Commentary: “Managing Pain Medications”: An Early Text in Pharmaco-Psychology

Mark Sullivan

INTRODUCTION

Very few clinicians in the 1970s were thinking about medication administration in the behavioral terms that Fordyce uses in this chapter. Fordyce’s primary innovation in this book was to apply the principles of operant conditioning to the care of patients with chronic pain. In this chapter, he applied operant conditioning principles to the management of pain medications. Although Chapter 9 is titled, “Managing Pain Medications”, its mission is narrower than this. It essentially reviews a strategy for shifting from pain-contingent to time-contingent dosing of analgesic medications and then tapering off these medications using a blinded pain-cocktail strategy. Fordyce aims this strategy at a broad population of opioid users: “This chapter concerns patients with medication problems, that is, addiction, habituation, evidence of a currently high level of medication intake, or history of previous heavy medication intake.” This is a very diverse clinical population, encompassing patients with opioid use, misuse, and abuse that we now think require quite different responses. For example, for patients who are abusing opioids and meet DSM-5 criteria for an Opioid Use Disorder, a complete opioid taper is largely futile. It has been shown that only a small minority of patients truly addicted to either prescription opioids or street heroin can remain abstinent from opioids. Even with a buprenorphine-assisted taper and counseling less than 10% of addicted patients are able to remain off opioids [13]. Opioid taper may be more successful for patients with opioid misuse and/or high dose use. We have a randomized trial underway to determine whether support with psychopharmacology and pain self-management training helps patients taper off their opioids [9]. In what follows, I will comment on Fordyce’s practical innovation of the time-contingent pain cocktail taper strategy and then his theoretical innovation of adding behavioral considerations to pharmacological considerations in opioid tapering.

EXPERIENCE WITH IN-PATIENT PAIN-COCKTAIL TAPER AT THE UNIVERSITY OF WASHINGTON MEDICAL CENTER

I was lucky to work with Bill Fordyce for a couple of years at the end of his career in the late 1980’s. We screened patients for our 3-week in-patient pain rehabilitation program that he had designed in conjunction with John Loeser. Our medical and psychological screening visits took all morning. They involved reviewing all the medical records sent with the patient and then taking a history and conducting a physical examination. The psychological evaluation involved interviews with patient and spouse (or significant other) separately. The patient would complete the MMPI which Bill would interpret. We then had a joint feedback session with patient, spouse, psychologist and physician. In this session we reviewed any need for further medical testing. We also determined whether the patient seemed a good candidate for our in-patient pain program. This was largely determined by the absence of unresolved medical issues and the patient’s willingness to suspend the search for a diagnosis and cure for their chronic pain problem and engage in a rehabilitation approach to their problem.

About half of the patients we screened were eventually admitted to our behavioral pain program. Many of these were injured workers sent by our state worker’s compensation system, the Washington Department of Labor and Industries. Bill was no longer working as a psychologist on the in-patient pain service by the time I started attending there in 1988. But we did follow the protocol that he lays out in this chapter for managing pain medications on our in-patient pain service.

Since patient reports of pain medication use prior to admission were thought to be unreliable, we allowed patients to use as much medication as they needed or desired during a 3 day “baseline period.” This baseline period had a medication strategy that was the exact opposite of what was to follow. “The patient and all relevant nursing personnel should then be informed that the patient is to remain on a strictly prn regimen for the baseline period.” Patients were allowed to ask for as much opioid as they wanted or needed. No effort was made to distinguish between needing the opioids and merely wanting them. Though Fordyce does not mention it in this chapter, I remember the only requirement for patients being that they needed to walk to the nurses’ station under their own power to request more medication. This was to prevent patients lying in bed asking for medication to maintain themselves in a state of obtundation. It is important to remember that many of the patients that we admitted in those days were severely deconditioned, if not bedridden, prior to admission, so getting to the nursing station was a challenge for some of them. We thought the daily opioid doses ingested by patients during this baseline period were quite high, up to 100 mg of morphine or the equivalent. But these doses pale in comparison to the opioid doses achieved in the early 2000s. The medical director of Medicaid for the State of Washington consulted with us in 2007 about how best to reduce opioid doses and risk for the over 300 Washington Medicaid clients on over 1000 mg of morphine or equivalent per day. By then, we did not have the luxury of admitting these patients to the hospital for opioid taper and detoxification. It turned out to be very difficult to taper these patients, who had high levels of medical and psychiatric comorbidity, without destabilizing them and prompting visits to the hospital or emergency department.

At the end of the medication baseline period, the patient’s average daily medication intake was calculated. “The 24-hour totals and the intervals between medication requests are averaged…. The guiding rule is to identify what kinds, amounts and frequencies of medications produce an optimal level of patient comfort….” Thus the baseline phase of Fordyce’s tapering strategy is purely pain-contingent. It allows the patient as much control as possible in determining how much opioid they take. “The patient is reassured that 24-hour totals will match or slightly exceed what he or she has been taking.” Once this liberal initial dose is established, the behavioral medication strategy shifts to a purely ‘time-contingent’ model where the clock, and not the patient’s pain level, dictates when opioids are administered. “The cocktail is taken at fixed time intervals, around the clock. The cocktail is not pain contingent.” We used a liquid pain cocktail containing methadone administered four times daily. The cocktail also included phenobarbital (if the patient was on sedatives; later clonazepam) [10], hydroxyzine, and acetaminophen. The hydroxyzine and acetaminophen patients were informed that active medications in the cocktail would be gradually tapered to zero. They agreed to not being informed about the exact doses of active ingredients on any given day.

This shift to a time contingent strategy had at least two purposes. First, it assured that taper was possible and not disrupted by pain increases that delay or abort the taper. Second, it was consistent with the overall “operant” focus of the program that prioritized increasing the capacity for activity over decreasing pain levels. “The objective in fading medications is not simply to eliminate potentially toxic or addictive agents from the patient’s regimen at the most rapid rate consistent with the avoidance of side effects from acute withdrawal symptoms, seizures, and the like. The objective is to provide opportunity for learning.” The learning that was encouraged concerning medications mirrored that of the overall operant behavioral program. Environmental reinforcements for pain behaviors were gradually removed in order to promote extinction of these behaviors. In the case of medications, the goal was to extinguish pain-contingent medication seeking behavior.

In the context of our inpatient pain program, the shift to a time-contingent medication strategy was merely a means toward complete taper off opioid medication [2]. Soon after the patient was placed on the timed pain cocktail, the dose started to be tapered, generally at a rate of 10% of the original dose per day. This is somewhat faster that the regimen described by Fordyce in this book: “The fading process seems to work better when carried out in a more gradual way. The period during which high or relatively high medication intake levels are reduced to zero, or close to it, appears from clinical experience to be approximately 7 to 10 weeks.” During this interval, the pain rehabilitation service had been moved from the inpatient rehabilitation unit where the program lasted 8 weeks to a dedicated pain unit where the program lasted 3 weeks. However, the goal of complete taper off opioids remained the same throughout this period [12].

Even the end of the taper was used as a learning opportunity for the patient. “The patient should be so informed when active ingredient levels in the cocktail finally reach zero. It is prudent to wait for 1 or 2 days to allow a demonstration of adequate functional performance without assistance from analgesics and the like…. The patient merits congratulations …” I remember waiting to inform the patient that their pain cocktail contained no active ingredients until they had a successful day in physical therapy. This allowed us to frame the cocktail taper as a success and to boost the patient’s confidence about being able to function in the future without taking opioids. This has been a successful strategy in my experience. Earlier this year, I successfully tapered a young woman with Asperger’s syndrome off opioids using the blinded methadone pain cocktail method. It took approximately 6 months and needed to be paused a number of times, but the blinding was valuable. She had tears of joy when she was told that she had functioned effectively for two weeks when her pain cocktail had included no opioids.

Bill Fordyce built his reputation as a behavioral psychologist and I remember him being somewhat skeptical of the benefits of psychopharmacology. However, in his instructions for pain cocktail taper, he singles out antidepressants as the sole psychotropic medication that should not be tapered. “The major and most common exception to the foregoing fading pattern concerns antidepressant medications.” In the nearly forty years that have passed since the publication of this book, antidepressant analgesia and the high rates of depression in patients with chronic pain have been thoroughly documented and described. Antidepressants may, in fact, be essential adjuncts to the effort to taper patients off opioids and benzodiazepines.

UNIQUE BLEND OF BEHAVIORAL PSYCHOLOGY AND PHARMACOLOGY BEHIND FORDYCE’S RECOMMENDATIONS

From Pain Contingent to Time Contingent Opioid Use

The operant conditioning paradigm provided Fordyce a very specific target for reform of patients’ medication regimens. This paradigm focused attention on the link between behavior and reward. Opioids and sedatives are reinforcing in most animal and human conditioning models, so it is easy to understand how they may function as rewards for behavior. Even though physicians see themselves as treating pain with opioids rather than rewarding pain behavior with opioids, the effect is the same. Fordyce sought to break the link between pain behavior and medication rewards by replacing pain contingent medication regimens with time-contingent medication regimens. This aimed to “… break the pain behavior-medication consequence relationship through extinction.” If the pain behavior was no longer rewarded with medication, the theory suggested that this behavior would gradually extinguish for lack of reward.

Fordyce thought that simply removing the reward of opioids did not provide the opportunity for learning that breaking the link between pain behavior and opioids did. “One way to proceed would be to simply withhold all pain medications, to compel the patient to quit cold turkey. That sometimes works, but it does not often produce lasting behavior change.” The turn to time contingent medication through the pain cocktail was not simply to aid in the taper off opioids and sedatives. It had the “… objective of preparing the patient to follow time-contingent medication regimens in the future to prevent recurrence of medication problems.” This future benefit of time-contingent medication regimens is more controversial than its benefit in medication taper, but Fordyce is clear that he thought it had this benefit. “Once operant conditioning techniques have been used successfully in bringing medication addiction, habituation, or high utilization under control, the methods become a tool for self-control by the patient… The patient now has a method by which to prevent future medication addiction or habituation.” To my knowledge no one has tested this idea. I personally doubt whether this translation of operant conditioning into self-control of future medication behaviors exists. I have had patients ask me to put them on time-contingent opioid regimens to assist them in tapering off opioids. I have also had patients with an addiction history ask that their access to opioids (e.g., in a post-operative period) be highly restricted and scheduled so as to limit the risk of addiction relapse. But I have never had a patient describe to me a period of time-contingent medication administration as a learning experience that they drew upon to avoid lapsing into using opioids in a pain-contingent manner. Perhaps this is because I have not asked.

Respondent Pain VS Operant Pain—Determining Which is Which

Perhaps the most controversial aspect of Fordyce’s behavioral approach to the care of patients with chronic pain was his focus on the operant conditioning of pain behavior. Physicians typically see pain behavior as a response to pain, not environmental contingencies. This is clearly true in the case of acute illness and injury, and the bulk of physician training focuses on the care of these acute conditions. Many physicians found Fordyce’s focus on the operant control of pain behavior through environmental contingencies to be disrespectful of the patient and his pain. Many considered it tantamount to saying that the patient was lying concerning his pain. This was never Fordyce’s intention and he became more open to the role of cognitive factors in pain behavior following the publication of this book.

It is clear from reading this chapter on pain medications that Fordyce considered pain and pain behavior to have respondent and operant components. The former was driven by antecedents, while the latter was shaped primarily by consequences. He thought that it was possible to determine whether reports of pain were respondent or operant by observing how these reports responded to blinded medication adjustments. “Once it is clear that there is not some new and perhaps unrelated illness event about which one must be concerned, the most direct way of checking whether the question is respondent or operant in character is to manipulate cocktail ingredients.” Pain behavior that was respondent would diminish in response to an increase in opioids in the blinded pain cocktail. On the other hand, “A finding that modifications in ingredient levels do not produce corresponding changes in pain behaviors suggests that the problem is operant.”

There are a number of problems with this strategy for distinguishing respondent and operant pain. First, it is unlikely that clinical pain behavior can be sorted into one or the other of these categories. Our modern understanding of pain and pain behavior sees them as the product of ascending (nociceptive) and descending (modulatory) activity in the nervous system. Thus, clinical pain has both respondent and operant components. Second, not all chronic pain is opioid-sensitive (e.g., thalamic pain syndrome). The fact that this pain does not diminish in response to a blinded increase in opioid dose does not prove that it is operant. Third, if manipulation of cocktail ingredients was not perfectly blinded, we would expect it to produce some sort of placebo response. It is well known that response to placebo does not diagnose pain as operant or respondent.

WHAT HAS ENDURED AND WHAT HAS NOT ENDURED OVER THE PAST 40 YEARS

End of In-Patient Opioid Taper

While there are still in-patient opioid detoxification programs that precede addiction treatment, these are no longer a standard part of chronic pain care. Opioid taper now generally occurs in outpatient settings. These tapers are usually slower than the minimum duration that Fordyce recommended: “… a period of at least 1 or 2 weeks is virtually mandatory.” Since these tapers are slower, one might expect that they are easier. But this is often not true. The pain cocktail tapers described in this chapter occurred in the controlled environment of an inpatient medical ward where all staff were educated in Fordyce’s behavioral principles. When the taper occurs in the outpatient setting, the patient is exposed to all the stresses and temptations of their home. The regulation of opioid intake is much looser. So a slower outpatient taper may be more difficult than a more rapid inpatient taper. However, it may also be more enduring in effect because the patient faces the challenge of living without opioids during the taper rather than at the time of hospital discharge.

End of Pain Cocktails

The blinded methadone pain cocktail is a powerful tool for separating fear of pain increase from actual pain increase. Frequently, patients resist an opioid taper because they are afraid of facing a pain flare-up with no opioids at hand. They have been overwhelmed by pain in the past and have no desire to go through that again. Nonetheless, many long-term opioid recipients are ambivalent about taking opioids and report that they “would like to stop if it were possible” [4]. The pain cocktail protects the patient from watching their pills go away one by one. Often, the pain increase is less severe and more temporary than they expected. The patient can arrive at the end of the pain cocktail taper with a sense of triumph.

However, the methadone pain cocktail has not been widely adopted since Fordyce’s book was published. There are multiple reasons for this. First, few pharmacies were willing to mix up the pain cocktail, especially those in rural areas… No drug company has stepped forward to offer commercial methadone cocktails. This would be difficult anyway due to the need to customize doses of multiple ingredients. Second, methadone pain cocktails can be dangerous, just like other formulations of methadone. Patients used to pain-contingent prn dosing often report that they cannot “feel” anything after their dose of methadone pain cocktail and may take more until they can feel it. Due to the long plasma half-life of methadone and due to the fact that the analgesic half-life is shorter than the plasma half-life, these extra doses could lead to fatal dose accumulation for the patient [5].

The Rationale for Long-Acting Opioids, OxyContin, and The Opioid Epidemic

Through the 1980s and into the 1990s, academic pain physicians agreed with Fordyce that time-contingent administration of long-acting opioids posed lower risk for opioid misuse, abuse, and addiction than pain-contingent administration of short-acting opioids. The latter were thought to be much more reinforcing of bad medication-taking behavior. Our lectures during this time period were full of recommendations to switch patients with chronic pain from short-acting to long-acting opioids delivered according to the clock and not according to the patient’s pain level. Methadone was the only long-acting opioid available at this time, so it was selected for use in the pain cocktail.

This line of argument was appropriated by the manufacturers of sustained-release opioids, most notably Purdue Pharma, the maker of OxyContin and MSContin. Based on pharmacological arguments about the benefits of stable blood levels of drug and psychological arguments similar to those made by Fordyce in this chapter, Purdue argued that OxyContin was less prone to abuse than short-acting opioids. We are all familiar with how badly this turned out. In 2007, Purdue paid a fine of $634.5 million for claiming the drug was less addictive and less subject to abuse than other pain medications [1]. By 2008, the US Centers for Disease Control and Prevention reported that there were more than 14,000 deaths per year from prescription opioid overdoses, many with sustained-release opioids [3]. As opioid use quadrupled in the past decade, so have misuse, abuse, overdoses, and deaths.

BACK TO THE FUTURE WITH THE CURRENT LONG-ACTING VS SHORT-ACTING OPIOID DEBATE

The debate about the risks and benefits of a long-acting, time-contingent opioid regimen compared to a short-acting, pain-contingent regimen rages on. Many of the purported benefits of time-contingent opioid regimens are not being confirmed in observational studies of long-term opioid users. In a study of patients receiving long-term opioid therapy for chronic non-cancer pain, we found that time-scheduled opioid dosing was associated with higher levels of patient opioid control concerns than pain-contingent dosing. This means that patients themselves felt that they were facing more risk of addiction with time-contingent rather than pain-contingent dosing. Patients with time-scheduled dosing also received higher opioid dosage than patients with pain-contingent dosing [11]. This suggests that something about time-contingent use is associated with opioid dose escalation. I suspect that this is due to the fact that the most commonly pain-contingently prescribed short-acting opioids are combined with acetaminophen or aspirin which imposes dose limits recommended for these drugs.

There have been relatively few studies directly comparing the benefits of short-acting and long-acting opioids [7]. Recent studies suggest that some medical risks of opioids, e.g., hypogonadism, may be increased with long-acting opioid use [8]. A recent randomized trial failed to show benefit of long-acting dihydrocodeine over short-acting dihydrocodeine [6]. Many patients with chronic pain appear to prefer short-acting opioids, though patients receiving long-acting opioid formulations over the long-term may report pain relief and improved quality of life. More controlled data is necessary to form definitive conclusions about opioid efficacy over the long-term. It is difficult to sort out the pharmacological factors from the psychological factors in these patient outcomes and clinical decisions. We can be thankful to Bill Fordyce for drawing our attention to some of the psychological complexities associated with opioid use and taper far before other clinicians were paying attention to these issues.

REFERENCES

  1. Associated Press. Purdue Pharma execs to pay $634.5 million fine in OxyContin case. May 10, 2007. http://www.cnbc.com/id/18591525.

  2. Buckley FP, Sizemore WA, Charlton JE. Medication management in patients with chronic non-malignant pain. A review of the use of a drug withdrawal protocol. Pain Aug 1986;26:153–65.

  3. Centers for Disease Control and Prevention. Policy impact: Prescription painkiller overdoses. November 2011. http://www.cdc.gov/HomeandRecreationalSafety/pdf/PolicyImpact-PrescriptionPainkillerOD.pdf.

  4. Howe CQ, Sullivan MD, Saunders KW, Merrill JO, Banta-Green CJ, Weisner C, Campbell CI, Von Korff M. Depression and ambivalence toward chronic opioid therapy for chronic noncancer pain. Clin J Pain Sep 2012;28:561–6.

  5. Layson-Wolf C, Goode JV, Small RE. Clinical use of methadone. J Pain Palliat Care Pharmacother 2002;16:29–59.

  6. Pedersen L, Borchgrevink P, Breivik HP, Fredheim OM. A randomized, double blind, double dummy comparison of short- and long-acting dihydrocodeine in chronic non-malignant pain. Pain 2014;155;5:881–888.

  7. Rauck RL. What is the case for prescribing long-acting opioids over short-acting opioids for patients with chronic pain? A critical review. Pain Pract Nov-Dec 2009;9:468–79.

  8. Rubinstein AL, Carpenter DM, Minkoff JR. Hypogonadism in men with chronic pain linked to the use of long-acting rather than short-acting opioids. Clin J Pain Oct 2013;29:840–5.

  9. Sullivan MD. Pilot Trial of Opioid Taper Support (POTS). University of Washington. National Institute on Drug Abuse (NIDA). ClinicalTrials.gov Identifier: NCT01883882. June 2013. http://clinicaltrials.gov/show/NCT01883882.

 10. Sullivan M, Toshima M, Lynn P, Roy-Byrne P. Phenobarbital versus clonazepam for sedative-hypnotic taper in chronic pain patients. A pilot study. Ann Clin Psychiatry Jun 1993;5:123–8.

 11. Von Korff M, Merrill J, Rutter C, Sullivan M, Campbell C, Weisner C. Time-scheduled versus pain-contingent opioid dosing in chronic opioid therapy. Pain 2011;152:1256–62.

 12. Sizemore WA. Behavioral aspects of managing medications for chronic pain not caused by cancer. In: Loeser JD, Egan KJ, eds. Managing the chronic pain patient: Theory and practice at the University of Washington Multidisciplinary Pain Center. New York, New York: Raven Press, Ltd, 1989. p.117–28.

 13. Weiss RD, Potter JS, Fiellin DA, Byrne M, Connery HS, Dickinson W, Gardin J, Griffin ML, Gourevitch MN, Haller DL, Hasson AL, Huang Z, Jacobs P, Kosinski AS, Lindblad R, McCance-Katz EF, Provost SE, Selzer J, Somoza EC, Sonne SC, Ling W Adjunctive counseling during brief and extended buprenorphine-naloxone treatment for prescription opioid dependence: A 2-phase randomized controlled trial. Arch Gen Psychiatry Dec 2011;681:238–1246.