Knowledge, Perceptions and Attitudes toward Chronic Pain and Its Management: A Cross-Sectional Survey of Frontline Pharmacists in Ontario, Canada

The treatment of chronic pain consumes a significant share of primary care. Community and family health team pharmacists frequently see patients with chronic pain, thus have the opportunity to improve their care. To assess the knowledge, perceptions, and attitudes of Ontario pharmacists, we invited 5,324 Ontario pharmacists, to participate in an online survey we developed using Qualtrics. The 31-question survey gathered demographic information, assessed pharmacists’ knowledge of three chronic pain conditions; chronic lower back pain (CLBP, eight true/false statements); chronic headache disorder (CHD, eight true/false statements) and painful diabetic neuropathy (PDN, seven true/false statements), and their attitudes toward and perceptions of patients with these conditions, and knowledge, attitudes, and perceptions of opioids in pain management. We received 688 responses (12.9%) and 392 pharmacists completed the survey. The mean age of respondents was 48.5 years and 48.5% were male. More than 50% of respondents were in practice for more than 20 years and 58.7% worked 25–40 hours per week. The mean knowledge scores were 4.5/8, 5.5/8, and 5.3/8 for CBLP, CHD, and PDN respectively. While 95% of respondents were aware of the increasing death rates due to opioid use, only half were familiar with the Canadian guideline for safe opioid prescribing for non-cancer use. Responses were compared based on gender, time in practice and location of practice. Pharmacists with more than ten years of experience scored significantly higher than those with less experience. Fewer differences were found in comparisons of gender and location of practice. Safe and effective care of chronic pain patients, particularly with opioids, will require additional pharmacist education.

Introduction knowledge, perceptions, and attitudes of primary care pharmacists in Ontario, Canada with respect to these chronic pain conditions.

Study design and population
The study, approved by the Institutional Review Boards at the University of Waterloo and the University of Toronto, is a cross-sectional survey of pharmacists providing primary care to patients in Ontario, Canada at community pharmacies, community health centers, and family health teams. The study population consisted of pharmacists registered in part A of the Register of Pharmacists with the Ontario College of Pharmacists (OCP), the province's regulatory body. Part A pharmacists are those that provide active patient care, dispense, sell, and compound medications, and supervise a pharmacy where medications are kept (OCP website, Pharmacy Act). Of the 13,072 Part A pharmacists in Ontario, 5,324 expressed an interest in participating in research during their initial registration with the OCP or when renewing their annual license. The OCP provided a list of names and email addresses of these pharmacists to the Ontario Pharmacy Research Collaboration, a multidisciplinary pharmacy practice research program, of which the authors of this study are members. All 5,324 pharmacists in OCP's registry were invited to participate in an online survey by email. The initial invitation and a description of the study's objectives were sent on September 15, 2014. Two additional reminders were sent by email on September 29, 2014 and October 13,2014. Pharmacists were given the opportunity complete the survey on paper and return it by fax or mail, but none had used this option. The link to the survey was active for 6 weeks and directed pharmacists to Qualtrics, an online survey administration tool (Qualtrics, Provo, UT, September, 2014). Respondents were anonymous; i.e., responses could not be linked to responder. At the end of the survey, respondents were provided another link to enter their name and contact details in a draw for an iPad Air in appreciation for their time, along with an opportunity to indicate interest in participating in future educational opportunities; however, this information was not linked to survey respondents.
The developed survey was pilot tested by five community and primary care pharmacists to assess wording, structure, layout, and readability, and to provide an estimate for completion time. The survey was modified based on their feedback. The instrument's face and content validity were determined through multiple separate reviews and discussions by the study's authors. It was estimated the survey would take 20 minutes to complete.
The final survey instrument consisted of 31 questions, several of which had sub-questions. Of these questions, seven assessed pharmacists' knowledge, five investigated attitudes and perceptions, and seven captured demographic characteristics. An additional twelve questions addressed pharmacy practice patterns, and pharmacists' self-efficacy in managing chronic pain, but these data are not presented in this paper.
Within the knowledge-assessment questions, were eight statements on CLBP, eight on CHDs and seven on PDN that responders were asked to indicate as true, false, or not sure. The other four knowledge questions were on opioid use as follows: one yes/no question asking respondents if they were familiar with the Canadian guideline for the safe and effective use of opioids in chronic non-cancer pain, one question where respondents were asked to provide the guideline's recommended "watchful dose" of opioids per day in mg morphine equivalents, and two true/false statements assessing knowledge about the rise of overdoses related to prescription opioids in Ontario and the potential for patients with a legitimate pain problem to become addicted to prescription opioids.
In two of the five questions addressing pharmacists' attitudes and perceptions of chronic pain, participants were asked to respond to questions on the importance of incorporating complementary treatment and importance of opioids in the treatment of CLBP, CHDs, and PDN on a five-point Likert scale (options as follows: always, very often, sometimes, rarely, never). The remaining three attitudes and perceptions questions asked participants to indicate how often under-dosing was the reason for ongoing pain for each condition on a five-point Likert scale (options as follows: always, very often, sometimes, rarely, never), how much improvement was expected during the course of illness for each condition on a five-point Likert scale (options as follows: full improvement, major improvement, moderate improvement, little improvement, no improvement), and how they felt when approached by patients with CLBP, CHDs and PDN, each on a three-point Likert scale (options as follows: burdened, neutral, comfortable). Respondents were permitted to answer as many questions as they desired and to skip over questions.

Data analysis
Study data were recorded using Qualtrics (www.qualtrics.com; Qualtrics, Provo, UT, September 2014) and analyzed using the Statistical Package for Social Sciences (SPSS; IBM Corp, Armonk, NY. Version 22; 2013). Descriptive statistics of survey respondents and survey results were calculated. Chi-square tests were used to examine the relationship between demographic variables (gender, years of practice, hours of practice, and population size) and pharmacists' attitudes, knowledge, and perceptions toward treatment of chronic pain. Odds ratios and confidence intervals were estimated for statistically significant variables.
A post-hoc calculation for sample size indicated a sample size of 385 was required to detect significant differences for a margin of error of 0.05 and an assumed standard deviation of 0.5.

Response rate
Of the 5,324 pharmacists asked to take the online survey, 52 participant email addresses were undeliverable. A further 95 respondents indicated they did not practice in either a community pharmacy or a family health team/community health center, thus were automatically excluded from the survey. A total of 668 (12.9%) pharmacists responded to the survey, of whom 392 completed the survey.
Of the questions assessing knowledge of CLBP, 89.5% of responding pharmacists provided a correct answer to the true/false question on the relationship of CLBP to injury, 79.0% on whether back pain gets worse before getting better, 59.1% on the role of bed rest, and 86.5% on the effect of chronic ibuprofen use on blood pressure. Fewer than half of pharmacists provided a correct answer to the true/false question on various descriptions of chronic pain (44.3%), the appropriate dose of morphine for low back pain relief (35.4%), the role of non-prescription analgesics as a first-line treatment option (38.0%), and whether patients should avoid painful movement (13.0%) ( Table 2). Of the questions assessing knowledge of CHDs, 95.9% of responding pharmacists provided a correct answer to the true/false question on factors and stress that contribute to chronic tension-type headaches, 95.9% on whether migraine patients should be asked if their headaches restrict activities, 94.6% on whether headache sufferers should guard against medication overuse, and 83.7% on whether the use of prophylactic drug therapy is recommended for chronic headaches. Respondents scored less well on the true/false question on use of combination analgesics and codeine as a first-line treatment (53.52% provided a correct answer) and appropriate use of serotonin antagonists (i.e., triptans) for headaches (52.8% provided a correct answer). Less than a third of respondents (30.3%) indicated correctly whether migraine is primarily a disease of the brain with a well-established neurological basis, and less than half (42.8%) indicated correctly whether a trial of non-prescription medications should be tried first before taking prescription medications among patients who suffer from severe migraine.
Of the questions assessing knowledge of PDN, 90.6% respondents provided a correct answer to the true/false question whether PDN results from damage to nerves in a hypoglycemic environment and 90.0% provided a correct answer to the question asking if all patients with diabetes will develop PDN. However, about two-thirds (65.3%) provided a correct answer to the question whether PDN results from both peripheral and central sensitization mechanisms. Most (83.6%) correctly answered the question whether optimal glycemic control is the cornerstone of treatment, but just less than half (49.0%) answered the question about tramadol being a reasonable initial treatment option correctly. Slightly more than three-quarters of respondents (77.6%) correctly answered the question whether a reasonable starting dose for gabapentin in a patient with severe renal failure is 600 mg three times a day and 78.6% correctly answered the question whether tricyclic antidepressants are effective ( Table 1).
Knowledge of opioid management. Slightly less than half (48%) of respondents were familiar with the Canadian guideline for safe and effective opioid use for chronic non-cancer pain. Almost all pharmacists (96%) were aware that inadvertent overdoses related to prescription opioids has been increasing in Ontario, and 72% correctly answered the question whether patients can become addicted to prescription opioids if they have a legitimate pain problem. Only 276 of 392 pharmacists who completed the survey provided responses to the request for the guideline recommended "watchful dose" of opioids per day in mg morphine equivalents. Of these respondents, 52% answered correctly that 200 mg morphine equivalents per day was the recommended dose, and 48% indicated inaccurate doses, ranging from 0 to 2000 mg morphine equivalents per day.
Perceptions and attitudes of pharmacists. Slightly greater than half (52%) of responding pharmacists considered complementary therapies (physiotherapy, acupuncture, and exercise) as "very important" to incorporate into treatment of patients with CLBP; however, only 21% and 22% of respondents indicated these complementary therapies as being very important to incorporate into treatment of patients with CHDs and PDN, respectively (p<0.001 across all comparisons). Slightly more than half of respondents indicated that under-dosing with analgesics was moderately important as the reason for ongoing pain, with 52% indicating this for CLBP, 53% for CHDs and 52% for PDN. Almost half of respondents (47%) indicated that opioids were "very important" or "important" to treat CLBP, whereas 23% indicated that opioids were "very important" or "important" to treat CHDs, and 24% indicated that opioids were "very important" or "important" to treat PDN. Few pharmacists expected full improvement from CLBP, CHDs, or PDN (3%, 3% and 1%, respectively); however, 24%, 39%, and 20% expected major improvement and 38%, 48%, and 55% expected moderate improvement in these disorders, respectively. When asked how they feel when approached by patients with CLBP, CHDs, and PDN, most pharmacists indicated they were comfortable (67%, 62%, and 58%, respectively). In contrast, few pharmacists indicated they felt burdened (7%, 4%, and 4%, respectively).

Subgroup analyses
Differences in responses between male and female respondents. Female respondents were significantly less likely than their male counterparts to answer the following question correctly, "Appropriate dose of morphine for pain is whatever dose relieves the pain as completely as possible," (56% males vs. 44% females, p<0.05) as related to CLBP (Table 3). Female respondents were significantly more likely than their male counterparts to answer the following question correctly, "Triptans should be reserved for patients who have failed at least two other prescription medications," (40% males vs. 60% females, OR = 2.0 (95% CI 1.37-3.07), p<0.001). Other responses by gender were not significantly different.
Differences in responses based on years of practice. Of the 23 questions on knowledge of different pain conditions, significantly more pharmacists with ten or more years of practice experience answered eight questions correctly (Table 3). Among pharmacists with ten or more years of practice experience, the odds ratios of correctly answering the questions, "Back pain usually gets worse before it gets better," "Chronic low back pain can be described as aching, burning, stabbing, tingling, dull, or sharp," "Appropriate dose of morphine for pain is whatever dose relieves the pain as completely as possible," were 2.56 (95% CI: 1.538-4.276, p<0.001), 2.116 (95% CI: 1.325-3.379, p<0.05), and 2.059 (95% CI: 1.241-3.417; p <0.05), respectively, compared with those with fewer than ten years of experience. Pharmacists who had ten or more years of experience were also more likely to correctly answer questions on factors that contribute to headaches and medication overuse than those with less than ten years of experience (OR = 3.597, 95% CI: 1.305-6.917, p<0.05 and OR = 2.532, 95% CI: 1.043-6.148, p<0.05, respectively). Similarly, pharmacists with ten or more years of experience correctly answered questions related to central and peripheral sensitization in PDN (76% vs. 24%; p<0.05), importance of optimal glycemic control in PDN (76% vs. 24%; p<0.05) and effectiveness of tricyclic antidepressants (69% vs. 31%; p<0.05) than did those with less than ten years of experience.
Differences based on population size. Seventy-five percent of pharmacists whose practice location was in an urban area (population >100,000) correctly answered the question, "Appropriate dose of morphine for pain is whatever dose relieves the pain as completely as possible," whereas among those who practiced in rural locations (population: <10,000-100,000) only 25% did (p<0.001). Other responses by size of city or town in which pharmacists practiced were not significantly significant.
Gender, years of practice, population size of location of practice, and opioid management. Differences in the proportions of pharmacists who answered questions correctly on knowledge of opioid management based on gender, years of practice, or population size of practice location were not significant

Discussion
Respondents had a good understanding of the pathophysiology and pharmacotherapy associated with CLBP, CHDs, and PDN, with an overall score of 66%. Nevertheless, there were statistically significant differences in the overall scores obtained in PDN (76%) as compared to CHDs (69%) and CLBP respectively (76% vs. 69% vs. 55%; p<0.001). The difference in the scores may indicate a greater familiarity with both headache disorders and diabetic neuropathy among pharmacists and a greater need to develop more educational content addressing the pathophysiology and pharmacotherapy of CLBP. For instance, less than 50% of the  respondents answered four of eight questions correctly. In the case of the question, "Patients should avoid all painful movements", only 13% answered correctly and only 38.0% answered the question, "Treatment with non-prescription analgesics should be first line." Furthermore, deficiencies were also noted with respect to opioid management. Only about half the respondents were familiar with the 2010 Canadian guideline for safe and effective opioid use for chronic non-cancer pain. While 70% of the 392 who completed the survey attempted to answer the question asking the recommended "watchful dose" per day of opioids in mg morphine equivalents, only half were correct, and incorrect answers ranged widely from 0 to 2,000 mg of morphine equivalents per day. If one were to assume that the remaining 30% did not have adequate confidence to answer this question, only 36% of the pharmacists were aware of the recommended "watchful dose" per day of opioids. The concept of the watchful dose presents a fundamental change in chronic pain treatment with opioids. Previously, it was acceptable to not limit opioid dosing if patients reported continued pain. However, the  Canadian guideline includes a recommendation for a watchful dose because "chronic non-cancer pain can be managed effectively in most patients with dosages at or below 200 mg/day of morphine or equivalent" and that "consideration of a higher dosage requires careful reassessment of the pain and of risk for misuse, and frequent monitoring with evidence of improved patient outcomes" [42]. The need for this limit is reinforced by the evidence that the odds of inadvertent overdose increases 3-fold with daily doses above 200 mg of morphine equivalents [17]. It is critical this recommended change is incorporated into frontline practice.
This study's findings are consistent with previous research indicating that clinicians are generally knowledgeable about opioids but are less adept at handling questions related to their use [21,[43][44][45][46]. This study's findings are also consistent with previous research that acknowledges a need for additional education on chronic pain and opioid use [21,47]. Moreover, it demonstrates a need for guideline recommendations to be incorporated into practice. This is a core theme in Canada's national strategy called "First Do No Harm: Responding to Canada's Prescription Drug Crisis" [48]. Citing healthcare practitioner knowledge gaps in pain management and addiction, the eight education-related recommendations in the national strategy all relate to the need for increased training for healthcare practitioners.
More than half of the respondents had more than ten years of practice experience-experience that appears to be associated with greater knowledge of CLBP, CHDs and PDN. This finding is consistent with the notion that experience has a substantial direct effect on knowledge in higher complexity employment [49], but it differs from a finding that physicians with more practice experience possessed less accurate factual knowledge [50]. It is possible, however, that respondents in our study were a self-selecting sample who were more interested and engaged in chronic pain care provision than the larger pharmacist population in Ontario.
About 95% of Canada is rural, but depending on the definition used, rural regions comprise 19% to 30% of the country's population. Of the province-wide professional licensing database, 36% of respondents indicated that they practiced in regions with fewer than 100,000 people. There was largely no difference in knowledge associated with CLBP, CHDs and PDN between urban and rural practicing pharmacists, despite challenges associated with rural clinical practice, such as reduced access to educational opportunities and less peer support [51].
Pharmacists overall were positive about managing these chronic pain conditions, with about two-thirds reporting an expectation of moderate to major improvement. Among the three conditions, complementary therapy was noted to have the strongest role in treatment of CLBP, similar to the proportion who indicated that opioids play a role in treatment. About half of the respondents indicated that under-dosing is a moderately important cause of poor pain control.

Strengths
This study provides insight into the state of practice of community-based pharmacists with respect to chronic pain care provision in Ontario. Despite a modest response rate, to our knowledge this is the largest survey of pharmacists' knowledge, attitudes, and perceptions of patients with chronic pain in primary care conducted in Canada. In comparison to Lalonde et al, our study had a larger sample size (392 vs 110 completed surveys), higher male to female ratio (0.6:1 vs. 0.9: 1) and a greater percentage of pharmacists with substantial practice experience (34% vs. 51%) [21].

Limitations
The low response rate and the possibility that participating clinicians in our study may have been a self-selecting sample because of their experience or interest in chronic pain, may not be representative of Ontario's pharmacists overall. Furthermore, there is no possibility of assessing the nonresponse bias inherent in a voluntary participation survey. Our survey may have included a higher proportion of pharmacists who manage chronic pain actively, inflating estimates of knowledge scores. Even with pilot testing and question refinement, survey questions may still be ambiguous, result in bias, and, as with all responses that require recall or estimation, responses may have inaccurately reflected practice reality.

Conclusion
Community pharmacists and pharmacists who are part of family health teams frequently see patients with chronic pain, thus have the opportunity to improve their care. However, this survey demonstrates that pharmacist contributions to the effective and safe treatment of these patients, particularly with opioids, appears to require additional education. Appropriate educational initiatives and pharmacist training to develop skills and provide tools may improve the care of chronic pain patients.
Supporting Information S1 File. Data of chronic pain knowledge and attitude. (XLSX) S1 Table. Representativeness of population of pharmacists who responded to survey (N = 392) to the Ontario pharmacist population at large.