See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224067421

Operant Learning Theory in Pain and Chronic Pain Rehabilitation

Article  in  Current Pain and Headache Reports · January 2012

DOI: 10.1007/s11916-012-0247-1

CITATIONS READS

88 6,602

4 authors:

Rena Gatzounis Martien G S Schrooten

Maastricht University Örebro University
20 PUBLICATIONS   177 CITATIONS    50 PUBLICATIONS   593 CITATIONS   

SEE PROFILE SEE PROFILE

Geert Crombez Johannes W Vlaeyen

Ghent University KU Leuven
576 PUBLICATIONS   32,928 CITATIONS    567 PUBLICATIONS   40,762 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

How do we sense our heartbeat? View project

Psychological models about pain View project

All content following this page was uploaded by Johannes W Vlaeyen on 06 August 2014.

The user has requested enhancement of the downloaded file.

Curr Pain Headache Rep (2012) 16:117–126
DOI 10.1007/s11916-012-0247-1

PSYCHIATRIC MANAGEMENT OF PAIN (MR CLARK, SECTION EDITOR)

Operant Learning Theory in Pain and Chronic

Pain Rehabilitation
Rena Gatzounis & Martien G. S. Schrooten &
Geert Crombez & Johan W. S. Vlaeyen

Published online: 20 January 2012

# Springer Science+Business Media, LLC 2012

Abstract The application of operant learning theory on Introduction

chronic pain by Fordyce has had a huge impact on chronic
pain research and management. The operant model focuses on Pain traditionally has been considered from a biomedical
pain behaviors as a major component of the pain problem, and perspective, viewed as a reflection of bodily damage. A rad-
postulates that they are subject to environmental contingencies. ical shift occurred in 1976, when Fordyce [1] proposed a
The role of operant learning in pain behaviors generally has behavioral model of pain as an alternative to the prevailing
been supported by experimental studies, which are reviewed in disease model. The behavioral approach does not disregard
the present article. Subsequently, the rationale, goals, and the role of biological factors in pain, but highlights the impor-
methods of operant behavioral treatment of chronic pain are tance of learning processes in the development and mainte-
outlined. Special attention is paid to three therapeutic techni- nance of chronicity. Fordyce’s distinction between factors that
ques (graded activity, activity pacing, and time-contingent initiate the pain problem and factors that perpetuate it was
medication management), which are discussed in detail with pioneering. Nowadays, the importance of associative learning
regard to their operationalization, effectiveness, and (possible) (operant and classical conditioning) in pain conditions (in-
mechanisms of action. Criticisms of the operant model are cluding chronic pain) via experiential, verbal, and observa-
presented, as are suggestions for the optimization of (operant) tional pathways [2, 3, 4•, 5–7] is widely recognized. Based on
behavioral treatment efficacy. the primacy of behavior and its environmental contingencies,
operant behavioral treatments for chronic pain have been
Keywords Operant learning . Operant conditioning . developed and are continued to be used. This article summa-
Chronic pain . Operant behavioral treatment . BT . rizes the operant theory of chronic pain and its applications in
Pain management . Pain intervention chronic pain management and rehabilitation.

R. Gatzounis (*) : M. G. S. Schrooten : J. W. S. Vlaeyen

Research Group Health Psychology, University of Leuven, Operant Learning Theory in Pain
Tiensestraat 102, Box 3726,
3000 Leuven, Belgium
e-mail: [email protected]
Unlike the traditional biomedical model, which focuses on
pain intensity as cause of disability and target of treatment, the
M. G. S. Schrooten : J. W. S. Vlaeyen operant model has pain behaviors at its core. Pain behaviors,
Department of Clinical Psychological Science, namely actions, verbalizations, or facial expressions that occur
Maastricht University,
in response to pain, might be adaptive when pain is acute, in
P. O. Box 616,
6200 MD Maastricht, the Netherlands the sense that they facilitate minimization of damage through
protecting one’s self from the source of pain or through elicit-
G. Crombez ing support from others by communicating distress and danger
Department of Experimental-Clinical and Health Psychology,
(protective and communicative pain behaviors, respectively;
Ghent University,
Henri Dunantlaan 2, [8, 9]). However, such behaviors can become dysfunctional
9000 Ghent, Belgium when they persist despite the absence of danger. Protective
118 Curr Pain Headache Rep (2012) 16:117–126

behaviors, particularly guarding, have been found to predict patient’s pain behaviors might become a discriminative cue
disability [10–12]. Moreover, the avoidance of activities driv- [25]. This may lead to a heightened elicitation of pain behaviors
en by fear of injury or reinjury can lead to the development of in the presence of the spouse. In contrast, a patient might
functional disability, depressive symptoms, and perhaps dis- engage in fewer pain behaviors in the presence of a person
use and physical deconditioning [13–16]. who has not been associated with regular reinforcement.
According to the operant learning theory [17], behavior is a Whether stimuli sharing common features with the discrimina-
function of its consequences. Specifically, behaviors of a tive cues also may act as such is not clear. If no reinforcement is
certain response class (ie, a cluster of functionally similar provided in the presence of a discriminative cue, extinction is
behaviors) are exhibited more frequently when followed by predicted to occur. However, a typical extinction curve involves
pleasant outcomes (positive reinforcement) or by the removal a temporary increase of the target behavior followed by a
of unpleasant outcomes (negative reinforcement); the inverse subsequent decrease.
happens when negative outcomes follow behaviors (positive To summarize, the operant model of pain focuses on pain
punishment) or when positive outcomes are removed (nega- behaviors as a central component of the pain problem, and
tive punishment); this is known as the law of effect. Different highlights the role learning plays in this respect. The small to
reinforcement schedules produce more or less powerful learn- moderate correlations between pain behaviors and pain inten-
ing. These basic principles, first introduced by Skinner [17], sity suggest that the exhibition of pain behaviors does not
were applied to pain and illness behaviors by Fordyce and his exclusively depend on how severe pain is. Numerous obser-
associates [1, 18–21]. Simply put, the model states that rein- vational studies that have focused on spousal responses as
forcement of pain behaviors leads to their maintenance, while learning factor in patients’ pain behaviors [25, 26] support the
punishment or nonreinforcement of healthy behaviors leads to model. These studies have been summarized by Newton-John
their extinction (Fig. 1). Outcomes that can function as rein- [26] and are not considered here. Instead, we now focus on
forcers of pain behaviors include the reduction of pain inten- experimental data on the operant model of pain.
sity [22], the inhibition of pain-related fear [3], the pleasant
feeling of rest [19], and the attention provided by significant
others [1, 8]. Experimental Validation of the Operant Model in Pain
A lack of perfect association between pain behaviors and
pain intensity would support the notion that factors other than Despite the great influence of the operant model on the man-
pain itself contribute to the maintenance of pain behaviors. agement of chronic pain, until recently few experimental
Indeed, a recent meta-analysis [23] found only a moderate studies had been conducted. An early investigation provided
correlation between exhibited pain behaviors and reported support for the model [27], but a subsequent attempt yielded
pain severity. Perhaps more interestingly, this correlation inconsistent results [28]. Since then, the interest in the exper-
was moderated by pain chronicity, in that it was significantly imental investigation of operant learning in pain seemed to
weaker among the chronic pain patients, who, presumably, diminish. During the past years, however, a number of rele-
had been exposed to learning factors for their pain behaviors vant experimental studies have been published.
for longer periods. In a typical operant conditioning paradigm, participants
Reinforcement of pain behaviors might be associated with receive a series of physically identical painful stimuli, and
stimuli that are regularly present when reinforcement is given. the consequences following behavioral responses to these
In operant learning terms, these stimuli are called discrimina- stimuli are manipulated. Baseline measurements of pain
tive cues [1, 24]. Discriminative cues signal the presence of a behaviors are obtained throughout an initial phase. During
context in which the probability of reinforcement is high. For a subsequent learning phase, target behaviors are reinforced
example, a solicitous spouse that frequently reinforces a or punished. In an up-training condition, the increase of pain

Fig. 1 Examples of operant conditioning procedures in chronic pain. others increase in frequency on second line. The pain decrease that occurs
Pain exacerbation contingent on activity functions as a punisher, leading during rest reinforces its occurrence, leading to higher frequency of rest
to a decrease of activity levels and/or avoidance of the specific activity on on third line
top line. Verbal or nonverbal pain expressions that elicit attention from
Curr Pain Headache Rep (2012) 16:117–126 119

behaviors is rewarded while their decrease is punished; the tightening the muscles around the eyes), but not for
inverse reinforcement pattern is followed in a down-training others. Although conditioning did not succeed for some
condition. Learning is assumed to have occurred when an participants, this study is the only one to date that has
increase in the frequency of the reinforced behaviors and a investigated operant conditioning of facial pain expres-
decrease in the frequency of the punished behaviors have sions. Other pain behaviors have not yet been investigat-
taken place. ed. The operant conditioning of electroencephalographic
Some studies have successfully brought verbal pain re- activity also has been addressed, but findings are incon-
port under operant control with the use of various rein- sistent [30, 31].
forcers. Jolliffe and Nicholas [29] used verbal statements In summary, several experiments that have used diverse
(such as “That’s it!”) to positively reinforce the increasing samples, painful stimuli, and reinforcers and punishers have
pain reports of healthy participants. Half their sample did shown that pain report can be brought under the control of
not receive any reinforcement. Painful stimuli were either environmental contingencies, even when the report is not
stable over trials or decreasing in intensity. As hypothesized, given verbally. To our knowledge, only one study showed
the reinforced group reported higher pain levels than the that facial pain expressions can come under operant control
control group; interestingly, this was the case even when [35••]. Remarkably, we also could not retrieve any study
lower-intensity stimuli were administered. Successful up- testing the modulation of pain behaviors in the presence or
and down-conditioning in healthy participants was demon- absence of discriminative stimuli after an operant condi-
strated by Lousberg et al. [30], who reinforced pain ratings tioning procedure. Although understandably difficult to
with monetary gains. Flor et al. [31] assigned pain patients operationalize, it would be interesting to see similar inves-
and healthy matched controls either to an up- or to a down- tigations targeting other pain behaviors, such as task inter-
conditioning group. The reinforcers comprised monetary ference and avoidance behavior because these measures
rewards and computer smileys. Electrocutaneous stimuli relate more to patients’ disability and are thus more
were administered at four different intensities above pain ecologically valid. In the following section, we consider
threshold. Both up-conditioned groups reported more pain how operant learning theory is applied in chronic pain
than the down-conditioned groups and, although clinical management.
status had no effect on learning, it did influence “unlearn-
ing,” as patients’ pain reports did not decrease as fast as
those of healthy controls during a subsequent extinction Operant Behavioral Treatment in Chronic Pain
phase. Rehabilitation
Besides verbal pain report, nonverbal reports also seem
to be conditionable. Hölzl et al. [32] instructed healthy Fordyce based his operant behavioral program on the suc-
participants to keep the intensity of tonic heat stimuli con- cessful treatment of one single patient (as reviewed by
stant by means of a trackball device. Decreases or increases Butler [36]). Purely behavioral interventions for chronic
in temperature served as intrinsic reinforcers and punishers pain have been found to improve functioning and reduce
of the participants’ responses. Each trial began with the final pain and pain behaviors, albeit the reported effects are
temperature of the previous trial. As expected, temperature rather modest [37, 38; but see 39••]. It has thus been argued
decreased in a down-conditioning group and increased in an that operant therapy solely is not a remedy for chronic pain
up-conditioning group. The degree with which the temper- [40]. Therefore, the state of the art in chronic pain rehabil-
ature decreased seemed to depend on the magnitude of the itation is the application of behavioral methods as one
reinforcement [33]. However, an attempt to replicate these component of a multidisciplinary program. For this reason,
findings in a patient sample was not successful, as comor- we first present the rationale of operant behavioral treat-
bidity appeared to affect learning. The fact that learning ment, and then separately discuss three therapeutic
was again observed among healthy participants, however, techniques.
led the authors to propose that patients’ pain “learning The aim of behavioral treatment, in contrast to interven-
history” might be too difficult to overcome under experi- tions stemming from the biomedical model, is not to dimin-
mental conditions [34••]. ish the pain experience, but rather to increase functioning
Whereas research to date has focused on pain reports, despite the pain [1, 24]. To achieve this goal, behavioral
facial pain expressions also have been shown to be influ- therapists attempt to decrease the frequency of pain behav-
enced by operant conditioning. Kunz and colleagues iors and increase the frequency of healthy behaviors by
[35••] used computer smileys to reinforce facial pain removing reinforcers from the former and adding them to
expressions to moderately intense heat pain. Conditioning the latter [19, 41]. When there are no healthy behaviors to
effects were observed only for two components of pain- reinforce, as is the case with severely disabled patients, the
indicative facial expressions (lowering the brows and generation of new healthy behaviors can be facilitated by the
120 Curr Pain Headache Rep (2012) 16:117–126

use of shaping, otherwise called reinforcement of successive patient [20]. Subsequently, the patient performs these exer-
approximations [1]. These are behaviors that progressively cises to tolerance level (ie, until pain or fatigue drives them to
approach a final target behavior. While reinforcement is stop), while their performance is recorded in distance units or
initially contingent on simple behaviors, it is gradually number of repetitions. After at least three observations, occur-
shifted towards more complex behavior patterns, leading ring over several days, baseline quotas are determined by
to the exhibition of the target behavior [42, 43]. Vlaeyen averaging the performance levels of the patient for each exer-
et al. [44] successfully used such a shaping procedure in a cise and lowering them by 10% to 25% (Fig. 2) [1, 20, 47]. In
chronic pain patient with standing and sitting intolerance. the main treatment phase, the patient performs the selected
Specifically, they divided the higher goals of standing and exercises to quota and receives reinforcement for doing so.
sitting into a hierarchy of smaller steps, and provided rein- Quotas are gradually increased over the course of treatment.
forcement each time one of them was achieved. The patient Continuous monitoring and, if needed, adjustment of the
progressively moved along this hierarchy of subgoals, quotas assist in preventing repeated failures of the patient to
reaching a significant increase in standing and sitting toler- meet them [20].
ance that was still visible at a 6-month follow-up assess- Intervention programs with GA at their core have re-
ment, despite the fact that no changes in pain intensity were ceived a great deal of attention by researchers. Compared
reported. to no intervention or to standard medical care, GA has been
The basis of a successful behavioral intervention is a thor- shown to be beneficial for conditions such as low back pain
ough behavioral analysis. According to Fordyce and col- and chronic fatigue syndrome [48, 49]. Studies on other
leagues [1, 19, 20], such a behavioral analysis should yield populations, however, have found small [50] or no effects
information regarding the usual time patterns of pain; activi- [51]. GA has been combined with workplace intervention
ties, postures, and positions associated with increased or for the reduction of return-to-work time for persons with
decreased pain; the patient’s pain behaviors and healthy low back pain. Two studies support the superiority of such
behaviors and the social environment’s responses to them; programs over usual care [47, 52]; however, others have found
pleasant and unpleasant outcomes that follow pain behaviors; no differences [53] or even GA to be counterproductive when
the impact of the loss of valued activities on the patient; and delivered after the workplace intervention [54]. Overall, there
the patient’s competencies and incompetencies to perform are indications that GA might not deliver better results than
their social, occupational, and other obligations. This infor- conventional exercise [55, 56], although it appears to be more
mation can be collected by means of interviews with the cost-effective [57].
patient and their significant other(s), diaries, electromechani- Efficient delivery of treatment requires the identification
cal devices and direct observation [45], and supplemented of its active components. It would thus undoubtedly be
with the use of questionnaires [18]. interesting to advance our knowledge regarding the exact
Once the behavioral assessment has been completed, thera- mechanism of GA. According to the operant model, this
pist and patient can set well-defined, quantifiable, realistic pertains to the quota-contingent reinforcement of increased
treatment goals [46] that are functionally relevant to the patient. activity levels. A second assumption is that performing
Because the final aim of therapy is to enable patients to perform activities with no adverse consequences (eg, further injury)
activities that are important in their everyday life, target healthy causes a reduction in fear of movement, thus increasing the
behaviors must be defined as such in the context of each probability of future engagement in activities and, therefore,
patient’s situation. Therapeutic techniques, such as graded of improved functioning. In support of this, pain catastroph-
activity (GA), activity pacing (AP), and time-contingent med- izing has been shown to decrease as a result of a GA
ication management, which we will now describe, are used to intervention [50], and to mediate the effects of both a GA
facilitate patients’ achievement of valuable life goals. Some and a physical training program for chronic low back pain
examples of goals, their translation into specific treatment [58]. Also, decreased hypervigilance to somatic symptoms,
targets, and their approach by operant behavioral treatment which is fueled by fear of pain [16], has been named as a
can be found in Table 1. mediator of GA treatment effects [48]. Moreover, regarding
reductions in disability among chronic low back pain
patients, GA has been found to be equally or less effective
Graded Activity than graded exposure [59, 60], the predominant treatment of
fear of pain, which aims at restoring functioning by expos-
The term “graded activity” was introduced by Lindström and ing patients to feared activities and thus disconfirming their
colleagues [47], although the technique had been described beliefs about the harmfulness of these activities [61]. Thus,
earlier [1, 20]. In an initial assessment period, clinician and it appears that GA might work in more ways than the one
patient select some target physical exercises according to their proposed by the operant model. These mechanisms might
pertinence to the pain problem and to activities relevant to the not necessarily be incompatible.
Curr Pain Headache Rep (2012) 16:117–126 121

Table 1 Examples of desired treatment outcomes, their translation into treatment targets, and operant behavioral techniques that can be used for
their achievement

Patient’s functional goal Specific therapy target(s) Operant treatment approach

Successful completion of basic daily Improved activity pattern Pacing

activities (eg, personal care, household tasks)
on all days, including “bad days”
Return to paid employment Increased sitting/standing tolerance Shaping
Increased activity level Graded activity
Improved sleep quality Increased lying tolerance Paced change of bodily position
Decreased sleep during the day Daily activity diary capturing contingencies
between sleeping patterns during the day
and sleep quality during the next night
Weekly visits to the cinema Increased sitting tolerance Shaping
Decreased reliance on medication Gradual decrease of medication use Time-contingent medication consumption
Improved gait Extinction of protective guarding behaviors Removal of reinforcement (eg, attention,
taking over the patient’s task) contingent on
guarding behaviors
Establishment of alternative, “healthy” gait Shaping of new walking pattern

Activity Pacing described in the discussion of GA [62]. Baseline break dura-

tion is also determined. Patients can use a timer as a reminder
AP involves breaking down an activity in smaller parts and of interrupting or resuming an activity, and a daily activity
alternating it with breaks [62–64]. According to several diary, which captures the time spent on target activities and
authors [46, 62], AP involves changing the contingency of rest, to obtain feedback regarding their progress in relation to
the breaks from pain or fatigue to a specific time point or the the goals initially set.
completion of a part of the activity. However, a clear consen- Unfortunately, AP has not yet been systematically investi-
sus regarding the definition of AP is lacking [65•, 66, 67]. gated. The recently published PACE trial (Pacing, graded Ac-
From the operant perspective, the rationale for AP is that tivity and Cognitive behavior therapy: a randomized
the continued performance of an activity to tolerance level Evaluation) [49] did not find AP to add to standard medical
exacerbates pain intensity, leaving the patient in need of care for chronic fatigue syndrome. No reference to changing
prolonged rest before being active again. The increase in pain the contingencies of the breaks is made in this trials’ AP arm
then functions as a punisher of activity, whereas rest is being treatment manual, though. A study on a small sample of oste-
negatively reinforced by pain reduction. After a while, the oarthritis patients showed that AP was more effective in reduc-
patient might again be active until pain becomes intolerable to ing fatigue (but not pain) when tailored to each patient [68].
make up for the lost time. Patients are then trapped in an over-/ AP appears to be commonly used and especially favored
underactivity cycle [46, 62]. To assist patients to take breaks in by patients [69]. For this reason, it might be worthwhile to
response to cues other than pain, the baseline level for the reach a consensus regarding its definition and to further study
target activity is firstly established in a way similar to the one its mechanism and the circumstances under which it might be

Fig. 2 Schematic
representation of the
number of repetitions
of a physical exercise during
baseline measurements, quota
setting, and graded activity
treatment (fictional data)
122 Curr Pain Headache Rep (2012) 16:117–126

effective. Besides the mechanism proposed by the operant cognitive processes [eg, [79, 80]. Learning theories in the field
model, alternatives should be considered. It is possible, for of pain can be refined by the application of such advancements.
example, that pain-contingent breaks reinforce the abrupt For example, it has been proposed that verbal learning
interruption of activities. Interruptions are known to have accounts for learning to fear, and thus to avoid, negative out-
adverse cognitive and affective consequences [70, 71], lead- comes that have never been directly experienced, such as
ing to higher disability. On the other hand, predictable breaks disability [81, 82]. Behavior acquired in this way is called
might facilitate preparation for activity resumption, thus in- rule-governed; “hurt equals harm” is perhaps the most widely
creasing functioning and self-efficacy. Moreover, relying on held rule amongst chronic pain patients. Rules can be provided
cues other than pain to initiate and terminate a break might by an external source, such as a physician [83], or be self-
limit the need to attend to the pain. As with GA, these generated [81]. Clinicians should carefully consider rule-
mechanisms are not necessarily mutually exclusive. Clinical governed behaviors because their slow extinction might com-
and experimental studies are needed to expand our knowledge promise treatment efficacy. A more recently developed treat-
on the effectiveness and mechanisms of this widely used but ment, which also has its roots in the operant and behavioral
poorly studied technique. tradition and has rule-governed behavior at its core, is Accep-
tance and Commitment Therapy (ACT) [84, 85]. The central
feature of the ACT analysis is that a person’s language history
Time-contingent Medication Administration can lead them to a state of “psychological inflexibility” such
that they are dominated by their pain experience. The aims of
The reduction or removal of the unpleasant feeling of pain ACT are essentially the same as those of the operant treatment
renders analgesic medication a powerful reinforcer. Some approach: to facilitate a person’s engagement with a range of
persons might therefore consume doses higher than sufficient, valued activities in the presence of pain, and to change the
with the risk to experience increased side effects and to control over behavior by altering the context. For a discussion
develop tolerance to the medication. Indeed, the prevalence on ACT, see [86].
of opioid misuse amongst chronic pain patients has been Secondly, the operant model has been criticized for overly
estimated to reach up to 50% [72–74]. As an intervention for simplifying the interactions between pain patients and their
persons using high medication doses, Fordyce [18] suggested partners [26, 87•]. These interactions have been placed in a
to change their use pattern, making medication contingent on broader interpersonal context by newer models [87•, 88, 89•].
time, instead of on pain, while gradually reducing the doses. Similarly, the view on facial pain expressions as a tool of
The time-contingent administration of medication is part of communication has been expanded to include their evolution-
several countries’ guidelines for low back pain management ary origins and function [9]. Variables such as marital satis-
[75]. However, although early data supported the superiority faction have been found to influence patients’ interactions
of the time-contingent regimen over the pain-contingent, or with (significant) others [26], and may thus be worthwhile to
pro re nata, method [76], a recent study showed that chronic consider in future studies.
pain patients consuming opioids in a time-contingent manner It has been shown that (operant) behavioral therapy is
used higher doses and were more concerned about their use superior to other treatments, including cognitive-behavioral
and the potential accompanying problems than patients using treatment, for the reduction of pain behaviors [37, 90]. This
opioids “as needed,” despite their comparable pain levels [77]. indicates that pain behaviors must be targeted directly to be
Because this study was cross-sectional, however, the possibil- reduced, and points to the usefulness of the approach. Despite
ity cannot be ruled out that the lack of differences in pain the effectiveness of behavioral interventions, however, effect
intensity can be attributed to the higher doses consumed by the sizes remain modest [37, 38]. One realistic possibility is that
time-contingent group. To obtain clear answers regarding the matching prospective patients to interventions to which they
effectiveness of the two different consumption regimens, it is are most likely to respond will increase treatment efficiency
important that controlled studies consider patients’ baseline [91, 92]. For example, it has been demonstrated that fibromy-
pain intensity. algia patients with higher levels of pain behaviors, physician
visits, catastrophizing thoughts, and physical impairment, and
whose partners are more solicitous, are more responsive to
General Considerations operant behavioral therapy than persons without these char-
acteristics [93]. Operant behavioral therapy might thus be
Despite the huge impact of the operant model on the under- more appropriate for this patient group. On the other hand, it
standing and management of chronic pain, it has been criti- is equally interesting to identify predictors of negative out-
cized for its exclusive focus on overt behavior and its neglect comes. High baseline functioning, for instance, has been
of cognitive variables [78]. Contemporary models in learning named as a contraindication for use of GAwith chronic fatigue
psychology have synthesized conditioning principles with syndrome patients [94]. Highly functioning patients might
Curr Pain Headache Rep (2012) 16:117–126 123

benefit more from reprioritization of activities, which involves studies reviewed in the present paper have provided support
a change in the tasks one performs, such as the substitution of for the role of learning in the exhibition of pain behaviors.
stressful, unpleasant activities with more enjoyable ones [94]. Learning is now widely accepted as an important contribut-
Treatment outcome researchers might want to consider the ing factor to the development and maintenance of chronicity
inclusion of such moderating variables in their designs [95] or in pain patients. Although the effectiveness of behavioral
to categorize patients to responders and nonresponders by treatments is generally supported by the data, it is important
means of cluster analysis. to investigate the mechanisms and the predictors of change
Equally important for the (cost-)effective delivery of treat- to maximize the efficiency of their delivery.
ment is the investigation of techniques’ mechanisms of
change [95, 96]. Awareness of the mechanisms will assist Acknowledgements This study was supported by the Odysseus Grant
clinicians in selecting the most appropriate technique for each “The Psychology of Pain and Disability Research Program” funded by
the Research Foundation—Flanders, Belgium (Fonds Wetenschappelijk
individual patient. For example, if GA is proven to work
Onderzoek [FWO] Vlaanderen).
through decreasing patients’ pain-related fear, then the use of
a technique that directly targets fear, such as graded exposure, Disclosures No potential conflicts of interest relevant to this article
might be preferable. As discussed earlier, our knowledge of the were reported.
mechanisms is still limited. Experimental studies, which allow
the manipulation of hypothesized mediators of treatment out-
comes, might be fruitful in this respect. References
The above suggestions only make sense when treatment
fidelity is guaranteed. Low therapist adherence to the princi- Papers of particular interest, published recently, have been
ples and practice of a therapeutic approach, low therapist highlighted as:
competence, and low differentiation between treatments with • Of importance
regard to crucial therapeutic elements constitute dangers for •• Of major importance
treatment integrity; despite its importance, however, treatment
fidelity is hardly considered in outcome studies in general 1. Fordyce WE. Behavioral methods for chronic pain and illness.
[97], and in studies on the effectiveness of behavioral treat- Saint Louis: Mosby; 1976.
2. Flor H, Birbaumer N, Schulz R, et al. Pavlovian conditioning of
ments of pain more specifically. Its assessment, though, is
opioid and nonopioid pain inhibitory mechanisms in humans. Eur
essential for inferring valid conclusions with regard to the J Pain. 2002;6:395–402.
effectiveness of treatments and/or techniques, especially 3. Goubert L, Crombez G, Peters M. Pain-related fear and avoidance: a
when these are delivered by professionals from different conditioning perspective. In: Asmundson GJG, Vlaeyen JWS,
Crombez G, editors. Understanding and treating fear of pain. New
backgrounds and disciplines. For example, low adherence
York: Oxford University Press; 2004. p. 25–50.
of physiotherapists to GA programs has been reported 4. • Klinger R, Matter N, Kothe R, et al. Unconditioned and condi-
[51]. Indeed, therapists might especially drift from behavioral tioned muscular responses in patients with chronic back pain and
“doing therapies” [98, p. 119] to avoid distressing the patient. chronic tension-type headaches and in healthy controls. Pain.
2010;150:66–74. This study experimentally showed how, through
A method for the assessment of the delivery of behavioral
a process of classical conditioning, one might come to exhibit
treatments for pain was recently reported by Leeuw and increased muscular reactivity to nonpainful stimuli.
colleagues [99]. They described the development and imple- 5. Schneider C, Palomba D, Flor H. Pavlovian conditioning of mus-
mentation of a protocol, which assesses treatment adherence cular responses in chronic pain patients: central and peripheral
correlates. Pain. 2004;112:239–47.
(occurrence of crucial therapeutic elements), treatment con-
6. Seymour B, O’Doherty JP, Koltzenburg M, et al. Opponent
tamination (occurrence of prohibited elements), and treatment appetitive-aversive neural processes underlie predictive learning
differentiation. The assessment of these factors can indicate of pain relief. Nat Neurosci. 2005;8:1234–40.
whether treatment was indeed delivered as intended. Further, 7. Goubert L, Vlaeyen WS, Crombez G, et al. Learning about pain from
others: an observational learning account. J Pain. 2011;12:167–74.
patient-related factors, such as the understanding of skills
8. Sullivan MJL. Toward a biopsychomotor conceptualization of pain:
learned in therapy and their application in the context of implications for research and intervention. Clin J Pain. 2008;24:281–
treatment (treatment receipt) and in daily life (treatment 90.
enactment), also have been proposed as important for inclu- 9. Williams ACC. Facial expression of pain: an evolutionary account.
Behav Brain Sci. 2002;25:439–56.
sion in treatment fidelity assessment [100].
10. Karsdorp PA, Vlaeyen JWS. Active avoidance but not activity pacing
is associated with disability in fibromyalgia. Pain. 2009;147:29–35.
11. Prkachin KM, Schultz IZ, Hughes E. Pain behavior and the devel-
Conclusions opment of pain-related disability: the importance of guarding. Clin J
Pain. 2007;23:270–7.
12. Sullivan MJL, Thibault P, Savard A, et al. The influence of com-
The operant model of pain has considerably influenced munication goals and physical demands on different dimensions of
chronic pain management and stimulated research. Many pain behavior. Pain. 2006;125:270–7.
124 Curr Pain Headache Rep (2012) 16:117–126

13. Leeuw M, Goossens MEJB, Linton SJ, et al. The fear-avoidance 36. Butler S. A personal experience learning from two pain pioneers,
model of musculosceletal pain: current state of scientific evidence. J J. J. Bonica and W. Fordyce: lessons surviving four decades of
Behav Med. 2007;30:77–94. pain practice. Scand J Pain. 2010;1:34–7.
14. Moseley GL, Nicholas MK, Hodges PW. Does anticipation of 37. Morley S, Eccleston C, Williams ACC. Systematic review and
back pain predispose to back trouble? Brain. 2004;127:2339–47. meta-analysis of randomized controlled trials of cognitive behav-
15. Verbunt JA, Seelen HA, Vlaeyen JS, et al. Disuse and decondi- iour therapy for chronic pain in adults, excluding headache. Pain.
tioning in chronic low back pain: concepts and hypotheses on 1999;80:1–13.
contributing mechanisms. Eur J Pain. 2003;7:9–21. 38. van Tulder MW, Ostelo R, Vlaeyen JWS, et al. Behavioral
16. Vlaeyen JWS, Linton SJ. Fear-avoidance and its consequences in treatment for chronic low back pain: a systematic review within
chronic musculosceletal pain: a state of the art. Pain. 2000; the framework of the Cochrane Back Review Group. Spine.
85:317–32. 2000;26:270–81.
17. Skinner BF. Science and human behavior. New York: Macmillan; 39. •• Eccleston C, Williams ACC, Morley S (2009) Psychological
1953. therapies for the management of chronic pain (excluding head-
18. Fordyce WE. Behavioural science and chronic pain. Postgrad ache) in adults. Cochrane Db Syst Rev. 2009;2:CD007407. This
Med J. 1984;60:865–8. article describes the latest Cochrane review on the effectiveness
19. Fordyce WE. Psychological factors in the failed back. Int Disabil of cognitive-behavioral therapy and behavioral therapy for the
Stud. 1988;10:29–32. management of chronic pain.
20. Fordyce WE, Steger JC. Chronic pain. In: Pomerleau OF, Brady 40. Sanders SH. Operant therapy with pain patients: evidence for its
JP, editors. Behavioral medicine: theory and practice. Baltimore: effectiveness. Semin Pain Med. 2003;1:90–8.
The Williams & Wilkins Company; 1979. p. 125–53. 41. Keefe FJ, Gil KM. Behavioral concepts in the analysis of chronic
21. Fowler RS, Fordyce WE, Berni R. Operant conditioning in pain syndromes. J Consult Clin Psychol. 1986;54:776–83.
chronic illness. Am J Nurs. 1969;69:1226–8. 42. Domjan M. The essentials of conditioning and learning. 3rd ed.
22. Jensen MP, Nielson WR, Kerns RD. Toward the development of Belmont: Wadsworth Publishing; 2004.
a motivational model of pain self-management. J Pain. 2003; 43. Sundel M, Sundel SS. Behavior change in the human services.
4:477–92. Behavioral and cognitive principles and applications. 5th ed.
23. Labus JS, Keefe FJ, Jensen MP. Self-reports of pain intensity and Thousand Oaks: Sage Publications, Inc.; 2005.
direct observations of pain behavior: when are they correlated? 44. Vlaeyen JWS, Groenman NH, Thomassen J, et al. A behavioral
Pain. 2003;102:109–24. treatment for sitting and standing intolerance in a patient with
24. Osborne TL, Raichle KA, Jensen MP. Psychologic interventions chronic low back pain. Clin J Pain. 1989;5:233–7.
for chronic pain. Phys Med Rehabil Cli. 2006;17:415–33. 45. Keefe FJ. Behavioral assessment and treatment of chronic pain:
25. Romano JM, Turner JA, Friedman LS, et al. Sequential analysis current status and future directions. J Consult Clin Psychol. 1982;
of chronic pain behaviors and spouse responses. J Consult Clin 50:896–911.
Psychol. 1992;60:777–82. 46. Otis JD. Managing chronic pain: a cognitive-behavioral therapy
26. Newton-John TRO. Solicitousness and chronic pain: critical review. approach. Therapist Guide. Oxford University Press; 2007.
Pain Rev. 2002;9:7–27. 47. Lindström I, Öhlund C, Eek C, et al. The effect of graded activity
27. Linton SJ, Götestam KG. Controlling pain reports through operant on patients with subacute low back pain: a randomized prospec-
conditioning: a laboratory demonstration. Percept Motor Skill. tive clinical study with an operant-conditioning behavioral ap-
1985;60:427–37. proach. Phys Ther. 1992;72:279–90.
28. Lousberg R, Groenman NH, Schmidt AJ, et al. Operant condi- 48. Moss-Morris R, Sharon C, Tobin R, et al. A randomized con-
tioning of the pain experience. Percept Motor Skill. 1996;83:883– trolled graded exercise trial for chronic fatigue syndrome: out-
900. comes and mechanisms of change. J Health Psychol. 2005;
29. Jolliffe CD, Nicholas MK. Verbally reinforcing pain reports: an 10:245–59.
experimental test of the operant model of chronic pain. Pain. 49. White PD, Goldsmith KA, Johnson AL, et al. Comparison of
2004;107:167–75. adaptive pacing therapy, cognitive behaviour therapy, graded exer-
30. Lousberg R, Vuurman E, Lamers T, et al. Pain report and pain- cise therapy, and specialist medical care for chronic fatigue syn-
related evoked potentials operantly conditioned. Clin J Pain. drome (PACE): a randomized trial. Lancet. 2011;377:823–36.
2005;21:262–71. 50. Geraets JJXR, Goossens MEJB, de Groot IJM, et al. Effectiveness
31. Flor H, Knost B, Birbaumer N. The role of operant conditioning of a graded exercise therapy program for patients with chronic
in chronic pain: an experimental investigation. Pain. 2002; shoulder complaints. Aust J Physiother. 2005;51:87–94.
95:111–8. 51. Ostelo RWJG, de Vet HCW, Vlaeyen JWS, et al. Behavioral
32. Hölzl R, Kleinböhl D, Huse E. Implicit operant learning of pain graded activity following first-time lumbar disc surgery: 1-year
sensitization. Pain. 2005;115:12–20. results of a randomized clinical trial. Spine. 2003;28:1757–65.
33. Becker S, Kleinböhl D, Klossika I, et al. Operant conditioning of 52. Lambeek LC, van Mechelen W, Knol DL, et al. Randomized
enhanced pain sensitivity by heat-pain titration. Pain. 2008; controlled trial of integrated care to reduce disability from chronic
140:104–14. low back pain in working and private life. Brit Med J. 2010;340:
34. •• Becker S, Kleinböhl D, Baus D, et al. Operant conditioning of c1035.
perceptual sensitization and habituation is impaired in fibromy- 53. Staal JB, Hlobil H, Twisk JWR, et al. Graded activity for low
algia patients with and without irritable bowel syndrome. Pain. back pain in occupational health care: a randomized, controlled
2011;152:1408–17. In this experiment, intrinsic reinforcers and trial. Ann Intern Med. 2004;140:77–84.
punishers were used for the operant conditioning of nonverbal 54. Anema JR, Steenstra IA, Bongers PM, et al. Multidisciplinary
pain reports, as measured with a behavioral discrimination task. rehabilitation for subacute low back pain: graded activity or
35. •• Kunz M, Rainville P, Lautenbacher S. Operant conditioning of workplace intervention or both? Spine. 2007;32:291–8.
facial displays of pain. Psychosom Med. 2011;73:422–31. To our 55. Veenhof C, Köke AJA, Dekker J, et al. Effectiveness of behav-
knowledge, this is the only experiment that has focused on the ioral graded activity in patients with osteoarthritis of the hip and/
operant conditioning of facial pain expressions. The authors used or knee: a randomized clinical trial. Arthrit Care Res. 2006;
a very delicate methodology. 55:925–34.
Curr Pain Headache Rep (2012) 16:117–126 125

56. Vonk F, Verhagen AP, Twisk JW, et al. Effectiveness of a behaviour 75. Koes BW, van Tulder MW, Ostelo R, et al. Clinical guidelines for the
graded activity program versus conventional exercise for chronic management of low back pain in primary care. Spine. 2001;26:2504–
neck pain patients. Eur J Pain. 2009;13:533–41. 14.
57. Smeets RJ, Severens JL, Beelen S, et al. More is not always 76. Berntzen D, Götestam KG. Effects of on-demand versus fixed-
better: cost-effectiveness analysis of combined, single behavioral interval schedules in the treatment of chronic pain with analgesic
and single physical rehabilitation programs for chronic low back compounds. J Consult Clin Psychol. 1987;55:213–7.
pain. Eur J Pain. 2009;13:71–81. 77. Von Korff M, Merrill JO, Rutter CM, et al. Time-scheduled vs.
58. Smeets RJEM, Vlaeyen JWS, Kester ADM, et al. Reduction of pain-contingent opioid dosing in chronic opioid therapy. Pain.
fear catastrophizing mediates the outcome of both physical and 2011;152:1256–62.
cognitive-behavioral treatment in chronic low back pain. J Pain. 78. Sharp TJ. Chronic pain: a reformulation of the cognitive-behavioural
2006;7:261–71. model. Behav Res Ther. 2001;39:787–800.
59. Leeuw M, Goossens MEJB, van Breukelen GJP, et al. Exposure in 79. Kirsch I, Lynn SJ, Vigorito M, et al. The role of cognition in
vivo versus operant graded activity in chronic low back pain patients: classical and operant conditioning. J Clin Psychol. 2004;60:369–92.
results of a randomized controlled trial. Pain. 2008;138:192–207. 80. Mitchell CJ, De Houwer J, Lovibond PF. The propositional
60. Bailey KM, Carleton RN, Vlaeyen JWS, et al. Treatments nature of human associative learning. Behav Brain Funct. 2009;
addressing pain-related fear and anxiety in patients with chronic 32:183–98.
musculoskeletal pain: a preliminary review. Cogn Behav Ther. 81. McCracken L. A behavioral analysis of pain-related fear responses.
2009;39:46–63. In: Asmundson GJG, Vlaeyen JWS, Crombez G, editors. Under-
61. Vlaeyen JWS, de Jong J, Leeuw M, et al. Fear reduction in standing and treating fear of pain. New York: Oxford University
chronic pain: graded exposure in vivo with behavioral experi- Press; 2004. p. 51–69.
ments. In: Asmundson GJG, Vlaeyen JWS, Crombez G, editors. 82. Pfingsten M, Leibing E, Harter W, et al. Fear-avoidance behavior
Understanding and treating fear of pain. New York: Oxford and anticipation of pain in patients with chronic low back pain: a
University Press; 2004. p. 313–43. randomized controlled study. Pain Med. 2001;2:259–66.
62. Harding VR, Williams ACC. Activities training: integrating be- 83. Lang EV, Hatsiopoulou O, Koch T, et al. Can words hurt?
havioral and cognitive methods with physiotherapy in pain man- Patient–provider interactions during invasive procedures. Pain.
agement. J Occup Rehabil. 1998;8:47–60. 2005;114:303–9.
63. Nielson WR, Jensen MP, Hill ML. An activity pacing scale 84. Dahl J, Wilson KG, Nilsson A. Acceptance and Commitment
for the chronic pain coping inventory: development in a Therapy and the treatment of persons at risk for long-term dis-
sample of patients with fibromyalgia syndrome. Pain. 2001; ability resulting from stress and pain symptoms: a preliminary
89:111–5. randomized trial. Behav Ther. 2004;35:785–801.
64. Rundell SD, Davenport TE. Patient education based on principles of 85. McCracken LM, Vowles KE, Eccleston C. Acceptance-based
cognitive behavioral therapy for a patient with persistent low back treatment for persons with complex, long standing chronic pain:
pain: a case report. J Orthop Sport Phys. 2010;40:494–501. a preliminary analysis of treatment outcome in comparison to a
65. • Gill JR, Brown CA. A structured review of the evidence for waiting phase. Behav Res Ther. 2005;43:1335–46.
pacing as a chronic pain intervention. Eur J Pain. 2009;13:214–6. 86. Hayes SC, Duckworth MP. Acceptance and commitment therapy
This article highlights the gaps in research with respect to pac- and traditional cognitive behavior therapy approaches to pain.
ing, and therefore, the need for a theoretical consensus regarding Cogn Behav Pract. 2006;13:185–7.
its definition. 87. • Cano A, Williams ACC. Social interaction in pain: reinforcing
66. Karsdorp PA, Vlaeyen JWS. On the validity of ‘activity pacing’: pain behaviors or building intimacy? Pain. 2010;149:9–11. The
comment on Jensen “Research on coping with chronic pain. The authors provided an alternative explanation of spousal responses
importance of active avoidance of inappropriate conclusions”. to pain patients’ verbal pain expressions.
Pain. 2009;147:305. 88. Hadjistavropoulos T, Craig KD. A theoretical framework for
67. Vlaeyen JWS, Karsdorp P, Gatzounis R, et al. The PACE trial in understanding self-report and observational measures of pain: a
chronic fatigue syndrome. Lancet. 2011;377:1834. communications model. Behav Res Ther. 2002;40:551–70.
68. Murphy SL, Lyden AK, Smith DM, et al. Effects of a tailored 89. • Hadjistavropoulos T, Craig KD, Duck S, et al. A biopsychosocial
activity pacing intervention on pain and fatigue for adults with formulation of pain communication. Psychol Bull. 2011;in press.
osteoarthritis. Am J Occup Ther. 2010;64:869–76. The authors synthesize elements from different perspectives to cre-
69. The ME Association. Managing my M. E. What people with ME/ ate a broad framework for the understanding of social interactions
CFS and their carers want from the UK’s health and social services. in the context of pain.
2010. Available at http://www.meassociation.co.uk. Accessed 29 90. Thieme K, Flor H, Turk DC. Psychological pain treatment in
Jun 2011. fibromyalgia syndrome: efficacy of operant behavioural and cog-
70. Bailey BP, Konstan JA. On the need for attention-aware systems: nitive behavioural treatments. Arthritis Res Ther. 2006;8:R121.
measuring effects of interruption on task performance, error rate, 91. McCracken L, Turk DC. Behavioral and cognitive-behavioral
and affective state. Comput Hum Behav. 2006;22:685–708. treatment for chronic pain: outcome, predictors of outcome, and
71. Trafton JG, Monk CA. Task interruptions. In: Boehm-Davis treatment process. Spine. 2002;27:2564–73.
DA, editor. Reviews of human factors and ergonomics. Santa 92. Vlaeyen JWS, Morley S. Cognitive-behavioral treatments for
Monica, CA: Human Factors and Ergonomics Society; 2008. p. chronic pain: what works for whom? Clin J Pain. 2005;21:1–8.
111–26. 93. Thieme K, Turk DC, Flor H. Responder criteria for operant and
72. Højsted J, Sjøgren P. Addiction to opioids in chronic pain patients: a cognitive-behavioral treatment for fibromyalgia syndrome. Ar-
literature review. Eur J Pain. 2007;11:490–518. thritis Rheum. 2007;57:830–6.
73. Martell BA, O’Connor PG, Kerns RD, et al. Opioid treatment for 94. Friedberg F. Does graded activity increase activity? A case study
chronic back pain: prevalence, efficacy, and association with of chronic fatigue syndrome. J Behav Ther Exp Psy. 2002;
addiction. Ann Intern Med. 2007;146:116–27. 33:203–15.
74. Turk DC, Swanson KS, Gatchel RJ. Predicting opioid misuse by 95. Kraemer HC, Wilson GT, Fairburn CG, et al. Mediators and
chronic pain patients: a systematic review and literature synthesis. moderators of treatment effects in randomized clinical trials. Arch
Clin J Pain. 2008;24:497–508. Gen Psychiat. 2002;59:877–83.
126 Curr Pain Headache Rep (2012) 16:117–126

96. Thorn BE, Burns JW. Common and specific treatment mecha- 99. Leeuw M, Goossens MEJB, de Vet HCW, et al. The fidelity of
nisms in psychosocial pain interventions: the need for a new treatment delivery can be assessed in treatment outcome studies: a
research agenda. Pain. 2011;152:705–6. successful illustration from behavioral medicine. J Clin Epidemiol.
97. Perepletchikova F, Treat TA, Kazdin AE. Treatment integrity in 2009;62:81–90.
psychotherapy research: analysis of the studies and examination of 100. Bellg AJ, Borrelli B, Resnick B, et al. Enhancing treatment
the associated factors. J Consult Clin Psychol. 2007;75:829–41. fidelity in health behavior change studies: best practices and
98. Waller G. Evidence-based treatment and therapist drift. Behav recommendations from the NIH Behavior Change Consortium.
Res Ther. 2009;47:119–27. Health Psychol. 2004;23:443–51.

View publication stats