repeated applications for honorary contract to different trusts.
One of the most important qualities required of health service researchers is persistence. Unrelenting efforts must be made to continually raise the profile of the study within the trust, and constant attempts must be made to contact and engage key staff.
The whole research team must be ready for frequent negotiation with the health-providing organisation under study. Logistical problems may require creative problem-solving skills. Although this can seem a very steep mountain to climb, it is achievable, and the result can be meaningful research.
Declaration of interest
We thank Tim Peters for help in data analysis, Glyn Lewis for helpful comments on the study and all the participating clinicians and trust ‘champions’ in Avon and Wiltshire Partnership Trust, Gloucestershire Partnership Trust, Somerset Partnership Trust and Devon Partnership Trust. We acknowledge Lilly Pharmaceuticals for their help in providing marketing training for the intervention.
This study was supported by an NHS South West Project Development Grant.