Oral versus Intravenous Opioid Dosing for the Initial Treatment of Acute Musculoskeletal Pain in the Emergency Department

Transcription

1 CLINICAL INVESTIGATION Oral versus Intravenous Opioid Dosing for the Initial Treatment of Acute Musculoskeletal Pain in the Emergency Department James R. Miner, MD, Johanna Moore, MD, Richard O. Gray, MD, Lisa Skinner, MD, Michelle H. Biros, MS, MD Abstract Objectives: The objective was to compare the time to medication administration, the side effects, and the analgesic effect at sequential time points after medication administration of an oral treatment strategy using oxycodone solution with an intravenous (IV) treatment strategy using morphine sulfate for the initial treatment of musculoskeletal pain in emergency department (ED) patients. Methods: This was a prospective randomized clinical trial of patients >6 years old who were going to receive IV morphine sulfate for the treatment of musculoskeletal pain but did not yet have an IV. Consenting patients were randomized to have the treating physician order either 0.1 mg kg morphine sulfate IV or mg kg oxycodone orally in a 5 mg 5 ml suspension as their initial treatment for pain. The time from the placement of the order to the administration of the medication was recorded. Pain was measured using a 100-mm visual analog scale (VAS) and recorded at 0, 10, 20, 30 and 40 minutes after drug administration. Results: A total of 405 eligible patients were identified during the study period; 328 (81.0%) patients consented to be in the study. A total of 158 patients were randomized to the IV morphine sulfate treatment group, and 162 were randomized to the oral oxycodone treatment group. Of the patients who were randomized to IV therapy, 34 were withdrawn from the study prior to drug administration; leaving 125 patients in the IV group for analysis. Of the patients who randomized to oral therapy, 22 were withdrawn from the study prior to drug administration, leaving 140 patients for analysis. No serious adverse events were detected. There was a 12-minute difference between the median time of the order and the administration of oral oxycodone (8.5 minutes) and IV morphine (20.5 minutes). The mean percent change in VAS score was larger for patients in the IV therapy group than those in the oral therapy group at 10 and 20 minutes. At 30 and 40 minutes, the authors could no longer detect a difference. The satisfaction scale score was higher after treatment for the morphine group (median = 4; interquartile range [IQR] = 4 to 5) than for the oxycodone group (median = 4; IQR = 2 to 5; p = 0.008). Conclusions: The oral loading strategy was associated with delayed onset of analgesia and decreased patient satisfaction, but a shorter time to administration. The oral loading strategy using an oxycodone solution provided similar pain relief to the IV strategy using morphine 30 minutes after administration of the drug. Oral mg kg oxycodone represents a feasible alternative to 0.1 mg kg IV morphine in the treatment of severe acute musculoskeletal pain when difficult or delayed IV placement greater than 30 minutes presents a barrier to treatment. ACADEMIC EMERGENCY MEDICINE 2008; 15: ª 2008 by the Society for Academic Emergency Medicine Keywords: pain, treatment, oxycodone, morphine, emergency department, emergency medicine Both adult and pediatric trauma patients often receive inadequate analgesia in the emergency department (ED). 1 6 There are a variety of barriers to pain treatment, many related to the assessment of pain and many related to treatment practices. Acute pain is, however, essentially an easy thing to treat. There are a From the Department of Emergency Medicine, Hennepin County Medical Center (JRM, JM, ROG, LS, MHB), Minneapolis, MN. Received June 4, 2008; revision received July 28, 2008; accepted August 1, Presented in part at the Society for Academic Emergency Medicine (SAEM) Annual Meeting, San Francisco, CA, May 2006 (adult data), and at the American College of Emergency Physicians (ACEP) Research Forum, New Orleans, LA, October 2006 (pediatric data). This work was funded in part by a grant from the Emergency Medicine Foundation. None of the authors have financial or other interests in the medications used in this study. Address for correspondence and reprints: James R. Miner, MD; jimminer@hotmail.com. ISSN ª 2008 by the Society for Academic Emergency Medicine 1234 PII ISSN doi: /j x

2 ACAD EMERG MED December 2008, Vol. 15, No variety of available pain medications and routes of administration. Given the wide variety of situations that require the treatment of pain in the ED, it stands to reason that having a wide variety of treatment options available would be beneficial. However, despite the frequency with which various pain medications are used, the relative efficacy and safety of different treatment practices have only been compared in a few studies. 7 It is common practice to place an intravenous (IV) line and treat the patient in pain with titrated doses of opioid until adequate pain control has been achieved A common barrier to this approach is the timing of, and potential difficulty in, the placement of an IV line. A critically ill patient usually receives an IV very quickly; but those who are less ill and are hemodynamically stable may get the IV placed less quickly, especially during periods of overcrowding when available resources are limited. 8,9 Because of the unpredictable nature of the demand for resources in the ED, treatment approaches that do not require the time-consuming placement of an IV may be useful starting points for pain treatment in some situations. Intravenous morphine sulfate at weight-based doses of 0.1 mg kg is a common treatment for moderate to severe musculoskeletal pain, with few reported side effects. 10,11 IV morphine is titratable, allowing customized delivery for individual patients. Oral opioids are also used frequently in the treatment of pain in the ED, often in combination with acetaminophen (e.g., Percocet). The titration of oral opioids is difficult due to variability in the absorption of the drug. Oxycodone has primarily been studied in cancer-related pain. However, in a 1998 study of postoperative pain, the same doses of IV oxycodone and morphine were used by patients with a patient-controlled analgesia pump, suggesting that the two drugs are equipotent in this type of pain when given intravenously. 12 Oral oxycodone has a bioavailability of 60% to 87% in cancer patients Therefore, using a bioavailability quotient of 0.8, we estimated that mg kg oxycodone will provide similar pain relief to 0.1 mg kg IV morphine at the time it reaches its peak effect. The objective of our study was to compare the time to administration after the order was placed, the side effects, and the analgesic effect at sequential time points after medication administration of an oral treatment strategy using oxycodone, with an IV treatment strategy using morphine sulfate, for the initial treatment of musculoskeletal pain in ED patients. Our null hypothesis was that there is no difference in the time of administration, the onset of detectable analgesic effect after administration, or the magnitude of analgesic effect between an oral treatment strategy using oxycodone solution and an IV treatment strategy using morphine sulfate. METHODS Study Design This was a prospective randomized controlled clinical trial of patients receiving treatment for musculoskeletal pain between June 1, 2006, and December 31, The institutional review board of Hennepin County Medical Center approved the study. Patients provided prospective informed consent or assent prior to enrollment. Parents or guardians provided consent for the enrollment of children into the study. In addition, patients less than 18 years and greater than 6 years of age provided assent for study enrollment. Study Setting and Population This study was performed at Hennepin County Medical Center, an urban county hospital with approximately 99,000 ED patient visits per year. In our ED, it is the usual practice to treat patients with severe acute musculoskeletal pain with IV morphine (0.1 mg kg IV, followed by 0.05 mg kg IV q10 minutes as needed tolerated) for pain control as soon as possible. Patients 6 years and older who were going to receive IV morphine for the treatment of musculoskeletal pain at the time of study entry, but did not yet have an IV placed, were eligible for the study and approached for enrollment. Patients who had received a pain medication within 30 minutes prior to enrollment, with an allergy to morphine or oxycodone, who were pregnant or nursing, had a history of significant hepatic or renal disease, had ongoing obstructive airway disease, were in respiratory distress, or who were clinically intoxicated were excluded. Eligible patients and the guardians of eligible children were approached to provide informed consent. Children 6 years of age or older were approached for assent. Study Protocol Consenting patients were randomized to treatment with either 0.1 mg kg morphine sulfate IV or mg kg oxycodone orally in a 5 mg 5 ml suspension. For each eligible patient, the next in a series of sealed, consecutively numbered envelopes was opened, revealing the treatment group to which the subject was assigned. Treatment group assignments were determined using computerized random number generation and were sealed in the envelopes prior to the start of enrollment in the study. Subjects were placed on a pulse oximetry monitor prior to drug administration, and naloxone was available at the bedside. The study was discontinued and rescue medications given at the discretion of the treating physician if it was felt that the patient was having increasing pain or insufficient improvement of pain prior to 40 minutes after administration of the medication. These conditions were considered treatment failures. Complications were defined as vomiting, an oxygen saturation of <97% at any time, a decrease in systolic blood pressure >20%, or withdrawal from the study due to increasing pain. Prior to drug administration, all medications the patient had received in the prior 6 hours were recorded. The time from the placement of the medication order by the treating physician to the time of the administration of the medication was recorded by trained research assistants. Blood pressure, pulse oximetry, and pain measurements were recorded at 0, 10, 20, 30, and 40 minutes after initiation of the treatment. Pain was measured using a 100-mm visual analog scale (VAS) containing the words no pain on one end and the words worst pain possible on the other. Subjects

3 1236 Miner et al. IV VERSUS ORAL TREATMENT FOR ACUTE PAIN could not see their previous VAS scores. Physicians were blinded to the subjects VAS scores. After the 40- minute time point, the treating physician was asked whether or not the patient had received adequate pain relief from the treatment given, whether or not the patient would require further pain medications, and whether or not they intended to continue using the current route of administration of pain medication. Patients were asked to complete a 5-point scale describing their satisfaction with their pain treatment, with 1 being not satisfied and 5 being completely satisfied. VAS measurements were administered to the subjects and the treating physicians by trained research assistants, who collected all study data. Data Analysis Data were initially entered into an Excel (Microsoft Corp., Redmond, WA) spreadsheet and then exported into STATA 10.0 (StataCorp, College Station, TX) for analysis. Sample size calculations were performed using StudySize 2.0 (CreoStat HB, Sweden). Descriptive statistics were used as appropriate. The absolute and the percent change in VAS scores and the time to administration were compared between groups at each time point using Wilcoxon rank sum tests. The proportion of patients in each group who had a 50% decrease in their pain score was compared using chi-square tests. Changes in pain VAS scores between time points were compared using Wilcoxon sign rank tests. The proportion of patients in each group who were deemed to have adequate pain relief by their treating physician and the number of complications in the groups were compared using Fischer s exact test. We chose to use the percent change in VAS scores in addition to the absolute change in VAS scores to account for the possibility that analgesic effects may be different depending on the magnitude of a patient s pain and to avoid overestimating pain relief in patients with high initial scores and underestimating relief in patients with low initial scores. To detect a 10% difference in the change of VAS scores between groups with a power of 0.8 and a significance of 0.05 using a Wilcoxon rank sum test, 122 patients were needed in each group. RESULTS A total of 405 patients were identified during the study period; 328 (81.0%) patients consented to be in the study. A total of 158 patients were randomized to the IV morphine sulfate treatment group and 162 were randomized to the oral oxycodone treatment group. Of the patients who were randomized to IV therapy, 26 were withdrawn from the study prior to initiation, leaving 126 patients in the IV group for analysis. Of the patients who were randomized to oral therapy, 16 were withdrawn from the study prior to initiation (Figure 1). Summary baseline data is presented in Table 1. There were no significant differences between the groups. Table 2 contains the medications patients had received prior to study enrollment, from 6 hours before arrival in the ED and more than 30 minutes prior to study enrollment while in the ED. There was no difference between the two groups in the number of patients 320 patients enrolled Randomization 158 IV morphine 162 Oral oxycodone 26 withdrawn before start 16 withdrawn before start 8 protocol violations 6 protocol violations 7 = no IV possible 6 = age < 7 years 1 = age < 7 years 125 for analysis 140 for analysis Figure 1. Study timeline and enrollment. IV = intravenous. receiving pain medications prior to study enrollment (p = 0.53). No changes greater than 10% occurred in heart rate or blood pressure at any of the study time points. The median heart rate after medication administration was 75 for the IV therapy group (range, 49 to 126) and 77 beats min for the oral therapy group (range, 55 to 129 beats min). The median systolic blood pressure after medication administration was 116 mm Hg, (range, to mm Hg). Data regarding the pain relief achieved at each study time point are displayed in Figures 2 and 3 and compared in Tables 3 and 4. The VAS score and the percent change in the VAS score were different from baseline at each time point by Wilcoxon sign rank test (<0.001 for each test). There was a decrease in the VAS scores after treatment compared to baseline at all time points for the IV therapy group. There was not a statistically significant decrease in pain for patients in the oral therapy group until 30 minutes after treatment. At the 10- and 20-minute time points, the IV treatment group had a larger decrease in pain than the oral treatment group. By 30 minutes after treatment, both groups reported the same reduction of pain. The number of patients who achieved a 50% decrease in their pain scores was larger for patients in the IV therapy group than those in the oral therapy group at the 10- and 20-minute time points. However, at the 30- and 40-minute time points, we could no longer detect a difference in the number of patients in each group with at least a 50% reduction in pain (Table 5). Two patients had oxygen saturations <93% transiently prior to the 10-minute time point. Both patients were in the IV morphine group, and their oxygen saturations returned to >93% within 1 minute with no treatment. There were more treatment failures

4 ACAD EMERG MED December 2008, Vol. 15, No Table 1 Baseline Data IV Therapy Oral Therapy p-value Number Age (years), median (range) 36 (13 to 80) 40 (10 to 89) 0.92 Gender (% male) Weight (kg), (median; IQR), 73.2 (75; 59 to 84) 74.7 (75; 60 to 90) 0.35 Time to medication after enrollment (min), 20.5 (14 to 31; 2 to 124) 8.5 (5 to 15; 1 to 60) <0.001 median (IQR; range) Medication dose (mg kg), median, (IQR) 0.1 (0.08 to 0.10) (0.11 to 0.14) Pulse oximetry (%), median (IQR) 98 (97 to 99) 98 (97 to 99) 0.48 Baseline VAS, median (IQR) 85 (72 to 98) 86 (72.5 to 98) 0.87 IQR = interquartile range. Table 2 Medications Received before Study Enrollment Prestudy Treatment IV Morphine Sulfate, before ED Arrival,* n Rx, in ED, n Oral Oxycodone, before ED Arrival,* n Rx, in ED, n Acetaminophen NSAID Oral opioid Parenteral opioid ED = emergency department; IV = intravenous; NSAID = nonsteroidal anti-inflammatory drug; Rx = prescribed. *Up to 6 hours prior to ED arrival. Treated in the ED more than 30 minutes prior to study enrollment. VAS score (mm) VAS Scores By Time and Group % Decrease in VAS score Change in VAS from Baseline By Time and Group Time Point (min) Oral Therapy IV Therapy Time Point (min) Oral therapy IV therapy Figure 2. VAS scores by time and group. VAS = visual analog scale. (rescue pain medications were provided) at the 10- and 20-minute time points in the IV therapy group than in the oral therapy group (Table 6). No other adverse events, including vomiting, were detected. There was a 12-minute difference (p = 0.001, Wilcoxon rank sum) between the median time between the order and the administration of oral oxycodone (8.5 minutes) and IV morphine (20.5 minutes; Table 1). The satisfaction scale score was higher after treatment for the morphine group (median = 4; interquartile range [IQR] = 4 to 5) than for the oxycodone group (median = 4; IQR = 2 to 5; p = 0.008). Physicians considered the pain treatment adequate in (68.8%) of Figure 3. Change in VAS from baseline by time and group. VAS = visual analog score. patients in the IV group and in (54.3%) of the oral group (p = 0.03). There was no difference in the rate at which physicians felt that further medications were needed after the 40-minute time point ( [54.4%] vs [54.3%]; p = 0.96) or in the number who planned to continue to use the current route of administration to give the patient further doses of pain medicine ( [54.0%] vs [54.3%]; p = 0.96). DISCUSSION The IV administration of morphine is a common approach to the treatment of pain. This is usually

5 1238 Miner et al. IV VERSUS ORAL TREATMENT FOR ACUTE PAIN Table 3 Figure Summary: Change in VAS from Baseline IV therapy group )22.9 ()6.8, )46.8) )11.2 ()2.2, )39.4) )31.0 ()15.0, )71.0) )43.7 ()15.6, )79.7) Oral therapy group )7.5 (0, )23.5)% )3.1 ()7.6, )25.0) )37.4 ()9.7, )72.3) )32.6 ()6.1, )73.9) p-value (Wilcoxon rank sum) < Data are reported as median % (IQR). IQR = interquartile range; IV = intravenous; VAS = visual analog scale. Table 4 Figure Summary: VAS Scores IV therapy 64.5, (44 to 82) 74.5, (44 to 97) 55, (25 to 71) 47, (17 to 70) Oral therapy 75, (58 to 94) 80, (60 to 97) 49, (22 to 74) 56.5, (17.5 to 80.5) p-value (Wilcoxon rank sum) < Data are reported as median mm (IQR). IQR = interquartile range; IV = intravenous; VAS = visual analog scale. Table 5 50% Change in VAS Score* Oral therapy group (7.8) (4.5) (43.2) (37.7) IV therapy group (23.8) (22.0) (38.9) (44.4) p-value < IV = intravenous; VAS = visual analog scale. *n total at measured time (%). Table 6 Treatment Failures (Received a Rescue Medicine, Cumulative): IV Morphine Sulfate* IV therapy group (7.2) (9.6) (19.2) (19.2) Oral therapy group (2.8) (2.9) (19.3) (19.3) p-value *n total (%); cumulative over duration of the study. IV = intravenous. accomplished easily; however, technical difficulties and limited resources can sometimes delay the placement of an IV line. The goal of this study was to determine if an oral loading strategy could achieve similar pain relief to an IV one and to measure the delay to achieve pain relief when an oral strategy is used in place of an IV strategy. To determine this, we measured the effectiveness of two routes of administration of bioequivalent doses of two pain medications and found that IV morphine sulfate decreases pain significantly 10 minutes after administration, while oral oxycodone provides equivalent pain relief 30 minutes after administration. From the patient s perspective, however, the goal of ED management is the relief of pain as soon as possible, and the time to achieve any pain relief after arrival to the ED is the most relevant measure. While patients in the IV therapy group achieved pain relief earlier after treatment than those in the oral therapy group (i.e., larger decreases in VAS pain scores at 10 and 20 minutes), the time to administration of the medications was significantly shorter for those receiving oral oxycodone. When the time to significant pain relief is added to the time required for drug administration (10 minutes minutes for the IV therapy group and 30 minutes minutes for the oral therapy group), the time at which pain relief was detected once the order had been placed was similar for the two groups.

6 ACAD EMERG MED December 2008, Vol. 15, No There are many factors affecting the time it takes to administer an ordered pain medication in the ED, such as current work demands on nursing and patient overcrowding. 8 In addition, the delivery of IV medications is affected by the ability to start an IV on the patient needing pain relief. In our study, seven patients were withdrawn from the study due to difficulty placing an IV and, in our busy ED, oral medications were delivered much more quickly than IV pain medications. We found that, considering inherent delays in IV starts, patients receiving either oral or IV medication achieved relief at similar times after the treatments were ordered. Despite the small difference in the overall time to any pain relief after the treatment was ordered, patients in the IV therapy group were more satisfied with their pain treatment than those receiving in the oral therapy group. It is possible that placement of an IV suggests to the patients that the reported pain is being taken more seriously or that more care is actually being delivered when an IV is placed. In addition, physicians perceived greater pain relief in patients who had received IV medication. However, more patients in the morphine group received rescue medications than in the oral oxycodone group. It is possible that the fact that patients appeared more satisfied influenced the physician s perception of the patient s pain relief. These perceptions may represent a significant drawback to the use of oral oxycodone as a routine replacement for IV morphine. We used oxycodone without acetaminophen so it would be possible to give further doses of the medication. Unlike IV morphine, however, the timing and rate at which oral oxycodone can be titrated is not known. We can, however, conclude, that we did not detect a difference in the rate of adverse events (two hypoxic events for morphine vs. no events for oxycodone), that mg kg oxycodone appears to be a safe initial loading dose for management of pain for ED patients, and that it has an effect similar to that of 0.1 mg kg IV morphine 30 minutes after administration. LIMITATIONS This study was not blinded because our intention was to compare two different clinical strategies and the timing of the administration of the treatments. We could have blinded this study by having all patients receive IV drug or placebo and oral drug or placebo. Our intent, however, was to study the clinical approach to pain treatment, not the actual drugs, whose efficacy has been long established. We initially intended to study all patients, including children less than 7 years old. Patient-derived VAS scores have not been found to be accurate in children less than 7 years, and we therefore would need to use physician-derived pain scales. Due to the problems of using physician-derived pain scales in a nonblinded study where the physician controlled the determination of treatment failure, it was felt that this would yield data that was difficult to interpret and these patients were excluded from our study. For both medications, the doses that we studied here appeared to be too low to provide pain relief in the majority of patients. Our rate of treatment failure with this protocol was high. This has previously been observed for IV morphine given at these doses. 10,11 The doses of medication actually given to our study patients were lower than the doses the protocol called for in both the morphine and the oxycodone dose. Our protocol controlled the ordering of the medication but not the actual administration of the medicine. This is similar to previous findings in pain medication research that have noted that pain medications tend to be given in smaller doses than they are actually ordered. 17 In light of the fact that there were few adverse events other than treatment failures (i.e., the need for rescue medications), larger doses of both medications should be considered. CONCLUSIONS The oral loading strategy using oxycodone solution provided similar pain relief to the IV strategy using morphine sulfate 30 minutes after administration of the drugs. When considering the time to any pain relief (time to drug administration + time to pain relief), we found the oral treatment group and the IV treatment group to be similar. When the time to administration was not taken into account, patients receiving IV morphine reported faster pain relief, a greater degree of pain relief at an earlier time, and greater satisfaction with IV morphine than with oral oxycodone. Oral oxycodone represents a feasible alternative to IV morphine in the treatment of severe acute musculoskeletal pain when difficult or delayed IV placement (greater than 30 minutes) presents a barrier to treatment. References 1. Hostetler MA, Auinger P, Szilagyi PG. Parenteral analgesic and sedative use among ED patients in the United States: combined results from the National Hospital Ambulatory Medical Care Survey (NHAMCS) Am J Emerg Med. 2002; 20: Miner J, Biros MH, Trainor A, Hubbard D, Beltram M. Patient and physician perceptions as risk factors for oligoanalgesia: a prospective observational study of the relief of pain in the emergency department. Acad Emerg Med. 2006; 13: Rupp T, Delaney KA. Inadequate analgesia in emergency medicine. Ann Emerg Med. 2004; 43: Todd KH, Ducharme J, Choiniere M, et al. Pain in the emergency department: results of the pain and emergency medicine initiative (PEMI) multicenter study. J Pain. 2007; 8: Todd KH, Sloan EP, Chen C, Eder S, Wamstad K. Survey of pain etiology, management practices and patient satisfaction in two urban emergency departments. Can J Emerg Med. 2002; 4: Wilson JE, Pendleton JM. Oligoanalgesia in the emergency department. Am J Emerg Med. 1989; 7: Chang AK, Bijur PE, Meyer RH, Kenny MK, Solorzano C, Gallagher EJ. Safety and efficacy of hydromorphone as an analgesic alternative to

7 1240 Miner et al. IV VERSUS ORAL TREATMENT FOR ACUTE PAIN morphine in acute pain: a randomized clinical trial. Ann Emerg Med. 2006; 48: Pines JM, Hollander JE. Emergency department crowding is associated with poor care for patients with severe pain. Ann Emerg Med. 2008; 51: Miner JR, McClain C. Time to administration of pain medications as a marker for overcrowding [abstract]. Ann Emerg Med. 2006; 48(4 Suppl 1): Bijur PE, Kenny MK, Gallagher EJ. Intravenous morphine at 0.1 mg kg is not effective for controlling severe acute pain in the majority of patients. Ann Emerg Med. 2005; 46: Birnbaum A, Esses D, Bijur PE, Holden L, Gallagher EJ. Randomized double-blind placebo-controlled trial of two intravenous morphine dosages (0.10 mg kg and 0.15 mg kg) in emergency department patients with moderate to severe acute pain. Ann Emerg Med. 2007; 49: Silvasti M, Rosenberg P, Seppala T, Svartling N, Pitkanen M. Comparison of analgesic efficacy of oxycodone and morphine in postoperative intravenous patient-controlled analgesia. Acta Anaesthesiol Scand. 1998; 42: Kalso E, Vainio A. Morphine and oxycodone hydrochloride in the management of cancer pain. Clin Pharmacol Ther. 1990; 47: Kalso E, Vainio A, Mattila MJ, Rosenberg PH, Seppala T. Morphine and oxycodone in the management of cancer pain: plasma levels determined by chemical and radioreceptor assays. Pharmacol Toxicol. 1990; 67: Leow KP, Smith MT, Watt JA, Williams BE, Cramond T. Comparative oxycodone pharmacokinetics in humans after intravenous, oral, and rectal administration. Ther Drug Monit. 1992; 14: Leow KP, Smith MT, Williams B, Cramond T. Single-dose and steady-state pharmacokinetics and pharmacodynamics of oxycodone in patients with cancer. Clin Pharmacol Ther. 1992; 52: Cleeland CS, Gonin R, Hatfield AK, et al. Pain and its treatment in outpatients with metastatic cancer. N Engl J Med. 1994; 330:592 6.