Dr. Michael Sullivan
Centre for Research on Pain, Disability and Social Integration
Current Programs of Research
My research focuses primarily on the psychological determinants of pain experience and pain expression. Separate lines of research address questions of interest from basic, clinical and applied perspectives. At the basic process level, my research is currently focusing on the nature of behavioural systems involved in the communication of pain. At the clinical level, I have recently launched two longitudinal studies examining psychological predictors of problematic medical outcomes. Recent developments in each program of research are described below.
Pain Catastrphizing: Basic processes
In 1995, we published the Pain Catastrophizing Scale (PCS; Sullivan et al., 1995); a measure of catastrophic thinking associated with pain. The PCS has since been translated into 11 languages and has been incorporated in clinical assessment protocols of hundreds of pain clinics in North America, Europe, Australia and Asia. The PCS has become the most widely used measure of catastrophic thinking related to pain, and has been used in population health studies in several countries (e.g., Belgium, Netherlands, US). Pain catastrophizing has been the subject of more than 200 publications and the rate of activity in this area of research continues to increase. Our initial papers describing the experimental and clinical applications of the PCS (Sullivan et al., 1995, 1998), and recent theoretical papers (Sullivan et al., 2001, 2008) have played a significant role in the direction of this research. Pain catastrophizing has emerged as the most powerful and robust psychological predictor of pain-related outcomes.
My research in this area has addressed the processes by which catastrophizing influences pain perception, and more recently, has explored the possible social functions of pain catastrophizing. Research from my laboratory has shown that pain catastrophizing contributes to pain-related outcomes independent of other cognitive or affective variables such as neuroticism, vigilance, self-efficacy, anxiety, fear or depression. Individuals with high scores on measures of pain catastrophizing report more intense pain (Sullivan et al., 1998), experience higher levels of emotional distress in response to painful stimulation (Sullivan et al., 1995), show deficits in their ability to divert attention away from pain stimuli (Sullivan et al., 1997), anticipate higher levels of pain (Sullivan et al., 2001), display more pain behaviour (Sullivan et al., 2002, 2004, Thibault et al, 2008) and infer greater pain in others (Martel et al., 2008). Pain catastrophizing has been shown to predict pain-related disability and activity intolerance better than pain itself (Sullivan et al., 1998; 2002; 2004).
Our theoretical work in this area has formed the basis of a number of studies specifically addressing the interpersonal functions of pain behaviour, and has led to the elaboration of the Communal Coping Model of pain catastrophizing (Sullivan et al., 2001, Gauthier et al, 2008). A series of studies is currently being conducted (SSHRC; 2006 – 2011) to test predictions of the Communal Coping Model of pain catastrophizing. The following broad questions are addressed in this program of research: a) Does pain catastrophizing increase the efficiency with which pain experience is communicated to others? b) Do the partners/spouses of pain catastrophizers provide more support in response to displays of physical or emotional distress? and c) Does the provision of support or solicitousness increase or decrease pain catastrophizers’ propensity to express high levels of pain and emotional distress? Results to date suggest that catastrophizers not only experience more intense pain but perceive more intense pain in others as well (Martel et al., 2008). Spouse observers are more accurate when inferring the pain of catastrophizers than the pain of non-catastrophizers (Gauthier et al, 2008).
A series of studies has also been planned to examine the manner in which pain behaviours are incorporated into narrative during the expression of pain. In these studies, pain patients and interviewers will interact in real time on video displays. Video records of facial displays of pain communication will permit fine grained analysis of the functional role of pain behaviour during verbal communication of pain experience. Manipulations of time delays in signal transmission will be used to influence interpersonal factors such as perceived interest of the interviewer and perceived effectiveness of communication of the pain patient.
Research on pain communication has important clinical implications. In the absence of empirical evidence concerning the meaning or significance of different behaviours associated with pain, clinicians must rely on their own intuitive inferences. In the domain of pain interventions, intuition-based practice has led to the frequent utilization of terms such as ‘functional overlay’, ‘chronic pain behaviour’, ‘symptom magnification’, or ‘secondary gain’. These terms are often pejorative in nature and more often than not, impede, as oppose to facilitate the implementation of appropriate interventions. Research aiming to elucidate of the meaning or significance of different pain behaviours might lead to the development of assessment or intervention tools that will improve the treatment experiences and treatment outcomes of individuals suffering from persistent pain conditions.
These studies form part of an ongoing program of research aimed at increasing understanding about the interactional and relational variables associated with pain experience. This work has prompted considerable discussion and debate amongst researchers from North America and Europe on interpersonal aspects of pain experience. The results of this line of research are calling for greater consideration of pain behaviour systems within theoretical models of pain (Sullivan et al., 2008). Advances in this domain will inform not only psychological models of pain experience and response, but might also have bearing on the elaboration of models that address links between psychological and physiological processes in pain experience. Beyond contributing to theoretical refinement, the ultimate goal of this research is to contribute to clinical interventions for pain-related disorders.
Pain Behaviour: Clinical research
In a series of studies (CIHR; 2004 – 2011), we are currently examining the determinants of pain behaviour in patients with musculoskeletal disorders. Pain behaviours are the various actions or postural displays that are enacted during the experience of pain. These include protective or pain management behaviours such as guarding, rubbing or activity discontinuation, and communicative behaviours such as grimacing or vocalizations.
Observation laboratories have been established at the Stewart Biological Sciences Building and the Constance Lethbridge Rehabilitation Centre. In a current study, participants with low back pain are videotaped while engaged in two performance tasks designed to simulate light domestic and occupational activities. A trained coder marks the position on the video record where pain behaviours occurred. Following completion of the two performance tasks, participants are asked to view the video record. The video record is stopped at positions where the participant displayed pain behaviour. Participants are asked to report the thoughts they experienced during each display of pain behaviour. Cognitions are coded by trained assistants according to standardized classification criteria.
The primary questions addressed by the research include: a) What type of cognitive activity is associated with the expression of different pain behaviours? b) What are the psychometric properties of a video-cued recall procedure for the assessment of cognition associated with pain behaviour? c) Does the assessment of pain behaviour-specific cognitive content enhance the clinical predictive utility of indices of pain behaviour? and d) Does the assessment of pain behaviour-specific cognitive content improve prediction of clinical improvement over existing trait-measures of pain-related cognition? By broadening the assessment of pain behaviour to include cognitive content, it is hoped that the predictive power of pain behaviour indices can be enhanced.
Current research is addressing the predictive value of pain behaviour in prospective analyses of work retention following rehabilitation for soft tissue injuries. As overt components of pain, pain behaviours are not only influenced by environmental contingencies, but also influence the perceptions and behaviour of individuals in the social (occupational environment). A high degree of pain behaviour might be reinforced by attention and concern of co-workers, and may lead to inferences of heightened severity of the pain condition; ultimately impacting negatively on the probability of work retention. Advances in this area, might lead to the development of interventions that might improve work retention rates following rehabilitation.From a clinical perspective, this research holds promise of contributing to new avenues for assessment and intervention for individuals suffering from persistent pain conditions. From a theoretical perspective, the proposed research might contribute to further refinement of biopsychosocial models of pain and disability.
Pain Catastrophizing and ohter risk factors for problematic medical outcomes
Relations between pain catastrophizing and pain-related outcomes have been observed in numerous clinical populations including individuals with work-related soft-tissue injuries (Sullivan et al., 1998; Sullivan and Stanish, 2003), whiplash injuries (Sullivan et al., 2002), patients undergoing aversive dental procedures (Sullivan and Neish, 1999), and patients with neuropathic pain conditions (Sullivan et al., 2005). Prospective studies have shown that pain catastrophizing predicts post-surgical pain (Pavlin et al., 2005), and duration of work absence following soft tissue injury (Sullivan and Stanish, 2003). Population studies have also shown that high levels of pain catastrophizing in asymptomatic individuals predict the future onset of musculoskeletal symptoms (Picavet et al., 2004).
We obtained a 5-year CIHR grant (2006 – 2011) to examine the role of pain catastrophizing, as well as other pain-related psychological variables (fear of pain, self-efficacy, expectancies) as predictors of problematic outcomes following total knee arthroplasty. This research project involves a collaboration of the Departments of Surgery of the McGill University Health Centre (Montreal), the Queen Elizabeth II health Sciences Centre (Halifax), and the Hopital Maisonneuve-Rosemont (Montreal). Initial papers emerging from this line of research suggest that pre-surgical levels of catastrophizing and fear of pain predict post-surgical pain and function (Sullivan et al., 2009).
In 2005, we obtained a 3-year FRSQ grant (2005 – 2008) to examine psychological predictors of poor recovery following whiplash injury. This program of research was a collaborative effort between the Departments of Psychology, Epidemiology and the School of Physical and Occupational Therapy at McGill University. Recently completed analyses indicate that individuals with high scores on measures of catastrophizing experienced pain symptoms for a longer period of time following injury (Sullivan et al., 2008). These findings highlight the importance of developing effective means of targeting catastrophizing thinking in early interventions for individuals with whiplash injuries.
In 2005, we obtained a 3-year IRSST grant to exame the relation between depressive symptoms in rehabilitation outcomes in injured workers. Previous research in this area suggested that depression might represent a risk factor for problematic recovery from soft tissue injury. Research also suggests that current methods of treating depression associated with pain are inadequate. This research project was a collaborative effort between the Departments of Psychology, Epidemiology and the School of Physical and Occupational Therapy at McGill University. Recently completed analyses indicate that while depression, compared to other psychosocial risk factors, was not the most powerful predictor of prolonged disability, depression was the best predictor of absenteeism following return to work.Our work in this area has been important in establishing pain catastrophizing and other psychological variables as risk factors for prolonged pain and disability. This work has also provided the platform for the development of intervention programs specifically designed to target catastrophic thinking about pain.
Perceived injustice: Scale development
Anecdotal reports and clinical studies strongly suggest that individuals’ perceptions of fault regarding their injury might have implications for their trajectory of recovery. In a series of studies (CIHR; 2005 – 2009), we developed a measure of perceived injustice associated with musculoskeletal injury. The scale items address respondents’ perceptions of the injustice of the situation they have lived, others’ lack of appreciation of the severity of their condition, and their belief that their losses are irreversible. Studies addressing the psychometric properties of the scale have been conducted in collaboration with colleagues in the US (Richard Gross, PhD) and Belgium (Geert Crombez, PhD). Results reveal that the scale has sound psychometric characteristics and contributes unique variance to pain-related outcomes, even when controlling for pain severity, pain catastrophizing and depression (Sullivan et al., 2008).
Of particular interest is the relation between scores of the measure of perceived injustice and progress in rehabilitation. In one study, high scores on the measure of perceived injustice were associated with more missed appointments at the treatment clinic, and less improvement on measures of physical improvement (Sullivan et al., 2008). There are indications that high scores on perceived injustice might be a marker for poor outcomes following rehabilitation treatment, and lower probability of return to work. Two recently completed studies showed that high levels of perceived injustice were associate with heightened expression of pain behaviour, greater self-reported disability, and the persistence of symptoms of post-traumatic stress disorder (Sullivan et al., 2009, in press).Avenues to be pursued include examination of the degree to which perceived injustice changes as a function of clinical improvement in rehabilitation, and the development of intervention strategies aimed at modifying perceptions of injustice.
Activity-related temporal summation of pain
Pain severity has been discussed as an important determinant of the trajectory of recovery following musculoskeletal injury. However, the role of pain severity as a predictor of work disability has been difficult to ascertain. Although considerable research has shown that initial pain severity of a predictor of the persistence of pain symptoms, the relation between pain severity and work disability has been less clear. Research findings have been mixed. There are indications that pain severity immediately following injury is a significant predictor of prolonged pain and return to work. Other investigations have reported that pain severity is not a predictor of return to work, and that pain reduction does not necessarily increase the probability of return to work. Even when significant relations between pain severity and occupational disability are found, pain severity rarely accounts for more than 10-20% of the variance in duration or severity of occupational disability.
It is possible that the disability-relevant dimensions of pain experience have not been adequately assessed in previous research. The bulk of research examining the relation between pain severity and occupational disability has relied on static measures of pain severity. Static measures of pain refer to assessments of pain that describe an individuals’ pain experience at one point in time. Static measures of pain may not provide the best index of an individual’s pain experience during physical activity, particularly the repeated nature of physical activity associated with the performance of occupational duties. Change in pain that occur as a function of repeated physical activity might be a more disability-relevant dimension of pain than static measures of pain.
Risk-Factor targeted interventions
It is becoming clearer that disability associated with persistent pain is not just a clinical problem, but a public health problem. Health Canada has reported that the cost of lost production due to pain is in excess of 32 billion dollars annually. Workers’ Compensation Boards (WCB) in Canada spend in excess of 6 billion dollars annually on indemnity benefits alone for musculoskeletal disorders. With wait lists as long as 2 years, pain clinics in the health system cannot deal with the magnitude of the pain problem in our country. Population health models of service delivery will be required in order to effectively manage the suffering and disability associated with persistent pain.
In our applied work, we have explored the viability of a population health model of treatment for injured workers (CIHR, WCB Nova Scotia; 2000 – 2005). Based on research on psychosocial risk factors for pain and disability, we developed a 10-week risk factor targeted intervention, known as the Pain-Disability Prevention (PDP) Program. The risk factors targeted by the PDP Program include fear of movement/re-injury, pain catastrophizing, perceived disability and depression. These psychosocial factors were chosen as targets of the intervention on the basis of research suggesting that these factors account for significant variance in return-to-work outcomes, and are amenable to change through intervention.
The PDP Program was implemented in 2001 in Nova Scotia, as part of a community-based approach to the management of pain-related occupational disability. Psychologists in communities across the province were trained to provide the intervention. Case workers of the WCB were trained in methods of identifying injured workers at risk for chronicity and determining suitability for referral to the PDP Program. A web site (www.pdp-pgap.com) was developed to conduct on-line risk assessments and facilitate communication between case workers and PDP trained clinicians. A manualized treatment approach including a client workbook and information video were developed to maximize fidelity to protocol. The primary objective of the intervention program was to facilitate return to work by increasing involvement in goal-directed activity and reducing psychosocial barriers to rehabilitation progress. By adding a psychosocial risk-factor targeted intervention to existing community-based treatment services (e.g., medical management and physiotherapy), the goal was to establish 'virtual' multidisciplinary treatment teams at the level of the community.
The PDP Program was implemented on a province-wide basis in Nova Scotia in August 2001 and in Prince Edward Island in January 2003. Over 100 psychologists in the Atlantic Provinces were trained to deliver the PDP Program, and more than 2000 pain patients were treated as part of the initial clinical trials. Promising results have been obtained in initial analyses of the effectiveness of the PDP Program. For the first 100 injured workers treated in the program, 60% of participants returned to work within 4 weeks of treatment completion, compared to an 18% base rate (Sullivan and Stanish, 2003). The PDP Program was featured in a Special Series of Secondary Prevention in the Clinical Journal of Pain in 2003, comparing international models of disability prevention. The PDP Program was featured once again in 2005 in a Special Issue of the Journal of Occupational Rehabilitation on models of disability prevention.
Prior to the development and implementation of the PDP Program, psychological services in North America were used primarily in the treatment of individuals whose condition had already become chronic, where the central goal was the management of chronic pain. The PDP Program was instrumental in increasing the use of psychological services in the early stages of recovery from injury, where the central goal is the prevention of chronic pain and pain-related disability.
The PDP Program was initially conceived as an evidence-based approach to occupational rehabilitation. After the first year of implementation in Nova Scotia, focus groups were held with psychologists who had provided PDP services to clients of the WCB. It was our view that while practice should be guided by evidence-based principles, consideration of the practice experience of the PDP clinicians would also be important in ensuring that the treatment program was achieving its best outcomes. As a result of this consultation, the PDP Program was modified in 2003, adopting a modular approach that could be better tailored to individual client needs. The current version of the PDP Program has been translated into French, and approximately 60 psychologists who provide services to work injured clients in Quebec have been trained.
In early 2005, all treatment-related materials pertaining to PGAP (Information Video, Treatment Manual, Client Workbook) were translated into French. To date, 500 physiotherapists and occupational therapists in Quebec have been trained to deliver PGAP. Recent results of clinical trials with the French version of PGAP indicate that PGAP increases return to work rates in individuals recovering from work injuries (Sullivan and Adams, 2009), and even in individuals who have been work disabled for several years (Adams et al., 2007).
We are also in the initial stages of developing a pilot project with the Department of Veterans’ Affairs Canada to assess the feasibility of implementation and the efficacy of the PDP Program and PGAP in the treatment of military members who suffer from persistent pain conditions. The community-based approach of the PDP Program and PGAP make these attractive interventions approaches for an organization whose (English and French) clients are distributed across a vast geographical area.
Over the past 3 years, we have been working closely with the Accident Compensation Corporation of New Zealand to examine ways in which PGAP might be implemented as part of a national strategy for disability prevention. Approximately 400 front rehabilitation professionals have been trained in New Zealand. A centralized data monitoring system has been established whereby all PGAP-relevant data from each clinic providing PGAP services in New Zealand is sent monthly to a main server. Country-wide data collection will permit examination of impact related questions on a population basis. To our knowledge, no other standardized rehabilitation program has been implemented on such a large scale.
In the US trial, a telephonic version of PGAP was used. The telephonic version of PGAP was procedurally identical to the face-to-face version except that all sessions were conducted by telephone. Although we considered that the telephone delivery method might compromise the impact of the intervention to some degree, accessibility was maximized in that anyone with a telephone could be a potential recipient of the service.
Community-based approaches to secondary prevention have a higher probability of being sustainable within the budgetary constraints of health care systems and third party payers. Although multidisciplinary pain treatment centres have been discussed as the treatment of choice for individuals suffering from persistent disability, these centres are typically situated in large cities and not readily accessible to individuals living in outlying or rural regions. Furthermore, the cost associated with treatment in multidisciplinary centres is often beyond the means of many third party payers. Through the establishment of ‘virtual’ multidisciplinary teams at the community-based level, the PDP Program and PGAP hold promise of making available the essential ingredients of multidisciplinary treatment to individuals who would have otherwise not had access to such services.