Multidisciplinary Collaboration in Professional Networks for PD A Mixed-Method Analysis

Abstract

Background:

ParkinsonNet, a nationwide organization with regionally oriented professional health networks in TheNetherlands, aims to improve the quality of Parkinson care. Facilitation of multidisciplinary collaboration is a key objective of ParkinsonNet.

Objectives:

This study examined whether the concept enhances multidisciplinary collaboration between healthcare professionals involved in Parkinson care.

Methods:

A regional network involving 101 healthcare professionals was newly established. Participants received two questionnaires. One aimed at documenting direct working relationships (‘connections’) between professionals and the other aimed at evaluating multidisciplinary team performance. Additionally, thirteen healthcare professionals were interviewed to identify barriers and facilitators for multidisciplinary collaboration. ‘Social network analysis’ focused on sub-networks around three community hospitals at baseline and one year after the implementation.

Results:

The number of ‘knowing each other’ connections increased from 1431 to 2175 (52% , p <  0.001) and ‘professional contact‘ connections increased from 664 to 891 (34% , p <  0.001). Large differences between sub-networks were found, positive changes being associated with a central role of neurologists and nurse specialists committed to multidisciplinary care. The perceived team performance did not change. Participants experienced problems with information exchange and interdisciplinary communication. Generally, participants were unaware of other healthcare professionals involved in individual patients and what treatments they provide simultaneously.

Conclusions:

ParkinsonNet partially enhanced multidisciplinary collaboration between healthcare professionals involved in Parkinson care. Crucial facilitators of this were a central role of nurse specialists and the commitment to collaborate with and refer to expert therapists among neurologists. Additional measures are needed to further improve multidisciplinary care across different institutions and around individual patients.

INTRODUCTION

Western societies age rapidly, which largely explains the rising number of people with chronic diseases and multiple co-morbidities [1]. In the US, chronic conditions already puts neurology healthcare services in high demand [2]. An illustrative example is Parkinson’s disease (PD), where patients become progressively disabled due to a mixture of cognitive, emotional and motor symptoms [3]. In 2005 the estimated number of PD patients in the 10 most populous countries was between 4.1 and 4.6 million; this number is expected to more than double by 2030 to between 8.7 and 9.3 million [4]. Conventional therapies provided by a neurologist, like pharmacotherapy or deep brain surgery, improve clinical outcomes in PD. Furthermore, the evidence to support the effectiveness of allied health interventions as an additional symptomatic treatment is growing rapidly [5, 6]. International studies show that the multifaceted nature of PD requires a team-oriented and personalized approach including physicians, nurses and therapists, but how such care should be delivered is under debate [7–10]. Insufficient expertise, poor collaboration and inadequate communication are recognized as important barriers of multidisciplinary care [11]. ParkinsonNet has been developed to tackle these barriers by standardizing and integrating the delivery of care provided by expert therapists [12]. The concept consists of professional health networks in the catchment areas of community hospitals. Within each network selected therapists are trained according to evidence-based guidelines. Neurologists are stimulated to refer patients to these skilled professionals. ParkinsonNet has nationwide coverage in The Netherlands, with 68 regional networks and 3,000 health professionals for an estimated 50,000 patients [12]. A cluster-randomized trial showed that ParkinsonNet reduced healthcare utilization and costs [13]. Furthermore, participants improved the adherence to guideline recommendations and treated higher patient volumes per year [11].

Despite the emphasis on multidisciplinary collaboration in Parkinson care in the Netherlands, PD patients experience a lack of collaboration between members of their healthcare team [14]. This study aimed to quantify the evolvement of multidisciplinary collaboration between professionals involved in Parkinson care, to investigate their perceived team performance and also to identify barriers and facilitators for multidisciplinary collaboration in the first year after the implementation of a new ParkinsonNet network.

METHODS

Implementation strategy

A regional network covering the catchment area of three community hospitals was newly established in 2011. Participants were selected based on their motivation and geographical location. All professionals received a two-day course focusing on the fundamentals of PD, conventional and allied health therapies, multidisciplinary collaboration and communication skills. Additionally, physical therapists, speech-language therapists, occupational therapists as well as dieticians received a two-day training course based on mono-disciplinary evidence-based guidelines. To improve the accuracy of dedicated referrals to allied health therapists, neurologists and nurse specialists received standardized referral forms with referral criteria. An online search engine, the ‘Parkinson healthcare finder’, included information of all network participants facilitating neurologists and nurse specialists to direct patients to expert therapists in the patients’ immediate environment. Furthermore, collaboration was facilitated by semiannual multidisciplinary meetings, an annual national conference and an internet-based communication platform.

Study design

We used a parallel, mixed-methods design, combining qualitative and quantitative methods [15]. At baseline and after one year, participants received a questionnaire to document their direct working relationships or ‘connections’ with other professionals in the network and a questionnaire on perceived team performance. Semi-structured interviews with a representative sample of participants were conducted prior to and one year after the implementation. The research protocol was submitted to and approved by a local ethics committee.

Measures

Social network analysis

The questionnaire on the direct working relationships listed the names of all participants. Health professionals were asked to tick a box for each name indicating whether this person was known to the participant and another box to indicate whether this person was involved in professional contact. Knowing each other was defined as ‘knowing the face, having talked to each other, or having heard of’. Professional contact was defined as ‘having had professional contact about at least one patient , including referral letters, e-mail, telephone and face-to-face consultation’.

Perceived team performance

The Integrated Team Effectiveness Model provides a multidimensional framework to describe healthcare team performance [16]. Based on this model, we applied the team performance questionnaire with 25 items in three domains; team psychosocial traits, team processes and team outcomes [17].

Semi-structured interviews

Interviewees were selected based on their profession, geographical location and setting to achieve maximum variation in the sample. At baseline and after one year, 13 professionals were interviewed, using a topic guide to identify barriers and facilitators for multidisciplinary collaboration. Verbal consent was provided by all interviewees.

Data analysis

Social network analysis

All participants were assigned to one out of three sub-networks based on their geographical location. Each sub-network represented the catchment area of one community hospital. All connections between health professionals were entered into squared data-matrices; two for ‘knowing each other’ and two for ‘professional contact’ at baseline and one year after the implementation, with the professionals in the rows and columns, and values in the cells to indicate absence (0) or presence (1) of a working relationship. If available, we replaced missing values of non-responders with the values provided by other individuals on the connection. If no substitution was possible, the missing value was replaced with a zero [18]. Separate and independent analyses were done for both ‘knowing each other’ and ‘professional contact’ connections. UCINET6, a software package for the analysis of social network data, was used to calculate the following network measures in each sub-network at baseline and after one year [19, 20]; First, the ‘number of connections’, an absolute measure for the total number of connections between health professionals in the sub-network. Second, ‘density’, a relative measure describing the proportion of all possible connections that are actually present in a network. A higher density means that more professionals know each other or work together. Third, ‘reciprocity’ of reported connections as an indicator of the reliability of the data. If person A indicates to have worked with person B, does person B confirm the connection withperson A?

Perceived team performance

Items on the team performance questionnaire provided a score between 0 ‘strongly disagree’ to 3 ‘strongly agree’. First, negatively phrased items were mirrored (2, 5, 6, 7, 8, 14, 15, 16). Second, an overall team performance score and subscale scores for team psycho-social traits (1, 3, 6, 8, 11, 17, 18, 19, 20), team processes (2, 5, 7, 10, 12, 14, 16) and team outcomes (4, 9, 13, 21, 22, 23, 24, 25) were calculated by averaging item scores. The overall team performance score and subscale scores ranged between 0 and 3, with a higher score meaning higher levels of perceived team performance. We used SPSS 18 to perform an independent student t-test comparing baseline scores with values from one year after the implementation. Overall satisfaction with multidisciplinary collaboration was measured on a 0 ‘poor collaboration’ to 10 ‘excellent collaboration’ scale.

Semi-structured interviews

The audiotape-recorded interviews were transcribed verbatim. Thereafter, we conducted an inductive analysis on barriers and facilitators of multidisciplinary collaboration. Inductive analysis involved the conceptualization of themes from the transcripts in Kwalitan 5.0. All care aspects were labeled independently by two researchers (ME, FN) using the Integrated Team Effectiveness Model as a conceptual framework. This was followed by axial coding, involving the classification of codes into recurring themes. Constant comparison analysis was used to identify a final set ofcodes [21].

Triangulation

We performed parallel data analysis, meaning that the social network analysis and the analysis of perceived team performance and interviews were carried out simultaneously. Findings were not compared until the interpretation stage giving equal weights to both types of data. Through the process of triangulation, relationships between qualitative and quantitative data were defined [22].

RESULTS

Participants

Participants were 101 healthcare professionals from 10 medical, nursing and allied health disciplines geographically scattered and based in either hospital, nursing home or primary care settings (Table 1).

Social network analysis

At baseline 95% (n = 96) and after one year 92% (n = 93) of the participants provided data on their direct working relationships. Table 2 shows that thenumber of ‘knowing each other’ connections increased 52.0% from 1431 to 2175 (p <  0.001) and ‘professional contact‘ connections increased 34.2% from 664 to 891 (p <  0.001). Further analysis on the professional contact connections showed large differences between sub-networks. In sub-network 2 and 3 the number of ‘professional contact’ connections increased significantly (p <  0.005), unlike sub-network 1, where the number of direct working relationships did not change. Figure 1 visualizes the ‘professional contact’ connections in sub-network 3. This figure shows the central position and large number of connections of neurologists and nurse specialists in the network one year after the implementation.

Perceived team performance

The team performance questionnaire was completed by 74% (n = 76) of the participants at baseline and 85% (n = 88) after one year. Overall team performance did not change,1.62 (SD 0.33) to 1.67 (0.33). Additionally, all three subscales did not show significant changes. The ‘team processes’ subscale received the lowest score [1.29 (0.41) to 1.35 (0.52)] and ‘team psychosocial traits’ received the highest subscale score [1.81 (0.34) to 1.86 (0.36)]. Overall satisfaction with multidisciplinary collaboration increased significantly from 4.9 (SD 2.1) to 5.9 (SD 1.7) (p <  0.005). This increase was significant for primary care professionals; 3.9 (SD 1.9) to 5.5 (SD 2.0) (p <  0.005), but not for professionals in community hospitals; 5.5 (SD 2.0) to 6.1 (SD 1.6).

Semi-structured interviews

Participants were interviewed at baseline and one year after the implementation (n = 13); three neurologists, three nurse specialists, three physical therapists, one occupational therapist, one dietician, one speech therapist and one rehabilitation specialist. The coding process produced 42 topics, allocated to the dimensions of the Integrated Team Effectiveness Model and divided into barriers and facilitators for multidisciplinary collaboration (Fig. 2). Interviewees stated that one year after theimplementation:

Nurse specialists play a pivotal role in the coordination of care and providing tailored information to patients.

Neurologists are vital in the referral of patients to expert therapists, yet their commitment to multidisciplinary care differed between the sub-networks. The neurologist in sub-network 1 retired within the first months after the implementation and the neurologist in sub-network 2 took up a sceptical position towards multidisciplinary health interventions. In sub-network 3 the neurologist managed a multidisciplinary team within the community hospital.

The number of referrals increased in sub-network 3, but lagged behind in sub-network 1 and 2.

Interviewees had increased their PD-specific expertise, incorporated guideline recommendations in daily practice and increasingly applied a holistic treatment approach. Interviewees were generally unaware of other professionals involved in individual patients and what treatments they provide simultaneously.

Triangulation

Social network analysis and interview data both indicated that the increase in professional connections in sub-network 3 is associated with a central role of the neurologist and nurse specialist. Moreover, the number of referrals lagged behind in the other two sub-networks, where the neurologists retired or did not commit to multidisciplinary health interventions. Additionally, team performance and interview data both showed that team processes need further improvement. Items on the team performance questionnaire with the lowest scores were: ‘no work agreements’, ‘no clearly defined roles and responsibilities’, and ‘patient data not shared with each other’. Interviewees acknowledged an absence of work agreements and indicated that communication and information exchange should be improved.

CONCLUSIONS

This study showed that the implementation of ParkinsonNet partially enhanced multidisciplinary collaboration. Moreover, the quality of this collaboration was facilitated by neurologists and nurse specialists committed to multidisciplinary health interventions. However, participants felt that one year after the implementation of the network, communication and information exchange could still be improved considerably. In the following paragraphs, we elaborate on these findings.

Key role for nurse specialists

Our study showed that nurse specialists play a vital role in facilitating regional multidisciplinary collaboration. Unfortunately, 55% of all PD patients across Europe have no access to specialized nurses [23]. International studies confirm that nurse specialists have a pivotal role in the coordination of Parkinson care, but also in providing patient education and emotional support [24]. Moreover, they have a positive effect on patients’ sense of well-being and quality of life [25]. Hereby, nurse specialists might reduce the workload of neurologists, enabling them to focus on their primary medical task. Importantly, the American Academy of Neurology predicted that the current shortfall of neurologists in the US, estimated at about 11% in 2012, will grow to 19% by 2025 [1]. Therefore, neurologists increasingly rely on supervised nurse specialists or physician assistants to provide follow-up care, provided that these professionals received sufficient neurological training [2]. As we showed, an alliance between nurse specialists and neurologists improves the quality of care for PD, which might well work similarly for other patients, e.g. those suffering from multiple sclerosis, stroke or epilepsy [26].

Neurologists who provide pharmacotherapy as monodisciplinary care has long been the gold standard in PD management. The challenge now is to combine conventional therapies with allied health interventions within an integrated approach [10]. Moreover, the European Parkinson’s Disease Association stated that patients have the right to be referred to a neurologist with a special interest in PD and who displays a commitment to multidisciplinary care [25]. However, multidisciplinary care is not a clear focus in medical training, and the attitude towards teamwork in healthcare differs between physicians [27]. To date, neurologists involved in ParkinsonNet are not obliged to participate in the training program. Therefore, neurologists who do not attend the training may not be fully informed about the effectiveness of allied health interventions. Moreover, neurologists are not being held accountable for the number of referrals. Non-commitment among certain neurologists might explain the variation in the number of direct working relationships between sub-networks in this study.

ParkinsonNet partially enhanced multidisciplinary collaboration

One year after the implementation of a regional network for PD, communication and information exchange across health professionals need further improvement. A longer timeframe and additional measures are needed to migrate from a multidisciplinary approach towards integrated care. Multidisciplinary care involves several health professionals working independently, not collaboratively, each being responsible for a specific patient care need. ‘Integrated care’ entails a holistic approach conducted by a team of health professionals [9]. In this study, multidisciplinary care was largely adopted, however integrated care was far from being implemented in daily practice.

Much can be learned from the Patient-Centered Medical Home approach (PCMH) [28, 29]. In this model a team of professionals provides care across the continuum of the healthcare system. A personalized care plan for each individual patient is at the heart of the approach with treatment goals, planned interventions and the identification of health professionals responsible for each intervention [28]. Again, the role of specialized nurses in the coordination of the team is essential. PCMH allows for different roles for neurologists; some might serve as the principal physician, others as specialist on referral to expert therapists [29]. Implementation of the PCMH means that patients and professionals are both recognized as experts of the disease and as equal partners in the healthcare team. For instance, neurologists are experts in the cause and the prognosis of the disease and in diagnostic and therapeutic procedures, however patients and caregivers know best about day-to-day decisions and experiences with the disease. Hence, the role of professionals should change from experts who care for patients to enablers who support patients to make their own decisions. Patients on their behalf have to learn how to manage their disease and how to interact with the healthcare team.

Strengths

First, some potential benefits of ParkinsonNet are subjective in nature and until now impossible to quantify [30]. ParkinsonNet has been evaluated in several studies, however variation in the quality of multidisciplinary collaboration was never assessed [12]. This study shows that social network analysis provides a useful tool to study and visualize connections between health professionals [19]. This method is applied in many scientific disciplines, including neurosciences, molecular life sciences and public health. The application in medical research is relatively new, although the first study, concerning the uptake of innovations by physicians, dates back to 1957 [31]. The challenge of social network analysis is achieving high response rates, as we managed to get in this study. Second, we used a mixed-method design to conceptualize the broad spectrum of multidisciplinary collaboration in healthcare. This research method is increasingly utilized in the evaluation of complex health interventions, because it capitalizes the strengths of both quantitative and qualitative approaches [15]. Here, qualitative data proved to be valuable in explaining differences between expected and observed results that would have been left undetected by quantitative methods alone.

Shortcomings

This study was not without shortcomings. First, we solely focused on the professional’s perspective and the patient’s view was not included. From previous studies we know that PD patients experience a lack of multidisciplinary collaboration within their care [14]. At the time of this study, a valid instrument to measure patient experiences in Parkinson care was lacking. Care experiences are now widely used to assess the quality of Parkinson care from a patient’s perspective and to provide health professionals with feedback on their performance [32, 33]. Second, there is no valid instrument to measure perceived team performance in PD care. Therefore, we used a team performance questionnaire originally developed to evaluate teamwork in COPD management. Based on the opinions of two PD experts [BB, MM] and one expert in chronic illness care [HJMV] the items on the questionnaire were found to be relevant for PD. Illustrative for the applicability in PD care is the utilization of the ‘not applicable’ answer category, which was marginally used by participants in our study. Unfortunately, the instrument did not capture information on the geographical location of participants in the network, meaning that no comparisons between sub-networks could be made on perceived team performance.

Practical implications

As a consequence of this study, neurologists involved in ParkinsonNet will be asked to agree on specific terms and conditions, including a commitment to collaborate with and refer to specialized therapists. Currently, we are creating a quality of care registry to provide information on the level of expertise and quality of care delivered by all participants. This registry holds information from health insurance companies (hip fractures and costs), health professionals (provided treatment and perceived collaboration) and patients (care experiences). The data is publicly available in an online Parkinson Atlas (www.parkinsonatlas.nl), thereby offering transparency on the quality of care for each of the 68 participating regions [12]. Additionally, the data will be used to credit professionals for providing better integrated, less fragmented care.

Broader impact for other neurological disorders

PD is a debilitating chronic disorder which benefits from a multidisciplinary approach. ParkinsonNet might be useful as a framework for the management of other neurological disorders, where it is equally important to improve specialized expertise and stimulate multidisciplinary collaboration between health professionals. Certainly, elements of the approach can be applied elsewhere, including the use of disease-specific multidisciplinary training and the adoption of a regionally oriented multidisciplinary approach. The concept might well serve as a model on how to offer patients access to specialized expertise in their immediate environment, and how to gradually develop multidisciplinary care. Finally, the concept is beginning to show its merits in other countries, including the US, where Kaiser Permanente has recently begun to implement the ParkinsonNet approach in its southern regions, with the long-term aim of scaling to other regions and other conditions.

CONFLICTS OF INTEREST

The authors have no conflict of interest to report.