Physicians Knowledge and Attitudes Toward the Use of Analgesics for Cancer Pain Management: A Survey of Two Medical Centers in Taiwan

Transcription

1 Vol. 20 No. 5 November 2000 Journal of Pain and Symptom Management 335 Original Article Physicians Knowledge and Attitudes Toward the Use of Analgesics for Cancer Pain Management: A Survey of Two Medical Centers in Taiwan Luo-Ping Ger, RN, MPH, Shung-Tai Ho, MD, and Jhi-Joung Wang, MD, DMSc Department of Medical Education and Research (L.-P.G.), Kaohsiung Veterans General Hospital, Kaohsiung; and Department of Anesthesiology (L.-P.G., S.-T.H.), National Defense Medical Center/ Tri-Service General Hospital, Taipei, Taiwan; and Department of Anesthesiology (J.-J.W.), Chi Mei Foundation Hospital, Tainan, Taiwan, Republic of China Abstract The purposes of this study were to examine the attitudes of physicians regarding the optimal use of analgesics for cancer pain management (CPM), to evaluate their knowledge and attitudes toward opioid prescribing, and to comprehend their perceptions of the barriers to optimal CPM. A survey was conducted on 356 physicians with cancer patient care responsibilities practicing in two medical centers in Taiwan. A total of 204 (57%) physicians responded, including internists (28%), surgeons (27%), oncologists (11%), anesthesiologists (10%), and other specialties (24%). The majority of physicians displayed significantly inadequate knowledge and negative attitudes toward the optimal use of analgesics and opioid prescribing. Multivariate analyses showed that the following six categories of physicians would be inclined to have inadequate knowledge of opioid prescribing: 1) those with perception of good medical school training in CPM, 2) those with perception of poor residency or fellowship training in CPM, 3) those with a medical specialty in surgery, medicine, or oncology (vs. anesthesiology), 4) those with limited clinical experience in cancer patient care (number of patients less than 30), 5) those with a limited aim of pain relief, and 6) those with an underestimation of analgesic effect. Additionally, physicians with inadequate knowledge of opioid prescribing and with hesitation to intervene earlier with maximal dose of analgesia would be inclined to have reluctant attitudes toward opioid prescribing. The most important barriers to optimal CPM identified by physicians themselves were physician-related problems, such as inadequate guidance from a pain specialist, inadequate knowledge of CPM, and inadequate pain assessment. The results of this study suggest that active analgesic education programs are urgently needed in Taiwan. J Pain Symptom Manage 2000;20: U.S. Cancer Pain Relief Committee, Key Words Physician, cancer pain management, opioids, knowledge, attitude, correlates Address reprint requests to: Luo-Ping Ger, RN, MPH, Department of Medical Education and Research, Kaohsiung Veterans General Hospital, 386 Ta-Chung 1st Rd., Kaohsiung, Taiwan, Republic of China. Accepted for publication: December 18, Introduction Cancer has been the leading cause of death in Taiwan since In a previous study, we showed that 69% of newly diagnosed cancer patients with cancer-related pain received in- U.S. Cancer Pain Relief Committee, /00/$ see front matter Published by Elsevier, New York, New York PII S (00)

2 336 Ger et al. Vol. 20 No. 5 November 2000 adequate medication (not consistent with the analgesic ladder approach recommended by the World Health Organization) or no pain medication at all. 2 Studies of patient-related barriers to cancer pain management (CPM) revealed that there are eight types of concerns on reporting pain and taking pain medication. The eight types of patient-related barriers are fatalism, fear of addiction, concerns regarding unmanageable side effects, concern about not being a good patient, concern about distracting one s physician from treating of underlying disease, concern about pain meaning disease progression, concern about becoming tolerant to pain medications, and fear of injection. Most of these concerns are important barriers to adequate pain medication. 3 5 However, our previous patient survey showed that only concern of disease progression was correlated to adequate pain medication. 6 Patient-related barriers are not extremely important to adequate pain medication in Taiwan. The ratios of daily morphine consumption (per million people) in Taiwan against other countries are 1 to 41, 1 to 25, and 1 to 4, in comparison with the United Kingdom, the United States, and Japan, respectively. The low consumption indicates that many Taiwanese physicians are still reluctant to prescribe morphine to control cancer pain. Hence, there are still many cancer patients suffering unrelieved pain in Taiwan. 7 Recent studies of medical students and physicians demonstrate a significant lack of knowledge in theory and application of analgesic drugs, particularly the use of opioids for CPM Insufficient physician knowledge and education in CPM has been suggested as one of the major factors contributing to inadequate pain relief in cancer patients throughout the world. 15,16 In order to work toward optimal CPM, it is essential to understand the knowledge, attitudes, and barriers of physicians. The purpose of this study was to examine Taiwanese physicians attitudes toward the optimal use of analgesics for CPM and their knowledge of and attitudes toward opioid prescribing. The correlates of their knowledge and attitudes toward opioid prescribing were evaluated. The physicians perception of the barriers to optimal CPM was also studied. The findings presented are based on a survey of physicians with cancer patient care responsibilities practicing in two medical centers in Taiwan. Methods Subjects The survey was conducted at two medical centers, Kaohsiung Veterans General Hospital (KS- VGH) and Tri-Service General Hospital (TSGH), in Taiwan. A questionnaire was sent to 356 physicians with cancer patient care responsibilities. The overall response rate was 57% (204/356). The physicians who responded included 193 males (95%) and 11 females (5%) and had a mean age of 36.4 years (SD 5.6, range 25 60). Instruments The questionnaire was designed to assess the physicians on the following aspects: 1) attitude toward the optimal use of analgesics for CPM (6 questions); 2) knowledge and attitudes toward opioid prescribing (16 questions); and 3) perception of barriers to CPM (14 questions). The questionnaire was organized in three sections. The first section consisted of 6 questions which assessed the physician s attitude toward the optimal use of analgesics for CPM. All of these questions were taken from a review of questionnaires used in similar surveys. 10,17,18 The second section contained questions for establishing two scales that assessed physicians knowledge and their attitudes to prescribe opioids, respectively. There were 11 questions, 10 negative and 1 positive, for the scale of knowledge to prescribe opioids. There were 8 negative questions for the scale of attitude to prescribe opioids. A 5-point Likert response format was utilized, ranging from strongly agree to strongly disagree, with a score of 1 to 5 for negative questions and with a score of 5 to 1 for positive questions. Content validity of these two scales was determined by an expert panel, which consisted of three experienced anesthesiologists and an oncology nurse. After assessing internal reliability, 3 of the 11 knowledge items were omitted from the knowledge scale because of their low correlation to the other eight items. The coefficient alpha was 0.81 for the knowledge scale (with only 8 items); if an item was deleted, the alpha values for the scale were in the range of 0.75 to For the attitude scale, the coefficient alpha was 0.80; the alpha values for the scale were also in the range of 0.75 to 0.80 if an item was deleted. The third section consisted of a list of 14 barriers to optimal CPM. Physicians were asked to

3 Vol. 20 No. 5 November 2000 Physicians Barriers toward Cancer Pain Management 337 indicate whether these CPM barriers existed in their hospitals. The barriers listed include factors related to health care system, health care professionals, and patients. These barriers were selected from potential barriers suggested in the published literature. 13,19,20 Statistical Analysis Several statistical methods were used in data analysis. In the univariate analysis, the association between categorical predictors and the outcome (knowledge or attitude to prescribe opioids) was tested for significance by a oneway-analysis of variance (ANOVA), Kruskal- Wallis one-way ANOVA, t-test, or Mann-Whitney U test. In the multivariate analysis, the simultaneous relationship of the multiple correlates to the knowledge deficits to prescribe opioids (or reluctant attitude to prescribe opioids) was assessed using a logistic regression model. The correlates in the model were classified as dummy variables (0,1); the score (0,1) of the dummy variables was determined by the results of univariate analysis. A P value 0.05 was considered statistically significant. Results Background Characteristics The personal and practice characteristics of the 204 physicians who responded to the questionnaire are presented in Table 1. Specialties included internal medicine (28%), surgery (27%), oncology (11%), anesthesiology (10%), and others (24%). The mean (SD) year in clinical practice was 10.4 (5.6) years, with a range of 2 to 24 years. The median number of cancer patients cared for per physician was 20 in the six months prior to the survey. Collectively, this group of physicians treated 11,266 cancer patients during that period. Attitudes toward the Optimal Use of Analgesics for Cancer Pain Management The responses of physicians to the attitude questions are shown in Table 2. Although the majority of physicians (66%) felt that the patients with pain were undermedicated, most of the physicians displayed a negative attitude toward analgesic medication. For example, only 31% of the physicians aimed for complete pain relief for the patients while the rest of physicians sought just to diminish the patients pain. The Table 1 Background Characteristics of Physicians Characteristics N % Sex Male 193 (95) Female 11 (5) Age (yrs) (21) (36) (28) (15) Seniority Head 30 (15) Attending 82 (40) Resident 92 (45) Medical specialty Medicine and subspecialties 57 (28) Surgery and subspecialties 56 (27) Oncology 22 (11) Anesthesiology 20 (10) Gynecology 14 (7) Family practice 15 (7) Others 20 (10) No. of cancer patients being cared for in the past 6 months (18) (23) (28) (21) (10) majority of physicians (80%) believed that only 20 60% cancer pain could be relieved with pharmacological pain relievers. An overwhelming majority (86%) would wait until the patient s prognosis was 6 months or less before they would start maximal analgesia for severe pain without consideration of tolerance. Additionally, 72% of the physicians believed that the incidence of addiction to narcotic pain relieving drugs was more than 1%; 30% of physicians even believed that the incidence was more than 10%. Furthermore, when a cancer patient requested increasing amounts of analgesics to control pain, only a minority of physicians (25%) believed the patient was experiencing an increase in pain, while the majority (75%) judged that the patient either had developed tolerance to drug or was psychologically addicted (Table 2). Knowledge of Opioid Prescribing and the Correlates The mean (SD) score was 2.91 (0.68; range ) for this scale. The responses of physicians to knowledge questions are shown in Table 3. The majority of physicians displayed substantial knowledge deficits of opioid prescribing. For example, when patients needed

4 338 Ger et al. Vol. 20 No. 5 November 2000 Table 2 Physicians Attitudes Toward the Optimal Use of Analgesics for Cancer Pain Management Attitudes N % Which of the following is true at your hospital? Most patients receive adequate pain treatment 62 (30) Most patients receive more pain medication than necessary 8 (4) Most patients in pain are undermedicated 134 (66) Which of the following is the most appropriate pain relief during the period of treatment at your hospital? Pain is completely abated 63 (31) Pain is diminished, noticed but not distressing 137 (67) Pain is diminished only during the maximal effect of analgesics 4 (2) What percentage of pain can be relieved with pharmacological pain relievers? 20% 7 (3) 40% 48 (24) 60% 109 (53) 80% 36 (18) 100% 4 (2) At what stage would you feel it is appropriate for a patient to receive maximal doses of analgesics for severe pain without consideration about tolerance? Prognosis 1 month 20 (10) Prognosis 1 3 months 88 (43) Prognosis 4 6 months 67 (33) Prognosis 7 12 months 20 (10) Prognosis months 9 (4) The incidence of addiction as a result of the legitimate prescription of narcotic pain-relieving drugs in cancer patients is: 10% 62 (30) 10 1% 85 (42) 1 0.1% 40 (20) 0.1% 17 (8) When a cancer patient requests increasing amounts of analgesic to control pain, this usually indicates: Patient is psychologically addicted 33 (16) Patient is experiencing increased pain 50 (25) Patient has developed tolerance to drug 119 (58) All of the above 2 (1) potent opioids, more than half of the physicians preferred meperidine. In addition, 54% of the physicians had the wrong idea (including 37% agreed and 17% had no any opinions) that meperidine caused less harmful effect in long-term opioid use. For patients with moderate or severe pain, 54% of the physicians wanted to prescribe meperidine (50 mg every 4 hours as needed). For persistent and severe pain, 59% of the physicians agreed to increase the dosage of potent opioids to every 4 hours as needed dosing. Furthermore, 42% of the physicians had the misconception (including 26% agreed and 16% had no any opinions) that a PRN dosing schedule could decrease the harmful effect of opioids. The majority of physicians (73%) believed that most patients preferred parenteral administration to oral; 66% of physicians believed that parenteral administration is more efficacious. Additionally, some of them (30%) would not let patients take morphine orally, even though their patients could eat food normally (Table 3). Various physician characteristics were evaluated by logistic regression to identify what kind of physicians would be inclined to have inadequate knowledge to prescribe opioids. The mean score (2.91) and median score (2.80) were out of a perfect mean score of A physician was classified as knowledge deficits to prescribe opioids if his/her mean score was equal to or less than 3.0 on the knowledge scale. Therefore, a binomial variable (inadequate knowledge vs. adequate knowledge to prescribe opioids) was defined and used as the dependent variable in the logistic regression. The multivariate analyses show that the following six categories of physicians would be inclined to have inadequate knowledge of opioid prescribing: 1) those with perception of good medical school training in CPM, 2) those with perception of poor residency or fellowship training in CPM, 3) those with a medical specialty in surgery, medicine, or oncology (vs. anesthesiology), 4) those with limited clinical experience in cancer patient care (number of

5 Vol. 20 No. 5 November 2000 Physicians Barriers toward Cancer Pain Management 339 Table 3 Physicians Knowledge of Opioid Prescribing Items Strongly agree Agree No opinion Disagree Strongly disagree 1. When patients need potent opioids, I would prescribe meperidine rather than morphine Meperidine causes less harmful effects (such as tolerance, addiction, or side effect) in long-term opioid use For cancer patients with moderate or severe pain, I would prescribe meperidine 50 mg q 4 h, PRN, IM For patients with persistent and severe pain, I would increase potent opioid dosage and administer it q 4 h, PRN Administering opioids in a PRN dosing schedule can decrease the harmful effect of opioids, such as tolerance, addiction, or side effect Most patients prefer the parenteral administration to oral administration Parenteral administration is more efficacious than oral administration in pain management The absorption of oral morphine in the GI tract is slow and incomplete. Even though my patients can eat food normally, I do not like them to take morphine orally Others: 1. When the patient s renal function is impaired, I don t like to prescribe meperidine for pain treatment Oral administration of morphine is more inclined to induce side effects of nausea and vomiting than parenteral administration Oral morphine absorbed from the gut is subject to first-pass metabolism in the liver and this would make 1 / 3 morphine ineffective patients less than 30), 5) those with a limited aim of pain relief, and 6) those with an underestimation of analgesic effect (Table 4). The odds ratio in Table 4 is an indication of the risk of a group of physicians having knowledge deficits in comparison with the reference group of physicians. For example, physicians with perception of good (or fair) medical school training in CPM had a 2.53-times higher risk for knowledge deficits of opioid prescribing compared with those with perception of poor school training. Attitudes toward Opioid Prescribing and the Correlates The mean (SD) score was 3.15 (0.58, range ) for this scale. The responses of physicians to each item are shown in Table 5. When prescribing opioids, most physicians (73%) agreed that they were very careful in the control of dosage and frequency for the prevention of drug tolerance and addiction. Some physicians (26%) did not object (agree or have no opinion) that the opioid dosage patients received must be much lower than the required dosage for the prevention of drug tolerance. Additionally, a significant number of physicians (34%) did not object to insinuating to patients or relatives that opioids were not good drugs and they had better bear the pain as much as possible. Many physicians (28%) had no objection to encouraging patients to bear severe pain and refuse the morphine injection. When patients experienced severe opioid side effects, the following percentages of physicians would prescribe opioids infrequently or with lower dosage, 63% for abdomen distention, 60% for nausea or vomiting, and 44% for severe constipation. Only 10% physicians would not prescribe opioids due to their belief of respiratory depression being a severe side effect (Table 5). Various physician characteristics were evaluated by logistic regression to identify what physician characteristics contributed to their reluctance to prescribe opioids. The mean score (3.15) and median score (3.10) were out of a perfect mean score of A physician was classified as reluctant to prescribe opioids if his/ her mean score was equal to or less than 3.0 on the attitude scale. Therefore, a binomial variable (reluctant attitude vs. inclined attitude to prescribe opioids) was defined and used as the dependent variable in the logistic regression. The multivariate analyses selected 4 of these predictors as significant (Table 6). The odds ratio in Table 6 is an indication of the risk of a group of physicians having a reluctant attitude in compar-

6 340 Ger et al. Vol. 20 No. 5 November 2000 Table 4 Logistic Regression Analysis of Various Factors Associated with Knowledge Deficits of Opioid Prescribing Variables Percentage of physicians in category Percentage with inadequate knowledge Coefficient SE Odds ratio (95% CI) Previous medical school training Poor Good or fair ( ) e Previous residency or fellowship training Good or fair Poor ( ) e Medical specialty Anesthesiology Oncology a ( ) e Medicine and subspecialties b ( ) f Surgery and subspecialties c ( ) f No. of cancer patients being cared for in the past 6 months ( ) f ( ) e The aim of most appropriate pain relief Pain is completely abated Pain is diminished, noticed but ( ) e not distressing d Belief of percentage of pain relieved by pharmacological pain relievers % % ( ) e a including hematology and radiation therapy b including family practice c including gynecology d including 4 (2%) whose patients goal was pain being diminished only during the maximal effect of analgesics. e P 0.05 f P 0.01 ison with the reference group of physician. For example, physicians in the Department of Medicine had a 5.92-times higher risk of reluctance to prescribe opioids compared with those in anesthesiology. Furthermore, as we put the status of physicians knowledge deficits to prescribe the opioids into the potential candidate predictors, only two significant predictors were selected in the multivariate analysis. Physicians with inadequate knowledge of opioid prescribing had a 9.02-times higher risk of reluctance to prescribe opioids compared with physicians with more adequate knowledge. Additionally, physicians with more hesitation in maximal dose of analgesia (expected prognosis 6 months) would have a 3.77-times risk of reluctance to prescribe opioids compared with physicians with less hesitation (expected prognosis 6 months). Identified Barriers to Optimal Pain Management Physicians were asked to rank a list of 14 potential barriers to optimal CPM in terms of how they might impede CPM in their hospital. Table 7 lists the percentage of 181 respondents (89%) who ranked each item as one of the four top barriers. The four most important barriers to adequate pain management identified were inadequate guidance of a pain specialist, inadequate staff knowledge of pain management, inadequate assessment of pain, and lack of psychosocial support services. Discussion This study showed that the majority of physicians (66%) recognized the problem of inadequate pain management in these two hospitals surveyed in Taiwan (Table 3). However, they displayed significantly more negative attitudes toward analgesic medication than physicians in other studies. 13,18,20,21,22 A large majority of physicians (80%) in this study underestimate the pain relief effect of analgesics in contrast with the physicians (30

7 Vol. 20 No. 5 November 2000 Physicians Barriers toward Cancer Pain Management 341 Table 5 Physicians Attitudes Toward Opioid Prescribing Items Strongly agree Agree No opinion Disagree Strongly disagree 1. When prescribing opioids, I would be very careful in the control of dosage and frequency for the prevention of drug tolerance and addiction The opioid dosage patients receive should be much lower than the required dosage for the prevention of drug tolerance When I prescribe opioids, I would insinuate to patients or their relatives that opioids are not good drugs and they had better bear the pain as much as possible When I find patients who bear severe pain and refuse the morphine injection, I would encourage their behavior I do not like to prescribe opioids, because respiratory depression is a very severe side effect For patients with severe nausea or vomiting, I would prescribe opioids infrequently or with lower dosage For patients with severe abdomen distention, I would prescribe opioids infrequently or with lower dosage For patients with severe constipation, I would prescribe opioids infrequently or with lower dosage %) in other studies. 18,21 Additionally, more physicians (86%) in this study were concerned about the rapid development of tolerance and delayed the use of maximal analgesia until expected prognosis was short ( 6 months) than physicians (23 31%) in the Western countries. 13,18,22 This concern had significantly induced their reluctance to prescribe opioids in our study (Table 6). Furthermore, more physicians (72%) in this study overestimated the likelihood of addiction (addiction rate 1%) than physicians (27%) in another study. 18 Because of the undue fear of tolerance and addiction, 73% of the physicians in this study indicated that they would be very careful in the control of opioid dosage and frequency. Therefore, 69% of them set a limited aim of pain relief at a diminished level. This result was quite different from Variables Table 6 Logistic Regression Analysis of Various Factors Associated with Physicians Reluctant Attitudes Toward Opioid Prescribing Percentage of physicians in category Percentage with reluctance to prescribe opioids Coefficient SE Odds ratio (95% CI) Medical specialty Anesthesiology Oncology a ( ) Medicine and subspecialties b ( ) e Surgery and subspecialties c ( ) e No. of cancer patients being cared for in the past 6 months ( ) e Patient-self being the best judge of pain severity Yes No d ( ) f Using maximal doses of analgesics for severe pain without care about tolerance Expected prognosis 6 months Expected prognosis 6 months ( ) f a including hematology and radiation therapy b including family practice c including gynecology d the best judge being physician (20%), nurse (7%), or patient s caregiver (14%) e p f p 0.05

8 342 Ger et al. Vol. 20 No. 5 November 2000 Table 7 Identified Barriers to Cancer Pain Management by the Physicians (n 181) Type of barriers No. (%) a Inadequate guidance of pain specialist 110 (61) Inadequate staff knowledge of pain management 103 (57) Inadequate pain assessment 98 (54) Lack of psychosocial support services 97 (54) Lack of neurodestructive procedures 56 (31) Physician reluctance to prescribe opioids 45 (25) Lack of access to wide range of opioids 41 (23) Patients relatives reluctance to permit patients to take opioids 37 (20) Excessive regulation of opioids in Narcotics Bureau, Department of Health, Republic of China 35 (19) Excessive regulation of opioids in pharmacy 34 (19) Patients reluctance to take opioids 28 (15) Nursing staff reluctance to administer opioids 18 (10) Patients reluctance to report pain 13 (7) Lack of equipment 9 (5) a Percentage of respondents who selected item as one of the top four barriers in the survey questionnaire. Ferrell et al. s study, which showed that 98% health care professionals correctly believed that patients should not have to endure pain. 20 The underlying reason for excessive morbid fear of opioids is in some way related to the historic and cultural background. In the nineteenth century, opium was widely abused in China. In order to correct this situation, the Chinese government has a very stringent policy toward any opioids. The message that such drugs are bad is deeply ingrained in Chinese people s mind, including those of the physicians. As a result, the legitimate medical use of analgesic drugs is overshadowed by undue fear of addiction, consciously or subconsciously. Although analgesic drug therapy is the mainstay of treatment, opioid use still remained an unsettled issue at these two medical centers in Taiwan. In comparison with physicians in the Western countries, many physicians in this study showed significantly inadequate knowledge of fundamental facts of opioid pharmacology, such as the choice of potent opioids (morphine or meperidine), 26 schedules (aroundthe-clock or PRN), 26,27 and routes (by oral or parenteral) (Table 3) The results of multivariate analysis showed that the root causes of knowledge deficits and reluctant attitudes were insufficient education (in medical school, residency, and fellowship training) and limited clinical practices (with medical specialty in surgery or medicine, and number of cancer patients less than 10). This study found that physician s self-evaluation of the quality of CPM training (either in medical school or in residency training) was correlated with their knowledge but not with their attitudes to prescribe opioids. It is interesting to note that those who perceived themselves as having good/fair school training actually are at higher risk of having inadequate knowledge. This phenomenon may be related to the fact that, in general, the medical education on CPM is rather poor in the medical schools in Taiwan. Hence, the perceived good/fair school education does not provide a solid foundation for accurate CPM knowledge. Because of the wrong perception, they might not actively participate in the ongoing CPM training programs provided for them. Therefore, they were inclined to have inadequate knowledge of opioid prescribing. A further study is needed to identify the root causes of this inverse association between school training and opioid knowledge. Medical specialty has a profound effect on the knowledge of and the attitude toward appropriate morphine use in CPM. 13,22,26,29 31 Our results were similar to the above finding except that anesthesiologists appeared more knowledgeable about opioid prescription in comparison to oncologists (Table 4). This study revealed that the majority of oncologists (73%) rated their medical school training as poor in comparison with other specialties (51%, , P 0.049) and almost all of them (96%) rated their residency or fellowship training as good in comparison with other specialties (64%, , P 0.003). It is interesting to note that if the quality of their previous medical training in CPM was not adjusted in the logistic model of knowledge deficits, oncologists were as knowledgeable as anesthesiologists on opioid use.

9 Vol. 20 No. 5 November 2000 Physicians Barriers toward Cancer Pain Management 343 However, when the quality of their previous medical training in CPM was adjusted, oncologists appeared to have worse opioid knowledge in comparison with anesthesiologists (Table 4). Therefore, both education and clinical practices had significant influence on physicians knowledge of opioid prescribing. Like medical specialty, patient volume was an another important variable. In the published literature, most of the studies supported the finding of Cleeland et al. that the number of cancer patients they cared for does not correlate with liberal attitudes toward CPM, in contrast to the finding of Levin et al. 17,26,30,31 This study found that patient volume was an another variable that affected the physicians knowledge of opioids. Physicians who treated more cancer patients would be inclined to have correct knowledge of and liberal attitude toward opioid prescribing. The implication of low patient volume is that the physicians are likely to rely heavily on school training, which is deficient to begin with. With more patients, the physicians have more opportunities to correct their wrong perception on CPM. Thus, clinical practices might be an important factor for physicians knowledge and attitudes where CPM education is inadequate. This study showed that physicians targeted level of pain relief was correlated with their knowledge deficits but not with reluctant attitudes toward opioid prescribing. Von Roenn et al. found that physicians who insisted on a total pain relief as the goal of pain treatment would treat pain aggressively earlier in the disease progression. 13 Therefore, we found that targeted level of pain relief was not in the logistic regression model since physicians liberal attitudes to intervene earlier with maximum analgesic therapy was already included in logistic regression model of the reluctant attitude (Table 6). The four most important barriers to adequate pain management identified by our physicians were all problems related to physician themselves, except for lack of psychosocial support services which was health care system related (Table 7). This result supported the results of our previous patient survey that most patients concerns were not the major barriers of adequate pain medication in Taiwan. 6 Additionally, this finding was comparable with Ferrell et al. and Furstenberg et al. in American studies and with Sapir et al. in an Israeli study. 20,28,32 This study was carried out at only two out of the nine major medical centers, and the external validity (representative) might be limited in Taiwan. The inverse relationship between medical school education and knowledge must be further studied to confirm the root causes for such an observation. Nevertheless, the results revealed that active analgesic education programs for CPM should be intensified in the basic medical education, in residency training programs, and in fellowship training programs. The focus of education should include proper assessment of cancer pain, basic pharmacology of the opioid analgesics, optimal use of analgesics for CPM, and prophylactic treatment of opioid side effects. With proper education, there is a better chance that physicians involved in clinical practices will have adequate knowledge and technique on CPM. In addition, further nation-wide health care professional survey is needed to provide additional support to push for effective CPM in Taiwan. In conclusion, most physicians displayed significantly inadequate knowledge and negative attitudes toward the optimal use of analgesics for cancer pain and opioid prescribing in Taiwan. This was especially true for physicians with insufficient education and limited experience with cancer pain. Many efforts are required to correct this situation. Acknowledgments Portions of this research were supported by grant No. DOH87-TD-1043, VGHKS89-12, and NSC B-075B-003. We thank Dr. Paul Lee for his linguistic review. References 1. Department of Health, Executive Yuan, Republic of China. Health statistics, Vol II. Vital statistics Taipei: Department of Health, Ger LP, Ho ST, Wang JJ, et al. The prevalence and severity of cancer pain: a study of newly-diagnosed cancer patients in Taiwan. J Pain Symptom Manage 1998;15: Ward SE, Goldberg N, Miller-McCauley V, et al. Patient-related barriers to management of cancer pain. Pain 1993;52: Ward SE, Hernandez L. Patient-related barriers

10 344 Ger et al. Vol. 20 No. 5 November 2000 to management of cancer pain in Puerto Rico. Pain 1994;58: Lin CC, Ward SE. Patient-related barriers to cancer pain management in Taiwan. Cancer Nurs 1995;18: Wang KY, Ho ST, Ger LP, et al. Patient barriers to cancer pain management: from the viewpoint of the cancer patients receiving analgesics in a teaching hospital of Taiwan. Acta Anaesthesiol Sin (ROC) 1997;35: Hsu CC, Li JH. The trends of requirements in medical opioid analgesics from 1987 through 1996 in Taiwan. Chin J Public Health (ROC) 1998;17: Peteet J, Tay V, Cohen G, et al. Pain characteristics and treatment in an outpatient cancer population. Cancer 1986;57: Cleeland CS. Pain control: public and physician attitudes. In : Hill CS, Fields WS, eds. Advances in pain research and therapy, Vol 11. New York: Raven, 1989: Weissman DE, Dahl JL. Attitudes about cancer pain: a survey of Wisconsin s first-year medical students. J Pain Symptom Manage 1990;5: Grossman SA, Sheidler VR, Swedeen K, et al. Correlation of patient and caregiver ratings of cancer pain. J Pain Symptom Manage 1991;6: Joranson DE, Cleeland CS, Weissman DE, et al. Opioids in chronic cancer and non cancer pain: a survey of state medical board members. Fed Bull Med J Licensure Discipline 1992;79: VonRoenn JH, Cleeland CS, Gonin R, et al. Physician attitudes and practice in cancer pain management: a survey from the Eastern Cooperative Oncology Group. Ann Intern Med 1993;119: Ferrell BR, McGuire DB, Donovan MI. Knowledge and beliefs regarding pain in a sample of nursing faculty. J Prof Nurs 1993;9: Marks RM, Sachar EJ. Undertreatment of medical inpatients with narcotic analgesics. Ann Intern Med 1973;78: Stjernsward J. Cancer pain relief: an important global public health issue. Clin J Pain 1985;1: Cleeland CS, Cleeland LM, Dar R, et al. Factors influencing physician management of cancer pain. Cancer 1986;58: Fife BL, Irick N, Painter JD. A comparative study of the attitudes of physicians and nurses toward the management of cancer pain. J Pain Symptom Manage 1993;8: Cleeland CS. The impact of pain on the patient with cancer. Cancer 1984;54: Ferrell BR, Dean GE, Grant M, et al. An institutional commitment to pain management. J Clin Oncol 1995;13: Ventafridda V, Tamburini M, Caraceni A, et al. A validation study of the WHO method for cancer pain relief. Cancer 1987;59: White ID, Hoskin PJ, Hanks GW, et al. Analgesics in cancer pain: current practice and beliefs. Br J Cancer 1991;63: Jacox A, Carr DB, Payne R, et al. U.S. Department of Health Services. Management of cancer pain. Clinical practice guideline, no. 9. U.S. Department of Health Services, Public Health Service, Agency for Health Care Policy and Research, Foley KM. Controversies in cancer pain: medical perspectives. Cancer 1989;63: Portenoy RK. Cancer pain management. Semin Oncol 1993;20(Suppl): Elliott TE, Elliott BA. Physician attitudes and beliefs about use of morphine for cancer pain. J Pain Symptom Manage 1992;7: Mercadante S, Salvaggio L. Cancer pain knowledge in Southern Italy: data from a postgraduate refresher course. J Pain Symptom Manage 1996;11: Furstenberg CT, Ahles TA, Whedon MB, et al. Knowledge and attitudes of health-care providers toward cancer pain management: a comparison of physicians, nurses, and pharmacists in the state of New Hampshire. J Pain Symptom Manage 1998;15: Larue F, Colleau SM, Fontaine A, et al. Oncologists and primary care physicians attitudes toward pain control and morphine prescribing in France. Cancer 1995;76: Elliott TE, Murray DM, Elliott BA, et al. Physician knowledge and attitudes about cancer pain management: a survey from the Minnesota cancer pain project. J Pain Symptom Manage 1995;10: Levin ML, Berry JI, Leiter J. Management of pain in terminally ill patients: physician reports of knowledge, attitudes, and behavior. J Pain Symptom Manage 1998;15: Sapir R, Catane R, Strauss-Liviatan N, et al. Cancer pain: knowledge and attitudes of physician in Israel. J Pain Symptom Manage 1999;17: