Integration of Complementary and Alternative Medicine Therapies into Primary-Care Pain Management for Opiate Reduction in a Rural Setting

Abstract

Background:

Opiates are no longer considered the best strategy for the long-term management of chronic pain. Yet, physicians have made many patients dependent on them, and these patients still request treatment. Complementary and alternative medicine (CAM) therapies have been shown to be effective, but are not widely available and are not often covered by insurance or available to the medically underserved.

Methods:

Group medical visits (GMVs) provided education about non-pharmacological methods for pain management and taught mindfulness techniques, movement, guided imagery, relaxation training, yoga, qigong, and t'ai chi. Forty-two patients attending GMVs for at least six months were matched prospectively with patients receiving conventional care.

Results:

No one increased their dose of opiates. Seventeen people reduced their dose, and seven people stopped opiates. On a 10-point scale of pain intensity, reductions in pain ratings achieved statistical significance (p = 0.001). The average reduction was 0.19 (95% confidence interval [CI] 0.12–0.60; p = 0.01). The primary symptom improved on average by −0.42 (95% CI −0.31 to −0.93; p = 0.02) on the My Medical Outcome Profile, 2nd version. Improvement in the quality-of-life rating was statistically significant (p = 0.007) with a change of −1.42 (95% CI = −0.59 to −1.62). In conventional care, no patients reduced their opiate use, and 48.5% increased their dose over the two years of the project.

Conclusions:

GMVs that incorporated CAM therapies helped patients reduce opiate use. While some patients found other physicians to give them the opiates they desired, those who persisted in an environment of respect and acceptance significantly reduced opiate consumption compared with patients in conventional care. While resistant to CAM therapies initially, the majority of patients came to accept and to appreciate their usefulness. GMVs were useful for incorporating non-reimbursed CAM therapies into primary medical care.

Introduction

Anumber of studies have disconfirmed the widespread belief that opiates are the best treatment for chronic pain. McGraw¹ has written about opiate-induced hyperalgesia as a distinct phenomenon. A study of patients with chronic back pain found hyperalgesia in treated groups compared with controls.² Other studies show the superiority of non-pharmacological methods over opiates. For example, Bronfort et al.³ compared chiropractic spinal manipulation (SMT), guided home exercises (HEA), and medication (inclusive of opiates) for acute and subacute non-specific neck pain among 272 people for 2–12 weeks. SMT had a statistically significant advantage over medication after 8, 12, 26, and 52 weeks (p < 0.010), as did HEA at 26 weeks (p < 0.02). Physician visits lasted 15–20 min, and included a brief history and examination.

Complementary and alternative medicine (CAM) therapies have been shown to be useful in the management of chronic pain.⁴ A study of changes in expectations about pain among patients using CAM therapies showed pretreatment conjectures about whether the CAM therapy could help,⁵ shifting over the course of treatment to be more inclusive of the need for broader lifestyle factors, long-term pain-management strategies, and attention to quality of life and wellness. A shift toward greater acceptance of chronic pain and the need for strategies to keep pain from flaring was also observed. Regardless of how participants evaluated the outcome of treatment, they reported increased awareness, acceptance of the chronic nature of pain, and attention to the need to take responsibility for their own health.

A 12-week pain education program at a VA Medical Center increased CAM utilization among patients with chronic pain.⁶ In another study, 228 individuals with chronic nonspecific neck pain were randomized to five groups receiving various doses of massage over four weeks or to a four-week period on a wait list. Thirty-minute treatments were not significantly better than the wait-list control condition, but 60-minute treatments two and tjree times weekly significantly increased the likelihood of improvement and reduced pain intensity.⁷

This paper reports the results of participation in group medical visits (GMV) inclusive of CAM therapies in rural New England. Advantage is taken of what is being called “real world data,” which are data collected in practice settings outside of academic medical centers.⁸ Consideration was given to how patients would react to a GMV model in which behavioral and CAM therapies were integrated, and whether this approach would be at least cost neutral so as to encourage administrators elsewhere to permit its implementation. The extent to which patients would reduce their opiate use when they were not directly made to do so, but rather when opiate reduction was implicit in the philosophy of the group, was also considered.

Methods

A substantial number of patients in the authors' primary-care clinic were receiving relatively high doses of opiates without the use of non-pharmacological treatments and sometimes without regular office visits. Another group of patients came to the primary-care practice after being placed on opiates by other physicians (usually specialists) and then being told to seek primary care elsewhere when the problems for which the opiates were started appeared no longer resolvable, or the opiate dose had become too great. These patients presented ethical and management dilemmas, since it seemed unreasonable to force them into withdrawal or methadone-maintenance programs when their opiate use had been physician created. It was recognized that the two paradigms had reached a crossover point: the earlier paradigm of pain as a sixth vital sign and the liberal use of opiates to eliminate that pain along with a developing paradigm that says that opiates are inappropriate for the management of non-terminal pain. It was deemed important to find a way to bridge the crossing of these two paradigms that would not increase patient suffering.

A GMV program was developed, in which patients could continue receiving opiates at their current dose if they attended a pain-management GMV at least twice monthly. They agreed to engage at least weekly in a physical activity. Routine refills and other questions would be managed in the 2 h group visit. The group consisted of a family doctor with training in behavioral health, a nurse, and a behavioral health specialist. Groups had a maximum of 12 participants. No patient would be forced to decrease their opiate dose, but no dose would be increased without recommendation from a pain specialist. CAM therapies were incorporated as a regular part of the GMV program.

An attempt was made to randomize patients to GMV or treatment as usual (TAU), but the patients would not accept randomization, almost uniformly preferring TAU. Most patients did not want to attend a group, stating that their pain was physical and that they only needed medication. They resented the implication that non-physical factors influenced their pain. They were not initially interested in CAM therapies. Therefore, a quality-improvement project was implemented in which one physician (L.M.M.) managed all his opiate-receiving patients in the GMV program. Through the use of the electronic health record, a matched, comparison population from the same practice was generated. This comparison population included patients who were seeing or had seen all providers at the practice, including L.M.M.

The weekly required physical activity consisted of physical therapy, yoga, exercise class, chiropractic therapy, osteopathic treatment, t'ai chi, or qigong. Patients agreed to use the same pharmacy. Early refills or replacement of lost medications were not allowed.

At the time of the commencement of this GMV program, there was a turnover of staff at the clinic. The attitude of the State Board of Medicine toward opiates was becoming less permissive. New practitioners recruited to the clinic did not want to accept new patients on chronic opiates or to start established patients on chronic opiate therapy. Therefore, new patients coming to the practice who were taking opiates had the choice of joining the GMV program, tapering off opiates, or finding other practitioners from whom to obtain their opiates.

At the beginning of treatment, no patient wished to reduce or stop opiates, and few were interested in CAM therapies. It was reasoned that 12 required sessions (over six months) in a GMV for pain, with a philosophy that non-pharmacological methods (CAM) were better long-term than opiates and a culture that encouraged reduction or cessation, would have an impact on these patients, regardless of their initial bias. Therefore, patients who completed at least six months of GMVs were compared to patients receiving TAU for the same length of time.

Within the psycho-educational component of the pain-management group, the idea was introduced that chronic opiate use actually increases pain,^9,10 and the aim was to increase emotional awareness and regulation so as to prevent requests for dose increases and to facilitate patients potentially spontaneously reducing opiate dosages. The pain contract included decreasing the dose of opiates by 10% in any month in which patients did not attend two GMVs and complete four physical activities per month.

During the GMV, the nurse took patients individually for medication reviews and refills. She wrote refill prescriptions, took blood pressures and other vital signs, and completed outcome measures. At the end of the GMV, the physician saw patients whose questions could not be answered or addressed in the GMV and signed prescriptions.

The GMV began with check-in, which was done in the format of a talking circle in which an object was passed around the circle and whoever holds the objects talks without interruption or comment from others, giving a short update of his or her status, mood, pain, and other life events that have happened.¹¹ Next followed a period of physical activity—yoga nidra, authentic movement, t'ai chi, stretching, or qigong. Then came guided imagery, mindfulness, and/or visualization. Following that was an exercise related to achieving one or more of the goals of dialectic behavior therapy, followed by general discussion. Homework assignments were made at the end of the group visit. The attitude of the leaders was to practice compassion and radical acceptance of the participants, their beliefs, and their struggles while still respectfully and politely maintaining that opiates are not the best strategy for managing chronic pain and actually appear to make chronic pain worse.

Outcome measures consisted of (1) change in opiate dosage in morphine equivalents, (2) change in scores on the My Medical Outcome Profile, 2nd version (MYMOP2), and (3) change in visual analogue pain ratings.

The MYMOP2 is a patient-centered, problem-specific outcome measure, and evidence suggests that it is a useful and sensitive measure of change in perceived symptoms and quality of life.^12
–14 The MYMOP2 has been used in clinical audits in the United Kingdom to improve patient care,¹⁵ to show that acupuncture benefited people with chronic illness,¹⁴ and to assess overall outcomes in a complementary-care clinic.¹⁶ In all these contexts, the MYMOP2 provided an opportunity to assess overall improvement in symptom severity, the degree to which symptoms restrict participation in desired activities, and overall quality of life in cases in which the symptoms themselves may be radically different. The use of symptom-specific scales in these cases would result in insufficient numbers of participants for comparison purposes and also prevent across-illness comparisons.

In the current study, participants used the MYMOP2 to choose two symptoms of greatest concern, along with one activity of daily living that was restricted or prevented by these symptoms. The client scored these items according to their severity in the last week using a seven-point Likert scale ranging from 0 (as good as it can be) to 6 (as bad as it can be). They also rated their overall feeling of well-being. Doses of medications and demographic data were obtained from the electronic health record. MYMOP2 forms were administered at the start of every GMV. The chi-square procedure in the SPSS Statistics for Windows v22 (IBM Corp., Armonk, NY) was used to compare initial opiate doses to doses achieved at the conclusion of treatment or the termination of the GMV and the numbers of patients reducing or stopping opiates. The initial and subsequent ratings on the MYMOP2 were compared for symptom severity, interference with activities of daily life, and quality of life using repeated-measures analysis of variance. Financial data collected by the practice were used to determine the profitability of group medical care versus conventional care. Comparisons with the TAU population were limited to opiate doses and number reducing or stopping opiates, since these patients did not complete the MYMOP2.

Comparison group

Two comparison groups were generated using the electronic health record. Forty-two patients were selected from the group practice using the electronic health record and matched prospectively with patients attending GMVs. Matching was done to obtain patients in the same age decile, with the same major diagnoses, same sex, and same dose within 25% in morphine equivalents. A second comparison group of 207 patients was generated for comparisons with the 207 patients who made initial consultations and were offered the opportunity to join the GMV.

Results

Table 1 presents the data on attrition, and Table 2 presents the demographics of the treatment population compared to the comparison population. Forty-two people attended GMVs for at least six months out of a total of 207 people seeking pain-management services. For those who stayed in the practice, no one increased the opiate dose. Patients who left GMVs before six months did not statistically significantly reduce opiate use. Patients attended group for an average of 7.4 months (SD = 6.9). Patients who attended group for more than six months continued in group for an average of 19 months. Eighteen people reduced their dose, and eight people stopped opiates altogether. On a 10-point scale of pain intensity, reductions in pain ratings achieved statistical significance (p = 0.001). The average reduction was 0.19 (95% confidence interval [CI] 0.12–0.60; p = 0.01). The primary symptom improved on average by −0.42 (95% CI −0.31 to −0.93; p = 0.02) on the MYMOP2. Improvement of the second symptom on the MYMOP2 was −0.20 (95% CI −0.08 to −0.40; p = 0.04). Improvement in activities of daily living was not statistically significant, but improvement in the quality-of-life rating was statistically significant (p = 0.007), with a change of −1.42 (95% CI −0.59 to −1.62). Forty-one people who joined the GMV program did not reach the six-month mark. Three moved out of the area, 11 found other physicians who would prescribe opiates without restrictions, and 26 left because they failed to maintain their pain contract and were being tapered off opiates and found other care.

Table 1.

Number of Patients Seeking Care

Number of patients receiving opioids for more than six months in the practice database	Number seeking care from L.M.M. after the start of this program^a	Number accepting group medical visits	Number attending for at least six months	Number who started group medical visits and later transferred to other providers
1951	207	84	42	37

Unexpectedly, at the commencement of this program, L.M.M. was the only physician in the practice accepting new patients who were on chronic opiate therapy. New patients who did not want to participate in the program had to find another practice from which to receive their opiate prescriptions. Established patients still receiving chronic opiates from other practitioners in the practice were included as comparison patients.

Table 2.

Demographics of the Population

Characteristic	Group visit model	Treatment as usual group
Participants	207	207
Mean age (SD), years	45.7 (18.1)	44.3 (13.1)
Women	124	124
Married or living with someone, %	59.9	63.3
College graduate, %	41.5	40.1
Current smoker, %	30.9	50.5
Mean body mass index (SD), kg/m²	29.5 (5.8)	27.9 (6.4)
Mean duration of pain (SD), years	9.0 (3.7)	9.2 (3.2)
Awake at night because of pain, %	44.0	51.7
Reported cause of pain, % (people may report more than one cause)
Trauma	38.6	25.6
Car accident	17.4	12.4
Recreational accident	13.8	12.5
Job accident	22.4	22.2
No apparent cause	18.7	25.7
Did not recall	1.3	8.1
Other	11.7	14.1

SD, standard deviation.

Participants were observed to change their attitudes toward pain and pain medication over time as they continued to attend GMVs. The repetitive stories told by other participants about reducing or eliminating opiates and the stories implicit in the CAM practices had an effect evolving over months. The leaders continually provided a gentle but consistent and persistent message that “you can learn to influence your pain.” Participants appeared to absorb this message over time. Few patients were initially enthusiastic about the GMV. Some were angry that we thought we could “teach them anything about pain.” Those who stayed, however, became enthusiastic participants, with some coming more often than required. Participation in the group through the GMV discussion also allowed those who were misusing medication or using illegally obtained medication to be identified quickly, which is more cost-effective and less coercive than random urine drug testing.

Comparison group

In conventional care, no patients reduced their opiate use, and 48.5% increased their dose over the two years of the study. Table 3 shows the differences in average opiate doses at baseline and at the conclusion of treatment (range 6 months–2.5 years). Statistically significant reductions in opiate dose occurred relative to the comparison group over the course of the treatment. More people reduced or eliminated opiates in the treatment group than they did in the comparison group. The duration of use prior to this contact was similar for both groups and was quite long.

Table 3.

Reduction in Opiate Dose (Morphine Equivalents) for Those Who Attended Pain Group Twice Weekly for Six Months Compared to Matched Patients from the Same Practice Who Received Conventional Care

	Initial average dose	Final average dose	Number stopping opiates	Number reducing opiates	Range and average years of use
Opiate reduction	82.1 (21.0)	32.4 (13.0)^*	8/42	18/42	1.0–17.1; 11.1 (5.6)
Conventional care	72.1 (21.7)	86.1 (21.5)^*	0/42^*	1/42^*	0.8–18.8; 9.8 (4.1)

Doses are in mg morphine equivalents.

p < 0.001.

In a second analysis, all 207 patients who made an initial consultation were matched with 207 comparison patients who had received TAU. It was assumed that none of the patients who refused to join the group or dropped out of the group reduced their opiate use. In the database, eight patients reduced their dose over the year. Chi-square analysis of these differences showed a significance of p = 0.05. Thus, even with the most conservative of assumptions, participation in the group made a difference.

Financial comparisons

In the United States, implementation of GMV programs must be financially justifiable. One must prove that income exceeds expenses. Calculations were made to address this question using data on average income and expenses for office visits and GMV for the authors' practice. In the GMV model, each patient received an office visit (CPT code of 99213) billing once per month and a group therapy billing (CPT code of 90853) at least twice per month. In the individual model, each patient is seen monthly for about 15 minutes with one office visit (usually CPT code of 99213), resulting in an income of $657.60 at a cost of $450 per month for practitioners (difference of $207.60 per month). This calculation is the same for both GMV and individual visits, since an office visit can be billed during a GMV when all the components of an office visit are charted, even if the patient is not taken out of group (pers. comm., Medicare of New England). Office visits were billed for only when refills were written or other concerns above what could be managed within the GMV were addressed, though an office visit and a group visit could have billed for each time, assuming the charting supported both.

GMVs averaged 12 people per session over 2 h. In this setting, the average income was $18 per person for a 2 h group. This generated on average $204 per meeting or $408 per month. Salaries for group leaders came to $320, which gave a net gain to the practice of $88 per month, which one hopes would cover overheads. Thus, improved care at a minimum broke even financially.

Discussion

The present results suggest that an approach grounded in the neurobiology of emotion and nociceptive perception and using CAM therapies in a group setting can reduce pain and medication use among a population of patients previously considered intractable by their physicians.

These results also indicate that the group structure as herein designed is financially viable and can serve more patients per month than the usual 15 min office-visit format, at about the same level of profitability. The physician co-leader reported enjoying seeing the patients in the 2 h GMV more than the same number of separate 15 min appointments. From his point of view, psycho-education about pain was delivered in a more effective and persuasive manner, and time existed to incorporate CAM therapies, which was not possible in 15 min individual appointments.

Pain is not simple. Patients and their physicians may not realize the extent to which emotional states are integrated into what is called pain. The same brain regions appear to manage both pain and emotion.^17,18 The insula, with its links to somatosensory cortex regions S1 and S2 and to the cingulate cortex, is a key element in feeling emotions¹⁹ and in feeling pain. Bilateral anterior insula²⁰ and rostral anterior cingulate cortex (ACC) are activated both when people feel pain and when they perceive a loved one feeling pain.²¹ Connected intimately to the insula, the ACC both regulates emotional processing²² and pain processing.²³ Activation of aMCC sites by nociceptive inputs from the thalamus induces fear of painful events and memory of earlier, similar painful events and trigger the re-experiencing of pain and consecutive behavior. The CAM interventions utilized may have been effective by introducing patients to practices and healing techniques that helped address this emotion–pain connection.^24
–26

Social relationships and the presence or absence of community has a strong influence on pain perception. The presence or loss of human relationships modifies physiological responses and affects general health.²⁷ Social interactions and relationships serve to regulate a person's internal biological systems.¹⁷ Evidence exists for a significant increase in maladaptive biological responses, bodily changes,^28
–30 and even increased mortality^31,32 after a major personal loss.

These findings suggest that increased awareness of one's emotions could facilitate pain management. At level I of emotional awareness,³³ emotional or external stimuli may cause only autonomous, vegetative, or endocrine reactions without any explicit cognition of the originally emotional content.¹⁷ At level II, affects with tendency for action without cognition or feelings are noticed. On these two levels, originally psychologically induced bodily sensations can either go unnoticed as bearing that meaning or can only be felt as a diffuse change in one's physical condition.¹⁷ For example, depression may initially present as headache or backache.³⁴

Some limitations to this study exist. Patient outcomes were known only until patients left the practice. As a GMV, patients who left before six months did not statistically significantly reduce opiate use. Patients who adamantly refused to consider non-pharmacological approaches to pain management left and found physicians more to their liking, though the number of physicians who will prescribe chronic opiates is progressively decreasing. It is not known what is different about the participants who continued to come to the GMV and benefitted compared with those who left the practice. However, it is known that these patients had been steadily increasing their opiates for some years, and, in conjunction with the implementation of the GMV, stopped that increase, and statistically significantly decreased their opiate use. Patients in the same practice receiving TAU continued slowly to increase their opiate dose. It is not known how representative these patients are of all opiate users, but they appeared to be representative of opiate users in primary care in the authors' small New England community of 12,500 people. Patients receiving opiates to prevent heroin abuse were not included. The sample size of patients is small, but nevertheless differences reached statistical significance, speaking to the potentially greater power of these CAM approaches with larger sample sizes.

While patients may not initially agree with the idea that chronic opiate therapy actually increases perceived pain and that non-pharmacological and CAM approaches have merit, over time they warmed to the idea and eventually accepted it for the most part. These results can encourage other practitioners to implement behavioral and CAM therapies as part of pain management GMV programs, irrespective of whether patients wish to attend. Requiring patients to attend GMV inclusive of CAM therapies for pain management as part of their pain contract resulted in reduction in opiate use that was accompanied by a statistically significant reduction in perceived pain, severity of patients' two most bothersome symptoms, and increase in quality of life. These results can encourage primary-care practitioners to include behavioral and CAM therapies as a required part of chronic pain-management strategies. This is important, because the number of physicians who will prescribe chronic opiates is progressively decreasing as public and governmental attitudes toward opiates change. GMVs also proved to be financially viable for introducing CAM therapies into primary care.

Footnotes

Acknowledgments

Internal funding was provided by the general operating fund of Coyote Institute, Inc.

Author Disclosure Statement

No financial relationships exist to be disclosed.

References

McGraw

. Known unknowns: A review of opioid-induced hyperalgesia. Mental Health Clinician, 2015; 5:136–143.

Giesecke

, Gracely

, Grant

, et al. Evidence of augmented central pain processing in idiopathic chronic low back pain. Arthritis Rheum, 2004; 50:613–623.

Bronfort

GRE

, Anderson

, Svendsen

, et al. Spinal manipulation, medication, or home exercise with advice for acute and subacute neck pain: a randomized trial. Ann Intern Med, 2012; 156:1–10.

Simpson

. Complementary medicine in chronic pain treatment. Phys Med Rehabil Clin N Am, 2015; 26:321–347.

Eaves

, Sherman

, Ritenbaugh

, et al. A qualitative study of changes in expectations over time among patients with chronic low back pain seeking four CAM therapies. BMC Complement Altern Med, 2015; 15:12.

Cosio

, Lin

. Effects of a pain education group medical visit program in Complementary and Alternative Medicine treatment utilization at a VA medical center. Complement Ther Med, 2015; 23:413–422.

Sherman

, Cook

, Wellman

, et al. Five-week outcomes from a dosing trial of therapeutic massage for chronic neck pain. Ann Fam Med, 2014; 12:112–120.

Garrisson

, Neumann

, Erickson

. Using real-world data for coverage and payment decisions: The ISPOR Real-World Data Task Force Report. Value Health, 2007; 10:326–335.

Ballantyne

. Opioids and other analgesics. In Drug Abuse and Addiction in Medical Illness. Springer, 2012:241–250.

10.

Ballantyne

, Mao

. Opioid therapy for chronic pain. New Engl J Med, 2003; 349:1943–1953.

11.

Mehl-Madrona

, Mainguy

. Introducing healing circles and talking circles into primary care. Perm J, 2014; 18:4–9.

12.

Hull

, Page

, Skinner

, et al. Exploring outcomes associated with acupuncture. J Altern Complement Med, 2006; 12:247–254.

13.

Patterson

. Measuring outcomes in primary care: A patient generated measure, MYMOP, compared with the SF-36 health survey. BMJ, 1996; 312:1016–1020.

14.

Patterson

, Britten

. Acupuncture for people with chronic illness: Combining qualitative and quantitative outcome assessment. J Altern Complement Med, 2003; 9:671–681.

15.

Rees

. Improving patient care in complementary medicine: using clinical audit. In: Lewith

, Jonas

, Walach

, eds. Clinical Research in Complementary Therapies: Principles, Problems, and Solutions. New York: Churchill Livingstone/Harcourt Brace, 2002:109–127.

16.

Harris

, Atkins

, Alwyn

. Evaluating a complementary therapies clinic: Outcomes and relationships. Complement Ther Clin Pract, 2010; 16:31–35.

17.

Gundel

, Tolle

. How physical pain may interact with psychological pain: Evidence for a mutual neurobiological basis of emotions and pain. In: Carr

, Loeser

, Morris

, eds. Narrative, Pain, and Suffering. Seattle, WA: IASP Press, 2005:87–112.

18.

Phillips

, Singer

. In search of common foundations for cortical computation. Behav Brain Sci, 1997; 20:657–683.

19.

Damasio

. Looking for Spinoza: Joy, Sadness, and the Feeling Brain. Orlando, FL: Harcourt, 2003.

20.

Lane

, Reiman

, Ahern

, et al. Neuroanatomical correlates of happiness, sadness, and disgust. Am J Psychiatry, 1997; 154:926–933.

21.

Singer

, Seymour

, O'Doherty

. Empathy for pain involves the affective but not sensory components of pain. Science, 2004; 303:1157–1162.

22.

Gundel

, Lopez-Sala

, Ceballos-Baumann

. Alexithymia correlates with the size of the right anterior cingulate cortex. Psychosom Med, 2004; 66:132–140.

23.

Devinsky

, Morrell

, Vogt

. Contributions of anterior cingulate cortex to behavior. Brain, 1995; 118:279–306.

24.

Lane

, Reiman

, Axelrod

. Neural correlates of levels of emotional awareness. Evidence of an interaction between emotion and attention in the anterior cingulate cortex. J Cogn Neurosci, 1998; 10:525–535.

25.

Derbyshire

. Exploring the pain “neuromatrix.”. Curr Rev Pain, 2000; 4:467–477.

26.

Vogt

, Berger

, Derbyshire

. Structural and functional dichotomy of human midcingulate cortex. Eur J Neurosci, 2003; 18:3134–3144.

27.

Hofer

. Relationships as regulators: A psychobiologic perspective on bereavement. Psychosom Med, 1984; 46:183–197.

28.

Horowitz

, Siegel

, Holen

. Diagnostic criteria for complicated grief disorder. Am J Psychiatry, 1997; 154:183–197.

29.

Parkes

. The first year of bereavement. A longitudinal study of the reaction of London widows to the death of their husbands. Psychiatry, 1970; 33:444–467.

30.

Clayton

. Mortality and morbidity in the first year of widowhood. Arch Gen Psychiatry, 1974; 30:995–998.

31.

Kaprio

, Koskenvuo

, Rita

. Mortality after bereavement: A prospective study of 95,647 widowed persons. Am J Public Health, 1987; 77:283–287.

32.

Lehman

, Wortman

, Wiliams

. Long-term effects of losing a spouse or a child in a motor vehicle accident. J Pers Soc Psychol, 1987; 52:218–231.

33.

Lane

, Schwartz

. Levels of emotional awareness: a cognitive-developmental theory and its application to psychopathology. Am J Psychiatry, 1987; 144:133–143.

34.

Simon

, VonKorff

, Piccinelli

, et al. An international study of the relationship between somatic symptoms and depression. New Engl J Med, 1999; 341:1329–1335.