Misconceptions that Hamper Assessment and Treatment of Patients Who Report Pain

Patient versus Caregivers/Family

Who is the authority on patients’ pain? Whose pain is it? Clinicians sometimes believe they know more about the patient’s pain than the patient does. No matter how appealing that belief may be, it is false. Nevertheless, privately or among themselves, clinicians may comment about a patient, “He doesn’t have as much pain as he thinks he does” or “The pain is not that bad,” implying that the clinicians are the true authorities on patients’ pain. In research cited earlier, clinicians’ tendencies to underestimate patients’ pain is noted.

No objective measures of pain exist. The sensation of pain is completely subjective. Pain cannot be proved or disproved. One definition of pain used in clinical practice says “Pain is whatever the experiencing person says it is, existing whenever he says it does” (McCaffery, 1968, p. 95). Statements by the American Pain Society (APS) have echoed the same approach to patients’ reports of pain by statements such as:

The gold standard for assessing the existence and intensity of pain is patients’ self-reports. No other source of information has ever been shown to be more accurate or reliable than what a patient says. Patients’ behaviors, the opinions of nurses and physicians delivering care, patients’ vital signs—none of these is as reliable as patients’ reports of pain and should never be used instead of what a patient says.

The Andrew-Robert survey presented in Box 1-1 (p. 18) illustrates what happens when clinicians do not adopt the patient’s self-report as the standard for assessment of pain intensity. This survey has been used by many nurse educators in hundreds of educational programs to explain the necessity of accepting the patient’s report of pain as the standard for assessment. The survey has become so familiar to staff nurse educators that it is often referred to simply as the Andrew-Robert survey. Several publications have reported the results of studies using this survey and modifications of it (McCaffery, Ferrell, 1991a, 1991b, 1992a, 1992b, 1994a, 1997a; McCaffery, Ferrell, O’Neil-Page, 1992). A summary of survey findings from 1990 to 2006 has been published by the originators of the survey (McCaffery, Pasero, Ferrell, 2007). The results of the Andrew-Robert survey as presented in Table 2-1 are based on the responses of 615 registered nurses who attended pain programs throughout the United States. The survey was administered to registered nurses before the pain conference began. In viewing the nurses’ responses to pain assessment (question 1), it is apparent that not all nurses understood that the patient’s self-report of pain is the single most reliable indicator of pain. Both of these patients reported their postoperative pain as 8, but 14% of the nurses did not record 8 for the smiling patient and 5% did not record 8 for the grimacing patient. Those nurses who did not record 8 falsified the record and made it impossible for the next nurse to evaluate previous treatment for pain.

The percentages of nurses who recorded what the patients said are high but there is no room for false entries in the chart. If these results were from a single institution, then one would have to say that up to 14% of the nurses were not recording what the patients said, and this is made more complex because we do not know who they are. Thus, all recordings of pain intensity would have to be viewed with some suspicion.

The Andrew-Robert survey is a quick and easy survey to use in small and large populations of nurses. The results of the survey help to generate discussion of common problems that arise in patient care. Another brief and more comprehensive survey of staff knowledge is titled Pain Knowledge and Attitude Survey and is available in Appendix A.

The survey findings illustrate what is sometimes seen in clinical practice—each person caring for a patient may have a different opinion about the intensity of a patient’s pain. Without a standard for assessing pain, chaos quickly ensues. For example, four different clinicians caring for the same patient may arrive at four different pain ratings, all of which are different from the patient’s own pain rating (and usually are underestimations of a patient’s pain). How do you resolve five different pain ratings so that intervention can be planned?

The need to establish the patient’s self-report as the standard becomes apparent. There appears to be no alternative. It is reassuring to realize that the validity and reliability of patients’ self-reports of pain are testified to in the numerous double-blind studies of analgesics, in which the patients’ pain ratings always determine the analgesic effect of the drugs being tested. Initial recommended doses of analgesics and equianalgesic charts have relied on the patient’s self-report for decades.

When a patient reports pain, the health care professional’s responsibility is to accept and respect the report and to proceed with appropriate assessment and treatment based upon the self-report. All reports of pain are taken seriously. When difficulties arise in accepting the patient’s report of pain, some of the strategies listed in Box 2-1 may be helpful.

Health Care Conditions that Influence Clinicians’ Judgments of Pain

Numerous factors influence clinicians’ tendencies to doubt patients’ reports of pain. This section discusses the effects of length of time caring for patients with pain, environmental cues that promote suspicion of patients’ pain reports, and the effects of the nature of one’s clinical practice.

Studies report conflicting results about the effects of length of time caring for patients with pain. As reported earlier, clinicians frequently underestimate patients’ pain ratings. This has been explained as being a process of habituation. Repeated exposure to patients’ pain seems to promote the development of insensitivity to pain (Kappesser, Williams, Prkachin, 2006). Yet other studies have not shown this (Everett, Patterson, Marvin, 1994; Hamers, van den Hout, Halfens, 1997). In a study of 50 registered nurses randomly selected from 2 general surgical units, two general medical units, and a combined intensive care unit/coronary care unit, participants completed an instrument that contained 60 vignettes describing a patient’s illness or injury (Dudley, Holm, 1984). Nurses were instructed to rate each patient on the degree of pain and psychologic distress. Nurses inferred significantly less pain than psychologic distress. Correlation analysis revealed a markedly weak and statistically insignificant correlation between years of practice and pain scores.

Other factors that may influence clinicians’ estimates of patients’ pain may be access to the patients’ self-reports and the atmosphere of suspicion surrounding particular patients. In one study, 60 physicians and nurses from the emergency department (ED) and 60 from the oncology setting watched videotapes of facial expressions of patients with shoulder pain undergoing range-of-motion exercises (Kappesser, Williams, Prkachin, 2006). Participants were divided into three groups and given different information: Group 1, videotape only; Group 2, the same videotape plus the patients’ reported pain intensity; and Group 3, the same videotape plus being told the patient pain reports and that some of the people were faking pain to obtain opioid drugs. All participants were asked to rate the patients’ pain, and participants in groups 2 and 3 were asked to state whether they had the impression that any of the patients had faked pain or hidden pain. Clinicians in all three groups underestimated patient pain intensity. The least discrepancy between clinicians’ and patients’ pain ratings occurred when the clinicians were given the patients’ self-reports but were not told that some patients were faking pain to obtain opioids.

The greatest discrepancy occurred when clinicians were given the patients’ pain reports but were told that patients might be faking pain to obtain opioids. This suggests that a clinical atmosphere of suspicion, such as sometimes occurs in trauma units or EDs, may cause an increase in clinicians’ doubting patients’ reports of pain. In all clinical areas clinicians need to be cognizant of the attitudes they convey when they discuss patient care with their colleagues. Managers and supervisors can be alert to stigmatizing labels and help staff to avoid them. For example, the term drug-seeking, which has no universally accepted definition, commonly conveys that a patient is addicted to opioids, abusing pain medicine, or manipulative (McCaffery, Grimm, Pasero, et al., 2005). In fact, the American Society for Pain Management Nurses (ASPMN; 2002) recommends that this term not be used because it creates bias and prejudice that subsequently have negative effects on pain management.

The nature of a particular clinical practice may influence clinician estimates of patient pain intensity. The effect of repeatedly inflicting pain on patients was studied by using brain scans to compare the responses of physicians who practice acupuncture with the responses of naïve participants while watching animated visual stimuli depicting body parts in both a nonpainful situation (being touched by a Q-tip) and a potentially painful situation (acupuncture [being pricked by needles]) (Cheng, Lin, Liu, et al., 2007). The acupuncturists rated these situations as being significantly less painful and unpleasant than did the naïve control participants.

In addition, the brain scans supported the notion that mechanisms are triggered in the areas of the brain that regulate emotions and cognitive control. The authors commented that without some regulatory mechanism, it is likely that clinicians would experience personal distress and anxiety that would interfere with their ability to perform. Acupuncturists know that they may be inflicting pain and seem to have learned throughout their training to inhibit the empathy-pain response.

Many other clinical practices, such as burn care and surgery, involve repeated infliction of pain. A study of burn care found that staff members who had spent more time working with burned patients believed débridement was less painful than did the staff members who had spent less time working with burned patients (Perry, Heidrich, 1982). To continue to manage pain effectively, such clinicians need to be aware of their tendencies to underestimate the painful impacts of what they do and to recognize that doing so may be an essential mechanism that allows them to continue to give care. Rather than be embarrassed or ashamed of this reaction, they can acknowledge it to themselves (and perhaps others) and try to compensate by obtaining and listening carefully to information from their patients.

Other examples are clinicians who perform venipunctures or bone marrow aspirations daily; they may become less sensitive to the amount of pain they are inflicting. Clinicians who work in a pain clinic several days a week may grow weary of and less sensitive to patients who continue to report pain despite their best efforts to relieve the pain. Again, to continue to practice effectively in such conditions it may be necessary to make a conscious personal inventory that honestly admits to having less empathy for patients as time goes by. Once this is acknowledged, the clinician can plan a systematic approach to obtaining information from the patient and ensure that it is acted upon.

Considerations When Doubts Arise

On occasion, accepting and acting on the patient’s report of pain are difficult. Because pain cannot be proved, health care professionals are vulnerable to being fooled by the patient who wishes to lie about pain. However, although accepting and responding to the report of pain will undoubtedly result in giving analgesics to some patients who do not have pain, doing so ensures that everyone who does have pain receives attentive responses. Health care professionals do not have the right to deprive a patient of appropriate assessment and treatment simply because they believe a patient is lying.

An important distinction exists between believing the patient’s report of pain and accepting the report. Following the recommendations of the clinical practice guidelines does not require that the clinician agree 100% with what the patient says. Clinicians are not required to believe a patient but are required to accept what a patient says, convey acceptance to the patient, and take the appropriate action. Clinicians are entitled to their personal doubts and opinions, but they cannot be allowed to interfere with appropriate patient care. Box 2-1 summarizes some strategies that can be used when the patient’s report of pain is not accepted.

Although accepting the patient’s report of pain occasionally results in being fooled, no stigma or blame should be attached to being duped. In any relationship, each party has certain responsibilities. Fault is assigned to the parties who fail to meet their responsibilities (Wesson, Smith, 1990). If the clinicians fulfills his or her responsibility to respond to all reports of pain with appropriate assessment and treatment, the clinician will be able to say, “Although I was probably fooled by some patients, I never failed to help those who did have pain. No one can find fault with my behavior or professional conduct.”

Furthermore, accepting the patient’s report of pain avoids an adversarial relationship. When a clinician conveys to a patient that a report of pain has not been accepted, it amounts to accusing the patient of lying. Understandably, this is upsetting and frightening to a patient who has asked the health care provider for help with pain. Much has been written in the literature about the distinction between suffering and pain. It is worth noting that in his analysis of suffering, Eric Cassell (1982) mentions that one source of suffering is physicians who do not validate a patient’s pain but rather ascribe it to psychologic causes or accuse the patient of faking. Clinicians ask patients to trust them. At some point clinicians must return the favor.

The term malingering may be encountered when a patient’s report of pain is doubted. Malingering can be defined as a conscious effort to produce symptoms such as pain for the purpose of deceiving or misrepresenting the facts, usually for monetary or other gains. Malingering seems to be suspected most commonly in patients with persistent pain, especially low back pain. A review of the literature prior to 1999 reveals 68 references to research of persistent pain malingering or disease simulation (Fishbain, Cutler, Rosomoff, et al., 1999). These studies conclude that malingering does occur, perhaps in 1.25% to 10.4% of patients, but these figures are not reliable because of the quality of the studies. This review concludes that there are no reliable methods of identifying malingering. The International Association for the Study of Pain notes that the process of identifying malingering is, in the final analysis, a legal, not a medical, process (Merskey, Bogduk, 1994).

On reflection, one might ask whether it is justifiable to be suspicious of all patients in an attempt to avoid being fooled by the few who lie. It is a burden to the clinician and an insult to the patient to wrestle with potential dishonesty at each encounter.

Acute Pain Model versus Adaptation

The acute pain model says that if the patient has pain, visible signs of discomfort, behavioral or physiologic, will be present. Examples of behavior usually expected of patients with pain include grimacing, rigid body posture, limping, frowning, or crying. Physiologically, elevated vital signs are commonly expected. Clinicians and laymen alike usually fail to appreciate that both physiologic and behavioral adaptation occurs, leading to periods of minimal or no signs of pain. Absence of behavioral or physiologic signs of pain does not necessarily mean absence of pain.

The acute pain model is of limited value for assessing pain. When pain is sudden or severe, behavioral and physiologic indicators may be present for a brief time. However, very quickly the patient may make an effort to cease behaviors such as crying or moaning. Behavioral adaptation or suppression of pain behaviors may occur because the patient values the stoic response or simply becomes exhausted. Physiologic indicators such as increased blood pressure (BP) or heart rate may also disappear. In a healthy individual, the body seeks homeostasis or equilibrium, returning to the former physiologic state despite severe pain. Or a patient may have a medical condition that causes low BP, such as hypothyroidism or dehydration, which has a much greater impact on vital signs than pain has. In such patients, sudden, severe pain may elevate the vital signs only briefly and minimally.

The APS addresses the misconceptions about the acute pain model by stating the following:

Responses Clinicians Expect of Patients with Pain

Nurses’ responses to the Andrew-Robert survey (see Table 2-1, p. 21) reveal that patients’ behavioral responses have significant effect on nurses’ pain assessments and treatment decisions (McCaffery, Pasero, Ferrell, 2007). The only difference between Andrew and Robert is their behavior—Andrew smiles and laughs with visitors, whereas Robert lies in bed and grimaces. This simple difference has a startling effect on nurses’ decisions about opioid doses.

Although the patients were exactly alike except for their behaviors, the nurses were more likely to increase the morphine dose for the grimacing patient. Both patients had received morphine, 2 mg IV, 2 hours before; half-hourly pain ratings had ranged from 6 to 8 out of 10 and were currently 8; and no clinically significant adverse effects such as sedation had occurred. The pain rating goal was 2. Nurses were given a choice of administering no morphine or 1 mg, 2 mg, or 3 mg IV. Morphine, 3 mg IV, was the correct choice for both patients because the previous dose of 2 mg was safe but ineffective. However, only 50.6% of the nurses would increase the dose for the smiling patient, whereas 64.3% would increase the dose for the grimacing patient. Both patients were undertreated, but at least 14% of the nurses knew that it was safe to increase the dose for the smiling patient, but they did not. Over the years since 1990, other similar vignette surveys conducted by these authors have shown similar findings, which show little improvement in nurses’ choices of opioid doses (McCaffery, Pasero, Ferrell, 2007).

The same discouraging results were found when the Andrew-Robert survey was adapted to reflect the needs of elderly patients in a long-term care facility (Katsma, Souza, 2000). The participants were 89 licensed nurses working in long-term care facilities. As in the original vignette, one patient was described as smiling and the other as grimacing. After receiving medication, hourly pain ratings were 6 to 8, and the patients showed no serious adverse effects. In response to the question about which analgesic dose they would now give, only 30% correctly selected a higher dose for the smiling patient, but 43% did so for the grimacing patient.

These same biases about behavior also exist in laypersons. The Andrew-Robert survey was revised to be appropriate for a nonnurse audience and was administered to 85 college students who were not enrolled as medical or nursing majors (McCaffery, Ferrell, 1996a). College students’ responses to assessment and relief of pain showed trends similar to those of practicing nurses. The smiling patient’s pain rating was accepted by 38% of the college students, whereas 55% accepted the grimacing patient’s report of pain.

Vital signs also influence nurses’ willingness to record a patient’s report of pain. A survey using the same format as the original Andrew-Robert survey was constructed; the only difference between the patients was their vital signs (McCaffery, Ferrell, 1992a, 1992b). One patient had low to normal vital signs, and the other patient had elevated vital signs. The responses of 166 nurses revealed that more nurses were willing to accept the report of severe pain from the patient with elevated vital signs than from the patient with low to normal vital signs.

As part of the Missoula Demonstration Project, a 15-year process to study end-of-life issues, surveys about pain knowledge and attitudes were sent to 942 nurses and produced 311 responses (Mayer, Torma, Byock, et al., 2001). The survey revealed that 93% of the nurses knew that vital signs were not reliable indicators of pain, but 75% said that vital signs moderately or greatly influenced their decisions to treat or not to treat pain.

In a small study designed to identify the criteria nurses use to assess postoperative pain, 10 nurses were interviewed about their pain assessments of 30 postoperative patients (Kim, Schwartz-Barcott, Tracy, et al., 2005). The strategy most frequently reported by the nurses relied on the patient’s appearance and drew on past experience of which physical signs to look for, such as facial expression, body movement, and heart rate. For example, nurses mentioned frowning or wincing as being indicative of pain and a patient’s being able to fall asleep as a sign of little or no pain. The researchers commented on the fact that the predominant strategy of looking for objective signs of pain was in marked contrast to current guidelines that emphasize the patient’s self-report of pain.

Expectations that certain behaviors indicate pain also influence the prescribing of analgesics. To identify factors that affect physicians’ decisions to prescribe opioids for persistent noncancer patients, the records of 191 patients referred to a pain center were examined to determine pain severity, physical findings, pain duration, age, gender, observed pain behaviors, reported functional limitations, and affective distress (Turk, Okifuji, 1997). Of all these variables, only observed pain behaviors were significantly related to receiving opioid prescriptions. In other words, patients who exhibited pain behaviors were most likely to receive opioids for pain relief. The extent of physical findings and the severity of the pain did not appear to influence the decision to prescribe opioids. Because opioids are prescribed for the purpose of pain relief, it would seem more logical to find that severity of pain determined prescription of opioids, but this was not the case.

Similar findings also were reported regarding patients with low back pain. Decisions about lumbar surgery were not made on the basis of physical pathologic conditions but rather on behaviors demonstrated by patients during their evaluations (Waddell, Main, Morris, et al., 1984). As summarized by Turk and Okifuji (1997), “physicians appear to believe that behavioral demonstrations of pain, such as limping and grimacing, indicate something important about the nature of the patient’s pain and the need for prescribing specific treatments such as surgery and opioids” (p. 334).

Patients’ Knowledge of Clinicians’ Expectations of Pain Behaviors

Interviews with patients who had used intravenous patient-controlled analgesia (IV PCA) for opioid administration after surgery revealed that a major reason for patients’ valuing PCA was related to the fact that PCA decreased the need to interact with the nursing staff regarding pain (Hall, Salmon, 1997; Taylor, Hall, Salmon, 1996a, 1996b). Not only did patients see PCA as being better than waiting for a nurse to administer opioids intramuscularly, they also believed it protected them from having to show distress to the nurses.

Apparently many patients with pain are aware of the behaviors expected of them, or they learn them quickly. Patients may learn from early childhood experiences with pain, television, the responses of their clinicians, and a variety of other sources how to behave to signal others that pain exists and that help is needed. It is a common observation that patients with pain change their behavior in the presence of clinicians and in other selected circumstances. Patients may appear to be calm and to read or have lively telephone conversations, but as soon as clinicians enter, patients may replace these activities with a solemn facial expression and may even grimace, moan, and restrict movement—just the behaviors clinicians want to see when patients report pain, but not necessarily the behavior patients prefer.

Consider what might happen if the patients in the Andrew-Robert survey were roommates. It would not take long for smiling Andrew to realize that grimacing Robert was receiving better pain relief, and the reason probably would be quite apparent to smiling Andrew. Andrew might take pride in looking energetic and happy in front of his visitors or may find that such distraction is very helpful as a coping mechanism. But if he wants better pain relief, he may decide to change that behavior, at least during the time clinicians are present. When clinicians see this change occur, they often regard patients as being manipulative, not realizing that it is the expectations that clinicians convey to patients that cause this behavior change.

Taking this hypothetical situation further, by succumbing to expectations of pain behavior to obtain relief, Andrew may begin to jeopardize his recovery. For example, both patients might be told to ambulate. Smiling Andrew may walk the hall until he begins to hurt and then go to the nurses’ station and ask for something for relief. The staff may feel that if Andrew can be this active, he could not hurt enough to require an analgesic. Grimacing Robert, on the other hand, may remain in bed grimacing rather than ambulating and may also ask for an analgesic. Robert may very well be more likely to receive the analgesic than Andrew. Andrew may then learn to stay in bed, prolonging his recovery time and increasing the risk of complications, but increasing his chances of receiving pain relief.

Comparable circumstances were observed in a study of cancer patients (Cleeland, Gonin, Hatfield, et al., 1994). The more active patients were more likely to receive inadequate pain management.

Because clinicians differ in how they expect patients to behave in response to pain, patients may have difficulty learning which behaviors will effectively convince which clinicians. An analysis of staff and patient behavior in an orthopedic unit revealed that the staff assessed pain by observing patients’ behaviors and that their expectations about how patients should express their pain varied within and between shifts (Wiener, 1975). Some, but not all, patients were adept at reading the explicit and implicit cues given by staff and changed their behaviors accordingly. Patients sometimes felt forced to use tactics that they believed were unacceptable but were expected by the staff and were necessary to obtain pain relief.

Expecting patients to behave in certain ways to verify their pain becomes more confusing when clinicians are especially particular about the intensity or type of behavior that a patient should display. Sometimes clinicians say that a patient does not appear to be in pain, but when a patient exhibits pain behavior, clinicians may say that a patient is making too much fuss about the pain.

Correction of the Misconceptions Concerning the Acute Pain Model

One of the most interesting aspects of the misconception that the patient’s behavioral response is more reliable than the patient’s self-report of pain is that it is totally illogical. Virtually every human being has had the personal experience of trying to hide pain and to function in spite of it, smiling when appropriate and even deliberately using humor as distraction. One of the oldest maxims in health care is that laughter is the best medicine. Why couldn’t the nurses and college students responding to the Andrew-Robert survey use this folk wisdom and their own personal experiences with pain to realize that patients with severe pain most certainly may smile and laugh? Nurses are probably the largest group of clinician advocates for nondrug pain relief measures, of which distraction and laughter are highly ranked. And no research has ever even suggested that smiling and joking are incompatible with feeling pain. Why clinicians seem to have such difficulty accepting and acting on reports of pain from smiling and active patients remains a mystery, at least to these authors.

A substantial amount of research refutes the value of the acute pain model, showing that neither behavioral indicators nor physiologic responses are dependably related to the intensity of a patient’s pain. Physiologic indicators appear to be even less valuable than a patient’s behavioral responses to pain.

Most nurses seem to have been taught to use elevated vital signs to assess or verify the presence of pain, especially severe pain. However, in literature reviews, investigators found very little research that supported using physiologic manifestations as specific indicators of pain (Herr, Coyne, Key, et al., 2006; van Cleve, Johnson, Pothier, 1996). In a study of 1063 patients admitted to the ED with conditions that could be verified to be painful, such as fractures and nephrolithiasis, no clinically significant associations were identified between pain scores self-reported on a scale of 0 to 10 and heart rate, BP, or respiratory rate (Marco, Plewa, Buderer, et al., 2006). In a study published in French, investigators found that absence of increased vital signs does not mean absence of pain (reported in Gelinas, Johnston, 2007).

Critical care nurses usually consider vital signs to be relevant indicators of pain, possibly because the signs are readily available, and it is often difficult to obtain pain reports from critically ill patients, whose behavioral responses may be compromised and who may be unconscious, intubated, or otherwise unable to communicate. In a study involving 30 conscious patients in an intensive care unit, no significant relationship was found between physiologic indicators and patients’ self-reports of pain, reinforcing the fact that physiologic indicators should not be considered primary indicators for pain assessment (Gelinas, Johnston, 2007). In other words, physiologic indicators were not related to patients’ reports of pain.

In another study of 755 patients, primarily in intensive care units, physiologic responses were monitored while the patients were undergoing tracheal suctioning (Arroyo-Novoa, Figueroa-Ramos, Puntillo, et al., 2007). Statistically significantly higher increases were noted in heart rate and systolic and diastolic BP, but these changes were not clinically significant. The authors suggest that methods of measuring these physiologic parameters may not be sensitive enough to capture the response to acute pain. For further discussion of the limited usefulness of physiologic measures in assessing pain, see the discussion of patients who are critically ill in Chapter 4, pp. 143-147.

The usefulness of physiologic measures of pain is further compromised by the presence of dementia in older adults. In a small study of adults 65 years old and older, 50 of whom were cognitively intact and 44 of whom had varying degrees of dementia, heart rate and other physiologic responses were monitored before and during venipuncture (Porter, Malhotra, Wolf, et al., 1996). Increasing severity of dementia was associated with the blunting of physiologic responses as measured by diminished heart rate increase in the preparatory phase and during venipuncture.

Thus, little research supports vital signs as being relevant indicators of pain, although they can be used as indicators of the need for further assessment of pain. A position statement by the ASPMN regarding assessment of pain in nonverbal patients (available at http://www.aspmn.org/Organization/documents/NonverbalJournalFINAL.pdf) suggests that physiologic indicators should not be used alone to assess pain but should be considered cues for further assessment of the possibility of pain (Herr, Coyne, Key, et al., 2006).

Considerable research demonstrates that the behavioral expressions of pain that clinicians expect to see in patients are often absent. In the late 1970s, investigators interviewed 102 adult patients with various types of pain, acute and persistent (Jacox, 1979). Many patients did not report pain and made strong efforts to conceal it. When the patients were asked whether they discussed pain with others, 70% said no or were ambivalent. When they were asked how they responded to pain, 66% said they tried to remain calm and not show pain. More than a decade later a similar finding was reported in a study of 45 patients with pain related to lung cancer (Wilkie, Keefe, 1991). Of the patients, 42% revealed that they coped with pain by trying not to let others know about it (Wilkie, Keefe, 1991).

Some researchers have attempted to use facial movements as a method of studying malingering. A review of the literature found that the results of these studies were inconsistent and concluded that it is unclear whether facial expressions of pain can be used as a reliable method for identifying malingering (Fishbain, Cutler, Rosomoff, et al., 1999). A few years after this literature review, facial expressions were studied in 40 patients with low back pain who were videotaped during rest and painful straight-leg raises (Hill, Craig, 2002). During painful movement they were asked to express their pain genuinely or to pretend that it did not hurt. Without moving, they were asked to fake pain. Although distinctions could be made between faked and genuine painful facial expressions, this research confirmed the difficulty of discriminating between them.

Patients with pain may deliberately engage in certain behaviors that are incompatible with those of the acute pain model but are helpful in coping with pain. In a study of 13 patients with pain related to advanced cancer, the patients reported that behaviors they used to control their pain included watching television (9 patients) or chatting with family and friends (Wilkie, Lovejoy, Dodd, et al., 1988). A questionnaire survey of 53 patients with persistent cancer pain asked them to identify and rate the effectiveness of the self-initiated, noninvasive pain control measures they used to cope with their pain (Fritz, 1988). Patients rated laughing as being the most effective.

Recently, researchers have identified that patients actually smile during painful situations, yet they did not find one tool for behavioral observation to assess pain that included smiling (Kunz, Prkachin, Lautenbacher, 2009). The facial expression of pain has been studied in both experimental and clinical research, and facial expression is often a part of behavioral observation tools to assess pain in both children and adults who cannot self-report. When these researchers viewed several videotapes of patients undergoing painful stimulation, one of the most unexpected findings in response to pain was the oblique raising of the lip that results in a smile. Their research revealed that the percentage of patients who smile at least once during painful stimulation ranges from 22% to 57%. A satisfactory explanation for this requires more research, but clinicians should not discount pain just because the patient smiles. Among other things, the smile could represent embarrassment over other behavioral responses to pain, an attempt to mask feelings of pain, or a willingness to endure the pain.

Sleep may be mistakenly equated with lack of pain, but even patients with severe pain may sleep. Some patients use sleep to help control their pain (Wilkie, Lovejoy, Dodd, et al., 1988). In one study, 100 patients were interviewed about the experiences of pain and sleep following abdominal surgery (Closs, 1992). Pain was the most common cause of sleep disturbance at night, demonstrating that pain occurs during sleep. Analgesics helped more patients get back to sleep than any other intervention. Also, about half of the patients felt that pain was worse at night.

Further, an appreciation of the fact that sleeping patients may have pain is demonstrated in analgesic research. When the effectiveness of analgesics is studied, trained observers ask patients to rate their pain at specific intervals, such as every hour, after the administration of the analgesic. When the observer finds the patient asleep at the time a pain rating is required, the observer awakens the patient to obtain a pain rating (Forbes, 1991).

This information does not necessarily indicate that in clinical practice sleeping patients should always be awakened to assess pain. If a patient has been given an analgesic and assessment of pain ratings afterward has shown that the analgesic has been effective, when the patient is given further doses, there usually is no need to awaken the patient to assess pain rating following a dose. Pain assessments can be made when the patient awakens. However, it may be wise to awaken the patient for the next analgesic dose if the analgesic lasts only 4 hours and the patient has continuous pain. This option can be discussed with the patient, explaining that if he or she is allowed to sleep beyond 4 hours, the pain may return to awaken him or her. If this happens, pain control is jeopardized. The patient must notify the nurse or obtain the analgesic on his or her own, take the analgesic, and then wait for it to be effective. In this scenario, a patient’s sleep is interrupted for longer than it would have been had the patient been awakened at 4 hours and given an analgesic before pain returned. Once this is explained to a patient, he or she may opt to be awakened or to wait to be awakened by pain. If the latter is chosen, a patient should be cautioned to notify the nurse immediately so the analgesic can be given. (If a patient’s sedation levels needs to be assessed, then it may be wise to awaken the patient. See the discussion of sedation levels on pp. 510-511.)

Very little research has been done to evaluate pain control using around-the-clock (ATC) dosing of analgesics versus administering them as needed, and the results are often inconclusive. However, one small study compared these two types of dosing schedules in 35 patients following abdominal surgery (De Conno, Ripamonti, Gamba, et al., 1989). Patients who received ATC analgesia had significantly better pain control, slept longer, and spent more time out of bed rather than lying down over the first postoperative week. Another study of medical inpatients compared ATC dosing of opioids to a control group and found that patients with ATC dosing reported lower pain intensity ratings than those reported by the control group yet did not take higher doses than the control group (Paice, Noskin, Vanagunas, et al., 2005).

Sedation is erroneously equated with analgesia. However, in a study of sedation and pain relief in the postanesthesia care unit, researchers found that opioid-induced sedation did not ensure adequate self-reported pain relief (Lentschener, Tostivint, White, et al., 2007). About half of the 26 patients who experienced opioid-induced sedation had persistently high pain scores in the postanesthesia care unit and during the initial 24-hour postoperative period. Further, morphine-induced sedation did not suppress the patients’ memories of early postoperative pain. Still another study confirmed that opioid-induced sedation could not arbitrarily be equated with analgesia (Paqueron, Lumbroso, Mergoni, et al., 2002). Of patients receiving intravenous morphine in whom morphine was discontinued because of sedation, 25% still had pain levels on the Visual Analog Scale (VAS) above 50. (In patients receiving morphine whose pain relief is unacceptable, a multimodal approach to analgesia should be considered; see pp. 9-10.)

Many sedating drugs, such as benzodiazepines and phenothiazines, are given to patients who are experiencing pain, but most of them provide no analgesia, and the resulting sedation may limit the amount of opioid that can be given safely to a patient in pain. Except for pain related to muscle spasm, benzodiazapines do not relieve pain. Further, available phenothiazines neither relieve pain nor potentiate opioid analgesia (APS, 2003).

When a patient’s report of pain is not accepted and acted on, an effective strategy is to ask, “Why is it so difficult to believe this patient has pain?” When the problem revolves around expecting the acute-pain model to exist, the answer is likely to be “He doesn’t look like he’s in pain.” When you find yourself thinking this or hearing someone say it, try asking, “How would this person have to act for us to believe he has pain?” A clinician might answer that a patient in that much pain would grimace or be less active. Unfortunately, expecting a patient to “act like he is in pain” may lead to those behaviors and contribute to manipulative behavior and physical harm such as disability or complications resulting from decreased function.

Lack of Physical Evidence of Pain

A previously suggested strategy for addressing clinicians’ reluctance to accept a patient’s report of pain is also useful here. Once again, try asking, “Why is it so difficult to believe this patient has pain?” If lack of a known physical cause is the reason, the answer is likely to be “There’s no reason for this patient to hurt.” A more accurate and appropriately humble response would be that we are as yet unable to establish the cause of the pain (Teasell, Merskey, 1997).

Statements that may help us reconsider our misconceptions include a reminder that the study of pain is a new and inexact science and that it would be foolish of us to think that we will be able to determine the causes of all the pains that patients report. It also may be helpful to articulate the underlying thought process, which is “We seem to be thinking that if there is pain, there is a cause. If there is a cause, we can find it. If we cannot find the cause, there is no pain.” Once it has been stated, we begin to recognize the absurdity of this idea.

Available assessment tools are not infallible and do not exhaust all possible means for determining the causes of pain. This has been especially true of chronic noncancer pain. In a study of 60 patients with persistent pain who had been referred to a diagnostic center, the overall rate of inaccurate or incomplete diagnosis at referral was 66.7% (Hendler, Kozikowski, 1993). In particular, neuropathic pain, which is often severe burning or shooting pain, tends to be underdiagnosed. It is not detectable by ordinary diagnostic tests because nerves, not muscles or other somatic structures, are involved.

Sometimes the physical cause of pain is known, but the pain is more intense or lasts longer than expected. In the 1980s (when hospital stays were longer than they are now), a study examined the belief that postoperative pain subsides rapidly over the first 3 days and is negligible by the fourth day. Research did not support this. Of 88 patients on a general surgical unit, 31% had pain that persisted beyond day 4, often related to being older or to complications such as infection (Melzack, Abbott, Zackon, et al., 1987). These patients typically received inadequate pain control because less effective analgesics were prescribed. Probably the staff believed there was no reason for these patients to hurt that much or for that long.

Clearly, the causes of pain cannot always be determined. This does not mean that the pain is absent or that clinicians are entitled to ignore a patient’s report of pain. Pain is subjective, and it seems rather easy to engage in faulty reasoning about it. An analogy that may clarify our responsibility to such patients is our response to patients with objective symptoms that have unknown causes. For example, if a patient vomits, we may not know why it has occurred, but because it is objective (an undeniable symptom), we treat it anyway. The cause of the vomiting is sought, but meanwhile treatment, such as antiemetics, is provided. Pain deserves the same respect as objective symptoms.

Belief that Noncancer Pain Is not as Painful as Cancer Pain

When persistent pain is not associated with a terminal illness, especially when the cause of the pain is unclear, that pain seems to be regarded as being more suspicious, less painful, or less in need of relief. Doubting the trustworthiness of the patient with persistent noncancer pain is thought by some to be at the heart of most treatment problems that arise with these patients (Richeimer, Case, 2004). For example, the patient’s motive for seeking care may be under suspicion. The cause of pain may not be identifiable or obvious, and the clinician may be suspicious that a patients is seeking opioids or disability compensation (Victor, Richeimer, 2005). Negative attitudes toward patients with noncancer low back pain have been recognized for many years (Burgess, 1980; Wiener, 1975).

Even when the causes of pain are known, non-life-threatening pain is less likely to be treated than is pain associated with terminal illnesses. In one study of physicians treating patients with cancer, inadequate pain management was more likely when pain was not attributed to cancer (Cleeland, Gonin, Hatfield, et al., 1994). A mail survey responded to by 368 physicians in Michigan asked them to identify their treatment goals for acute pain, cancer pain, pain due to terminal illness, and persistent noncancer pain (Green, Wheeler, Marchant, et al., 2001). Although their goals for pain relief for terminally ill patients and patients with cancer pain were similar, the goals were significantly lower for patients with persistent noncancer pain.

Surveys of laypeople also reveal differing attitudes toward various types of pain. A telephone survey of 1000 Americans asked whether high doses of analgesics should be prescribed for any of these three conditions: severe persistent pain, cancer pain, and rheumatoid arthritis (RA) (The Mayday Fund, 1998). Approximately 80% supported high doses for cancer pain, approximately 70% supported high doses for severe persistent pain, but only about 50% supported high doses for RA. Perhaps both the public and clinicians tend to believe that some types of pain should be tolerated or are not very painful. In any case, the study reflects the tendency to provide less aggressive analgesia for one type of noncancer pain.

However, non-life-threatening conditions can result in very intense and prolonged pain. In a survey of 204 people with persistent noncancer pain, respondents revealed that their average length of time in pain was 9.5 years, with a range of 6 months to 74 years (Hitchcock, Ferrell, McCaffery, 1994). They were in pain an average of 80% of the time. For 30% the usual intensity of pain was severe, with a rating of 4 to 5 out of 5.

A common temporal classification of pain types is acute, cancer, and nonmalignant. Turk (2002) proposes that the mechanisms underlying cancer pain and nonmalignant pain are no different and that it makes no sense to discriminate between cancer pain and nonmalignant pain. He says that to do so results in paying insufficient attention to knowledge gained about cancer pain, and he suggests that greater effort be made to classify and treat pain according to the underlying mechanisms causing that pain as opposed to basing all treatment of persistent pain on whether the pain is caused by cancer. As a consequence, both cancer pain and persistent nonmalignant pain are now more often grouped together and referred to as persistent or prolonged pain.

Treating noncancer pain in the same way as cancer pain does raise legitimate concerns about the prolonged administration of medications. Certainly the chronic use of opioids remains controversial to some extent because of questions about the occurrence of addictive disease, hyperalgesia, hormonal changes, and other outcomes. (See Addiction on pp. 32-42 and Opioid Analgesics in Section IV for discussions of these conditions.) Prolonged use of nonsteroidal antiinflammatory drugs is of even greater concern than the use of opioids because of the possibility of life-threatening gastrointestinal bleeding or adverse cardiovascular events. (See Nonopioid Analgesics in Section III for a discussion of this subject.) Thus, long-term treatment by pharmacologic measures poses many challenges but cannot be ignored as a possible source of relief for persistent pain.

Implication that Anxiety or Depression Is the Cause

In a previously described study, 50 registered nurses completed an instrument that contained 60 vignettes describing patients’ illnesses or injuries (Dudley, Holm, 1984). Nurses were instructed to rate the patients on the degree of their pain and psychologic distress. Nurses inferred significantly less pain than psychologic distress, whereas in actuality, patients were experiencing both. Unfortunately, patients are then likely to receive psychologic support but not the analgesic needed to decrease pain.

When the physical cause of pain is unknown or seems insufficient to account for the severity of pain a patient reports, clinicians may attribute the pain to the patient’s emotional state and cease treating the pain. A comment that suggests this is so is, “The patient is just upset.” It is interesting that in a survey of the public’s attitude toward stress and pain, 95% of respondents agreed with the statement that stress increases pain (The Mayday Fund, 1998). Actually, evidence that stress increases pain is limited.

An erroneous and simplistic view promoted around the middle of the 20th century was that physical and psychologic causes of pain were mutually exclusive; that is, that pain is caused by either organic or psychologic factors (IASP ad hoc Subcommittee for Psychology Curriculum, 1997). Gagliese and Katz (2000) state that Melzack’s gate control theory argues against the simplistic thinking that leads to categorizing pain as being either organic or psychogenic. The authors believe that medically unexplained pain is not caused by psychopathology and that thinking that separates mind and body should be abandoned. Trying to differentiate between psychogenic and physical pain is usually fruitless. Pain caused solely by psychologic factors is rare, as is pain caused solely by physical causes. Most pain is a combination of physical and psychologic factors and is best treated as such. The subsequent discussion concerns the as yet unclear relationships between pain and anxiety or depression.

The International Association for the Study of Pain defines pain as “an unpleasant sensory and emotional experience which we primarily associate with tissue damage or describe in terms of such damage, or both” (Mersky, Bogduk, 1994, p. 210). By definition, pain is always unpleasant and always subjective, so pain is always an emotional experience. But what are the relationships between pain and various emotions? A common assumption is that anxiety or depression makes pain worse, but this is not always true.

One review of the literature found that research findings are conflicting and inconsistent regarding the relationship between depression and pain and between anxiety and pain (Zimmerman, Story, Gaston-Johansson, et al., 1996). Another literature review pointed out that a high proportion of patients with persistent pain have some kind of depressive syndrome; however, the depression may precede, follow, or develop concomitantly with the persistent pain (Dellemijn, Fields, 1994).

For the purposes of discussion, it is helpful to consider how anxiety and depression affect coping with pain as well as how they affect the intensity of pain. Although certain levels of anxiety are actually helpful in mobilizing appropriate coping mechanisms, high levels of anxiety and possibly any level of depression may adversely affect a patient’s ability to cope with pain.

Roughly 50% of patients with persistent pain also have depression or an anxiety disorder (Weisberg, Boatwright, 2007). It has been relatively well established that persistent pain is often the precipitating factor for mood disorders and for anxiety disorders, yet the interaction between persistent pain and such disorders is less well understood (Weisberg, Boatwright, 2007). Depression may adversely affect coping by causing a patient to be less motivated to engage in activities and try new ideas for rehabilitation or treatment.

Pain-related anxiety or pain-related fear may also immobilize patients. Such fear or anxiety may cause hypervigilence and avoidance of activities that a patient fears may cause pain. This concern with avoiding painful activities may result in a patient’s having difficulty freeing attention to focus on nonpainful activities (Vlaeyen, Crombez, 2007). One approach to helping these patients to improve their functioning has been exercise and graded activity programs in which patients gradually increase their physical activities. Another approach is verbal reassurance, but this can have an opposite effect unless done carefully. Simply telling patients not to worry, that they do not have a severe disease, or that their tests have shown negative results may increase their distress. Fearful patients may become more puzzled because they still hurt and no explanations have been offered. Such patients need credible explanations of why they hurt (Vlaeyen, Crombez, 2007).

How anxiety and depression influence pain intensity is an even more complicated issue. Anxiety certainly is associated with many types of pain, and depression is common in patients with persistent pain. However, the cause-and-effect relationship is unclear. Does anxiety cause pain or is it the result of pain? Is depression the cause or the result of pain? Some studies show a relationship between depression and pain, and some do not. Likewise, anxiety has been correlated with increased pain in some studies but not in others. Anxiety unrelated to pain may actually decrease pain, possibly because the anxiety increases the production of endorphins (Janssen, Arntz, 1996).

In a study of 120 patients with cancer pain, no difference was found in pain intensity measures or functional status between the depressed patients and those who were not depressed (Grossman, Sheidler, Sweeden, et al., 1991). In other words, the depression associated with pain did not alter a patient’s report of the intensity of the pain. In another study of patients who had cancer, the presence of depression, hostility, and anxiety did not correlate with the effectiveness of attempts at pain relief (Cleeland, 1984). Cleeland warns us that when patients do not respond to analgesics, clinicians should be wary of blaming this result on the patients’ depression rather than on inappropriate analgesic therapy.

The relationship between preoperative-state and postoperative-state anxiety and pain magnitude has been addressed by a few studies. (State anxiety is situational, whereas trait anxiety is a general level of anxiety.) A study of 96 patients after coronary artery bypass graft surgery examined the relationship between postoperative-state anxiety and the perception of postoperative pain. The results showed that even in the range of low to moderate state anxiety, higher anxiety postoperatively was associated with higher pain intensities (Nelson, Zimmerman, Barnason, et al., 1998). The correlations between postoperative-state anxiety and postoperative pain ranged from low to moderate but were nevertheless statistically significant. However, again, the direction of causality has not been established. In other words, although increased pain and increased anxiety may occur together, it is not always clear whether the anxiety causes the pain or the pain causes the anxiety.

An exploratory study of 24 patients undergoing varicose vein surgery examined the relationship between pain ratings and preoperative state anxiety (Terry, Niven, Brodie, et al., 2007). Pre- and postoperative-state anxiety and pain ratings in the postoperative period were significantly positively correlated. However, the actual postoperative pain ratings were relatively low, with a mean VAS rating of 28.1 mm on a 100-mm linear scale. Nevertheless, the authors point out that the findings show the need to decrease anxiety in the clinical setting whenever possible. Suggestions about how to reduce anxiety were not included.

Another small study of women undergoing termination of pregnancy investigated the links between preoperative-state and trait anxiety and pain magnitude (Pud, Amit, 2005). About 1 hour before the procedure the women completed Spielberger’s State-Trait Anxiety Inventory. Postoperatively, patients were asked to mark a 100-mm VAS at 50, 30, and 60 minutes. State anxiety was able to predict pain magnitude at 15 minutes, and trait anxiety predicted pain magnitude at 30 minutes. The authors explained this finding on the basis that state anxiety was expected to disappear when patients felt the immediate threat had been removed, whereas patients with trait anxiety needed more time to relax. The authors suggest that patients with high anxiety may require additional information relevant to the impending pain along with effective pain-relief protocols.

The belief that anxiety causes pain is reflected in the common practice of combining anxiolytics and opioids. Based on a literature review, Pud and Amit (2005) found that two methods, information giving and relaxation techniques, were beneficial in relieving anxiety and pain but that the use of pharmacologic approaches to anxiety such as benzodiazepines was more controversial. In one study, lorazepam (Ativan) preoperatively did not significantly relieve pain (Wiebe, Podhradsky, Dijak, 2003). The findings in a study of postoperative patients who had access to IV PCA morphine and midazolam (Versed) separately revealed that use of midazolam did not influence pain scores or amount of PCA morphine used (Egan, Ready, Nessly, et al., 1992). In another study of postoperative pain, the administration of diazepam (Valium) preoperatively was shown to have an ongoing antianalgesic effect on morphine analgesia (Gear, Miaskowski, Heller, et al., 1997). In a double-blind study, combinations of various doses of midazolam and meperidine were administered to 150 patients with postoperative pain and, once again, midazolam did not significantly enhance the analgesic effect of the opioid (Miller, Eisenkraft, Cohen, et al., 1986).

No doubt exists that pain results in considerable distress for many patients, causing anxiety, depression, and hostility and interfering with all domains of quality of life (e.g., Ferrell, Grant, Funk, et al., 1997; Zimmerman, Story, Gaston-Johansson, et al., 1996). Until the relationships among pain and anxiety, depression, and other emotional states are clarified, the most practical initial approach to patients who are both in pain and anxious or depressed probably is to assume that pain causes these emotional responses rather than to assume that the emotional responses cause or intensify pain. Anxiety and depression appear to be normal responses to pain. When a patient is both in pain and anxious, initial intervention probably should be aimed at reducing the pain. The APS (2003) states, “In anxious patients with pain, opioid titration should precede treatment with benzodiazepines” (p. 46). Pain relief may well reduce the anxiety and minimize the need for a benzodiazepine. Likewise, for patients who are in pain and depressed, the most logical initial approach probably is to relieve the pain, which may then reduce the depression. If anxiety or depression persists following pain relief, other interventions, such as behavioral and pharmacologic approaches, are indicated.

Seeking Drugs versus Seeking Pain Relief

Perhaps the most common reason for not accepting and acting on a patient’s report of pain is the belief that the patient is or will become addicted to an opioid. A previously suggested strategy for addressing clinicians’ reluctance to accept a patient’s report of pain is to ask, “Why is it so difficult to believe this patient has pain?” Suspicion of addiction is often the answer. Clinicians may say, “The patient is drug seeking,” “He just wants drugs,” or “He’s getting addicted.” At present, in many clinical settings, the term drug seeking, though poorly defined, is used to mean addiction to opioids, abuse of opioids, or manipulative behavior (McCaffery, Grimm, Pasero, et al., 2005). Although patients seek drugs for many legitimate purposes, such as treating an infection, diabetes, or heart disease, somehow the term drug seeking has become associated with seeking opioids for reasons clinicians believe are inappropriate.

Questions, discussed later, that may be helpful in clarifying the confusion that surrounds beliefs about addiction are:

Definitions Related to Addiction

Tolerance to opioids and physical dependence on opioids are not the same as addiction to opioids, but these three terms are often confused. Following are the definitions proposed by the American Academy of Pain Medicine (AAPM), the APS, and the American Society of Addiction Medicine (ASAM) (AAPM, APS, ASAM, 2001; APS, 2003):

Each of these conditions—physical dependence, tolerance, and addictive disease—is a separate entity requiring different treatment. One may occur alone, any two together, or all three together. Physical dependence and tolerance are a result of repeated administration of the opioid and should be expected if regular daily doses of opioids are taken for 2 to 4 weeks or longer. Addiction, however, is a rare consequence of using opioids for pain relief.

Pseudoaddiction, as the name implies, is a mistaken diagnosis of addictive disease. The term was first used and the behaviors described by Weissman and Haddox (1989). Patients with undertreated pain may manifest behaviors very similar to those typical of addictive disease, such as escalating demands for more or different medications, repeated requests for opioids before the prescribed interval between doses has elapsed, illicit drug use (such as obtaining opioids from others), and deception (APS, 2003). Pseudoaddiction can be distinguished from true addictive disease in that the behaviors resolve when pain is effectively treated.

Likelihood of Occurrence of Addictive Disease

To address concerns about the likelihood of the occurrence of addictive disease in patients who receive opioids for pain relief, two separate questions must be addressed:

Addiction as a Stigmatizing Label: Societal versus Medical Views

An understanding of some of the stigma clinicians associate with addiction arises from confusion related to society’s general views of addiction (the moral and criminal model), which are at odds with the validated view of addiction as a disease (the disease model; Table 2-2) (Compton, 1999). Taken to its extreme, the moral model of addiction views substance abuse as a shameful behavior in which good and moral persons do not participate. It is seen as a behavior that is the individual’s fault and that an individual can freely choose that behavior or not. Because addicts in this model appear unwilling to adhere to the so-called virtuous path, they are viewed as being at fault for any negative consequences arising from their drug use and commonly are considered less worthy of care and concern than patients suffering a “faultless” disease (Malone, 1996).

Somewhat different, but resulting in the same negative, judgmental view of addiction, is the criminal or criminal justice model. It is illegal in this country and many others around the world to use certain drugs for the purpose of experiencing their psychoactive effects, so persons possessing or ingesting drugs are criminals and are subject to punishment and imprisonment (Compton, 1999). The predominance of the criminal justice view of addiction is reflected in the high percentage of Americans incarcerated in federal (59.5%) and state prisons (22.3%) for drug offenses (Office of National Drug Control Policy [ONDCP], 1997). Although the use of alcohol is legal, laws against drunk driving and public intoxication criminalize excessive alcohol use too. It is interesting that addiction is the only recognized disease for which its sufferers risk arrest rather than medical care when impaired by its symptoms. Thus, in the criminal justice model, like the moral model, persons suffering addictive disease are characterized as bad, unethical, and deserving of criminal justice or social sanctions.

Unlike the moral and criminal justice models, the medical model views addiction as a disease, specifically as a disease acquired by the brain. Like any disease, addiction has identifiable risk factors and a demonstrated pathophysiologic basis, can be diagnosed according to a well-described cluster of signs and symptoms, follows a predictable pattern of progression, and can be managed and treated. Unlike the moral and criminal justice models, the medical model of addiction places no blame or negative social value on the sufferer. The focus is on understanding the factors that result in the disease’s development and progression and on evaluating methods of prevention and treatment. Punishing and stigmatizing addicted individuals are not viewed as being effective methods of intervention for addiction in the disease model, whereas treatment approaches, including pharmacotherapy, cognitive-behavioral skill training, group support, and lifestyle change are (Compton, 1999).

Unfortunately, drug abuse and addiction often are treated as crimes rather than as public health problems (Cotton, 1994). Consequently, clinicians judgmentally regard addicts as bad people rather than focusing on addiction as a bad disease that requires treatment. So the label of addiction is stigmatizing and should be avoided. The term addiction should be used only to refer to a brain disease that requires appropriate treatment.

When a patient is referred to as being addicted or drug seeking, it is helpful to ask, “Has the patient been diagnosed as being an addict? If so, how does this affect our plan for pain relief? How do we plan to treat the addictive disease?” If a patient is addicted, he or she still deserves the best possible pain relief that can be provided safely. The following position is unethical: I can provide safe and effective pain relief for this patient, but I will not because I do not believe the patient deserves it.

Misconceptions about the Meaning of Positive Placebo Responses

Occasionally, when the question is asked, “Why is it so difficult to believe this patient has pain?” the answer is that the patient reported satisfactory pain relief after receiving a placebo. Pain relief resulting from a placebo is mistakenly believed to invalidate patients’ reports of pain. This may result in the patient being deprived of pain-relief measures.

Common misconceptions about pain relief resulting from a placebo are that patients are malingering (lying about pain) or the pain is psychogenic rather than being physical in origin. The fallacy of this is apparent in the results of research using placebos as controls in the evaluation of potential analgesic drugs. Beginning in the 1950s, numerous placebo-controlled analgesic studies were conducted with postoperative patients. Thus we know the answer to this question: What percentage of patients will report adequate relief of pain after having received a placebo injection the day after abdominal surgery? The answer commonly quoted is 36% (Beecher, 1955; Evans, 1974; Goodwin, Goodwin, Vogel, 1979); however, no fixed percentage of the population responds to placebos (Wall, 1992).

In fact, researchers find that the belief that overall, the positive placebo effects occur approximately one third of the time in any individual circumstance is incorrect (Moerman, Harrington, 2005). Beecher’s (1955) study (as reported in Turner, Deyo, Loeser, et al., 1994), which proposed that on average, one third of the time a placebo would relieve pain, was based on a review of 15 studies of patients suffering a variety of painful conditions; and the placebo response rate ranged from 15% to 58%. In trials of antisecretory medications for peptic ulcer disease, one study examined the 4-week endoscopically verified healing rates in 117 control groups and showed that the placebo response ranged from 0% to 100% (Moerman, 2000).

If, on average, one third of patients who have obvious physical stimuli for pain (abdominal surgery) report pain relief after a placebo injection, clearly placebos cannot be used to diagnose malingering, psychogenic pain, or any psychologic problem and cannot be used to justify denying the patient pain relief. Quite simply, all we can conclude is that sometimes placebos are effective in relieving moderate to severe pain of known physical origin.

Although it is well documented that positive placebo responses occur, why placebos relieve pain remains poorly understood. Many theories have been proposed, including operant conditioning, faith, anxiety reduction, expectation, and endorphin release (Finniss, Benedetti, 2007; Grevert, Albert, Goldstein, 1983; Levine, Gordon, Fields, 1978; Lidstone, de la Fuente-Fernandez, Stoessl, 2005; Price, Chung, Robinson, 2005; Richardson, 1994). Yet none has been proved, and the mystery remains. It is interesting that in studying the role of endogenous opioids in producing a positive placebo response, researchers have demonstrated that placebo-induced respiratory depression may occur and can be reversed by naloxone (Finniss, Benedetti, 2005).

Misconceptions about Using Placebos for Pain Relief

Because placebos can provide pain relief in a rather large portion of patients, why wouldn’t they be a legitimate option for pain relief? One reason placebos are not appropriate as pain treatment is that a positive placebo response cannot be predicted in an individual, and a placebo may be effective for one person at one time and not another (Sullivan, Terman, Peck, 2005). Although deceitful use of placebos is not recommended, efforts to enhance legitimate therapies with positive placebo effects is encouraged (Brody, 1997). This may be as simple as saying to the patient when an injection of morphine is administered, “This medication usually is a very effective way to relieve pain, and we can adjust it as necessary.” Placebo effects occur with both inert substances such as saline injections and with proven analgesics such as opioids.

Another reason for not using placebos for pain relief is that they may produce harm as well as benefit. Research shows that placebos may mimic active drugs, may produce adverse effects (such as respiratory depression, as mentioned earlier) and toxic reactions, may worsen symptoms (negative placebo effects), and may directly affect many body organs (Benedetti, Amanzio, Baldi, et al., 1998; Lavin, 1991; Wolf, Pinsky, 1954). Therefore, substituting a placebo for an opioid cannot be justified on the basis of avoiding adverse effects, such as sedation or preventing addiction.

Unfortunately, placebos are most likely to be used in patients the staff regard as being difficult. The most extensive study of placebo use was undertaken in the 1970s (Goodwin, Goodwin, Vogel, 1979). Researchers surveyed placebo knowledge of and use by 60 physicians and 39 nurses in a university teaching hospital. Most of the physicians (78%) had ordered placebos, and most of the nurses (82%) had administered them. The reasons given for administering placebos revolved around proving the patients “wrong” and administering to patients who were disliked, who did not obtain pain relief by standard treatment, or who had frustrated or angered the staff. Not only does this deceitful use of placebos in place of appropriate therapy violate patients’ rights to the highest quality of care possible, it clearly poses a moral, ethical, and professional danger to the clinicians.

Perhaps the most important reason for not using placebos in the assessment and treatment of pain is that deception is involved. The clinician lies to the patient. Deceit is harmful to both patients and health care professionals. When discovered (and it usually is), it may permanently damage a patient’s trust in health care professionals.

Legal and ethical considerations involved in the deceitful use of placebos in pain management include liability for fraud, malpractice, breach of contract, and medical negligence (Fox, 1994). Nurses should examine their nurse practice acts and the policies of their state boards of nursing for guidance. For example, in 1997, the Board of Registered Nursing in California specifically stated that the use “of placebos for management of pain would not fulfill informed consent parameters” (p. 12).

Placebo use violates the rights of all patients who receive them outside of the context of approved clinical trials in which patients have given informed consent to participate. Even under these circumstances, placebo use may not be justified when the drug being studied could be evaluated by comparing it with an effective medication instead of a placebo (Hampton, 2006; Rothman, Michels, 1994; Sullivan, Terman, Peck, 2005).

There are no individuals for whom and no conditions for which placebos are the recommended treatment. Literature review shows no basis for the assumption that placebo pain medication is useful to patients (Kleinman, Brown, Librach, 1994). Therefore, the use of placebos inevitably deprives patients of more appropriate methods of assessment or treatment. In other words, there is always a better way to assess and treat patients with pain than to administer a placebo.

Recommendations against the Use of Placebos

Avoidance of deceitful use of placebos is recommended in the APS clinical guideline “Principles of Analgesic Use in the Treatment of Acute Pain and Cancer Pain”: “Do not use placebos to assess the nature of pain. . . . The deceptive use of placebos and the misinterpretation of the placebo response to discredit patients’ pain reports are unethical and should be avoided” (APS, 2003, p. 37).

Nursing organizations have issued placebo policy statements that strongly recommend against the use of placebos in pain management. The Oncology Nursing Society (ONS) (McCaffery, Ferrell, Turner, 1996) developed the first position statement. Soon afterward, 27 or more professional organizations, such as the National Association of Orthopaedic Nurses (NAON), endorsed the statement. The ONS position statement focuses on patients with diagnoses of cancer, but the article accompanying its publication verifies that the facts and ethical issues presented are equally applicable to other patients with pain, regardless of their diagnoses.

The ASPMN (2004) adopted a position statement on placebo use similar to that of ONS, the primary difference being that the opposition to placebo use in pain management applies to all patients, regardless of age or diagnosis. The ASPMN’s position statement says, “placebos should not be used by any route of administration in the assessment and/or management of pain in any individual regardless of age or diagnosis” (p. 1). Box 2-2 summarizes the key elements to consider concerning the use of placebos in clinical practice.

Debate about Use of Placebos in n-of-1 Trials

In 2005, the APS issued a position paper about placebo use that was in agreement with the statement by the ASPMN except for one thing—to find the least invasive or damaging treatment for patients. The APS states that “trials involving alternation between active and placebo treatments in a single patient (known as “n-of-1 trials”) might provide clinically useful and scientifically valid information” (Sullivan, Terman, Peck, et al., 2005, p. 216). n of 1 literally means the number of subjects in the trial is 1. In such cases, treatment decisions might be made on the basis of a positive placebo response, and a patient might be denied active treatment. The APS states that ethically, such treatment requires a patient’s consent, so a patient should be informed prior to the treatments that a placebo will be used as one of the treatments.

As an example, a patient who is being evaluated for an anesthetic block or injection would be told that a placebo would be used, and then the patient might receive a placebo injection of saline and later an injection of local anesthetic. If the patient responded positively to the placebo, active treatment using a local anesthetic might be denied. This does not seem consistent with other parts of the APS position that include the statement that “placebo response does not prove that pain is not ‘real’ or unworthy of any real treatment” (Sullivan, Terman, Peck, 2005, p. 216). The APS also states that placebo use might “delay specific treatments, thereby leading to disease or symptom escalation” (Sullivan Terman, Peck, 2005, p. 216). Thus, the APS seems to contradict its own statement about making a decision to withhold treatment based on a positive placebo response.

The ASPMN does not agree with the APS’s support of the n-of-1 trial for some of the same reasons stated earlier by the APS as well as for other reasons (McCaffery, Arnstein, 2006). The ASPMN’s position statement on use of placebos in pain management (2004) says that the ASPMN “supports the use of placebos only in Institutional Review Board (IRB)-approved clinical trials” (p. 1). Further, this policy states that “placebo use for the assessment and/or treatment of pain, including the evaluation of response to pain treatments, constitutes fraud and deception” (p. 4). The ASPMN agrees with the need to identify patients who will not benefit from treatments that are irreversible or invasive, but it does not agree that the n-of-1 trial accomplishes this. There is no research that provides strong support for this method of determining whether to provide a patient with active treatment for pain relief.

N-of-1 trials pose several problems. The sequence of injections (placebo versus local anesthetic) is not random; clinicians are not blinded as to which is being injected; the trials are not done with the oversight of an IRB; and the effects when a placebo is administered might be caused simply by the puncture or the instillation of fluid (McCaffery, Arnstein, 2006). Also, patients may respond positively to placebos once and never do so again. Further, patients who do not get pain relief from the local anesthetic the first time it is administered might respond later after proper dosing and titration. And, finally, the placebo response tends to be more common and more pronounced with highly invasive procedures such as spinal nerve blocks. In one study of seven patients with complex regional pain syndrome who were receiving a nerve block with saline followed by a local anesthetic, significant pain relief occurred after saline in six of the seven patients (Price, Long, Wilsey, et al., 1998). Thus, in n-of-1 trials involving invasive procedures, the positive placebo effect is very likely to occur and would not seem to be an adequate reason for denying patients active treatment.

An even greater danger exists when n-of-1 trials are supported. It opens the door to uses of placebos in situations other than invasive procedures. For example, hospitalized patients with cancer pain caused by widespread metastases might be taking large doses of opioids and, under the conditions of an n-of-1 trial, could be given placebos. If patients reported pain relief, the opioids might be discontinued (McCaffery, Arnstein, 2006).

Institutional Policies Concerning Placebos

For the protection of patients and staff, written institutional policies about placebo use appear to be essential. The current frequency of deceitful or inappropriate placebo use in the assessment and treatment of pain is unknown, but it undoubtedly exists. Articles on this topic have appeared regularly through the years in the professional literature, and misunderstandings persist (Arnstein, 2006; Fox, 1994; Fox, McCaffery, Ferrell, et al., 1995; Grace, 2006; Haddad, 1993; Kleinman, Brown, Librach, 1994; McCaffery, Arnstein, 2006; McCaffery, Ferrell, 1994a; McCaffery, Ferrell, 1997b; McCaffery, Ferrell, Pasero, 1998; McCaffery, Pasero, 1995; Oh, 1994; Wall, 1993).

Surveys of nurses reveal that they are susceptible to becoming participants in deceitful use of placebos. Surveys completed by 601 nurses during 1995 revealed that a significant percentage of these nurses exhibited a high level of readiness to participate in deceitful placebo use (McCaffery, Ferrell, 1996b). About one of five nurses (19%) thought pain was a clinical problem for which placebos could be used. Even more respondents (27%) endorsed placebo use for anxiety; 25% believed placebos were appropriate in patients whose symptoms (pain) were questionable. Clearly, if placebos were suggested or prescribed, many nurses would participate in the action.

All hospitals and other health care agencies should have written policies that address unethical use of placebos. When policies are in place, communication among health care providers is less difficult. If placebos are prescribed, the nurse may use the existing placebo policy as a point of departure for communicating concerns about the patient’s quality of care and for explaining why a nurse will not administer placebos, even when a prescription exits.

The ASPMN’s (2004) position paper concerning placebos states that a policy is necessary to ensure that no patient receives a placebo unless it is in the context of an IRB-approved clinical trial. This should include mechanisms for reporting a clinician who violates the policy and the actions to be taken to censure such an individual. The policy should also provide a means of protecting any person who reports inappropriate use of placebos or refuses to participate in the use of placebos outside the context of an IRB-approved clinical trial. Other guidelines for developing a policy may be found in the ASPMN (2004) position paper and in Box 2-2.

Other Hidden Biases and Misconceptions

Biases and misconceptions that lead to inadequate pain management are not new. They are evident in recent history. In the late 1800s, many physicians and scientists believed that cultural and intellectual development were accompanied by increased sensitivity to pain (de Moulin, 1979). In 1892, Mitchell said, “The savage does not feel pain as we do.” Mitchell suggested that increased capacity for feeling pain began at the end of the 18th century. Van den Berg (1963) was more exact, stating that pain sensitivity increased between 1780 and 1845. Racial and class prejudices were also apparent. In the Old South of the 1700s and 1800s, black people were considered insensitive to pain and underwent limb amputation without anesthesia. However, upper-class white women were believed to have “delicate nerves,” and the same surgeon would use ether or chloroform to deliver a baby (Pernick, 1985). No doubt many of these physicians believed that they were practicing medicine in professionally and ethically acceptable ways. Clinicians today probably believe the same thing about their clinical practices, but what will be said of our pain management in the year 2099?

Do we know when our biases are showing? Following is a brief overview of research findings related to some biases that are likely to escape conscious awareness but influence clinical practice: