- (2008) Volume 5, Issue 1
Research Fellow, Royal College of Nursing Research Institute, University of Warwick, UK
Previously, Research and Development Fellow, Royal College of Nursing Institute, Oxford, UK
Received: 11 May 2007
Accepted: 11 December 2007
This paper reports on the process of collecting information to develop recommendations for a national UK guideline for the short-term management of disturbed/violent behaviour in adult psychiatric inpatients and emergency departments. Part of this information was gathered using focus groups undertakenwith service users. The views of African- Caribbean individuals were a particularly important part of this information gathering, as numerous reports and inquiries have demonstrated that black people are more likely than others to have negative experiences of mental health services. Twenty-four mental health service users and nine staff, all, except one staff participant, of African- Caribbean origin, took part in focus groups. Data were transcribed and content analysis was conducted independently by three researchers. Findings revealed four overarching themes: voicelessness, powerlessness, inappropriate treatment and control. These themes informed the generation of recommendations. This paper shows that focus group data ensure that service users’ voices are heard and that those voices can contribute to the development of guidelines alongside other data. In doing so, service users’ voices can help to improve the sensitivity and quality of guideline recommendations.
African-Caribbean mental health, focus groups, national guideline development, violence and mental health
This paper reports on an approach to the development of sensitive and clinically appropriate recommendations for a mental health guideline. The UK’s National Institute for Health and Clinical Excellence (NICE) commissioned theNationalCollaborating Centre forNursing andSupportiveCare(NCC-NSC) toproduceaguideline on the short-term management of disturbed/violent behaviour in adult psychiatric inpatients and emergency departments. The description ‘short-term’ refers to any violent incidentwithin a 72-hour period. This is a sensitive topic requiring consideration of the rights of patients and staff involved in the violent incident.
The importance of this guideline’s topic to African- Caribbean service users in the UK is widely recognised by healthcare professionals (Fernando et al, 1998; Ndegwa, 2000; Sasidharan, 2003). The national census of inpatients in mental hospitals and facilities in England and Wales found black Caribbean, black African and other black groups were more likely (by 33–44%) to be detained under the Mental Health Act 1983 when compared with the average for all inpatients (Commission for Healthcare Audit and Inspection, 2005). In particular, African-Caribbean men are overrepresented in mental health services (Pereira et al, 2006), and black men are eight times more likely to be detained under the Mental Health Act (Audini and Lelliott, 2002). They are more likely to access care through a criminal justice agency, involving negative experiences, coercion and disengagement, and ethnic explanations for ‘negative pathways to care are not fully explained by differences in diagnosis, social circumstances and the involvement of others’ (Morgan et al, 2005). A number of high-profile inquiries have raised concerns about the nature and quality of the treatment and care that African-Caribbean people receive in mental health services. One example is the inquiry into the death of David ‘Rocky’ Bennett. Bennett was an African-Caribbean who died in a secure unit in 1998 following restraint after he hit and injured a female nurse (Independent Inquiry into the Death of David Bennett, 2003).
Guideline methodology continues to develop (National Institute of Health and Clinical Excellence, www. nice.org.uk/aboutnice/howwework/developingnice clinicalguidelines/clinicalguidelinedevelopment methods/theguidelinesmanual2006/; The Guideline International Network, www.g-i-n.net/index.cfm?fuse action=membersarea; The AGREE Collaboration (www.agreecollaboration.org/pdf/aitraining.pdf))within the context of translating evidence into clinical practice. While research evidence is crucial there is also a need for clinical expertise, local data and patient experiences and preferences (Rycroft-Malone, 2004; Jarret, 2004). In the work reported here, the guideline development group took into consideration multiple forms of evidence. First systematic reviews collated all relevant research with no restrictions on study design. Second, expert reports were considered, and third, the findings fromthe focus groups were examined.
Focus groups are used extensively in research, but less so in guideline development, to collect information or investigate individual responses to different situations or policy initiatives, particularly those of a sensitive nature (Grbich, 1999). They are useful for encouraging participation from people who are reluctant to be interviewed individually or who feel they have nothing to contribute, or to tap into a specific cultural context (Kitzinger, 1994). Studies show that focus groups are useful and appropriate when working with ethnic minority groups (Hennings et al, 1996; Winslow et al, 2002). This paper presents an account of the conduct of the focus groups convened to inform the development of the guideline and the data obtained. It then moves on to discuss how the guideline was developed.
The rationale for conducting the focus groups was to consult with African-Caribbean service users giving or receiving care in psychiatric inpatient settings. The key objectives were to:
• gather data to inform the NICE guideline development of recommendation formulation
• provide an opportunity for the ‘voice’ of this ethnic group to be heard in the guideline process.
We conducted the three groups over a period of two months in 2003. We were involved as observers in the groups and took supplementary notes to clarify and supplement the taped discussion. There was an additional independent observer from the Patient Involvement Unit (PIU), which provides support to the guideline development process on behalf of NICE for the involvement of service users in the guideline process. There were two observers with the facilitator in each focus group.
Neither of the authors was African-Caribbean and so they identified two black and minority ethnic (BME) advocacy organisations: one had a diverse and transient population, the other a long-established and marginalised population. Two nominated facilitators from these organisations identified African-Caribbean participants with inpatient experience of mental health services.We consulted with these facilitators to develop participant information and a topic guide for the focus groups based on guideline-specific topics. Written and oral information was given to the people selected by the facilitators; informed written consent was obtained via the nominated facilitators. Participation in the focus groups was voluntary. The facilitators organised venues familiar to the participants, and lunch appropriate to their needs. The BME facilitators conducted one service user focus group each, with nine and 15 participants, respectively (Table 1). One of the authors and the PIU member of staff facilitated the staff focus group of nine participants (Table 1). Participants were informed that all information received would remain confidential and anonymous. They were also assured that care and future treatment would not be affected as a result of their participation. Participants were also asked to respect the confidentiality of other participants.
The BME facilitators in both organisations offered one-on-one discussionswith any participantwho wanted to further explore issues raised before or after the focus groups. This was part of the support service provided by the BME advocacy organisation. The BME facilitators were experienced advocates within their organisation. Specific focus group training was not provided as this was not a research study but a consultation with African-Caribbean mental health service users. Facilitators were selected from within their organisations and the groups conducted on their premises.
The NCC-NSC decided formal ethical approval was not required because the focus groups were not considered to be a research study. However, ethical issues such as confidentiality were taken into account. Information leaflets on topics for discussion, and consent forms were given to the organisations several months in advance and the facilitators purposively sampled participants fitting the inclusion criteria. Participants and the groups were allocated a personal identifier (number and a letter) to maintain confidentiality. Topic areas covered in the discussions included the ward environment, predictions of violence, interventions, de-escalation and observation techniques, restraint and seclusion including training, rapid tranquillisation, staff and service user relationship, and use of interventions in emergency departments. Three or four questions were asked under each topic area focusing on service users’ experience, observation of other service users, and their views on staff use of practices and techniques.
In order to present views from both sides of the therapeutic relationship we sought the perspectives of African-Caribbean staff working in the mental health sector. We had already established a network of contacts within the sector via our stakeholder list, and snowballed for appropriate members for a third focus group. We sought members of staff who were African-Caribbean. One participant was white, and was included because he had extensive experience with African-Caribbean service users. Nine participants agreed to take part (Table 1). All participants had some expertise with black service users in psychiatric inpatient settings or emergency departments. Aims and objectives of the focus groups with the topics to be covered were provided to participants before the group. Their views were requested on the use ofmanagement of disturbed behaviour interventions, relationships between staff and service users, the inpatient environment and its impact on disturbed behaviour. Written consent was obtained from all participants for the focus groups to be taped and transcribed and for anonymous data to be used in a publicly available report.
Table 1 :Details of focus group participants
Thematic content analysis was used to analyse focus group transcripts. Thematic content analysis is described as linking together patterns of experience and behaviour, creating categories of main and subthemes (high or low order; Aronson, 1994). We used a stageby- stage method of analysing transcripts from semistructured open-ended interviews proposed by Burnard (1991). Participants were given individual codes, and material was not personally identifiable.
Table 2 summarises the main themes from the three groups. The service user themes combine the data from both groups while the staff data were analysed separately. Themes were generated from verbatim transcripts of the focus group data. These themes emerge and are derived from exhaustive categories taken from the focus group data. A systematic approach was used (Burnard, 1991) that iteratively clustered the themes and built up to the main (high-order) themes. The low-order themes are subthemes to these main themes. Powerlessness/voicelessness was identified as a main theme because service user participants expressed an inability to influence their care or to have their needs met.
‘When you got to staff members and say ‘‘I’m feeling a bit worse’’, it’s ‘‘Go and sit down’’, ‘‘Go away’’, they don’t want to hear when you’re actually acknowledging that you are actually feeling worse, they’re not prepared to hear.’ (participant 2, group A)
‘The system is failing us.’ (participant 5, group A)
Service user participants also expressed experiencing racial stereotyping and the inappropriate use of disturbed behaviour management techniques – hence the main theme of inappropriate treatment/control.
‘On the ward, African-Caribbean people, we’re seen as a lot more aggressive rather than this is how we culturally are, so we’re treated this way.’ (participant 1, group A)
‘I was restrained by a particular nurse who had a technique where he put my arm behind my back and crunched my fingers together until the knuckles swapped over and wrench it in a different direction so it would swell up like when you do boxing and I couldn’t usemy fingers at all ... I plait my little girl’s hair, so I wasn’t able to plait her hair for her and it was making me more upset.’ (participant 7, group A)
The healthcare professional main themes of black experience also concurred with the service users’ experiences, particularly fear.
‘One interesting dynamic is that one of staff fear the patients more than the patients fear each other.’ (participant 26, group C)
They particularly expressed frustration with the systems:
‘We know from whatever little data is available that black patients are more likely to be restrained, secluded, medicated and so on, so that the whole process of heavy-handed management is an issue for black patients.’ (participant 31, group C) These findings are consistent with the report of the Sainsbury Centre for Mental Health (2002), which included 31 focus groups with African-Caribbean mental health service users. Key themes from this report include circles of fear, inhumane and unhelpful care, problematic care pathways, and conflict between professionals and service users not being addressed.
Table 2 :High- and low-order themes derived from the transcripts of the focus groups.
The guideline development process was constrained by the guideline timelines. Time for preparation and data analysis of the focus groups was limited, and the emphasis was on using our time effectively to achieve our aim. Much of the material gathered in the focus groups did relate specifically to the experience of BME service users but, as there was no comparison group of non-black service users, it could be suggested that some of the views expressed could relate to all mental health service users irrespective of their ethnicity.
It is acknowledged that the focus group data represent a small group of individual views that are not directly verifiable. However, they are consistent with other focus group data of African-Caribbean mental health service user experiences undertaken by the Sainsbury Centre for Mental Health (2002). If the focus group material had been inconsistent with other findings, further data collection would have been required.
The recommendations were developed to improve the overall experience of care by shifting the focus away from custodial and control techniques towards an approach that responded to the needs of the individual service user. The development process endeavoured to be as transparent as possible.
The guideline development group was composed of 12 expert clinicians representing medicine (4), nursing (2), pharmacy (1), allied healthcare professionals (1), emergency medicine (1), physical management training (1), economics (1), legal (1), and two service users. These service users were not from a BME group. Service user and professional BME representatives were specifically co-opted to increase representation; they were not involved in the focus groups.
The guideline development group drew on all three sources of information: systematic reviews, expert reviews and the focus group data. A modified nominal consensus technique was used for recommendation development, amendment and ratification (Trickey, 1998; Rycroft-Malone and Duff, 2000). This technique has been used in the development of similar guidelines with a variable research evidence base that included other forms of evidence, for example patients’ experiences (Trickey, 1998; Rycroft-Malone, 2004). The guideline was peer reviewed and went through two stakeholder consultation processes.
An independent external facilitator chaired the consensus meetings. This person was familiar with guideline development, but not experienced with the particular guideline topic under development. The consensus process was facilitated by computerised voting consoles, which assured anonymity and allowed percentages to be quickly calculated. Consensus was set at 80% unless a significant number within the group all voted against a recommendation. For example, if all the psychiatrists had voted against a recommendation, even though 80% agreement was achieved, consensus would not have been reached. No abstentions were allowed. The computerised voting system permitted an inclusive, constructive discussion before the vote. It included microphones that required switching on and off, and members wishing to be heard were allocated in turn to speak on the screen. The wording and amendment of the recommendations were facilitated by this process. Changes were viewed as they were made, allowing an iterative refinement. This helped to ensure that all members of the guideline development group, including two service users, were able to contribute equally.
Table 3 shows examples of the links established between the systematic review evidence, extracts from the focus groups, and recommendations made. Themes from the focus groups echo the literature which suggests an imbalance of African-Caribbean representation (Commission for Healthcare Audit and Inspection, 2005), in mental health service environments, and that African-Caribbean individuals experience conflict and fear (Sainsbury Centre for Mental Health, 2002).
All the evidence obtained during the guideline development process was taken collectively into account to generate recommendations. The different information sources gathered for this guideline were not weighted. The guideline development group gave due consideration to all the information presented to them. The information was presented in a written review or orally via expert representation. The purpose of Table 3 is to illustrate and provide a structure to show that concerns raised in the focus groups fed into guideline recommendations.
The guideline development deliberations on recommendations involved much debate on the management of the physical intervention (restraint) and the use of medication (rapid tranquillisation). Focus group themes about inappropriate treatment/control, uncaring and powerlessness referred to participants’ experiences of receiving medication and physical interventions during disturbed or violent episodes. In particular they expressed concern about the overuse of the techniques, a lack of alternatives, and a readiness by staff to use medications, physical intervention (restraint) and seclusion (isolation).
Table 3 :Focus group: participant quotes with links to the evidence and recommendations.
‘It’s about controlling people ... it’s keeping them in a zombiefied state so that their job is made easier and then they feel powerful.’ (participant 1, group A)
‘I just think that there should be something where they constantly train people in restraint ... people tend to get hurt, the patient tends to get hurt.’ (participant 6, group A)
It was also expressed by participants that management of violence techniques were particularly overused on black people.
‘There was a women that was constantly being picked on with regards to restraint and she was a Jamaican.’ (participant 7, group A)
‘Most of the time, if you’re black, you’re going to get bent and twisted up and carried into seclusion.’ (participant 23, group B)
As a result, the guideline development group used this opportunity to address racial and other issues of equality. Furthermore, an emphasis on staff and service users’ safety was also incorporated into guideline recommendations. The recommendations emphasise continuous risk assessment and the ongoing use of deescalation techniques in the care pathway for managing short-term disturbed/violent behaviour. Table 3 identifies specific recommendations in relation to focus group concerns. There is also an overarching recommendation to ensure that responsibility for racial and other equality issues is addressed at the organisational level. Additionally consultation with local BME groups is promoted (Royal College of Nursing, 2006).
Focus groups as a qualitative method to gain patient opinion in health research have become increasingly popular (Webb and Kevern, 2001; Barbour, 2005). However, focus groups often do not report participant interaction and the group dynamic is an important part of the method (Webb and Kevern, 2001). The limitations of the guideline timelines did not allow a fuller exploration. Consensus of opinion in the group is an emergent property, not an individual participant opinion, and although inference can be drawn this does not indicate the strength of an opinion (Barbour, 2005). Caution must also be used against attempting to generalise focus group findings (Sim, 1998). We used a multi-method approach within the guideline methodology, and sought the perceptions of care of a specific population. The overall negative opinion of care expressed by the service users in the focus groups could indicate the group selecting negative rather than positive experiences to fit in with an emergent group trend. Data analysis needs to be more thorough and effective, and inclusive of group interaction and dynamics to determine whether this might actually be the case (Barbour, 2005). An example is shown in Table 3 under inappropriate treatment and control. Members of the group responded to one individual by murmuring agreement. This happened on a number of occasions, and a better description of the group processes could have enabled this information to be used more effectively in illustrating one aspect of focus group experience, which is the recognition of experience between group members.
Focus group studies need to be congruent with the qualitative tradition and context appropriate (Hughes and Dumont, 1993). The guideline consultation focus groups were conducted as part of the guideline development methodology and had specific aims and objectives. The task of a guideline is to ensure quality clinical practice based on evidence at best, or at the least use of formal consensus techniques. The literature did not provide information or evidence on the more subtle and sensitive issues of racial stereotyping, but the focus group findings channelled information into the guideline development process to fill the gap. The importance of this work illustrates the requirement for guideline development to be flexible and responsive, to meet the clinical need for best practice in the absence of good-quality and relevant evidence. It also needs to avoid tokenism in gaining the service user perspective.
In addition, this experience of conducting focus groups as part of guidelinemethodology has highlighted the importance of establishing transparent methods to report sources of evidence, their context and their synthesis.Wehave clearly stated that there was no formal weighting of evidence and that consensus methods were used to develop guideline recommendations. A refined conceptual framework for consideration of evidence and context to be applied to the development of health policy recommendations acknowledges the role of agreement-based consensus methods alongside rule-based grading schemes (Dobrow et al, 2006). The sensitivity given to the context, as in this instance, of health policy will promote the improvement ofmethods for utilising evidence in development of health policy (Dobrow et al, 2006).
Respondent validation was conducted via the black advocacy organisations; no significant changes were required and participants seemed satisfied with the draft report. The analysis was not conducted by an independent assessor. However, as the focus group observers did not attend all the groups they did analyse the taped data of all three groups individually. The two guideline developers and the PIU representative discussed the key findings and agreed themes together. This was done through a process of presenting individual themes and collectively agreeing common themes. The findings were consistent with other reports such as Breaking Circles of Fear, which conducted 31 focus groups with African-Caribbean service users (Sainsbury Centre for Mental Health, 2002). This report, the Independent Inquiry into the Death of David Bennett (2003), and the focus group material provided an overall picture of an untherapeutic environment that is threatening and fearful for both staff and service users. It also suggests that there is a perpetuation of cycles of violence leading to a culture of violence on some wards.
This paper’s key messages and the implications for guideline development (Box 1) are the value of focus groups as part of a strategy in guideline development to engage the voice of African-Caribbean mental health service users to inform the generation of clinically sensitive recommendations. The primary focus throughout the development process was on the staff and patient relationship as crucial to the short-term management of disturbed/violent behaviour. Erosion of the therapeutic relationship was a principal concern. Focus groups informed the development of recommendations to address staff behaviour and attitudes by ensuring that the African-Caribbean service users’ voices made a direct contribution to the process. The purpose of this paper has been to show the rationale for using focus groups in guideline development and the impact on aiding recommendation generation. We suggest an evaluation of focus group work across guidelines to ascertain the advantage of using focus groups, as an additional form of evidence to aid guideline development.
We would like to thank Joan Field-Thorn, BlackOrchid, Bristol and Pauline Abbott-Butler, Footprints UK, London for facilitating the focus groups on our behalf of the NCC-NSC. We would also to acknowledge and thank Jane Cowl, Patient Involvement Unit, NICE for her collaboration with the authors on this project. The authors would also like to thank Dr Jo Rycroft-Malone for comments on drafts of the paper.
The recommendations were generated by consensus techniques with the NICE Guideline Development Group listed below:
William Bingley, University of Central Lancashire Tony Bleetman, British Association for Accident & Emergency Medicine
Frank Corr, Royal College of Nursing
Jane Cronin-Davis, Clinical Effectiveness Forum for
Allied Health Professionals
Donna-Maria Fraher, Independent Consultant
Phil Hardy, Institute of Conflict Management
Sophie Jones, National Voices Forum
Kevin Gournay, Institute of Psychiatry (Chair)
Edwin Gwenzi, Institute of Psychiatry
Sue Johnston, Royal College of Psychiatrists
Stephen Pereira, National Association of Psychiatric
Intensive Care Units
Peter Pratt, Royal Pharmaceutical Society of Great
Britain
Rick Tucker, Nursing & Midwifery Council
Aki Tsuchiya, SchAAR, Health Economist
