Improving decision-making in caring for people with epilepsy and intellectual disability: an action research project


D. Vallenga:


Title. Improving decision-making in caring for people with epilepsy and intellectual disability: an action research project.

Aim.  This paper is a report of an action research study aimed at improving decision-making concerning risks in people with epilepsy and intellectual disability and the empowerment of participants to sustain this improvement.

Background.  Residents of long-stay units in specialized epilepsy centres suffer severe epilepsy and are often intellectually disabled. Professional caregivers have to find a balance between risk-taking and protection, as both can have a negative effect on quality of life. Clients, their representatives and caregivers are involved in this decision-making process.

Method.  A participative action research project was conducted. Data were gathered continuously over a 22-month period in 2004–2006 by interviews, observation, written reports of meetings and personal stories.

Findings.  The action research resulted in an ongoing process of improvement. Problems in decision-making about risk were clarified. The importance of consensus about risks and commitment to risk management between all involved became apparent. Phases in risk management were discerned and used to adjust caregiving to clients’ needs and capacities Assessment of client risk became more systematic and ‘steps to an individual framework for decision-making’ were developed as a tool for a systematic approach.

Conclusion.  Using an action research model to improve decision-making provided the preconditions for improving decision-making and risk management and suggesting a way to improve this process in this and other institutions.

What is already known about this topic

  • • Approximately 30% of people with intellectual disability also suffer from epilepsy.
  • • People with severe refractory epilepsy are at greater risk than others of suffering accidents as a consequence of seizures.
  • • Both risk-taking and protective measures can have a negative effect on quality of life.

What this paper adds

  • • A systematic approach oriented to individual clients enables decisions about risks to be tuned to the severity of their seizures and to contextual and personal influences.
  • • A dialogue with the client is a precondition for including their experiences, preferences and capacities in decision-making.
  • • Commitment to risk management, reflection on decisions made and open communication are preconditions for continuity of care.


Residents of the long-stay units in specialized epilepsy centres often have multiple handicaps besides severe epilepsy. In many cases, intellectual disability is involved, often combined with physical and sensory handicaps. These factors increase the risk of injury (Veendrick-Meekes 1997, Vallenga et al. 2004). Decision-making about risks is a complex process burdened with uncertainty. Taking risks can lead to dangerous, sometimes life-threatening situations, while protective measures, varying on the individual level from medication and supervision to a total restriction of the client’s mobility, can be restrictive, intrusive and detrimental to quality of life.


Risk of accidents in epilepsy and intellectual disability

Approximately 14–44% of all people with intellectual disability also suffer from epilepsy (Bowley & Kerr 2000). People with epilepsy are at greater risk of accident than people without epilepsy. This is mainly because of the relatively small group with refractory epilepsy and frequent tonic, tonic-clonic or atonic seizures (Buck et al. 1997, Neufeld et al. 1999, Commission on Epilepsy, Risk and Insurance 2000). People with both epilepsy and intellectual disability often belong to this group. The underlying condition provoking the intellectual disability may also lead to severe, refractory epilepsy (Vallenga et al. 2004). The frequency and severity of seizures makes management of the epilepsy difficult (Tobias et al. 1994).

Protective measures

The risks to which people with epilepsy are exposed are not only dependent on the aetiology of the epilepsy and seizure type but also influenced by the circumstances in which they live and the measures applied to protect them. In the epilepsy centre in which the study took place, protective measures on an individual level included (poly)pharmacy, supervision, avoidance of risk-associated activities, protective headgear and mobility restrictions. In some cases measures accumulated (Vallenga et al. 2006), possibly because single protective measures cannot guarantee safety (Deekollu et al. 2005).

Decision-making and intellectual disability

Decision-making is described by Gumpel (1994, p. 195) as encompassing the interpretation of a stimulus, generation of response options, analysis of the feasibility of those options and execution of an ‘appropriate and executable’ response choice. People with intellectual disabilities are vulnerable in making decisions because of generally impaired cognitive functioning and their reduced capacity for step-wise decision-making (Hickson & Khemka 1999). They are dependent on nurses and their personal representatives, who decide whether or not to involve them in decision-making (Biesaart & Hubben 1999, 2000).

Decision-making about protective measures

In 2003, we studied the process of risk-evaluation and subsequent decision-making concerning specific risks and the application of protective measures in the care for people with epilepsy and intellectual disability (Vallenga et al. 2006). We found that decisions to apply protective measures or to adjust them to the current situation were often taken after accidents so severe that interventions were necessary to prevent recurrence and enable healing, or on the basis of the cumulative effect of repeated accidents. Circumstantial factors could also trigger the implementation of protective measures. In these situations, it was almost inevitable that considerations of protection outweighed quality of life. It became clear that a systematic approach to risk-management and regular evaluation tuned to fluctuations in seizure severity was lacking, although observation and evaluation are seen as essential for successful injury prevention (Forjuoh et al. 2001).

The study

The study presented here was the third step in a project to investigate and improve decision-making about risks in the care for people with epilepsy and intellectual disability (hereafter ‘decision-making’). A literature study and a case study describing this decision-making process and the factors influencing it have been published previously (Vallenga et al. 2004, 2006). For this stage of the study, the research questions were: ‘Is it possible to improve decision-making using action research?’ and ‘To which improvements can this lead?’


The aim of this part of the study was to improve decision-making concerning risks in people with epilepsy and intellectual disability and to empower participants to sustain this improvement.


Improving decision-making is not a technical issue to be resolved using a technical approach (McNiff & Whitehead 2006). As our case study showed, technical knowledge is a factor but not the primary factor involved (Vallenga et al. 2006). In decision-making personal factors of clients, their representatives and caregivers are also important, including anxiety about possible seizures and injury, attitudes to care-giving, responsibility and personal competencies (Vallenga et al. 2006). For this reason, the participative research method known as action research was chosen to study and improve decision-making.

Action research was developed from Habermas’ critical theory, which is based on the assumptions that:

  • • It is important to understand a situation in order to change it.
  • • Social situations are created by people, and can be deconstructed and reconstructed by people.
  • • Taken-for-granted situations need to be seen in terms of what has brought them into existence, especially in terms of relationships of power (McNiff & Whitehead 2006, p. 41).

Understanding a situation in order to change it is the primary aim in critical theory. ‘How can the situation be changed?’ is the central issue in action research (McNiff & Whitehead 2006).

Action research involves a cyclical process first described by Lewin in 1946. In action cycles (Figure 1), phases of planning, action, observation and reflection are looped in an ongoing process involving collaboration and interaction between practitioners and the researcher (Holter & Schwartz-Barcott 1993, Hoogwerf 2002). Interview techniques were used for preparing and involving the nurses, the team management, the clients and their representatives, forming the action research groups, and investigating the points of view of those who were only partly involved (Table 1).

Figure 1.

 Action cycles.

Table 1.   Chronological overview of activities prior to starting and during the action cycles
ActivityPersons (n)ThemesGoals
  1. AL, Agate Lane; RR, Ruby Road.

Activities preceding the action cycles
Group meeting nursing staff and team managementAL: 14
RR: 8
Results of the previous study (decision-making about risk)
Need for improvement of decision-making
Informing, preparing and involving nursing staff and management
Letter to all representatives23Results of the previous study (decision-making about risk)
Need for improvement of decision-making
Invitation group meeting
Informing, preparing and involving representatives
Group meeting nursing staff, team management and representativesAL: 17
RR: 14
Results of the previous study (decision-making about risk)
Improvement of decision-making
Introducing the method of action research
Informing, preparing and involving nursing staff, management and representatives
Several interviews with/or observation from small groups of clients (two to four in a timeframe of 4 weeks).AL: 12
RR: 11
The clients experience of risk
Protective measures
To assess clients’ possibilities for participating
Focus group
 nursing staff
 team management
 client representatives
AL 14
RR: 13
The balance between taking risk, using protective measures and quality of life. A fictional case was used to discuss:
 The need of analysis of accidents
 Transparency in decision/making
 Capabilities of the client to  participate in decision-making
 Communication between nursing  staff and representatives
To create support and involvement
Goal setting
Forming the action research groups
Activities during the action cycles
Focused interviews with nursing staff AL12The method of action research
Nursing staff experiences with risk management
Testing the nursing staff’s knowledge of the method of action research
Seeking nursing staff’s personal experience with risk management for more precise goal setting
Focused interviews with nursing staff RR7The method of action research
Experiences during the start of the project
Testing the nursing staff’ knowledge of the method of action research
Group meeting AL nursing staff, team management, representatives (action cycle 3)6The progress of action research projectInforming and involving representatives
Group meeting RR nursing staff, team management, representatives (action cycle 5)9The progress of action research projectInforming and involving representatives,
Discussion about progress
Adjusting goals
Questionnaire representatives AL12
Risk management and care plan
When is a decision a good decision?
Investigating the points of view of representatives
Focus group AL nursing staff, team-management, representatives (action cycle 6)28The results of the projectDiscussing the results of the action research project
Focus group RR nursing staff, team management, representatives (action cycle 7) 17The results of the projectDiscussing the results of the action research project


Two teams, fictionally named Agate Lane (AL) and Ruby Road (RR) participated in the project as action research groups. It should be noted that only nursing team members, (registered nurses and nursing assistant staff) fully participated in the action research groups, whereas in decision-making as it occurred in practice clients and their representatives were also involved (Vallenga et al. 2006). Because of their intellectual disabilities, most clients were not capable of discussing risks and associated choices and therefore could not participate in the action research groups. In RR, five clients were capable of discussing the risks to which they were exposed and the decisions made to cope with these risks but they were not fully capable of understanding the process as a whole. Overseeing a stepwise process over time exceeded their capacities. Matching the complex process in which nurses and client representatives are involved to their capacities was not feasible. The decision to not include these clients was made after discussing this problem with a psychologist.

The client representatives were parents or siblings, assigned officially in accordance with Dutch legislation. They participated together with nursing staff in focus groups before the action cycles started and later at preset times. While the RR representatives considered it sufficient to be informed about the project, the AL representatives expressed their desire to participate in the project. However, agenda difficulties made it impossible for them to participate to the same degree as the nursing staff. Instead, at preset times, they were informed about advances made and asked to formulate their viewpoints in a questionnaire.

Figure 2 shows characteristics of the teams, clients and the situation at the beginning of the study.

Figure 2.

 Characteristics of teams, clients and the initial situation.

Action research process

The overall goal of the study was determined by the management of the long-stay units of the epilepsy centre. Attainable goals tuned to the specific situation in the units were set on the basis of focus group interviews. Representatives and teams participated in these focus groups.

During a 22-month period in 2004–2006, data were gathered continuously, using a range of research strategies. The primary researcher (D.V.) facilitated the change process. Focus group interviews were used to determine the goals for the project within the context of improving decision-making and later to discuss the findings of the action cycles. Small groups of clients were interviewed to assess their possibilities for participating in the project. Individual focussed interviews were carried out with all nursing staff to assess the impact of information given to them and their experience of decision-making. A questionnaire was used to assess the viewpoints of representatives about developments in the AL team. The researcher wrote a personal diary, including observations and reflections on discussions and meetings. Written reports were made of all interviews and personal stories told by participants. Reflection and planning was carried out in action research meetings which were part of monthly team meetings. The results of each action cycle were used to plan the next cycle.

Data to evaluate the project and results were gathered by focussed interviews in RR. In AL, an external researcher (De Munck 2006) held focussed interviews, participated in conversations and discussions, made observations and wrote field notes and a report.

Ethical considerations

The Medical Ethics Committee of the epilepsy centre approved the study.

In action research, collaboration between researcher and participants gives rise to ethical dilemmas, including how to ensure voluntary participation, informed consent, shared decision-making, anonymity, confidentiality and how to resolve conflicting needs (Walker & Haslett 2002). Our improvement project was part of a study of risk-management in people with epilepsy and intellectual disability. It was initiated by the management, which formulated the overall goals on the basis of the previous study. This provided the context for the action research project. However, when using action research, goals and actions are determined by the members of the action research group and conflicts may arise if goals not shared by the management are suggested.

The primary researcher therefore first informed the management about the methodology of action research. They proposed units for participation, implicitly agreeing with the goals and chosen method. Subsequently, informing and generating commitment from nursing staff, the clients and their representatives was important. Nursing staff as a group were explicitly asked during the first information meeting whether they wanted to participate. In later interviews, participants were asked individually about their views on the project. Representatives received written information and were informed orally about the project during the first group meeting. All agreed to participate.

The primary researcher and the teams discussed the dilemmas which could arise such as conflicting goals or fading commitment. As also noted by Walker and Haslett (2002), continuous reflection and open communication created a context in which participation was always open to discussion. The action research group decided that there should be dissatisfaction about cooperation, actions or goals, they should immediately inform each other and look for a solution; however, this situation did not occur.

Twenty individual interviews were conducted with nursing staff; in 17 cases permission was given to record the interviews. The focus group participants also agreed to make audio-recordings. When participants appeared hindered by this, notes were taken by one of the participants and reports were written immediately after the event.

Data analysis

Data were analysed continuously because the results of each action cycle were used to plan goals and actions for the next cycle. Recordings made of individual interviews were analysed by repeated listening and later an external researcher (J.D.M.) made verbatim transcriptions and thematically coded and analysed them (De Munck 2006). Data from each action cycle were structured thematically and analysed using the constant comparative method (Boeije 2002). Subsequently, data and analyses of the various action cycles were compared to each other and finally the complete data sets from RR and AL were compared with each other. Within the action cycles, this analysis process was carried out together with the nursing staff by using the computer program MAXqda (Kuckartz 2001).


The rigour of the study was enhanced by using multi-method triangulation (Polit & Hungler 1991, Meijer et al. 2002). This included methodological triangulation using interviews, observation and written documents and researcher triangulation. An external researcher analysed the transcripts to validate the primary authors’ interpretations, and these were found to be coherent. Reports, analyses and findings of each action cycle were submitted to the participants for confirmation. After the last action cycle, the external researcher observed decision-making in AL in its improved form. In RR, final interviews were undertaken 6 months after the last action cycle.


The findings of each part of the study were discussed in the action research group meetings and, on the basis of this reflection, new goals were set. Actions and goals were closely related to problems encountered in daily practice and to different aspects of decision-making. This sometimes led to goals which were broader in focus than only decision-making. In what follows, the findings will be discussed for each residential unit, as each unit developed a separate path.

The action research project in Ruby Road

The timeline for RR is shown in Figure 3. The results of the focus group before starting the action cycles showed that a methodical approach was lacking in all aspects of care delivery. The desire for transparency in decision-making and communication became apparent. A representative noted that it would be impossible to change decision-making without applying a methodical approach. This viewpoint was shared by the other focus group members. They decided to make improvements in work methods related to risk-management the goal of the whole project.

Figure 3.

 Timeline for Ruby Road.

Improving assessments

In the first three action cycles, the action research group used one case to formulate questions and develop instruments for systematic observation. In the second cycle, they uncovered the risks because of epilepsy, and in the third, the risks because of behavioural problems and the consequences of the underlying medical condition for one specific client. At that point, the action research group realized the necessity of including circumstances and clients’ personal preferences in the assessment. The insights gained in this case were applied in the third action cycle to assess risks to which other clients were exposed. The results of these first three action cycles showed that a systematic approach to risk analysis was necessary, and that risks associated with epilepsy needed to be seen in conjunction with risks associated with other disabilities including intellectual disability and in the context of clients’ preferences, possibilities and personal circumstances (Figure 4). This should be an integral part of professional nursing care, but was often neglected. There was an individual care plan for each client which was reviewed every 2 years. Nevertheless, the findings showed that care plans were not always followed.

Figure 4.

 Risks, possibilities and clients’ wishes and circumstances.

Consensus and commitment

In the fourth and fifth action cycles, the action research group realized that the changing perception of the risks to which the client was exposed should lead to changes in care-planning and delivery. That consensus was important to achieve changes in care delivery became apparent when the goals of these action cycles were not achieved. Nursing staff did not oversee the consequences of their decisions about caregiving. Consensus about risk and risk-management alone was not enough; nursing staff needed to accept the consequences of their decisions for themselves and to adjust not only their view of the client but also their work attitudes and routines. One nurse noted: ‘First I had to accept the client’s limitations and now I have to change my behaviour. I will have to let old habits go and accept new ways. I’m not sure I can do that.’

It became clear that besides consensus about risks there must also be consensus about management of these risks and the consequences for nursing staff. To realize effective changes, commitment to risk-management and addition of new agreements to care plans was essential. As nursing staff became fully aware of these phases in risk management they were more able to adjust their behaviour to the risks and also to the clients’ capacities, and this in turn led to more client-oriented care (see Figure 5).

Figure 5.

 Phases in risk-management.

Improving work processes

The wish for consensus necessitated improvement of work processes. In the last three action cycles, all nursing staff worked on this supported by the team management. The researcher’s support became less essential. Nursing staff introduced changes in the report system, applied a methodical cycle to each meeting and practised giving feedback. The action research group discussed the insights gained in action cycles with representatives and included them in exercises so that they could experience the benefits of systematic risk analysis. For client representatives and nursing staff, this was a positive and emotional experience.

The observation and reflection on work processes showed that increasingly decisions about risk were made not only in reaction to incidents but that the clients’ broader situations were included in the analysis. Nursing staff increasingly sought measures suited to a client’s wishes in the situation at hand and ways of reducing the restrictive aspects of protection. For example, after a client was injured during a seizure in the bathroom, nursing staff did not immediately react by limiting clients’ mobility and increased personal assistance in the bathroom as would have happened previously. Instead they looked for ways to maintain mobility and respect privacy and decided to place safety frames around the toilet.

At the end of the last action cycle the action research group evaluated the project and concluded that the major goals had been achieved. Caregiving had become more client-centred, and decisions were now made after a more systematic analysis of the clients’ situation. They reviewed not only occurrence of seizures and accidents, but also the circumstances in which they occurred, including the behaviour, wishes and capacities of clients and their representatives. The drive for consensus became an integral part of caregiving as nursing staff committed themselves to decisions and worked according to the plans accepted by all.

A focussed interview with two RNs representing the team 6 months after the last action research meeting showed that decisions were no longer made by staff acting alone, but care was taken to involve other staff members. The limited possibilities for sharing and discussing viewpoints while at work led to the development of a written system to discuss choices and proposals. The team learned to apply a methodical cycle to team meetings, which led to more effective and goal-centred meetings. Decisions made were evaluated at subsequent meetings.

In this interview, the nursing staff explained what effects the project had had on their attitude to working as a team. Analysing client problems as a team, seeking improvements and sharing decision-making had resulted in a team which was less task-oriented and more coherent in observing and evaluating client problems. Nursing assistants felt that their viewpoints were now heard on a structural basis and felt an increased autonomy. Team reflection on decisions made also stimulated their personal reflection and increased objectivity. Reflection had become a more integral part of the job.

The action research process in Agate Lane

The timeline for the action research process of AL is shown in Figure 6. The project in AL started approximately 8 months after the clients had moved from the institution to housing in a village. It became apparent that team members differed in their readiness to permit activities associated with risk, causing hesitancy to act in the caregivers. Participants in the focus group decided to make the improvements in decision-making in daily practice the goal of the project.

Figure 6.

 Timeline Agate Lane.

Problem clarification

Problem clarification was the goal of the first three action cycles. Analysis of interviews held in the first action cycle and personal stories written in the second cycle showed that nursing staff from AL had to decide daily about client participation in common but risk-bearing activities such as cooking and navigating village traffic. Care plans predated the relocation and had not been re-evaluated to tune risk-management to the new situation. There were considerable differences in the choices made by nursing staff. Decisions were complicated by various factors: the unpredictability of seizures; impulsive and unpredictable behaviour of clients; and the influence and authority of representatives, which led to nursing staff accepting more risk or protective measures than they personally wanted to take responsibility for. Nursing staff were uncertain about who would be accountable in case of serious accidents or injury. The conclusion was that making decisions about common but risk-bearing activities was hampered by a lack of consensus about the extent to which clients should be protected and unresolved responsibility in the case of accidents.

When is a decision a good decision?

In a team meeting during the third action cycle, the roles of nursing staff and representatives in decision-making were explored in our search for tangible points in which this process could be improved. There was low consensus among nursing staff and between them and client representatives about the extent to which protective measures should be applied. The lack of consensus led to discussions, questions and uncertainty, even after a plan had been drawn up. The second question addressed was ‘How do I know that I have made a good decision? The results may always be bad’. In a subsequent meeting, the action research group discussed the last question again with the management and client representatives. A good decision was defined as one taken on the basis of accurate and recent information so that measures taken related to risks experienced at that time. The clients in AL were not capable of taking decisions about risk or overseeing the consequences, and these decisions were made for them by their representatives and usually carried out by nursing staff. Consensus between those involved was therefore important. Decisions about risk-taking had consequences for daily care, and for this reason flexibility and room for individual choice suited to the actual circumstances were considered important. This room for individual choice could be created by making decisions from a perspective that included the wishes and goals of clients, their representatives and the nursing staff.

Individual guidelines

On the basis of these findings, the action research group decided to make an individual guideline for decision-making for each client. To achieve this, during the fourth action cycle nursing staff started to analyse the risks associated with the client’s epilepsy, the risks of the intellectual and other disabilities in the context of living in AL, and the client’s wishes. Subsequently, they reviewed which measures were called for in different foreseeable situations to protect the client from accidents. On the basis of this analysis, an individual guideline was drawn up. The guideline specified a minimal level of protection which always had to be applied and included examples of possibilities within the boundaries of this basic protection and of situations when increased protection should be applied. These data were discussed in the team and, after reaching consensus, the individual guidelines were discussed with representatives, adjusted where necessary, and included in the care plan. These guidelines were the framework for decision-making, within which nursing staff could make individual choices if the situation called for this. The steps taken to devise individual guidelines for decision-making are listed in Figure 7.

Figure 7.

 Steps to an individual guideline for decision-making.

In the sixth action cycle, the representatives’ viewpoints were assessed using a written questionnaire. The questionnaire results showed that they wanted all risks to which the client was exposed to be analysed and included in the care plan; that consensus was important to them, but when unattainable they reserved the right to make the decision; that decisions must be based on recent and correct information, tuned to the situation; and that the wishes and preferences of the client had to be the primary consideration.

The results of the action research project in AL and the results of the assessment of the representatives’ viewpoints were discussed in a focus group in which nursing staff, management and representatives participated. Agenda problems had prevented client representatives from participating in the study to the same extent as the nursing staff, and the focus group revealed disappointment among representatives about this. What they described as a missed opportunity necessitated further improvement in communication between them and the nursing staff.

In AL, the improvement process also did not end when the study was completed. Observation 6 months later showed that the individual guidelines were being used. They had been discussed with representatives and included in the care plan; they had become a context for caregiving in daily practice and had clarified the limits of responsibility of those involved. Caregiving became more client-centred. Furthermore, the team attributed the improved internal communication to the action research project.


The major goal of the study was to improve decision-making and to empower participants to sustain this improvement. The findings showed that in both action research groups, participants had gained insight in analysis and evaluation of the risks to which clients were exposed. They experienced the need for systematic analysis of risks, consideration of contextual influences and the importance of communication with clients and their representatives about their wishes and preferences. They learned to work methodically and in a goal-centred way, to share insights and decision-making and to discuss solutions tuned to daily practice and clients’ wishes. These findings indicate that the decision-making process had been improved.

In the discussion of validity of action research, some authors see the pragmatic approach in which research is combined with action as validity-enhancing (Hope & Waterman 2003). Besides the rigorous application of the cyclical approach, methodological triangulation and researcher triangulation were used to reinforce validity.

Because of the involvement in the study of only two residential units from one epilepsy centre, the transferability of its findings to other care settings cannot be empirically substantiated. However, it was the small scale of the study that created the conditions for progress, including involvement of all participants, intensive contacts of the researcher with the actors, and the constant tuning into positive changes that lead to growing insight in factors which could improve risk management. The description of phases in risk management and the outline for an individual guideline can be used as techniques in other situations.

It is unlikely that the recognition of phases of risk management and creation of an individual guideline alone will improve decision-making. People go through their own individual processes, and knowledge derived from the technical paradigm alone is not enough to bring about change (Holter & Schwartz-Barcott 1993, Hoogwerf 2002, McNiff & Whitehead 2006). The participative approach in action research, with its continual reflection and mutual goal-setting, is a learning process producing changes in insight, skills and behaviour. Action research can deepen understanding of the use of instruments which support decision-making. Further research could show under which conditions the use of these instruments leads to positive results.

Decision-making about risk involves a sometimes high level of uncertainty. A positive outcome is never guaranteed and the results of even a well-made decision can be tragic (Vallenga et al. 2006). The policy of the organization in which the study took place was that these decisions should be made by consensus between representatives, the physician or psychologist responsible for management of clients’ care, and nursing staff. Nevertheless representatives, while seeing consensus as important, maintained that where consensus could not be reached they should make the decision. This is in accordance with Dutch legislation controlling the use of protective measures to protect the rights of clients in residential care (Vallenga et al. 2006). The study showed that consensus between all involved was not enough in decision-making about risks. The consequences of a decision can affect the behaviour of people involved in caregiving. Therefore, after reaching consensus about risks and the management of these risks, commitment to risk management is necessary. Consensus about the definition of commitment is lacking (Meyer & Herscovitch 2001), but Brickman’s (1987) definition of commitment is well suited to our findings: ‘...a force that stabilizes individual behaviour under circumstances where the individual would otherwise be tempted to change that behaviour’ (Meyer & Herscovitch 2001, p. 302). Clarity in roles and responsibilities and open communication are prerequisites for good decision-making (Fisher & Fisher 1999, Vallenga et al. 2006).

As in our previous study (Vallenga et al. 2006), client participation in decision-making was low. As other people with intellectual disabilities, the clients of AL and RR depended on their representatives and caregivers and their voices were not always heard in decisions made (Hickson & Khemka 1999, Biesaart & Hubben 2000, Vallenga et al. 2006). Some were partially, others fully, unable to participate in decision-making. During the action research, caregiving became more client-centred, and the insight that clients should be involved in decision-making grew.

Client autonomy, emphasis on self-determination and freedom of choice have become important issues in care for people with intellectual disability (Gennep & Steman 1997, Celia & Fisher 2003). However, this concept of autonomy is not meaningful when used in relation to people with severe intellectual disability. Therefore, the concept of autonomy was enlarged to a relational model focussed on the social context in which people live (Agich 1993). People do not make their choices independently, but in continual interaction with each other (Meininger 2001). A continuing dialogue between caregivers and clients is necessary to create relational autonomy and make the concept of dialogue operational as ‘to continuously tune in to the clients’ (Roelink et al. 2002, p. 175) In line with Roelink, we found that continuous assessment of the actual situation of clients, including their wishes and preferences, was necessary to fine tune decision-making.


Using an action research model to improve decision-making provided preconditions to improve decision-making and risk management in the institution involved and suggested a way to improve this process in other institutions. Given the data, we conclude that a systematic approach can improve the quality of risk-management. Discerning the phases in risk management and developing steps towards an individual guideline for decision-making appeared useful tools for this systematic approach. Individual guidelines created the conditions to tune decisions about risks to the severity of seizures and to contextual and personal influences at any time. A dialogue with each client is a precondition for involving the experiences, preferences and capacities of the client in decision-making. Commitment to risk management is a precondition of continuity. This commitment requires reflection on decisions made and on the effects of changed care-giving and open communication with all involved, especially client representatives.

Author contributions

DV, BHGML and MHFG were responsible for the study conception and design and the drafting of the manuscript. DV performed the data collection and data analysis. IYT and PAJMB obtained funding and IYT provided administrative support. MHFG made critical revisions to the paper. MHFG, IYT and PAJMB supervised the study.


The authors would like to thank the clients, their representatives and the participating teams at Kempenhaeghe for their ongoing and enthusiastic collaboration and Jo de Munck for his research contribution.