The Behavioral Analysis of Pain

FIGURE 1 Schematic criteria for patient selection

It is also difficult medically to assess how much pain or functional limitation can reasonably be expected from a given type or amount of physical findings. Conclusions about that issue usually represent a weighing of judgments about physical findings and assessments of the patient’s age, general health, pain tolerance, and extent of hypochondriacal tendencies—from none to many. The last two, as noted previously, seem more prudently viewed as pain behaviors and not as personality traits. Judgments about them will not be drawn reliably from observations about a kind, location, and amount of tissue damage. Judgments about them based on knowledge of and observations about the patient as a person are helpful. But if those data take one back into the motivational model perspective, they are fraught with imprecision. Predictions as to what a person will do (including how much activity is to be expected in the presence of a given amount of physical findings) based on judgments about what the person has (personality traits) are simply less precise than those based on careful observations of what the person does, that is, behavior consequence relationships. It is therefore sometimes desirable to round out a disease model analysis of a pain problem with an additional behavioral analysis. When there is doubt as to whether functional limitations are appropriate to the physical findings, a behavioral analysis may help by providing negative evidence; that is, the amount of limitation observed seems not to be accounted for in operant terms and may therefore more certainly relate to the body damage factor. This negative evidence may encourage additional disease model evaluative studies. Alternatively, the amount of functional limitation may bear systematic relationships to environmental consequences. The evidence would then suggest that significant degrees of the limitations relate to nonorganic factors. They are largely or partially operant in character.

If a disease model evaluation identifies the problem as predominantly one of respondent pain and points clearly to a course of action, there is rarely any need to proceed with a behavioral analysis.

The next point to be made about when to do a behavioral analysis is that the behavioral analysis has nothing whatever to say about whether there is respondent pain. This is a point of fundamental importance. Establishing what appear to be tight relationships between the occurrence of pain behaviors and what appear to be reinforcing environmental consequences does not by itself indicate that there are not physical findings to be found to account for the pain. A behavioral analysis is not a substitute for a disease model analysis. A behavioral analysis can only indicate the viability of learning factors as an alternative explanation for the pain problem. Even in situations in which pain behavior-reinforcing consequence relationships are close and persuasive, all or part of the pain problem may in fact be a result of some organically-based factor. Behavioral analysis is neither a diagnostic nor a treatment panacea. It should follow, not precede, a disease model analysis.

Situations fitting one of the categories set forth in Figure 1 indicate the potential utility of a behavioral analysis. This assumes that the pain problem has existed for several months or longer or is the reappearance of a previously long-standing pain problem or that the patient has previously displayed lengthy episodes of pain of a roughly similar nature, although perhaps at different body sites. Conditioning requires opportunity for practice. If there has been inadequate opportunity for learning to occur, there is no point in looking for conditioning effects.


The experienced chronic pain patient has undergone many diagnostic evaluations. Understandably, the patient has a set of expectations as to the kinds of questions that will be asked as well as the kinds of non-interview types of examinations that are to be expected. Alterations in the diagnostic process, particularly the introduction of different lines of questioning, may produce or increase patient apprehensiveness or even suspicions. This can raise serious difficulties in the interview process. The scope and nature of this problem needs to be fully understood.

The chronic pain patient is, by definition, a loser in the health care process. If that process had been successful, the pain would not have persisted to its present chronic state. It follows that the patient enters these later or chronic stage evaluations with some mixture of fear, anger, resentment, suspicion, and sheer resignation. The problem is really more severe. As was noted earlier, pain often occurs in relation to causes that are invisible to the eyes of observers. For one to state that he or she cannot perform a task because “my back hurts” may raise few problems in the minds of those around the patient the first time or two. However, if the problem persists and endures beyond prescribed treatment programs, there is increasing probability of growing doubt or skepticism by observers. Sometimes it is only that the patient may fear that others doubt the reality of the pain. In either case, patients often develop concern or suspicion that others question the authenticity and severity of the pain and associated impairment. It is in that context of suspicion and resentment that diagnostic interviews often are perceived by the patient.

The problem is sometimes even more pointed. It was noted earlier that a disease model view of pain encourages the idea that there are essentially two kinds of pain, organic and nonorganic. The former is often accompanied explicitly or implicitly by such qualifiers as “real” and “physical.” The nonorganic type, more often called psychogenic, usually carries such explicitly or implicitly stated qualifiers as “imaginary,” “unreal,” “all in the head,” “hysterical,” or “hypochondriacal.” These two characterizations of pain have accrued a good deal of currency, both by physicians and lay people. Most chronic pain patients who have undergone a number of diagnostic examinations and treatment programs report having received allusions or direct statements from physicians that all or part of the pain problems was psychogenic with all of its socially unattractive connotations. This may truly be the case, or it may be only that patients think it is. The frequency of this feeling in patients is easily confirmed. One may comment to the patient at the outset, “I suppose by now one or more people have said or implied to you that some or all of your pain is imaginary or all in your head.” Such a statement leads virtually without exception to a vividly affirmative comment by the patient.

Another factor influencing how patients view the diagnostic interviewer is that the patient with much operant pain is often a chronic loser in life. Sternbach and associates (1973, p. 136) have described a type of pain patient often encountered who “Because he has little likelihood of future gainful employment … seems to be a loser in the game of life.” Whether the patient’s difficulties focus on employment or on other matters for which avoidance learning helps to maintain pain behaviors, he or she is likely to have had many unsuccessful life ventures and to he apprehensive, suspicious, and excessively guarded in relating interview information. Among other discomforts they may face, there is the threat of loss of the relative haven of illness, with no tenable alternatives yet in sight. If those among this group for whom it is appropriate knew that legitimated retirement was a viable potential alternative, they might feel differently.

The search for valid information from which to judge the nature of chronic pain problems and what to do about them is burdened with these special difficulties. It follows that the gatekeeper to the evaluation process needs to be skillful to cope with the communication problems surely to be encountered.

A behavioral analysis of a pain problem explores alternatives to a disease explanation. The patient may assume that this analysis represents a challenge to the reality or authenticity of the pain. That is unfortunate, as well as untrue. The issue should be met squarely. That proves surprisingly easy to do.

If there are indications in the situation that the patient does have these concerns, as a preliminary to the analysis, it is worthwhile to take the time to point out one’s entering premises to the patient. There are four major points that experience suggests deserve attention. These can be summarized as follows:

   1.  The idea that there are two kinds of pain, (1) organic, physical, or real and (2) nonorganic, psychogenic, or imaginary, is nonsense. Pain is what one feels when one hurts. Of course it is real. The proper question is not, Is it real? The proper question is, What are the factors that influence or control the pain?

   2.  Pain, like virtually every other body process, is subject to the influence of learning or conditioning. A pain problem does not begin because of learning, but it may come to be maintained or controlled by learning factors. When that happens, it still feels the same, but it is now controlled by different factors.

   3.  Learning or conditioning is automatic. If conditions are favorable for learning to occur, it will occur, whether we want it to or not. Learning can neither be turned off nor on. We can only help ourselves to get into or out of situations where learning can occur.

   4.  If what you, the patient, now have is a significant amount of learned or conditioned pain, you are suffering more than you need to. There is something that can be done about it. If learning has occurred, unlearning can also occur.

Those points and some suggested ways of describing and illustrating them are provided in more detail in Chapter 8. A few moments devoted to these issues at the outset of evaluation of a chronic pain patient are a good investment toward patient cooperation and understanding and toward greater precision of information exchanged.

Much of the diagnostic information needed must come in part from the patient and in part from a key family member, usually the spouse. Additional useful and often critically important information will come from measures of patient activity level. The simplest and most direct way to obtain that information is to have the patient complete daily diary records (see Appendix A) before the workup. If diary recording cannot begin until the time of the behavioral analysis, the final conclusions will often need to be delayed until completed diaries have been examined.

Diaries provide an additional advantage. Patients should record directly on the diary forms the kinds, amounts, and times of pain medication intake.

The importance of the spouse accompanying the patient to the workup and of diary data being completed before diagnostic and treatment recommendations can be made should be spelled out to the patient at the first consideration of a behavioral analysis of the pain problem. Experience with a pain clinic (Bonica, 1973) and with operant treatment methods go beyond that. Experience suggests that virtually all chronic pain evaluations would be bolstered and facilitated by the requirement that the spouse be available during the workup process and that 1 to 2 weeks of diary data be completed and returned by the time of first workup. When such a policy is adopted, it is wise to be sure the point is clear to the patient. Patients are not accustomed to having family members participate in the workup. As a consequence, patients are not infrequently casual about arranging for the spouse to come along or to come at another time. The same casualness is sometimes applied to diary recording.

The patient and the spouse should be interviewed separately. A few moments should be taken with the two of them before proceeding to point out that the separate interviews are not a matter of questioning the accuracy or honesty of one against the other. Instead, with a problem as complex and long-standing as chronic pain, it simply makes sense to expect that each of them will be able to see things about the problem that the other has not recognized. They are both so caught up in the problem that there are bound to be some differences in their perspectives and their observations. Evaluation is too important to ignore that and to pass up the opportunity to develop all the information that can be developed to determine how help can be provided.

The issue as to which questions to direct to the patient and which to the spouse is one for which little firm guidance can be provided. It is largely a matter of judgment as to the adequacy, depth, and precision of the information that has been obtained. When in doubt, ask the spouse as well as the patient. One certainly would ask both about the time patterns of the pain, how pain is displayed or communicated, what activities and events systematically do and do not occur (that is, time out), when pain occurs, and what actions are taken to ease or relieve pain. In addition, if an operant treatment program is being considered, both patient and spouse need to be queried about probable and feasible posttreatment activities.

Patient evaluation may in a given case be concerned only with the narrower issue of defining the problem. Alternatively, the evaluation mission also may be to identify possible treatment methods. These two workup missions will be linked here to avoid needless repetition. The evaluator whose charge is only to determine whether the patient’s pain problem could he accounted for in behavioral terms can omit elements of the interview guide, although the two missions’ diagnosis and treatment are not entirely separable. The concern about the interplay between patient and his or her environment inevitably leads one into some of the more treatment-oriented questions and topics. They will, however, be separated as much as seems practical in the behavior analysis guide.

The issues to be determined by a behavioral analysis of a pain problem can he formulated as follows:

   A.  The search for evidence indicating or contraindicating operant pain, that is, systematic and effective

       1.  direct positive reinforcement of pain behavior

       2.  indirect positive reinforcement of pain behavior by avoidance learning

       3.  punishment of or failure to reinforce patient well behavior

   B.  The search for reinforcers holding promise of effectiveness in treatment

   C.  Identification of feasible and accessible posttreatment target behaviors

   D.  Determination of the readiness of family and the patient’s environment to support change by

       1.  reduction in reinforcement of pain behavior

       2.  effective reinforcement of well behavior

The key element in searching for operant pain is systematic pain behavior-environmental consequence sequences (that is, repeated consistently and with time and consequence relationships capable of yielding effective reinforcement).

The next step is an interview guide. The procedure will be to present the key issues or questions, accompanied by explanations designed to relate them to material presented earlier in the book about the development of operant pain (Chapter 3), operant rudiments or behavior change technology (Chapter 4), or identification of treatment goals (Chapter 5).

The items presented are intended as guidelines. There are innumerable ways by which one might explore by interview for given bits of information. Furthermore, the special circumstances of the patient, of the context in which the interview occurs, and information developed within the interview all may lead one to abandon or to expand a given area.

Before commencing the step-wise assessment of issues set forth in the interview guide, it is usually a helpful shortcut to ask the patient to describe a typical day. (“When do you awaken? Do you get up? What happens during the morning and the afternoon? How ordinarily do you spend your evenings? When do you usually go to bed?”) For the patient who stays home and is relatively inactive because of pain, determine also whether he or she dresses, how much time is spent reading or watching TV, and how often he or she goes out of the house.

The foregoing information should permit one to estimate roughly how many hours of each 24 is spent reclining, sitting, or moving about. That information, coupled with the diary data, will provide a useful check point against later items as well as being a convenient shortcut to several questions.

   A.  Time pattern (core issue: What is going on around the patient, or what activities preceded pain behaviors?).

       1.  Episodic pain (example: When do you have pain? Are there periods of several days or weeks free of pain, followed by episodes in which pain occurs almost around the clock?). Some patients will report episodes in which pain is virtually unremitting for several days or longer, followed by intervals of almost total relief. That kind of time pattern rarely bears systematic relationship to the environment unless it can be shown that pain episodes consistently follow certain patient activities or certain environmental events. The housewife who scrubs her floor once a month and then has 4 days of pain is illustrative. If pain seems to occur only when those kinds of physical demands occur, the episodic character of the time pattern is incidental.

             Occasionally pain episodes coincide with the activities of others in the environment. Pain-free intervals may coincide with the spouse being away, as with the husband who travels or the wife who visits relatives.

             Periods of around-the-clock freedom from pain when activity and circumstances in the environment are essentially normal are likely to be indicative of respondent and not operant pain. Systematic pain behavior-environmental consequence relationships are missing.

             Consistent relationships between pain episodes and certain activities or environmental arrangements indicate that operant pain should be considered. More detail will be needed as to what these relationships may be.

       2.  Nocturnal time patterns (example: Is there any particular or consistent time of night when the pain is better or worse? What happens with the pain when you go to bed at night, turn out the lights, and begin to go to sleep?).

             The patient who reports that the reduction in sensory stimulation accompanying lights out leads to increased pain is more likely to have respondent pain because reduced distraction opens the gate to peripheral noxious stimuli or because the reclining body position is specifically productive of pain.

             One exception that occurs occasionally is the long-time pain patient who has evolved a pattern of sleeping or resting all day while the spouse is at work or is quietly supporting the apparent need for full-time rest. Nighttime may, for that patient, be boredom time, an antidote for which is the movement and attention generated by displays of pain. Patients will sleep sooner or later. It is common for pain patients to report that each night is marred by pain to the point where but a few moments or a handful of hours of restless sleep were gained. More detailed exploration with the patient or the spouse often indicates that the amount of nighttime sleeping is underestimated or that it is compensated for by ample daytime sleep. If those alternative rest patterns cannot he identified, it is more likely to be respondent pain.

             Nocturnal awakening (example: How often do you awaken at night because of the pain? How many times do you awaken each night in an average night? How many nights each week are you likely to awaken became of the pain?—each month?).

             The overlapping questions are essential because of the frequency with which the first response is vague or offhand: “Oh, several times every night.” It often comes out that awakening is much less or more frequent than initially implied. Assuming that the patient’s or spouse’s reports indicate frequent or consistent awakening during the night, find out what happens when the patient awakens. Precision is needed here, as in all behavioral analyses. General or vague responses such as, “Oh, I get up and have a smoke,” “I sit up for a few minutes,” “I toss and turn for awhile before going back to sleep,” often omit key systematically occurring events. If one is not satisfied that the question is receiving the kind of detailed thought it deserves, an often helpful remedy is to pose it this way: “If I were there when you awakened, what would I see and hear?”

             There are several things to watch for. The patient who, on awakening, takes private action to relieve the pain (aside from medication, which will be considered in a moment) or to seek distraction from it is more likely to have respondent pain. The patient may quietly steal from the bed and go to another room to read, pace, sit up, or whatever. This arrangement is to be contrasted with the patient whose awakening activities also awaken the spouse, whether deliberately or ostensibly accidentally, and then lead to systematic spouse action. It would be a mistake to assume that spouse action must be obviously beneficial. It may be obvious as in providing a back rub, bringing the heat pad, bringing medications, or simply providing reassurance or expressions of concern. In some situations, the sheer act of awakening and disturbing the spouse has more subtle reinforcing effects for the patient. To awaken the spouse may, for example, serve to reassure the patient that the spouse continues to be aware of the extent and duration of the patient’s suffering. There may even be sadistic pleasure at harassing or disturbing the spouse, particularly if the patient is resentful at being impaired when the spouse sails through life unburdened by chronic pain. The key thing is systematic spouse action, whatever its form. When there is systematic spouse action, there is always the possibility that it represents pain-contingent social reinforcement as well as perhaps direct palliative relief.

             Frequent nocturnal awakening may only mean that something else awakened the patient and, in retrospect, the patient incorrectly attributes awakening to pain. The interestingly frequent illustration of this is the patient who, on awakening “from pain,” always goes to the bathroom to empty his or her bladder. When, under further inquiry, it becomes evident there is consistent and reasonably frequent bladder urgency, day or night, the alleged nocturnal pain is sometimes an innocent mislabeling. Perhaps it was only that, having awakened because of bladder urgency, the patient was then reminded of and experienced the pain. Under those circumstances, that part of the pain pattern may he said to have come under control of other factors. Nocturnal awakening by itself does not support the inference of respondent pain.

             Nocturnal medications. Another issue to explore in regard to consistent nocturnal awakening is pain medications. A two-step approach to this issue seems to work better. First determine the frequency and time-spacing of the awakening (example: How soon after going to sleep do you awaken? How long is it until the next time you awaken? Are there fairly consistent times at which you awaken or are likely to be awake, such as 1 AM or 4 AM?). Once it has been established that the patient is often awake at certain approximate times, ask if medications are taken at those times. Sensitivity must be exercised here, for the patient may be apprehensive or guilty about the amount of medication being consumed. The responses one gets are potentially burdened with patient concerns about how the interviewer will react. This is also one of the topics about which spouse observations may prove particularly helpful. If it can be shown that the patient takes medications in a consistent time pattern (for example, every 3 to 4 hours), addiction or habituation must certainly be considered. But more than that, the nocturnal awakening from pain may not be that at all. It may be that the systematic consequence to awakening is the reinforcing effects of medication. Under those conditions, the pain behavior of awakening and taking medication may be substantially or even totally under control of addiction or habit. There wi1l be more about that later.

       3.  Diurnal time patterns (example: Is there any particular time of day when pain is likely to be better or worse? What is the pain like when you first awaken in the morning and get out of bed? What is the trend of that across the day? Does pain tend to get better or worse as the day goes along?).

             Pain that is specific to activity or movement can often be expected to occur or increase in daytime and diminish or disappear at bedtime. It follows that a pattern of pain only in the daytime is not necessarily operant.

             Daytime pain time patterns do not often have significance except when they point to specific pain-activity or pain-social reinforcement relationships. Those relationships will be explored more directly further along. Examination of the trend of the pain across daytime hours usually quickly indicates whether it is activity related. Unremitting pain or lack of variation when the day’s activities vary suggests respondent pain. A trend toward lessening pain as the day wears on also argues against rest or time out from activities as sustaining operant pain.

             Systematic time periodicity to increases of pain should be checked. If, at the time of the increases, the patient takes medication, the problem may be addiction or habituation, which, in turn, may indicate operant pain and a minimal respondent component.

             To sum up the time issue, without systematic pain behavior-environmental consequence relationships, there can be no operant pain. Time patterns, that indicate such relationships may reflect operant pain, respondent pain, or a mixture of the two.

   B.  Pain behavior (core issue: How do those around the patient know when pain is experienced or is severe?). If the patient’s pain and associated functional limitations remain a totally private matter, systematic social reinforcement cannot occur. Although it seems a straightforward matter, it is often difficult to obtain precise information from the patient about this point (example: How do people around you know when you are having pain? How do they know when the pain is severe? How do people know what activities you cannot do or must limit because of your pain?).

          Perhaps the most common initial response to this topic is, “I don’t let people know I am hurting!” or “I don’t say a word. I keep it to myself.” Such a response is of course more attractive because of its higher social desirability. In addition, however, people who have long been experiencing pain often will have developed ways that may seem to mask the occurrence of pain but in fact signal to others that it is being felt. These signals are often subtle. Even the spouse may no longer be aware that they are occurring or that they successfully elicit systematic spouse responses. This issue becomes of major import in the matter of spouse retraining, a topic of Chapter 12. It is also essential now to pinpoint whether there is ample opportunity for people in the patient’s environment to make systematic responses. They must receive cues to do so. It is often helpful to ask, “If I were in your living room when you were present and were experiencing pain (or severe pain), what would I see you do and hear you say?” The pain behavior may be nothing more than not doing something else. For example, the patient may recline on the sofa when, were there no pain problem, some other activity would be underway at that time. The key element is that those around the patient can reliably discriminate pain-free states from those in which pain is experienced or anticipated.

          Pinpointing pain behaviors does not bear directly on the question of how much is respondent and how much is operant. The information helps to identify the consequences, social and otherwise, to those pain displays, as will be explored next. Whether those consequences are systematic and reinforcing remains to be determined.

   C.  Environmental responses (core issue: How do those around the patient respond when pain is signaled?).

       1.  Direct reinforcement of pain behavior. Patients commonly interpret this question to be, “Do others help ‘baby’ or protect me?” Occasionally the question is interpreted as aimed at determining whether the spouse is loyal in the face of adversity or is fed up with the pain problem. The task is made easier by clarifying that the information being sought is description and not judgment (example: What I’m after here is not whether what others do is good or bad, helpful or harmful, or anything like that. It is just that when a patient problem has lasted a long time, people often develop ways of behaving or responding to the pain of which they no longer are aware. Sometimes knowing what those are can help—can make a difference).

             The most common first response is, “They don’t do anything.” Taken literally, people cannot do nothing. The point should be pursued further (example: If I were there when your husband/wife came into the room and could see that the pain was bad, what would happen? What would I see and hear him/her do? What would you do next?).

             The major thing to watch for here is consistency. The content or quality of the spouse behavior when pain is signaled is a second-order question. Data from the patient and, separately, from the spouse indicating much variability or almost random behavior precludes effective social reinforcement for pain behavior or punishment of efforts toward activity or well behavior. On the other hand, consistency of spouse response to pain behavior increases the possibility of social reinforcement as a significant factor.

             It can hardly be overemphasized that the content or quality of the spouse response may have systematic effects on the patient, regardless of whether the interviewer thinks that spouse behavior is reinforcing, aversive, or whatever.

             There are some obvious and direct things to watch for. The spouse may be immediately solicitous or helpful; he or she may reassure or console, offer medications, or step in to take over a pain-producing task. The reinforcement may be more subtle. The spouse may consistently change the subject (“I am trying to divert his/her attention—get his/her mind on other things”), may turn toward (or away from) the patient, without saying anything, or may become angry or disgusted and express that feeling openly. All but the last named clearly can have positive reinforcing value. Negatively toned spouse affect in response to pain displays is also capable in several ways of being reinforcing. The negative affect may be followed consistently by helpful behaviors. The angry response may reassure the patient that the spouse is still aware of the pain problem. The patient may delight in bedeviling the spouse. The possibilities of the subtleties are endless.

       2.  Discouraging activity (punishing of well behavior). After direct reinforcement of pain behavior, the punishment of well behavior must next be considered (example: When you try to do things and your husband/wife knows you are having pain or expects that you will have pain if you continue with what you are doing, what does he/she say and do?).

             Discouragement of activity may await a cue from the patient that pain is present and thus be a response to pain behavior. Matters may have progressed to the point where the spouse asserts protective action in anticipation of pain. Seeing the patient begin some action, or even simply signaling that action is being considered, may be enough to elicit spouse response or intervention. That response may be to step in to take over the contemplated task or in some other fashion to discourage the patient from proceeding. Routine assumption of tasks by the spouse, without a pain behavior display as a cue or reminder, will be dealt with later. At the moment, the concern is with direct responses to pain behavior.

             The content or emotional tone of spouse response is of secondary importance. If it is systematic, it may he an effective punisher or inhibitor of activity, whether presented solicitously or with bitter overtones.

             Discrepancies between patient and spouse are common in response to those areas. That is to be expected. Interchanges between them about pain have been going on long enough to lead inevitably to some routinized responses about which one or both parties may have become unaware. Conflicting accounts must be weighed. It becomes a matter of clinical judgment as to where the greater accuracy rests.

             In this section the spouse has been the target and not other people around the patient. One occasionally finds situations in which others in addition to, or instead of, the spouse play key roles. That is uncommon. The exceptions are mainly to be found relating to adolescent children in the household or to parents of the patient. There are always potential exceptions. But experience indicates that neighbors, coworkers, and relatives or immediate family members living outside the household almost always play, at most, secondary and noncrucial roles. Contact is too infrequent for them to be sources of effective social reinforcement for pain behavior or for discouragement of activity and well behavior. They may play a role, but it is rarely enough to sustain the pain problem if the spouse is not also contributing.

             The single patient is, of course, another exception. It is usually a simple matter to judge whether there is someone in the picture with whom the patient has sufficient contact to derive sustaining social reinforcement for operant pain. When such a person emerges in the picture, he or she should be called in for an interview in the same fashion as is done routinely in the case of a spouse.

   D.  Pain activators (core issue: What increases pain or causes it to occur? There is a closely related issue: What is the reinforcement value or importance to the patient of the activities that bring on or aggravate pain?).

          At this point the concern is with kinds of movement or physical demands that may increase pain. Later these will be studied again in larger units that combine these activators. “Taking care of the housework” or “going to work,” for example, represent clusters of bending, lifting, and walking. The issues here have complex interrelationships. One way of posing the central question is to explore the relative costs of pain versus activities. If onset or increase in experienced pain results in termination of a high-value, high-strength activity for the patient, it might be said that the cost of pain is high. Cost here is used in a psychological sense, although it might be economic as well. The patient must, because of pain, limit or forego reinforcers that ordinarily accrue to the relinquished activity. His or her pain costs that much. Furthermore, continuing to engage in the activity for a distinctive period of time or to a significant amount, after onset or increase of pain, supports the idea that the pain costs a great deal. The patient is wi1Jing to suffer to gain the reinforcers relating to the activity. The cost-benefit ratio of pain to activity is high. Stated another way, the payoff or reinforcement for pain behavior is low.

          A quick termination or militant avoidance of the activity when pain begins or increases suggests that pain behavior (in this instance, stopping to rest or avoiding a particular activity) is relatively more reinforcing. Logically, there are two possibilities here. One is that the pain is so severe or uncomfortable that the patient is willing to undergo the reinforcement deprivation resulting from limited or avoided activity to gain relief from the pain. The second is that the amount of reinforcement from the activities is not great. In the latter case, rest and time out readily yield equal or greater reinforcement than do the activities. The pain-activity cost-benefit ratio may be said to be low.

          One must determine what activities increase or aggravate pain. The relationships between amount performed and stopping or limiting one’s effort must also be determined to judge whether contingent reinforcement relationships exist. Finally, one needs to determine the potential reinforcing value (high or low, positive or aversive) of the activities.

       1.  Roster of activators (example: What are the things that produce or increase pain? What makes your pain worse?).

             Care again should be exercised to gain accurate detail. General comments (“Any kind of movement,” “Any kind of bending or lifting”) should be elaborated. The housewife can be asked to indicate a typical list of household chores that produce or increase her pain. The wage earner can develop a similar set of details. The list need not be exhaustive. It should be broad enough to help later in pinning down what activities are indeed avoided.

             Patients occasionally will report that the pain keeps them from doing so many things that they no longer know what aggravates the pain. They can be asked to recall which activities previously engaged in brought on the pain. At the other extreme is the patient who steadfastly claims that all activities increase pain. He or she should he asked to discriminate a few that produce greater amounts of pain or more rapidly or consistently aggravate pain.

             Patients often fail to distinguish among pain, weakness, and fatigue. An activity, limited in relation to the pain problem, may now be curtailed because of weakness or fatigue resulting from infrequent or insufficient exercise. Or perhaps it is limited as a by-product of heavy medication intake and associated lethargy or discoordination. The error works both ways. The patient may state that an activity is curtailed because to continue with it is to produce an increase in pain, when in fact it is weakness and fatigue that now are experienced. Alternatively, the patient may have limited an activity for a lengthy period because to do otherwise would have resulted in significant aggravation of pain. Long periods of disuse lead to ease of fatigue. Subsequently, when the activity is attempted, fatigue or weakness are experienced and reported by the patient, and he or she must stop. For the purposes of behavioral analysis, the distinction between pain, on the one hand, and weakness or fatigue, on the other, are usually unimportant. It is the systematic relationship between doing or not doing (behavior) and rest or other consequences (contingent reinforcement) that matters. It follows that questions about activity or nonactivity can be posed such that pain, weakness, and fatigue are essential equivalents.

             Once a representative roster of pain activators has been identified, these should be sampled to learn how soon pain follows. Determine how much is done before pain begins or before the patient stops or modifies the activity because of pain, weakness, or fatigue.

       2.  Activity-pain time interval (example: When you do an activity leading to pain, how long is it from when you begin until the pain begins? When do you stop—when the pain begins? before it begins? If you stop before the pain begins but the pain comes later, how long is it after you stop before the pain begins?).

             Pain that begins many minutes or even many hours after an activity has terminated is not likely to be reinforced by rest, by time out from the activity, or by contingent social consequences. Rest and attention both occur too late to be considered effective reinforcers. The patient who no longer engages in an activity reportedly because of pain it produces, although delayed by several minutes or hours, may have time out or avoidance learning sustaining pain behavior. Always assuming that medical evaluation has raised doubt about whether physical findings account for the functional limitations reported, the question of whether this represents avoidance learning can best be judged by assessing the reinforcing value of the activity itself, apart from the issue of pain.

             If the patient persists in engaging in the activity, when to do so surely produces delayed pain, it is almost certainly respondent pain. There is no time-out reinforcement.

             Enquire about how much of each of a sample of the pain activators ordinarily is attempted before halting, resting briefly, or significantly restricting how the task is performed. One can then weigh the relative costs of pain behavior versus rest or time out. If little is attempted and it is reported that pain, weakness, or fatigue quickly lead to termination, the distress is highly aversive, the activity is not reinforcing, or both. The patient who imposes pervasive functional limitations on himself or herself such that only limited amounts of activity are attempted, providing a clear basis for this from the preceding medical evaluation, may well have much operant pain. The prime exception to this is the patient for whom this much inactivity is a striking contrast to premorbid behavior. Such a patient as that shows by history that activity is highly reinforcing.

             Particular attention is needed here in regard to patient-spouse and patient verbal report-diary data discrepancies. The spouse may not observe how much the patient does but may only infer it. Patients and spouses sometimes form word habits describing much or little activity when in truth what they are describing is what used to be the case but is no longer true.

             To this point, in regard to pain activators, information has been developed on what they are, on the time relationships between activity and reported pain, and on amounts of the activity leading to significant pain or functional impairment. The matter of the reinforcement value (high or low) of those activators remains and will be explored later.

   E.  Pain diminishers (core issue: What actions or events, in addition to responses of others, are pain contingent?).

          The objectives here are specific. The responses of others to pain behaviors have been explored. How much they reinforce pain behaviors or discourage well behaviors should be clear. If not, that matter can be reopened here. In addition, however, it is important to determine whether rest is a reinforcer and is pain contingent (example: When the pain becomes bad, what can you do to ease or relieve it? What helps?).

       1.  Rest as a pain diminisher. If the response indicates rest (reclining, sitting, some form of taking it easy) helps, the time relationships should be identified. This was accomplished in part in the immediately preceding section (D.2) when it was determined how soon rest begins after the start or increase of pain. The information to be added now is the length of time it takes for rest to have a beneficial effect. Rest taking many hours to be felt as significantly easing the distress is less likely to maintain operant pain. Furthermore, if the patient arises after a period of rest and resumes the activity that led to pain, the cost of rest must be nearly as high as the cost of pain. Rest in that pattern is not likely to sustain operant pain, although it still might relieve respondent pain. Finally, rest used only episodically, to help control pain, is not a systematic reinforcer.

       2.  Medication as a pain diminisher (example: How many times each day do you take pain medication? About how many hours between doses? ‘What do you take? How much difference is there between amounts of daytime and nighttime medication?).

             It is often difficult to gain sound data about pain medication intake from chronic pain patients. Errors come from many sources. Some patients understate what is taken out of concern that they will be criticized or to disguise secret or secondary sources of medications that, if revealed, would risk being lost. Other patients err by overstating what they consume. This phenomenon seems like the heavy smoker who always carries a spare pack “just in case.” Some patients taking much medication overstate current consumption in an effort to provide a safety factor by seeking prescription of an ample supply. Other patients are toxic or sedated to the point that they are incapable of accurate recall. Yet others routinely substitute large amounts of ethanol (alcohol) as a supplemental source of sedation and so, although honestly reporting medication intake, seriously underestimate the amount of sedation they consume. In regard to interview sources, clinical judgment from close questioning of patient and spouse and cross comparisons with diary data and medical records of prescriptions are about all that can be done. As will be noted in the section on treatment, direct inpatient observation is sometimes necessary.

             The major issue here concerns habituation or addiction or systematic, pain-contingent sedation. Medications taken in a consistent time pattern (for example, each 3 to 4 hours), day and night, make habituation or addiction a strong possibility. That in turn is a strong indication that pain behaviors may be under control of the medication habit rather than physical causes. Other sources of relief encountered with some frequency, such as heat pads, massage, a vibrator, or hot baths do not often maintain operant pain except as they also entail selective and pain-contingent attention or time out from aversive activities. The visible use of these other sources of pain reduction may better be thought of as pain behaviors in their own right. They serve as cues from the patient to others that there is a pain problem for which special help or protection may needed. The persistent use of a cane, long after it no longer has ample respondent basis, illustrates the point.

   F.  Tension-relaxation (core issue: Does tension increase and relaxation decrease pain, or does the opposite occur?).

          This subject is better dealt with separately. There are many reasons why tension and emotional distress might increase pain and why relaxation or escape from tension might decrease it. An opposite picture occasionally emerges: tension decreases and relaxation increases pain. Increased pain associated with emotional or muscle tension can be a purely respondent phenomenon. However, if rest eases the pain and permits the patient to avoid the source of emotional distress, the conditions for avoidance learning exist. The pain behaviors may yield time out from situational or emotional problems because they require the patient to rest to gain relief.

          The restless, driven, life-is-a-racing treadmill patient with chronic pain may yield the opposite picture. For such a person, rest and idleness may be threatening. Activity is a critically important source of reinforcement. When a pain problem occurs and a lightening of the workload appears desirable, the patient may work against himself or herself. Prescribed rest creates distress, and so the patient prematurely or excessively resumes activity, thereby worsening the pain problem. This is what might be called operant abuse of a respondent pain problem. The central behavioral problem now is to help the patient to reduce activity, not increase it.

          If relaxation, like rest, fails to ease pain, there is less likelihood that it is an operant pain problem. If relaxation does ease pain, it may still be a respondent pain problem. In this case, relaxation training, for example, biofeedback, training in relaxation or self-hypnosis, may prove helpful. Alternatively, relaxation that eases pain should also be considered as a potential form of rest or time out and judged in that light.

          The behavioral analysis has now completed exploring whether pain behavior receives direct positive reinforcement from attention, rest, or medication. The question of whether functional impairment was a direct result of zealous compliance with physician prescription will also have been clarified. The question of the direct punishment or discouragement of activity or daily chores (well behavior) has also been explored. The groundwork has been laid for exploring the third possibility, indirect positive reinforcement of pain behavior through time out or avoidance learning. The roster of pain activators and pain diminishers serves as a foundation for exploring what activities the patient no longer engages in and their positive or negative reinforcing properties.

          There are other problems the patient may have alone, in marriage, or in other vocational or social ways. Everyone has problems. Establishing that the patient has problems or displays what some may characterize as a neurosis, deviant personality, interpersonal difficulties, or whatever, is not enough. For these problems to influence the pain problem or have functional relations to it, it must be shown that there are contingent relationships. Pain behavior must lead to time out from or to reduction of the other problems. If those pain-contingent relationships do not exist, the other problems are unlikely to be relevant to the pain problem.

          There are many forms in which the timeout or indirect reinforcement effect may exist. They usually can ultimately be reduced to some combination of these: the patient now does something because of the pain problem that was not done before, quits doing something previously done, the spouse now does something not done before, or the spouse quits doing something previously done. In each of the four possibilities there is a pain-contingent relationship. What is (or is not) done might be anything from taking tranquilizers to changing jobs or recreational pursuits.

          The material covered to this point also tells the extent to which rest, attention, and medication, the three naturally occurring and fairly readily programmed consequences in treatment settings, are likely to be effective reinforcers, should a treatment program begin. Any of them which has been sustaining a patient’s operant pain is, by definition, a high-strength behavior. As such, it can, if properly reprogrammed, serve to strengthen well behavior.

   G.  Changes in activity as a result of pain (core issue: What changes in the patient’s and spouse’s ways of life have occurred, and are those changes costly or reinforcing?).

          Distinguish vocational from leisure pursuits and explore them separately. They often show different amounts of change and may have widely different reinforcement characteristics. Leisure pursuits tend to reflect what a person likes or wants to do. That is another way of stating that leisure is reinforcing or provides access to yet other effective reinforcement. Vocational pursuits provide access to reinforcers, but sometimes the reinforcement is only that it makes possible desired leisure time pursuits. For these reasons information about a person’s leisure time activities tends to shed more precise light on what is reinforcing for that patient. It is important also to determine the premorbid frequency or baseline of the activities so that change can better be assessed. Patients sometimes report they no longer do something because of the pain that they had not done for years anyway.

          The question of how much of an activity was done before onset of the pain problem or before reduction of the activity because of pain has an even more important implication. Behavior that has been occurring at a high rate has, by definition, been receiving some form of effective reinforcement. Once one knows what behaviors have been reduced in rate, it becomes possible to assess the cost in reinforcement deprivation from the pain problem for the patient and the spouse. Loss of an activity usually also means loss of the reinforcers that sustained it. That is the case, of course, unless it is also shown that the reinforcers previously sustaining premorbid behaviors continue to be available. In that event there is little or no deprivation of reinforcers.

          The sustaining reinforcement for leisure activities usually will become evident from the nature of the activities. Leisure activities tend to be reinforcing in their own right or because the spouse or other family members richly reinforce them. It is, for example, usually easy to determine what the reinforcers are for a housewife who reports that she used to do much gardening before her back pain. She may have been reinforced by the exercise and sunshine, the pride or esthetic values of a pretty garden, either in her own eyes or those of family and neighbors, or by the access that gardening provided to a garden club in which she could socialize with others.

          The key reinforcers sustaining vocational efforts are not always so apparent. It is well recognized that what it is about a job that provides the payoff to the worker is often a complex mixture of many factors. To name a few—money, career enhancement, status, exercise, sense of accomplishment, opportunity to interact with (or to not have to interact with) people, opportunity to be indoors out of the cold or outdoors in the fresh air—all play a role. Some sense of what these reinforcers have been should be developed to determine how much of a deprivation now exists. There should also be exploration of whether significant changes in the reinforcing characteristics of the job occurred at or shortly before onset of the pain problem. Work or leisure that has become less reinforcing or less accessible may change from attractive to aversive, and avoidance learning may follow.

          The issue here benefits from being approached in two different ways. One is to enquire about what was done but now occurs on a reduced and perhaps zero frequency. The second is to enquire what would be done in the future if there were no significant limitation because of pain. Contrasts between the two approaches are common. Those discrepancies help to identify what the patient and spouse expect and want and what they will reinforce, which are vital issues in setting treatment goals and posttreatment target behaviors. The slightly different approach reflected in asking what one would do in the future sometimes leads patients to describe lucidly and spontaneously what is reinforcing and what is not. The wage earner or the housewife often will respond to this question by elaborating many pleasant activities but make no mention of work or housework. These omissions often are made after earlier assertions about the pleasures of work. They sometimes reveal more about what is truly reinforcing than the patient expects.

       1.  Patient activity modifications (example: What are the things you no longer do because of the pain? About how many times a day, week, month—as indicated—did you do those things?).

             The patient who reports that he or she is continuing to go to work or to do the expected daily chores should also be queried as to whether there is any change in what is done on the job and whether others have taken over parts of the load. This latter point also may lead to further description of how others respond to displays of pain behavior.

             Next, explore what the reinforcers were for the job (example: Did you enjoy the job? What was it about the work that kept you at it?). When all the data are in, it is usually a simple process to judge how much those reinforcers continue to be available in spite of the pain. If they are, there is no deprivation, no cost, and no avoidance of aversive conditions through pain behavior. Operant pain is then not being supported in this fashion.

             Next, be sure there has not been a significant change in the work or leisure situation, independently of the pain problem. Several wives who were pain patients have described situations in which social and leisure pursuits were markedly disrupted by husband transfers to other cities. Old friends (and social reinforcers) were no longer available. The patient who had limited skill at forming new friendships then suffered a major deprivation. In similar vein, it is not uncommon for wage earners to report significant job changes at about the time of or before the onset of pain. A job that previously was reinforcing became less so and perhaps even aversive. If there has been no significant change in vocational or leisure activities because of the pain, there is no opportunity for time-out effects. The likelihood of operant pain being supported in that fashion is accordingly diminished.

             Where changes in activities have occurred, the challenge is to judge their net cost in reinforcement to the patient. This task is usually not difficult so long as it is kept in mind that reinforcers are much more than money and that leisure activities are an easier way to determine the effective reinforcers for a person. Pain that significantly changes leisure pursuits is indeed almost always costly to the patient unless it can be shown there continues to be effective reinforcement from the altered leisure pursuits.

       2.  Pain-free activity plans (example: If you had no significant pain problem, what would you do that you do not now do?).

             This point usually can be covered briefly because of the great overlap with the preceding item. The additional information developed here, if there is any, sometimes reveals that the patient or spouse is content with the status quo. No activity change is anticipated or desired. If that is the case, spouse responses have limited relevance to the pain problem. Alternatively, the pain problem provides one or both of them with considerable reinforcement.

             Responses to these items should be checked to determine whether they reflect realistic assessments of the future. It is sometimes easy to confuse what one would like to do with what one is likely to do. The central point of the activity modification items is to determine what has been given up or reduced because of pain and would be resumed were the pain problem eased.

       3.  Spouse activity modifications (to the patient: “What are the things your husband/wife used to do but no longer does because of your pain problem? About how many times a day/week/month were they done?” The same question is then asked of the spouse, both about the patient and about the spouse).

             The issues here are as in G.2: What are the costs in reinforcement to the patient for pain behavior? These items amplify data obtained directly from the patient. More importantly, however, they also explore that vital source of reinforcement and punishment or non reinforcement for a patient, the spouse.

             No significant change in spouse activity because of the patient’s pain problem usually means that the spouse is not indirectly reinforcing pain behavior or indirectly discouraging well behavior—direct reinforcement, perhaps, as explored in item C, but not indirect. The spouse is not gaining time out from some aversive activity because of the patient’s functional impairments. The spouse is also not undergoing a deprivation of reinforcement from loss of access to a previously reinforcing activity.

             The importance of this issue should not be overlooked. Clinically, one finds repeated examples of a spouse systematically reinforcing pain behavior or discouraging well behavior because the pain problem has produced reinforcing activity changes for the spouse. An earlier example described a husband with secondary impotence who reinforced his wife’s pain behaviors with enormous effectiveness. Maintaining her pain problem yielded him time out from the now threatening activity of intercourse. He was apparently unaware of what he was doing, but that is beside the point. One need not be aware of a contingent reinforcement relationship for it to be effective, and awareness does not by itself reduce effectiveness.

             In another case, the solicitous and now lonely mother of a 19-year-old son actively reinforced his pain behaviors and sabotaged treatment programs that threatened her with his emancipation. In both illustrations, patient pain behavior led to the successful avoidance of aversive events in and for another person, not the patient.

             Evidence that a spouse gains indirect reinforcement from pain behavior increases the likelihood of operant pain. It also means that changes in that situation will be essential for treatment to succeed. Evidence that a spouse suffers a loss of reinforcement as a consequence of the pain problem argues against the spouse contributing to operant pain. That evidence also indicates that the spouse is more likely to be an effective reinforcer of increased patient activity level although he or she may need hell> in learning how to do that. Treatment is likely to move ahead more easily.

       4.  Spouse pain-free activity plans. This item should require no further elaboration. It is the twin to G.2, the patient’s pain-free plans.

       5.  Sexual intercourse (example: What changes in your sexual life have been caused by the pain problem? How often each week do you have intercourse? How often did you before the pain problem began?).

             Sexual problems occur with sufficient frequency in relationship to chronic pain to warrant separate treatment. The central issue here is to determine indirect reinforcement of pain behavior or indirect punishment of well behavior. Changes in frequency or quality of sexual activity in a marriage probably—although not certainly—have major reinforcing properties. Therefore, when changes in sexual activity result from a pain problem, whom does it cost and whom does it reinforce? Time out from intercourse is not likely to be a reinforcer helping to maintain operant pain if both patient and spouse experience sexual deprivation. But if one of the marital pair experiences intercourse as aversive, the opportunity for strong reinforcement exists, and the matter should be explored.

             Sex-related questions presented as but one of a series of enquiries regarding activities potentially influenced by pain do not often raise interview problems. Information about the frequency of intercourse, before the pain problem and presently, should be elicited from patient and from the spouse.

             Questions about frequency of intercourse can be followed with an enquiry regarding the quality of the relationship. (“How would you describe the sexual adjustment in your marriage?” “Is there much difference between you and your husband/wife in how often intercourse is desired and in how much it is enjoyed?”) This brief enquiry, when directed to both the patient and the spouse, appears sufficient in most cases to provide a judgment as to how likely intercourse is either aversive or reinforcing to one or both members of the marriage.

             When major problems emerge in regard to sexual functioning, there remains the question of what to do about them. Clearly, when sexual dysfunction appears to help maintain a pain problem, the sexual problems need to be dealt with if treatment gains are to be maintained. Reduced sexual activity is sometimes related to side effects of pain medications. Reducing medication intake usually handles that problem. Other sexual problems may not be so straightforward. They may represent gaps in effective well behavior for which no solutions are in sight. If that is the case, they may prejudice treatment success. Severe impotence, frigidity, or disabling problems of sex-role identity may be examples.

Earlier in this chapter, four potential objectives of the behavioral analysis were noted: evidence for or against operant pain, the search for reinforcers helpful in treatment, identification of posttreatment target behaviors, and determination of family readiness to support treatment and behavior change. The data obtained from this kind of interview, often supplemented by such additional explorations as will be set forth immediately following, provide the basis for judging the extent to which learning or conditioning can account for the pain behaviors and associated functional impairments. The data also provide much, although not all, of the essential information for the other three objectives. Those, too, will be spelled out in more detail in the following sections of this chapter.


It will be evident that parts of the interview guide presented above are also found in social history or psychiatric diagnostic interviews. Data from those sources should be drawn on when available. They are not, however, adequate substitutes for a behavioral analysis in most cases because of differences in focus and because they do not explore certain key areas. There are additional behavioral samples that can prove useful in assessing the probable reinforcing or nonreinforcing characteristics of pain or functional impairment to the patient. These will be dealt with briefly.

Diagnostic Evaluation of Diary Data

Diary data are simple, inexpensive, and useful. It is a tactical error not to make routine use of diary data in the evaluation of chronic pain. Analysis of the diary data is easy.

The patient is instructed to record medication intake on the diary form: what, when, and how much. That information may not be precise nor complete and so should be viewed cautiously. A quick scanning of one or two weeks of diary forms should suggest the approximate frequency of medication intake and the rough average of hours between doses. Consistent time spacing, unless on a prescribed basis, suggests addiction or habituation and warrants more careful examination. Random and infrequent intake may reflect lack of a medication problem or lack of judicious recording.

The patient will also have been instructed to record (see Appendix A) how much of each day is spent sitting, standing or walking, and reclining and when during the day each of those occurs. At the bottom of each daily sheet is space for totaling the hours in each of the three categories. The combination of sitting and standing, known as uptime, and rest provides a quick index of activity level. A weekly total uptime normally is approximately 112 hours (7 days x 16 hours). A quick totaling of uptime for a 1- or 2-week diary sample should be made. Weekly uptime totals of less than approximately 80 hours indicate that the patient has considerable functional impairment, finding rest or time out effective reinforcers. Patients with more than approximately 80 hours of uptime per week are much less likely to find rest reinforcing.

Occasionally uptime figures are distorted because the patient spends many hours sitting. That can be checked by noting the sitting and standing or walking totals or, even more easily, simply by scanning morning and afternoon hours to see how much is entered in the sitting column. Patients who sit much of each morning or afternoon without compensatory pacing at night are also likely to find rest reinforcing and to experience time-out benefit from pain behavior.

Psychological Tests

Some but not all psychological tests have the advantage that information may be obtained in minimal professional time. Moreover, some of these tests have yielded considerable empirical data bearing on issues pertinent to the behavioral analysis of chronic pain. It is too much to expect that there be a comprehensive review here of an of those data. In Chapter 2 reference was made to the MMPI and to data drawn from it in describing a variety of patient groups. That test is the most commonly used instrument in the context of pain and so deserves some attention here (Butcher, 1969; Dahlstrom and co-workers, 1972; Mischel, 1968; Welsh and Dahlstrom, 1956). It will be the only so-called personality test discussed.

The MMPI can be used with English-reading, American-educated people who are age 16 or over and above mental retardation levels of intellectual functioning. The test requires but a few moments to explain to the one who is to take it. That explanation can be done satisfactorily by an appropriately oriented clerical person, office nurse, or other health care professional. The test can be scored by a clerical person (with practice in 20 to 30 minutes), or it can be mailed to scoring services for processing for a fee. It yields scores on a series of scales. The most commonly used scales (three validity and ten clinical) are reported on a graph or test profile. The MMPI processing services to which answer sheets may be sent also usually report a narrative test commentary, which is a printout from computer-based experience data. Those printouts vary in their relevance to questions about chronic pain.

The relationship of personality and personality tests to behavioral perspectives needs a brief consideration. The distinction has been drawn repeatedly in this book between what a person has (his or her personality) and what he or she does (behavior). Most personality tests, including the MMPI, originated in the context of motivational or personality conceptual schema. The MMPI scales, for example, although constructed empirically, soon came to be described and thought about as states of the person or as personality traits. The descriptions of a person derived from the MMPI and other personality tests aim principally at characterizing what he or she has. It is assumed that knowing what a person has permits one to predict and understand what that person does. The long history of studies of how well, in fact, personality tests predict behavior indicates only modest effectiveness (Butcher, 1969; Dahlstrom and associates, 1972; Mischel, 1968; Welsh and Dahlstrom, 1956). Aggregates of subjects meeting predetermined selection criteria often yield statistically reliable trends in behavioral prediction. However, those trends often prove to be clinically trivial in the individual case. Knowing, for example, that 586 of 1000 subjects scored high on a given test scale and displayed more signs of depression, excitability, right parietal headaches, or whatever was the subject for study and that the other 414 subjects had lower scores on the scale and less often displayed the criterion behavior does not permit one to judge wisely whether a particular patient belongs in one criterion group and not the other. For example, the studies reported in Chapter 2, which compared various pain-related criterion groups on personality tests, indicated trends or tendencies of the target populations. Not all subjects had high scores on a given scale. Inferences drawn from those studies were based mainly on mean test scores of criterion groups.

An alternative way of viewing MMPI data is to consider patterns of responses as reflected in test profiles and the configuration of high and low scale scores. These may be seen as indicating high-frequency or low-frequency behaviors. That is, a person with a particular score configuration may display a higher rate of some class of behaviors than a person having a different score configuration.

The point is illustrated by an MMPI issue highly central to chronic pain. Scales 1, 2, and 3 (Hypochondriasis, Depression, and Hysteria, respectively) are usually involved in studies of pain. This was true of studies cited in Chapter 2 (Fordyce and associates, 1975; Janis, 1958; Pilling, 1967; Sternbach and co-workers, 1974). A high score on scale 1 might, to the personologist, indicate the person has hypochondriasis and a low score that he or she does not have it. Behaviorally, the person with the high scale 1 score might be expected to complain readily about bodily distress. That is what is done—the behavior. To this point the personologists and the behaviorist are saying essentially the same thing. Where they may part company is that the behaviorist views those high-frequency behaviors as somewhat more related to reinforcement patterns of the immediate environment. Were the environment to begin to honor different contingency relationships to patient behavior, the behavior would be expected to change. What he or she does changes. The personologist may still see the person as having hypochondriasis, since alleged underlying personality characteristics were not changed.

This is not simply an intellectualized conceptual debate having minimal pertinence to the subject at hand. The part about this issue that makes a difference in using instruments such as the MMPI relates to the extent to which the patient’s pain behaviors are operant or respondent and, whichever is the case, what might be done about it. Those are the questions one asks of test data. The central questions from the MMPI, as they relate to the evaluation of chronic pain, might be posed in this fashion:

   1.  How ready is the person to signal pain?

         If scale 1 is high, those around the patient are more likely to receive repeated cues that pain is being experienced. They may or may not reinforce such behavior. The behavior may or may not be a direct consequence of physical findings.

   2.  How much does\pain behavior cost the person?

         Indications of emotional distress (elevations on F scale and scale 2) suggest that pain has a high emotional cost for the patient or that pain is occurring in someone already having significant emotional distress. In the latter case, pain, instead of costing, may yield time out from sources of distress and therefore be reinforcing.

         Indications that the person has been restless and energetic (elevations on scales 4 or 9 or both) or persistent and self-demanding (elevations on K plus scales 2 and 7) suggest that inactivity from pain (if that is the effect of the pain) probably has a high personal cost. Inactivity is not likely to be reinforcing to that person.

         This second question relates only indirectly to whether the pain is respondent or operant. The question explores reinforcing or nonreinforcing characteristics to impairment from pain.

   3.  Are there other problems for which pain behavior may yield reinforcing time out?

         Indications of marked anxiety (elevations on F, 2, and 7) or of depression (elevations on F and 2, low score on 9) may point to long-standing emotional or interpersonal problems, or the distress may be a product of the pain and related problems.

         Indications of severe disorganization (for example, psychotic states, delusional thinking, or deficient reality testing) (there are a number of possible test profile configurations that may occur here, but perhaps the most common are elevations on F, 2, and 8—perhaps also 6, with a low K and with 7 lower than 8) immediately raise the possibility that inactivity or functional restrictions from pain may yield rich reinforcement in the form of time out from emotional threats. That would not refute the presence of respondent pain. It would indicate that operant pain may be present.

   4.  How probable are addiction or habituation to medications?

         Indications of tending toward addiction or habituation (a configuration of scale 1 elevated and higher than 2, which is higher than 3. If scale 4 is also elevated, whatever its height relative to the others, further confidence in the inference is added) help to assess adequacy of patient information about medication intake. It also raises the possibility that it is addiction that now maintains pain behaviors and encourages the precautionary use of time contingent pain medication regimens for that patient.

         Test score indications of loosening or disorganization of thought processes (preceding item, no. 3) (scale 8 elevated and higher than 7; low K) increase the chances that toxicity is present.

   5.  How likely is the patient to be responsive to attention or rest?

         This question concerns both evaluation and treatment planning. Probable high sensitivity to attention and the actions of others is found particularly in anyone or combination of three score configurations: elevated 3, elevated 4 and 9; low 10. A high scale 10 is a strong contraindication that attention and social responsiveness of others serve as effective reinforcers for the patient.

   6.  How likely are increases in activity to be reinforcing?

         This issue essentially repeats the second part of item 2, that is, it is concerned with measures of energy, restlessness, and activity. It is restated because it bears directly on treatment planning. If other evaluation indications are that there is much operant pain, including reduced activity level, there remains the question of how much help the patient (and his or her immediate environment) will need to maintain treatment gains. In addition to the configurations pointing toward restlessness or surplus energy (high 4 and 9), toward persistence and self-demanding productivity (elevated K and elevated scales 2 and 7), one can add a test profile in which scale 3 is elevated and distinctly above all other scales. Such a person often has a history of job effectiveness and responsibility. He or she often finds activity and productiveness reinforcing.

The foregoing is not intended as a cookbook on how to use the MMPI in relation to evaluating and managing chronic pain. Test profiles need to be viewed in their totality to derive the wisest inferences. The information presented is intended instead to direct attention to the kinds of questions that need to be considered and test analysis methods for dealing with them. They are to be contrasted with approaches that seek to categorize the patient as “hypochondriacal,” “conversion reaction,” and the like. The data reviewed in Chapter 2 indicate that those kinds of test-based labels fail to distinguish consistently patients with significant physical findings from those without. What is needed is an analysis of the interaction between patient behavior (of which test scores provide one source of description) and environmental contingencies. The latter are, at most, only touched on by test scores.

Selective psychological testing can play other productive roles in patient assessment and management. For example, as part of treatment planning, vocational test appraisals provide helpful information from which to identify appropriate target behaviors. Perhaps the most important and commonly relevant additional contribution of selective testing relates to the question of cortical deficits. Cortical deficits enter in in two ways. One concerns addiction and toxicity; the second concerns other problems for which pain behavior yields time out.

It was noted earlier that significant numbers of older patients reporting pain problems for which physical findings are not clear and definitive are found to have cortical deficits. This also may happen with progressive neurological diseases (for instance, multiple sclerosis). Restrictions in social or vocational activities because of pain yield time out from revealing cortical deficits and the ensuing personal embarrassment.

There is an abundance of evidence (Diller, 1971; Reitan and Davison, 1974) to indicate that an appropriate battery of psychological tests can identify with practical validity the presence of interference in cortical or intellectual activity. That kind of specialized testing deserves to be a frequently used part of the evaluation process in matters of chronic pain. As with any other evaluation procedure, effectiveness of these procedures will depend in large measure on the skill and experience of the person who analyzes the data. If it is a computer memory bank that serves as the repository of the knowledge, the right questions must be asked of that memory bank.

The behavioral analysis guide presented in this chapter serves to identify whether there is a significant amount of operant pain in the picture, what the problems are, and who besides the patient may be involved in perpetuating them. The matter of what to do about these problems involves patient selection. The next task is to transfer evaluation data into patient selection criteria. It is the topic of the next chapter.

Commentary: The Behavioral Analysis of Pain 38 Years On

Lance M. McCracken

For anyone who has significant chronic pain, the notion that this experience is not fundamentally a physical problem, or that it could be analyzed as partly a behavioral problem, may be very strange ideas indeed. One can only imagine that these were very bold notions to propose in the way they were almost four decades ago. Yet, today such notions have become more or less commonly accepted among clinicians and researchers, even if occasionally misconstrued. The purpose of this short comment is to examine what was said almost four decades about how one might conduct a “behavioral analysis of pain.” From there I will attempt to repeat a few of the valuable and forward-looking contributions offered at that time and then briefly demonstrate how far we have come since then.


One of the core ideas offered by Fordyce is that once actions that reflect pain and disability, or “pain behaviors,” are emitted, whatever their initiating source, opportunities for conditioning and learning immediately begin [1]. These happen through the influence of consequences that occur in the environment of the person with pain and, in turn, in the environment of significant others who interact with that person. Here patterns of behavior related to pain and disability are potentially shaped, elaborated, and eventually maintained by a mix of reinforcing and punishing consequences that systematically relate to these patterns. Fordyce’s guide for behaviorally analyzing chronic pain sets out an extended series of highly practical steps. With these one is able to identify the presence, extent, and character of the situations and consequences that can begin to play their role in maintaining or increasing the suffering and deactivation that are so common, and in many ways quite natural, for those who seek treatment for chronic pain.

Some of the features of Fordyce’s guide that ought not to be forgotten include the skillful way he distinguishes, and positions, the behavioral analysis as a highly practical addition to disease and personality model approaches, without in any way attempting to pick a fight with these, or exclude them. This open and integrative posture would greatly benefit us today, if more of us could follow it. Second, his advice regarding a proper attitude to adopt while conducting assessments is both compassionate and practical. He demonstrates great empathy for how the patient is likely to be feeling, recognizing the potential for delegitimizing the patient’s suffering or provoking suspicion, and he suggests ways to frame the assessment so that it enlists greater trust. His advice to confront doubts directly, to teach patients not to dichotomize pain into physical or psychogenic, and to essentially “normalize” processes of learning in relation to pain-related behavior, ought to strike a chord for all of us today. Third, his clinical assessment methods are exceptionally incisive. He reminds us how to take socially desirable responding into account as the way to uncover an accurate sense of behavior patterns. His use of skillful questioning, such as to ask “if I were there…what would I see and hear?” includes a quite creative use of perspective taking that is really rather ingenious. Finally, among other valuable nuggets, his guide is filled with numerous commonsense methods and principles for identifying aspects of patient behavior that may need psychological intervention. His recommendation to collect diary data to supplement the interview is one of these, as is his advice that the reinforcing quality of spouse responses lay not in the interviewer’s sense of what is reinforcing or aversive, but in observed relations between the response and patient behavior. Here even an angry or irritated spouse response can be reinforcing.


Fordyce is amazingly consistent and thorough in describing an assessment model where the job is to find the presence of “operant pain,” by finding a role for (a) systematic, repeated, and relatively immediate reinforcing consequences, mostly including pain reduction, rest, medication effects, and social responses, the latter mostly from the spouse, and (b) patterns of punishing consequences or a lack of reinforcing consequences for “well behavior.” As thorough as this analysis is, today’s readers are likely to find limitations in it and a number of processes that are absent or underappreciated. For one, as much as Fordyce is sensitive to the notion of not wanting to dichotomize pain, his entire guide sits on a frame of distinction between respondent and operant pain. Even when he says that these can blend or interact with each other, it is difficult to not perceive this dichotomy strongly in the background.

Another limitation is perhaps an ironic one: Fordyce adopts a relatively strong focus on pain. This may at first seem like an odd statement to make regarding the behavioral approach to pain. The distinction being raised here may be a subtle one, so let me see if I can explain. The behavioral approach of Fordyce clearly focuses on the behavior of the person affected by pain rather than on their experience of pain itself. This may seem to de-emphasize pain, but it may not go far enough. Ironically, pain stays important in this analysis, because most of the focus is on “pain behavior.” Here behavior is selected for analysis, and targeted for change, by virtue of its relation to pain. Perhaps this should not surprise the reader of a chapter called “The behavioral analysis of pain.” Nonetheless, an element that is less emphasized here is a focus on behavior that bears no relation to pain, behavior that might be identified, shaped, and integrated into daily functioning because it serves goals and healthy functioning. A behavioral analysis designed today would certainly deal with pain and pain behavior directly; and it would focus on values and goals-based behavior, within the same analysis, bringing these direct constituent parts of healthy social and physical functioning more to the center of interest. To be clear, this is merely a point about relative emphasis, or perhaps about breadth, and not an attempt to say that these were ignored completely in the analyses as described or in the practice of the time.

Perhaps no one will be surprised that an operant analysis circa 1976 did not include much focus on thoughts and beliefs, or activities in the mind, and included only minimal focus on emotions. Today these issues would surely receive a greater emphasis as well.

The operant model of Fordyce is wonderful for its simplicity and is likely very user-friendly as it asks the person doing the behavior analysis simply (!) to look for systematic behavior-consequence relations. Today a more complex, complete, flexible, and dynamic operant model exists. I find a couple of important elements missing from Fordyce’s operant response. For one he defines an operant as primarily “being subject to influence by consequences that follow their occurrence” (p. 54). At another earlier point he says “our behavior tends to be more influenced by the immediate and the immediately anticipated environment than we often recognize” (p. 7). I feel these statements are open to misinterpretation or may mislead. One issue missing is that an operant, particularly within behavior analytic or Skinnerian approaches, is a behavior pattern where the coordinating circumstances lay not in the consequences that may come next, as these have not happened yet and can play no role, but in consequences experienced in the past. It is not “consequences that follow” (italics added) but consequences that have followed. A key element that is not clear enough in Fordyce’s operant conceptualization is the situation where the behavior occurs. Technically speaking, in an appropriately conceived operant model, historical consequences determine the strength or probability of different behaviors by virtue of their link to the particular situations of interest. Here the coordinating influence of historical consequences is transferred to the situation where the behavior and consequences have happened in the past, it is certainly not held in consequences that may follow. And, if anticipated consequences are the key controlling factor, this is actually not an operant – this now becomes a kind of cognitively or verbally regulated behavior, like rule-governed or instruction-based behavior. For Skinner, and others following his work, the operant is not defined in behavior-consequence relations, but in what is called the “three term contingency:” (a) situation or discriminative stimulus, (b) response, (c) consequence [9]. This is a very powerful concept because it calls our attention to ways to create change that include not only creating changes in “c” or in relations between “b” and “c” but in the relations between “a” and “b” too.

Readers who conduct their own behavioral analyses in the present day may find another missing element in the Fordyce guide. The guide is rather exclusively focused on consequences and focuses relatively little on patient skills, competencies, or deficits. Today, for example, a clinician might be interested to learn how well a patient with chronic pain is able to perform behavior patterns that show qualities of openness to experience, the capacity to shift and maintain awareness focused in the present, to step back from overwhelming or restricting thoughts, to experience that one is not simply made up of one’s thoughts and feelings, to clarify and follow one’s values, to follow patterns goal-directed action that are sensitive, persistent, and yet flexible. We call these capacities “psychological flexibility” [3,6,7], and it is based in the same operant roots as the work of Fordyce.

If skills or capacities required for effective action do not exist within a person’s competencies, they must be built incrementally, and then become basic behavioral elements in other behavior patterns. As I say, this competencies-based view and this notion of basic foundational elements, or basic behavioral repertoires, required for higher order behavior patterns are mostly absent from Fordyce’s guide.

Once again, not to take anything away from the extraordinary accomplishment it is to have proposed the behavioral analysis of pain at a time when most of those currently working in the field were either still in school or not born yet, there is at least one missing piece worth mentioning. Fordyce categorized behavior primarily into two types: pain behavior and well behavior. If one were to fully take on the clinical role of helping a person with complex, disabling, chronic pain to improve their currently devastated daily functioning, one might need more categories.

Left out of the dichotomy of pain and well behavior are many other basic functionally-considered patterns of pathological and of healthy functioning, patterns that are no less “operant” than those emphasized by Fordyce. These could include various examples of what can be called “generalized response classes” – and Fordyce seems not to have focused on these. A couple of examples of these that he could have included, and were well known within the operant approach of the time, include behavior coordinated by consistency between talking and doing (aka, “say-do” correspondence), and so-called “problem-solving” behavior. Skinner presented a rather well-developed conceptualization of “self” in which he recognized that in a social world our behavior meets reinforcing consequences when we act consistently with what we say, and meets punishing consequences when we do not. Through these kinds of processes general patterns of behavioral coherence develop. Such patterns certainly could trap a person in a chronic pain role where the verbal behavior “I have pain” needs to entail overt behavior patterns reflecting “I cannot do activities,” generally, both across situations and over time. “Problem-solving” is another good case in point. Humans are great problem solvers, and this becomes a persistent, generalized, and often over-extended category of behavior, because it is so useful. It is over-extended and sometimes becomes a problem because in some settings it does not work. The point to notice about these categories of “social consistency” and “problem-solving” behavior is that they are rather normal, may not appear as pain behaviors as such, and may not appear in current easily observable contingencies. These are long-standing, highly generalized, fully integrated, and persistent patterns that have developed over a lifetime.


Today there are numerous approaches to the behavioral analysis of problems like chronic pain that are just as “behavioral” as Fordyce’s, just as true to the operant approach, and yet they carry refinements and additional user-friendly features learned over 40 or so years. These include Acceptance and Commitment Therapy (ACT)[3,7], Functional Analytic Psychotherapy [12], and Focused Acceptance and Commitment Therapy (FACT)[10]. In fact, in FACT, an approach designed particularly for settings where time is very limited, the “behavioral analysis” interview is reduced to just four focused questions: (a) “What are you seeking?” (b) “What have you tried?” (c) “How has it worked?” and (d) “What has it cost you?” [10].

Fordyce clearly appreciated avoidance behavior and even understood that avoidance of distress could be a part of this picture. What he did not perhaps fully appreciate is how large this part might be. Today within approaches extending from the early operant approaches avoidance-related processes are given a much larger role than ever before. Today we have a process called “experiential avoidance” deemed a fundamental process of human behavior pathology [4] that is much broader than the behavioral process Fordyce seemed to have in mind, in at least two ways. For one it is considered a process of being unwilling to experience and therefore attempting to limit one’s contact with of a wide range of thoughts, feelings, sensations, urges, images, memories, or generally any psychological experiences, when to do so constitutes an unhealthy behavior pattern. Second, where Fordyce focused on overt actions of avoidance, the concept of experiential avoidance opens up the range of interest to any avoidance activities a person can do, including covert acts of suppression or denial, engagement in other painful activities, impulsive over-engagement in activities, or even such positive-appearing behaviors as distraction or relaxation.


Historically clinical psychology has often included two separate perspectives or camps, one focused on environmental influence, including operant approaches, and the other focused on cognitive influence, including cognitive theory and models and cognitive therapy. Another way to describe these two camps is that they either emphasize environment-behavior relations or mental mediation of behavior as the prime elements of their causal models. It is clear in which camp the work of Fordyce falls. There is also a longstanding tradition of dividing treatment methods within Cognitive Behavioral Therapy into either “behavioral,” such as behavioral activation and goal-setting, or “cognitive,” such as behavioral experiments and cognitive restructuring. Today both of these sets of influences can be incorporated into an approach that is essentially operant.

Today’s functional contextual wing of what used to be called the operant behavioral approach has new, interesting, and already productive ideas about thoughts, beliefs, rule-governed behavior, and general cognitive influence [2]. They essentially have picked up where Skinner left off with his analyses of verbal behavior and have identified new principles and created new models that extend the notion of the operant into the realm of mental activity. Although a complete description of this work is beyond the scope of this commentary, at least the one or two core features of it can be described. One core feature is called “derived arbitrarily applicable relational responding” or AARR. According to this notion humans have the capacity to relate events to each other in a way that does not require learning from direct contact with relations between the events or direct reinforcement (also called “environmental contingencies”), and in a way that does not depend on the formal properties of the events. From this relational responding a person is able to respond to one event in terms of another event. Responding to a pain sensation as if it is a catastrophe or equal to something that is unbearable might be examples of this. Through these acts of relating increasingly complex networks of relations can be derived and psychological functions or influences can be transferred between the elements of these networks (See Törneke [11] for a relatively accessible account of this model). This generalized capacity for relational responding appears to be learned, essentially from repeated shaping within the social-verbal community, and contextually sensitive, and hence is operant in nature. This places cognitive responses like evaluative or judgmental responding and instructional behavioral control into the frame of operant responses. This functional contextual approach to cognitive processes is already presenting itself as a fully viable approach to such phenomena as implicit cognition and associative learning [5] and more complex psychological phenomena, such as self and perspective taking [8].


There have been many ingenious ideas and practices that could have saved us a great deal of trouble in rediscovering them, if only we could have learned them when they first appeared. The work of Fordyce includes many of these. Nonetheless the field still progresses, probably more in fits and spasms than in a steady process of measured steps. Fordyce saw things many of us are still not seeing. With what we have been able to understand he offered a perspective on chronic pain that radically shifted how we define the problem. In many ways he proposed ideas that were more radical than the ones that guide most of what we do today, and in this way he was way ahead of us. Even so his guidance and advice from nearly 40 years ago cannot be the last words. Much has happened in those years, and there are still those busily working with the “behavioral” wing of CBT, those who might like to regard themselves as working within a functional contextual tradition, who probably maintain our closest links with Fordyce’s behavioral analysis of pain. They bring to us today such potentially powerful new ideas as experiential avoidance, a notion of generalized response classes, and “relational responding,” an operant view on cognition.


  1.Fordyce WE. Behavioral methods for chronic pain and illness. St Louis: CV Mosby, 1976.

  2.Hayes SC, Barnes-Holmes D, Roche B. Relational frame theory. A post-Skinnerian account of human language and cognition. New York: Kluwer Academic / Plenum Publishers, 2001.

  3.Hayes SC, Strosahl K, Wilson KG. Acceptance and Commitment Therapy: An experiential approach to behavior change. New York: Guilford Press, 1999.

  4.Hayes SC, Wilson KG, Strosahl K, Gifford EV, Follette VM. Experiential avoidance and behavioral disorders: a functional dimensional approach to diagnosis and treatment. J Consult Clin Psychol 1996;64:1152–68.

  5.Hughes S, Barnes-Holmes D, Vahey N. Holding on to our functional roots when exploring new intellectual islands: A voyage through implicit cognition research. J Context Behav Sci 2012;1:17–38.

  6.McCracken LM, Morley S. The psychological flexibility model: a basis for integration and progress in psychological approaches to chronic pain management. J Pain 2014;15:221–34.

  7.McCracken LM, and Vowles KE. Acceptance and Commitment Therapy and mindfulness for chronic pain: model, process, and progress. Am Psychol 2014;69:178–187.

  8.McHugh L, Stewart I. The self and perspective-taking: Contributions and applications from modern behavioral science. Oakland, CA: New Harbinger Publications, 2012.

  9.Skinner BF. Science and human behavior. New York: Free Press, 1953.

10.Strosahl, K, Robinson P, and Gustavsson T. Brief interventions for radical change: Principles and practice of Focused Acceptance & Commitment Therapy. Oakland, CA: New Harbinger Publications, 2012.

11.Törneke N. Learning RFT: An introduction to Relational Frame Theory and its clinical applications. Oakland, CA: Context Press, New Harbinger Publications, 2010.

12.Tsai M., Kohlenburg RJ, kanter JW, Kohlenberg B, Follette WC, and Callaghan GM. A guide to Functional Analytic Psychotherapy: Awareness, courage, love, and behaviorism. New York: Springer, 2009.

Commentary: The Behavioral Analysis of Pain–From a Contemporary Perspective

Michael K. Nicholas

This perspective considers Fordyce’s approach to the assessment of patients in chronic pain. The historical context in which he developed his approach is balanced with a reflection on how we might view his methods almost 40 years later and possible lessons we could still benefit from re-learning.

In this chapter, Fordyce describes what he saw as the key elements of a behavioral analysis of a patient’s pain complaints [5]. In common with other sections of the book, he links each element to relevant underlying learning principles, but throughout he provides extensive clinical guidance and commentary based on his considerable experience in the field. As such, he provides a valuable text for both the novice assessor and experienced pain clinicians who must regularly review their assessment methods. Besides guidance for the clinician on how the assessment should be conducted and what to look for in the patient’s responses, Fordyce also provided examples of advice that could be directed to the patient, especially when s/he appears apprehensive about undertaking a behavioral analysis of his/her pain–not an uncommon situation given the widely-held concerns about possible psychological contributions to pain experience. In this context, it should be realized that when he wrote his book Fordyce was working in a medical rehabilitation setting, rather than a psychiatric or psychological clinic. As a result, his assessments were always conducted as part of a multidisciplinary evaluation of a patient’s presenting pain complaints. At the time this had become standard practice in the Department of Rehabilitation Medicine at The University of Washington Hospital where Fordyce was working. Appropriately, Fordyce has always readily acknowledged the contribution by his (multidisciplinary) rehabilitation colleagues for many of the ideas and methods outlined in the book. Many of these colleagues were included in their 1973 publication of an (uncontrolled) trial of operant methods in their inpatient service [6].

Interestingly, it is worth noting that the approach advocated by Fordyce and his rehabilitation colleagues was consistent with Bonica’s original vision for a multidisciplinary pain service at the same hospital where he was in the Department of Anesthesiology [3]. It will be recalled that arguably Bonica’s landmark achievement in clinical pain practice was his championing of a shift in medical practice with patients suffering from chronic pain conditions that moved away from the traditional ‘serial referral’ approach to a multidisciplinary approach. In the serial referral approach patients were typically referred from one independent specialist to another once each had done or tried what they were specialized to do (e.g., surgery, nerve blocks, cortisone injections, etc.) without success. Bonica found this mode of operating slow and inefficient, especially if he had further questions for his consulting colleagues. In response to this situation, Bonica started to arrange regular meetings with his colleagues to discuss the patients they had in common and to arrive at a degree of consensus on their diagnoses and a treatment plan. Bonica found this multidisciplinary mode of operating much more effective and efficient, especially in the treatment of complex cases. Initially, the disciplines participating at Bonica’s clinic were from medical or surgical specialties, but over time other health disciplines were included as well.


As noted in earlier chapters, Fordyce saw no difficulty or contradiction in accepting both a biomedical (disease model) approach alongside a behavioral perspective. In fact, he argued that medical assessments and possible diagnoses should always be undertaken as the first step in evaluating a new patient with chronic pain. He acknowledged that it was possible the medical findings could adequately account for a patient’s pain behaviors–describing them in this context as respondent behaviors, albeit to unobserved or even inferred “afferent impulses” (p. 207)[5]. Here it should emphasized that Fordyce was not so concerned with exploring the possibility of respondent (or classical) conditioning as others have explored in the years since 1976 [9,11,19,26], but rather his focus was on primarily operant learning mechanisms, where the consequences of a behavior could effectively reinforce it and thereby maintain the behavior subsequently, largely independently of how it might have started [21].

In the event that disease model accounts could not be made with confidence, Fordyce proposed an operant behavioral evaluation should be undertaken. However, how this determination might be made was, and remains today, a matter of considerable uncertainty. For example, some authors talk of ‘medically unexplained symptoms’ (MUS), their likely etiology and possible management, but these accounts assume it is possible to detect all possible physical pathology [17]. Interestingly, current accounts of chronic pain generally accept that pain may persist due to central nervous system (CNS) changes secondary to an original peripheral (nociceptive) event (like an injury), which may well have resolved, but these CNS changes are not readily demonstrable in an individual patient [28]. But neuroplastic mechanisms were not established in 1976, so Fordyce could only make general statements like: ‘It is therefore sometimes desirable to round out a disease model analysis of a pain problem with an additional behavioral analysis.’ (p. 208) [5] thereby avoiding the question of ‘when?’ However, he also acknowledged: ‘A behavioral analysis is not a substitute for a disease model analysis. A behavioral analysis can only indicate the viability of learning factors as an alternative explanation for a pain problem’ (p. 209) [5]. This implies there is room for both approaches in a single case as they have a different purpose and may account for different aspects of a patient’s presenting problems – a key feature of more recent biobehavioral models of chronic pain [4]. Thus, today, even when there is a clear case of neuropathic mechanisms contributing to a person’s pain, a psychological assessment and intervention is routinely encouraged [7].

In this context it should be remembered the prevailing view in 1976, and extending to quite recent times, held that consideration of psychological contributors to pain was only required when medical explanations were insufficient – an approach often referred to as diagnosis by exclusion (i.e., if no clear medical explanation was available, there must be a psychological explanation [14]. While Fordyce was clearly influenced by this dictum, he did differentiate between his (operant behavioral) approach and the other major psychological perspectives of the time that were essentially intrapsychic (described by Fordyce in terms of ‘motivational’ and ‘personality-based’ accounts). In particular, he looked for relationships between pain behaviors and environmental events, especially interactions between the patient’s behavior and those in his/her social environment, such as a spouse. He argued that predictions about what a person will do that are based on careful observations of what a person does (i.e., behavior consequence relationships) are likely to be more precise, relative to what might be achieved with predictions based on his/her personality traits.

In classic operant behavioral parlance, Fordyce was interested in the stimulus conditions (e.g., when) under which a pain behavior (such as limping or groaning) occurred and the consequences that often followed the behaviors (such as attention from others or avoidance of an unpleasant activity). But in contrast to what he termed “motivational” accounts, he was not implying the patient might perform a behavior in order to gain attention, but rather the behavior in a given context was maintained by a consequence, like attention from significant others. The corollary of a behavioral or learning explanation for recurring pain behaviors was that they required repetition over time. Fordyce maintained that without these there was ‘no point looking for conditioning effects’ (p. 208) [5].

Interestingly, the term ‘motivational’ has recently been revived in accounts of chronic pain phenomena, but these new uses of the term differ markedly from those referred to by Fordyce. Van Damme et al. [25], for example, argued that “pain is never an isolated event, it always occurs within a context of goal pursuit” and consideration of the person’s goals was important to appreciating their response to the experience of pain, such as the interruptive effect of pain on behavior. One effect of this line of research has led to more direct consideration of pain patients’ broader goal framework (i.e., not just in terms of pain relief, but also the achievement of a range of important life ambitions, regardless of pain). This perspective might also be read into one of Fordyce’s reported statements that “folks that have something better to do, don’t suffer as much”, but experimental research in 1976 was not considering motivational accounts of pain in these terms.

With these considerations in mind Fordyce turned himself towards the conduct of a behavioral analysis in clinical practice.


The methods Fordyce describes include a semi-structured interview, an interview with a patient’s significant other or spouse (preferably without the patient being present), as well as what he termed ‘supplemental “behavioral” data’ (p. 225)[5]. The latter might include data from previous psychiatric diagnostic interviews, the patient’s own pain diary (daily behavioral recordings by the patient for at least 1–2 weeks), and data from psychological tests (primarily the MMPI). The primary focus of these methods was to identify the presence of what Fordyce termed ‘operant pain’ present in a particular case (i.e., evidence that reinforcement contingencies may be contributing to a patient’s presenting pain problems). In addition, the behavioral analysis was expected to provide a summary of the specific problems presented by a patient, and some indication of who, besides the patient, may be contributing to their maintenance. In the main, these findings were expected to contribute to treatment planning – as they ought to be susceptible to change.

The topics covered in the interview included a description of a typical day at present for the patient, the time pattern of pain during the day, nocturnal patterns in the pain and the patient’s responses, the timing of pain-related medication consumption, the responses of significant others when they believed the patient was in more pain, their awareness of activities that seemed to aggravate or reduce their pain, and so on. In exploring the relationships between the patient’s pain behaviors and his/her social environment in this way, rather than rely just on a patient’s memory, Fordyce also used the patient’s daily diary recordings to prompt closer examination of these relationships.

A critical point to note here is how the interviewer was encouraged to ask these questions. Specifically, Fordyce encouraged the adoption of what has been called a Socratic style (after the famous Greek philosopher/teacher who believed in teaching by asking questions rather than simply giving students information [24]. The approach advocated by Fordyce entailed asking questions that begin with words like ‘how’, ‘when’, ‘what’ as these tend to result in more factual or specific responses than questions that begin with words like ‘why’, which run the risk of putting a patient on the defensive and feeling s/he has to somehow explain him or herself. Interestingly, exponents of cognitive therapy have also taken a similar position to interviewing [2]. But these cognitive methods were still in the early stages of development in the mid-1970s.


The use of the MMPI is probably the most contentious source of data today, and in reality it is now rarely used outside some US clinical and medicolegal contexts. Despite it having once been the most frequently used psychological test in the assessment of patients with chronic pain [20], its use has fallen over time due to a number of factors, including the time required to complete it making it impractical for most clinics, but more importantly, its questionable value in clinical pain contexts [8, 12]. Many authors have also considered there are better ways (i.e., quicker and more valid in the populations tested) to gather the sorts of information it was hoped the MMPI might provide. Flor and Turk [4], for example, argued the primary use of the MMPI was to determine the presence of psychopathology, but even the value of this use is debated [10, 12], and with the advent of DSM-V, out of date. The MMPI was also considered a means of predicting treatment outcomes (and therefore suitability for treatment), but on the basis of the accumulated evidence that provides little support for this role, Flor and Turk concluded ‘it is inappropriate to use the MMPI as the basis for making a decision to include or exclude patients from a pain rehabilitation program’ (p. 258) [4].

One might also wonder about Fordyce’s support for the use of a measure like the MMPI, especially in view of his comments noted earlier in this chapter that personality-based data on pain patients were likely to prove much less precise than careful observations of what a person does. To be fair, he did refer to data from the MMPI as ‘supplementary’ to the behavioral interview (and observations), rather being critical to the assessment process. He noted, for example, that studies had shown that MMPI profiles were inconsistent at distinguishing between patients with significant physical findings from those without, and that test scores compared poorly with an analysis of the interaction between patient behavior and environmental contingencies. However, he did seem to feel the data from the MMPI provided a means of confirming impressions he gained from interviewing his patients, as well as a generator of possible hypotheses he might raise with a patient. On the other hand, he did not have available the range of more pain-specific psychological measures that have been developed since 1976, such as those evaluating cognitive constructs like catastrophizing [22], self-efficacy [18]; fear of re-injury [9], acceptance [13], and sense of injustice [23]. It is quite likely he would have found these much briefer and more relevant for the assessment of chronic pain patients than the MMPI, but the development of these cognitive constructs in relation to pain was only in its infancy in 1976 (see Bandura [1] as an example).


In looking back at Fordyce’s approach to behavioral analysis, one might be struck by the apparent lack of attention to such individual level details in current pain management papers. These days it seems chronic pain patients are typically described in terms of demographics and mean scores on scales of various dimensions, rather than the more functional approach that characterized Fordyce’s key methods. While our current approaches have many benefits in terms of characterizing the populations being studied or treated, and there have been major improvements in research methodology [15], the (mean) results of many recent studies using broadly cognitive-behavioral types of treatment have increasingly been described as trending towards small effect sizes [27]. This has led to a number of leading researchers to argue that the time has come to revisit the more individualized (n = 1) types of studies that characterized the early behavioral studies generally [16]. In this sense, it would seem very timely that we re-discover the work of Bill Fordyce.


  1.Bandura A. Self-efficacy: toward a unifying theory of behavioral change. Psychol Rev 1977;84:191–215.

  2.Beck AT, Rush AJ, Shaw BF, Emery G. Cognitive therapy of depression. New York, Guilford, 1979.

  3.Bonica JJ. The management of pain. London: Kimpton, 1953.

  4.Flor H, Turk DC. Chronic Pain: an integrated biobehavioral approach. Seattle: IASP Press, 2012.

  5.Fordyce WE. Behavioral methods for chronic pain and illness. St Louis: CV Mosby, 1976.

  6.Fordyce WE, Fowler RS, Lehmann JF, Delateur BJ, Sand PL, Trieschmann RB. Operant conditioning in the treatment of chronic pain. Arch Phys Med Rehabil 1973;54:399–408.

  7.Haanpää M. The assessment of neuropathic pain patients. Pain Manage 2013;3:59–65.

  8.Helmes E, Reddon JR. A perspective on developments in assessing psychopathology: a critical review of the MMPI and MMPI-2. Psychological Bulletin 1993;113:453–71.

  9.Kori SH, Miller RP, Todd DD. Kinesiophobia: a new view of chronic pain behaviour. Pain Manage 1990;3:35–43.

10.Leavitt F, Garron DC, McNeill TW, Whisler WW. Organic status, psychological disturbance, and pain report characteristics in low-back-pain patients on compensation. Spine 1982;7:398–402.

11.Lethem J, Slade PD, Troup JDG, Bently G. Outline of a fear-avoidance model of exaggerated pain perception–I and II. Behav Res Ther 1983;21:401–409.

12.Main CJ, Spanswick CC. Personality assessment and the Minnesota Multiphasic Personality Inventory. 50 years on: do we still need our security blanket? Pain Forum 1995;4, 90–96.

13.McCracken LM, Vowles KE, Eccleston C. Acceptance of chronic pain: component analysis and a revised assessment method. Pain 2004;107:159–166

14.Merskey H. Beware somatization. Eur J Pain 2000;4:3–4.

15.Morley S. Efficacy and effectiveness of cognitive behaviour therapy for chronic pain: Progress and some challenges. Pain 2011;152:S99–S106.

16.Morley S, Williams A, Eccleston C. Examining the evidence about psychological treatments for chronic pain: time for a paradigm shift? Pain 2013;154:1929–31.

17.Morris R, Gusk L. Assessment and immediate management of patients with medically unexplained symptoms in primary care. Psychiatry 2009;8:179–83.

18.Nicholas MK. The pain self-efficacy questionnaire: taking pain into account. Eur J Pain2007;11:153–63.

19.Philips HC. Avoidance behaviour and its role in sustaining chronic pain. Beh Res Ther 1987;25:273–80.

20.Pincus T, Callahan LF, Bradley LA, Vaughn WK, Wolfe F. Elevated MMPI scores for hypochondriasis, depression, and hysteria in patients with rheumatoid arthritis reflect disease rather than psychological status. Arthritis Rheum 1986;29:1456–66.

21.Skinner BF. Science and Human Behavior. New York: Free Press, 1953.

22.Sullivan MJL, Bishop S, Pivik J. The pain catastrophizing scale: development and validation. Psychol Assess 1995;7:524–32.

23.Sullivan MJL, Adams H, Horan S, et al. The role of perceived injustice in the experience of chronic pain and disability: scale development and validation. J Occup Rehabil 2008;18:249–61.

24.Turk DC, Meichenbaum D, Genest M. Pain and behavioral medicine: a cognitive-behavioral perspective. New York: Guilford Press, 1983.

25.Van Damme S, Legrain V, Vogt J, Crombez G. Keeping pain in mind: a motivational account of attention to pain. Neurosci Biobehav Rev 2010;34:204–213.

26.Vlaeyen JWS, Kole-Snijders AMJ, Boeren RBG, van Eek H. Fear of movement/re-injury in chronic low back pain and its relation to behavioural performance. Pain 1995;62:363–72.

27.Williams ACdC, Eccleston C, Morley S. Psychological therapies for the management of chronic pain (excluding headache) in adults. Cochrane Database of Syst Rev 2012;11:CD007407

28.Woolf CJ. Central sensitization: implications for the diagnosis and treatment of pain. Pain 2011;152, S2–S15.

Only gold members can continue reading. Log In or Register to continue

Jul 9, 2018 | Posted by in Uncategorized | Comments Off on The Behavioral Analysis of Pain

Full access? Get Clinical Tree

Get Clinical Tree app for offline access