Exploring physicians ’ comfort level with opioids for chronic noncancer pain

1Department of Anesthesiology, University of Calgary, 2Evaluation and Research Consultant, Calgary, Alberta Correspondence and reprints: Dr Martin N Scanlon, Calgary Pain Centre, Calgary Health Region, 160–2210 2nd Street Southwest, Calgary, Alberta T2S 3C3. Telephone 403-943-9900, fax 403-209-2954, e-mail martinscanlon@calgaryhealthregion.ca MN Scanlon, U Chugh. Exploring physicians’ comfort level with opioids for chronic noncancer pain. Pain Res Manage 2004;9(4):195-201.

T he use of opioids in managing chronic noncancer pain (CNCP) in adults has been a controversial issue in medical literature.In 2003, Gardner-Nix (1) cautioned that omitting opioids in CNCP management without clinical and scientific justification, stops patients from receiving much needed relief.Gardner-Nix stated that "Opioids can be beneficial and safe as a component of the long-term management of non-cancer pain" (1).Eighty (38%, n=209) of Gardner-Nix's (1) CNCP patients (followed for two to 15 years) experienced reduced pain with concomitantly increased physical function compared with preopioid treatment.
A 2002 original article (2) concluded that in Canada, chronic pain is undertreated because physicians underutilize major opioids in spite of growing empirical evidence supporting their use in the management of CNCP.The literature is replete with articles describing evidence and recommendations in favour of using opioids and the most effective ways to use them (3)(4)(5)(6)(7)(8).A recent survey (8) conducted with Canadian physicians (palliative care and family physicians [FPs]) revealed that approximately 28% were unwilling to prescribe short-or long-acting opioids for the management of chronic pain, even as a third-line treatment.According to some experts, the role of opioids in the management of chronic pain is not well-established (9).Russell and Aaron (10) acknowledged the benefits of judicious use of opioids under extraordinary circumstances but, by and large, expressed skepticism toward them.The majority of responding physicians in a 1994 American study (11) reported that they were not overly concerned by potentials of tolerance, dependence or addiction when prescribing opioids.The College of Physicians and Surgeons of Alberta have published guidelines to help manage local cases (12).
There is a need for primary care physicians to increase their comfort level with opioids to adequately treat patients suffering from moderate to severe chronic pain.A recent study (8) reported that approximately 60% of participating physicians felt that "pain management could be enhanced by improved physician education" (8).Further, the need for education among clinicians was recognized by an international panel.It concluded that, while progress has been made, much remains to be done to correct unfounded fears and misconceptions that impede the provision of opioid analgesia when it is indicated (4).Turk et al (11) considered physicians' education in pain management to be inadequate, both during medical school and in residency programs.Weinstein et al (13) concluded that new strategies are needed to educate physicians to manage CNCP.Goldman's 1991 article (14) painted a negative picture of physicians dealing with CNCP patients (eg, inept).
The present article is part of a larger study conducted in 2002/2003 to explore multiple aspects of chronic pain management in Calgary, Alberta.One such aspect was physicians' comfort in managing opioids including Tylenol No 3 with Codeine (Janssen-Ortho/McNeil Consumer Healthcare, Canada), morphine/MS Contin (Purdue Pharma, Canada), Percocet (Bristol-Myers Squibb, Canada)/OxyContin (Purdue Pharma, Canada), Dilaudid (Abbott Laboratories, Canada)/Hydromorph Contin (Purdue Pharma, Canada), fentanyl patch and methadone.The study was funded by the Physician Partnership Steering Committee at the Calgary Health Region (CHR), and the unpublished report is titled "Physicians' and patients' perspectives on chronic pain management in the Calgary Health Region (CHR): Exploring the next step toward chronic pain management in the CHR".The present article contributes to the debate on physicians' readiness to manage chronic pain with special emphasis on opioid therapy.The purpose is to share and explore findings from a larger study about physicians' comfort level in managing opioids including Tylenol No 3 with Codeine, morphine/MS Contin, Percocet/OxyContin, Dilaudid/Hydromorph Contin, fentanyl patch and methadone.Although data are collected from various specialties, the main focus is on the CHR's FPs because they are the gatekeepers of the CNCP among Calgarians.

METHODS
One hundred forty-seven FPs and 142 specialists were randomly selected from the College of Physicians and Surgeons of Alberta's 2002 Medical Directory (15) using the MINITAB Statistical Software (Minitab Inc, USA).Approximately 893 FPs (in the directory, this group of physicians are also referred to as nonspecialists) and 661 specialists were considered to be eligible for this study (ie, two sampling frames were constructed, one for the FPs and one for the specialists).Some may consider family practice to be a specialty in its own right.Of the 50 types of specialists listed in the medical directory, 22 types were arbitrarily selected for the present study.The two sample sizes from the two subpopulations were expected to be representative at a 95% CI with a sampling error of ±0.05.This was contigent on a 60% to 70% response rate.Theoretically, the response rate should be 100%, but the CI was not the sole factor in deciding on an efficient sample size.Other factors such as variability of the important variables, need for subpopulation analysis and cost were also considered (16).
A 20 min survey was designed to explore several questions related to CNCP in the CHR (eg, to what extent can FPs and specialists deal with chronic pain without the intervention of pain specialists?and how well prepared do these physicians feel in managing chronic pain patients?).The present paper focuses on a subquestion of the above, namely, how comfortable do CHR physicians feel in prescribing opioids and what factors influence their comfort level?The complete survey was composed of 23 closed-or open-ended questions and additional subquestions.
Thirty-two FPs (12 responded, 37.5% response rate) were invited to pretest the survey for clarity.Participants had the flexibility to complete the survey by telephone, fax, e-mail or regular mail.
One hundred fifteen FPs (excluding the above 32 invited for pretesting) and 142 specialists were telephoned to conduct the pretested survey.Physicians who declined to participate in the full survey were asked to supply information on their type of practice, number of hours devoted to direct patient care, number of patients in practice and significance of chronic pain in their practice.These parameters were compared between the respondents and nonrespondents to examine selection bias.
In the full survey, physicians were asked to indicate their level of comfort in managing six opioids: Tylenol No 3 with Codeine, morphine/MS Contin, Percocet/OxyContin, Dilaudid/Hydromorph Contin, fentanyl patch and methadone in noncancer chronic pain management.For the sake of comparison, the comfort level for nonsteroidal anti-inflammatory drugs (NSAIDs) was also solicited.The comfort level was measured on a four-point ordinal scale as a proxy measure for drug use by physicians.Responses were coded numerically from four to one.• Handle myself (code 4): I handle all aspects myself; • Somewhat uncertain (code 3): I would read/check the CPS (Compendium of Pharmaceuticals and Specialties) first; • Specialist's advice (code 2): I would like advice and direction from a specialist, and would then manage the patient myself; and • Specialist manage (code 1): I would not use this myself, but would like a specialist to manage it.
The coded data were entered for statistical analysis by the Statistical Package for the Social Sciences 11.01 for Windows (SPSS Inc, USA).
Using comfort level as an ordinal dependent variable, the Mann-Whitney U test was performed to determine differences in comfort level between male and females (independent variable).In addition, Spearman's rank correlation coefficients (rho) were computed between comfort level and length of practice, number of chronic pain patients in the practice, and self-reported five-point level of significance of chronic pain in physicians' practice (extremely significant, very significant, moderately significant, mildly significant and not at all significant).Eight variables emerging from the larger survey were mentioned here: comfort with the seven drugs, type of physician, significance of chronic pain in physicians' practices, number of pain patients in practice, weekly hours devoted to direct patient care, sex, length of practice and the province of graduation.Data concerning physicians' sex, length of practice (length between graduation year and the time of the survey) and province of graduation were all obtained through the medical directory (15).

RESULTS
Forty-nine FPs and 45 specialists (Table 1) consented to partake in the 20-min survey (long survey).The response rates were 42.6% and 31.6%,respectively.
Thirty-six of 49 (73.5%)FPs and 31 of 45 (68.9%) specialists graduated in Canada (Alberta, Ontario, Saskatchewan, Quebec, British Columbia and Manitoba).The majority of both the FPs (53.0%) and specialists (48.8%) graduated in Alberta (University of Calgary or University of Alberta, Edmonton).FPs (n=42) reported managing anywhere from six to 1200 chronic pain patients (mean=84.2,SD=189, median=30, mode=10).Specialists (n=24) reported managing anywhere from one to 400 chronic pain patients (mean=62.6,SD=118.9,median=10, mode=5).The forms of chronic pain included backache, headache, fibromyalgia, pelvic pain, and neuropathic pain with backache prominent within both the subgroups.The Mann-Whitney U test revealed that, compared with the specialists, FPs express a higher level of comfort with six of the seven drugs in question.Methadone appeared to be the exception because the specialist subgroup had a higher mean ranking (Table 3).The Mann-Whitney U test revealed no significant (α>0.05)difference in comfort level between male and female FPs when prescribing NSAIDs, Tylenol No 3 with Codeine, fentanyl patch and methadone.However, when managing morphine/MS Contin, Percocet/OxyContin and Dialudid/ Hydromorph Contin, male FPs reported a greater level of comfort compared with their female counterparts.Table 4 presents the frequency distribution broken down by sex within the two subgroups.Table 5 presents statistically significant (α<0.05)Mann-Whitney U results.No significant sex differences were found among specialists when managing any of the seven drugs.Only eight female specialists participated in the study.
Based on a Spearman correlation test (at α≤0.05), there were no relationships (neither positive nor negative) between the comfort levels with the seven drugs being tested and the two independent variables in either of the two subgroups  (number of patients managed and the length [years] of practice in Canada).Among the specialists' subgroup, the Spearman's rho test revealed a positive relation, albeit weak, between the significance of pain in physicians' practice and the comfort with methadone (Spearman's rho=0.035and α=0.03).See Table 6 for the significance of chronic pain in physicians' practices.
Analysis of the 31 short surveys (14 FP and 17 specialist nonrespondents) revealed that approximately 57.2% of FPs reported either extremely or very significant influence of chronic pain in their practice compared with 42.9% reported by the nonrespondents.The nonrespondent specialists responded similarily to the 45 specialists who responded to the long survey.Striking differences were noted within both Handle myself 17 ( subgroups between nonrespondents and repondents regarding the length of practice (FPs' nonrespondents mode=28 years and respondents mode=11 years; specialists' nonrespondents mode=13 years and respondents mode=7 years).The majority of respondents and nonrespondents, both FPs and specialists, graduated in Canada and, particularly, in Alberta.Regarding sex, the FPs from the nonrespondents' group had a male to female ratio of 2:1 (compared with 1:1 for the respondents) and specialists had a ratio of 3:1 (compared with 4:1 for the respondents).One should note that surveys were sent to 59 male and 56 female FPs (male to female ratio 1:1); in contrast, surveys were sent to 108 male and 34 female specialists (male to female ratio 3:1).The term 'sent' encompasses invitation (and rejection) by phone calls.

DISCUSSION
In view of the low response rates and the potential that the College of Physicians and Surgeons of Alberta Medical Directory is incomplete, the following discussion cannot be widely generalized (eg, this does not apply to every single physician practicing in Calgary, Alberta).This discussion is only based on the FPs subgroup data, which was deemed more homogeneous than the specialists subgroup.Because of heterogeneity among the specialists' subgroup (Table 1), data emerging from it are not amenable to sound discussion.Nonetheless, it serves as a comparator or quasi control group for the findings from the FP subgroup.Within the FP subgroup, the respondents were found to be younger, to have devote more hours in direct care of chronic pain patients and to have experienced a higher significance of CNCP in their practice compared with nonrespondents.This by itself has the potential to threaten generalizability.The following, however, serves as a forum for discussion among health care professionals and policy makers.
A percentage of FPs need a pain specialist's input (advice or talking over management) in managing opioids.This percentage was approximately 12% for morphine/MS Contin and Percocet/OxyContin, 26.5% for Dilaudid/Hydromorph Contin, 38.8% for fentanyl patch and 98% for methadone.In comparison, no FPs reported the need for pain specialist involvement when managing NSAIDs and, with the exception of a single respondent, the same was true for Tylenol No 3 with Codeine.Evidently, there is a conspicious ommission in physician training.Except for methadone, reasons for these current statistics are not clear.(In Alberta, the use of methadone is restricted and regulated.)In a 2003 Canadian survey (7), 29% of physicians preferred NSAIDs and 16% preferred acetaminophen as first-line treatment for moderate to severe chronic pain, despite the negative impacts.
Unlike the 2003 study (7), the present survey did not attempt to define the reasons for the FP level of comfort with the various medications.However, in the case of methadone, it is easy to surmise that the requirement for a special license to prescribe this drug greatly discourages FPs from using this medication.In addition, there is a lack of colleagues/teachers capable of helping others increase their comfort level with this drug.
Previous studies (3,7) have identified the following barriers to prescribing opioids: • potential for addiction; • potential for patient misuse/abuse; • fear of side effects; • fear of audit by regulatory body; and • cost.
It is interesting that many of these issues are perhaps as applicable to the management of Tylenol No 3 with Codeine as with the other drugs, but the overwhelming majority of respondents in the present study were quite comfortable handing Tylenol No 3 with Codeine themselves.Similarly, 98% of FPs were comfortable handling NSAIDs, even though this class is associated with greater morbidity (eg, gastrointestinal side effects) than opioids.Indeed, the Alberta College's guidelines for treatment of CNCP support the use of opioids (11).In Alberta, a triplicate prescription program for opioids (not for Tylenol No 3 with Codeine and NSAIDs) signifies college scrutiny and time commitment.This might also decrease the FPs' comfort level with opioids.To add to this, the FPs' fee structure is not perceived to be conducive to the time and efforts spent.
The present study suggests that male FPs in the CHR feel more comfortable handling at least three (of seven) of the drugs in question: morphine/MS Contin, Percocet/OxyContin and Dialudid/Hydromorph Contin.These findings are preliminary and do not allow a definitive conclusion that male FPs perceive less of a risk in prescribing opioids than their female counterparts.However, this hypothesis should be tested much more rigorously.
FPs' low comfort level with opioids may be due to a lack of knowledge.Previous surveys have suggested that FPs want more continuing medical education (eg, the purpose for the Triplicate Prescription Program) in chronic pain management (2,7).If it is not obtained during training (medical school and residency), this knowledge may be difficult to obtain later.There are very few chronic pain specialists to provide hands-on and problem-based education.However, unless FPs learn to increase their comfort level with opioids, these same specialists, whose services are already in great demand, can expect to be continually overburdened by referrals for cases of opioid management that might be managed independently by more knowledgeable FPs.An estimated 8.7 million people suffer from CNCP.Therefore, FPs throughout Canada will be expected to bear the responsibility of caring for them.A lack of sufficiently trained physicians could result in an unmanageable situation.

FUNDING:
Physician Partnership Steering Committee, Calgary Health Region and Department of Anesthesiology, University of Calgary, Calgary, Alberta.Exploring physicians' comfort level with opioids Pain Res Manage Vol 9 No 4 Winter 2004 201 Table 2 presents descriptive statistics for the length of service of both the FPs and specialists.
Figures 1 to 7 compare FPs' and specialists' comfort profiles (frequency distribution of responses) for seven drugs individually.Exploring physicians' comfort level with opioids Pain Res Manage Vol 9 No 4 Winter 2004 197

TABLE 5 Significant
Mann-Whitney U results detecting sex differencesDrugMean rank Mann-Whitney U α α value

TABLE 6
Significance of chronic pain in physicians' practices