Radiotherapy for Bone Metastases: A Critical Appraisal of Outcome Measures

Transcription

1 208 Journal of Pain and Symptom Management Vol. 17 No. 3 March 1999 Review Article Radiotherapy for Bone Metastases: A Critical Appraisal of Outcome Measures Robin Dawson, BA, B Soc Wk, David Currow, B Med, MPH, FRACP, Graham Stevens, MD, FRACR, Graeme Morgan, FRCP, FRACR, and Michael B. Barton, MB BS, FRACR Division of Radiation Oncology (R.D., M.B.B.), Westmead Hospital, Westmead, Australia; Department of Palliative Care (D.C.), Nepean Hospital, Penrith, Australia; Department of Radiation Oncology (G.S.), Royal Prince Alfred Hospital, Camperdown, Australia; and Department of Radiation Oncology (G.M.), St Vincents Hospital, Darlinghurst, Australia. Abstract Pain from bone metastases is a common problem in patients with advanced cancer, and radiotherapy plays an important role in its palliation. Single fraction treatments are often prescribed, but there is no clear consensus on this issue and clinical practice shows significant variability. This situation is unsatisfactory for all parties the patient, the clinician, and the health care administrator. Randomized trials may use poor outcome measures and this contributes to practice variability. The credibility of outcome studies is often reduced due to poor study design, small sample sizes, and the use of endpoints that are both unreliable and unsuitable. The endpoints used have been narrowly defined, the patient s perspective has generally been overlooked, and quality of life has only once been used as an endpoint. A review of the current literature suggests that instruments specific to bone metastases are required. These must be based on patient experience, and rely on self-report. In addition, there is a need to understand the relative priority that patients attribute to treatment outcomes. The use of better instruments and methodologies in future trials will enhance the credibility of results and reduce practice variations. J Pain Symptom Manage 1999;17: U.S. Cancer Pain Relief Committee, 1999 Key Words Radiotherapy, palliative radiotherapy, bone metastases, measures of outcome, quality of life, practice variations Address reprint requests to: Michael Barton, MB BS, Radiation Oncology Department, Westmead Hospital, Westmead, 2145, NSW Australia. Accepted for publication: June 19, Introduction Thirty to fifty percent of all patients receiving radiotherapy are treated with palliative intent and treatment of bone metastases accounts for at least one-fifth of all radiotherapy treatments. 1,2 Given the number of patients involved, the use of radiotherapy for bone metastases is a very significant public health issue. Metastatic bone disease frequently causes morbidity, including pain, spinal cord compression, and pathological fracture. Radiotherapy is effective in relieving pain and preventing complications. The challenge for the radiation oncologist is to determine the treatment regimen that will deliver rapid and enduring pain U.S. Cancer Pain Relief Committee, /99/$ see front matter Published by Elsevier, New York, New York PII S (98)

2 Vol. 17 No. 3 March 1999 Radiotherapy for Bone Metastases 209 relief with the minimum morbidity, disruption, and cost. There have been numerous attempts to identify the optimum dose and fractionation. The weight of evidence indicates that neither a dose response relationship nor any significant advantage characterizes the use of multiple fractions. 3,4 Yet, despite the published evidence, clinical practice continues to show marked variation, ranging from a single fraction to 10 or even 20 fractions for the same presenting condition. 5,6 Reluctance to adopt single fractions may be due to a number of factors, including reservations about the measures of response selected. Further, this variability suggests that treatment decisions are currently being made on an ad hoc basis. Clearly, there is a need to improve patterns of practice and to define the optimum fractionation schedule for effective palliation of bone metastases, while at the same time minimizing cost to the patient and the community. The first step is to review the measures of outcome that have been used, and to assess their suitability and relevance. The basic question is as follows: are outcomes in this field being assessed by the right criteria? This review considers the adequacy of endpoints used to date in clinical trials to assess the effectiveness of different fractionation schedules for palliation of bone metastases. This will be contrasted with the realities of clinical practice, where decision-making is based on a range of factors that are seldom if ever mentioned in the research literature. Finally, the implications for future research are considered, outlining the next steps required to define relevant measures of outcome for bone metastases. Literature Review A Medline search of bone metastases, radiotherapy, and quality of life identified a wide range of relevant literature. This is summarized in Tables 1 and 2, which focus on measures of outcome. The following points are noted: The majority of studies are retrospective reviews of medical records and their reliability is questionable. In some of the prospective studies, patients were not randomized; in such studies, variability in selection factors is a concern. The confounding effects of concurrent therapy are not always controlled. Small sample sizes are common. There are only eight studies with a sample greater than 200, of which only five are prospective studies. The majority of the studies rely on assessment of pain relief by the physician, which is open to bias. Only one study 7 attempted to validate the chosen pain scale. In view of the rapidity of relief sometimes seen, it is noteworthy that only two studies 8,9 appear to have measured pain prior to treatment. The range of outcome measures is wide and shows little consistency. Apart from pain and pain relief (response), other measures used appear somewhat arbitrary. Many of the studies measure pain response at several points in time, yet the details of timing are often unclear. The results reported seem to relate more to the timing of follow-up appointments than the timing of actual response. Omissions in existing research are also worth noting: There is a marked paucity of input from the patient. Quality of life and well-being are subjective endpoints, yet the patients perceptions are seldom sought, even for the relatively straightforward endpoints such as pain, toxicity, use of analgesics, and performance status. Although self-report was used for six studies, this was in most cases limited to pain scales. Measures of outcome have focused narrowly on pain intensity. Few studies have incorporated toxicity, and even fewer have measured performance status, although the latter is widely used as a measure of treatment effectiveness in clinical practice. Given that the focus of care is palliative, quality of life has a higher priority than quantity of life, but quality of life has only been used as an endpoint in one trial to date. 9 The aim of treatment is always palliation, yet treatment has not generally been assessed against this goal. Patient priorities

3 210 Dawson et al. Vol. 17 No. 3 March 1999 Study Patients Selfreport Timing Method Other therapy controlled Randomized Response Table 1 Prospective Studies Pain Time to response Duration of response Outcome measures Fracture Toxicity Analgesia Performance status QOL Retreatment Survival Cole Daily 28 days monthly 6 mth Gaze et al , 3-4 wk, 3mth Hoskin et al Pre 2, 4, 8, 12 wk Madsen , 4, 6, 8, 12 wk Okawa et al Unclear Poulter et al. a , 5, 9, 12 mth b Price et al , 1, 2, 4, 8 wk Price et al Daily 4 wk weekly 8 wk Salazar et al. a Weekly 8 wk then monthly Tong et al Weekly 2, 4, 6 mth a HBI: Hemibody radiation. b Time to new disease.

4 Vol. 17 No. 3 March 1999 Radiotherapy for Bone Metastases 211 Table 2 Retrospective Studies Outcome measures Pain Study Patients Method Time to response Duration of response Performance status Retreatment Timing Response Toxicity Analgesia Survival Allen et al mth b Arcangeli et al Unclear Barak et al Unclear Burmeister and Probert a Unclear c b Garmatis and Chu Unclear c Gilbert et al and 12 mth Hoskin et al. a Unclear Jensen and Roesdahl wk, then monthly Kuban et al. a Unclear Martin Mithal et al wk Nag and Shah a Day 1 1, 3, 6, 12 mth Needham and Hoskin wk Penn wk, 6 mth, annually 4 yr Schocker and Brady sites 30 days Vargha et al Not stated Wilkins and Keen a Unclear Zelefsky et al. a Unclear a TBI: Hemibody irradiation. b Recurrence. c X-ray response.

5 212 Dawson et al. Vol. 17 No. 3 March 1999 in terms of treatment and health outcomes have not been examined. Thus there is a bias in endpoint selection towards what is readily measurable, and more complex palliative outcomes have been overlooked. Outcomes to date have typically been assessed by the doctor rather than the patient, have focused narrowly on pain, and have minimized the importance of the patient s subjective health assessment. The most important data derive from five major prospective, randomized studies of radiotherapy for bone metastases. The conclusions of the Radiation Therapy Oncology Group (RTOG) study 10 have been questioned, and it incorporated no evaluation of toxicity. RTOG did not measure either pain or pain relief, and noncompliance and protocol variations were marked. In contrast, both studies from the Royal Marsden Hospital 7,8 have the benefit of well-defined response criteria, use of patient self-assessment, use of a validated pain scale, and a precise schedule for evaluating response. The study by Price 7 has a wider range of measures, greater reliance on patient input, and the control of concurrent therapy which could influence outcomes. A recent study for Edinburgh Hospital was also comprehensive, featuring a wide range of outcome measures, assessed by both patient and doctor. 9 The latter studies represent an improvement, but still offer only a limited body of reliable evidence available on which to base treatment decisions. Further large trials are needed to clarify the many uncertainties in this field and define the appropriate palliative schedules. Measures of Outcome Response Although all studies report results in terms of response, there is no standard definition of this outcome in radiotherapy trials. The term typically implies the extent to which pain relief is achieved, but use of analgesia, 12 a combination of pain relief and analgesia, 7,13,14 or performance status 12,15 have all been used as measures of response to radiotherapy. The use of differing response criteria significantly inhibits comparison between studies. The concept of response requires validation. It is uncertain whether a partial response is of any value to the patient. For instance it is difficult to believe that a minor reduction in analgesic dosage, as distinct from cessation of all analgesics, is a worthwhile outcome of treatment. To address such issues, patient perceptions about the relative value of alternative outcomes must be evaluated. Timing of Response Evaluation Many studies give no precise details about the timing of evaluation, making both the interpretation and comparison of results difficult. In retrospective studies, classification of response can only be simple, and thus imprecise and insensitive. Prospective studies have the potential to be more precise in their definition of response, as well as their measurement schedule. The key issue is that the assessment schedule should be based on the occurrence of relief. Discrepancies can occur when authors select and report measures of response from different points in time. One may report results in 4 weeks, whereas another presents results at 6 months. This may explain why a recent metaanalysis could make no sound estimate of either the speed of onset of relief, or the duration that could be expected. 4 Pain Assessment Pain is a subjective experience and the quantification of pain is not straightforward. Patients with metastatic disease may have pain at more than one location, and the nature of the pain may differ from site to site. A hierarchy of pain often occurs, and as pain is controlled at one site, the focus can shift to pain at another site. The pain may be related to activity levels, being aggravated by movement or weight bearing. Ideally, pain evaluation would take account of duration as well as intensity. A reliable assessment of pain relief should take account of concurrent treatment that could affect pain, or pain from other unrelated causes or sites. Control for these variables is not universal in studies of palliation radiotherapy. Among the prospective studies, only three forms of self-report have been used. Madsen 13 opted for a 5-point scale (none, slight, moderate, severe, unendurable), Gaze et al. 9 used a 5-point scale with different descriptors (none, mild discomfort, moderate pain, severe pain, and intractable pain), and Price et al. 7 chose a

6 Vol. 17 No. 3 March 1999 Radiotherapy for Bone Metastases point scale (none, mild, moderate, severe). The major RTOG studies 10,11 used this 4-point scale, the only validation of which involved 26 patients. 7 In the RTOG studies, many different clinicians recorded the pain level as mild or moderate, suggesting the possibility of classification errors. The methods of pain measurement used to date have been simplistic, focusing exclusively on pain intensity. There has been no effort to distinguish among types of pain, such as neuropathic and mechanical, or acute and chronic. There has also been no assessment of the emotional reaction to the pain, and the impact of that pain on psychological, emotional, and social well-being. 16 Although numerous pain scales have been tested and validated in other fields, particularly nursing, existing investigations of bone pain makes no reference to this work. Experimental pain research has shown that subjects can discriminate up to 21 levels of pain, 20 but these sophisticated pain scales have never been tested in studies of palliative radiotherapy. It is possible that the use of very simple measures could yield inaccurate results, which overestimate the treatment effect. Multidimensional pain measurement scales, such as the widely used McGill Pain Questionnaire, which assesses sensory, affective, and evaluative aspects of pain, also have not been tried. 21 Patients beliefs about pain 22 and the data obtained from the regular use of pain charts have never been collected. 23,24 It would be worthwhile testing such concepts and instruments in the field of palliative radiotherapy for bone metastases. Physician or Patient Assessment? In studies of palliative radiotherapy outcome assessment has predominantly rested with the doctor. This is undesirable on several counts. First, the data are second hand proxy measures, which are inferred by the doctor from patient responses rather than recorded directly by the patient. There is obvious potential for misinterpretation, which has been observed When patients and doctors assessments of quality of life are compared, they correlate poorly. When asked to rate the same patient, considerable variation occurs among doctors, even for the single observer-rated Karnofsky scale. In one study, doctors and patients achieved agreement on only 54% of occasions. 27 In a comparison of nurses and patients rating of pain levels, only one-third of the nurses ratings agreed with patients perceptions. 28 Such studies indicate that quality of life, as well as pain, are best assessed by the patient rather than the physician or nurse. Second, physician assessment should be avoided due to the potential for bias. This is particularly important because it is impossible to double-blind fractionation trials. When response rates are compared for different radiotherapy regimens, it is worth noting that the complete response rates are relatively low in studies that have relied on patient rather than physician assessments of pain. 7,13 This raises the question whether physicians and patients assess pain differently. Use of Analgesics Use of analgesics is a common outcome measure. Some studies simply note only whether the use of analgesics has decreased; others record both the type of analgesic (opioid or nonopioid), as well as the frequency of use. 7,29 Several prospective studies have combined pain severity scores and analgesic use to determine response, 11,30 but there is some debate about whether this increases the sensitivity of results. The major trial by the RTOG 10, which involved 1016 patients and used physician assessments of pain relief, concluded that pain relief was not directly related to the dose and number of fractions. Blitzer s reanalysis of the data, 14 which used analgesic use and retreatment as endpoints, indicated the reverse, and concluded a dose response relationship existed. This finding highlights the sensitivity of such studies to the endpoints chosen, and confirms the need to include measures of outcome other than pain response. Acute Toxicity In palliative radiotherapy, the side effects of irradiation are generally considered to be both mild and brief, which may explain why toxicity is seldom included as an endpoint, and generally poorly reported. It is also true, however, that there is a wide individual variation in response to radiation, and side effects depend on the site being treated. A full evaluation of the effectiveness of treatment requires that both

7 214 Dawson et al. Vol. 17 No. 3 March 1999 side effects and tumor effects be considered. In the event that multiple and single fractions are equivalent in terms of tumor effect, then the discriminator becomes the level of side effects. Four prospective studies evaluated acute toxicity. The results were inconclusive. Three 7,9,29 found no relationship between the incidence of acute side effects and the number of fractions, and one noted that nausea was more common among those receiving single fractions. 31 The latter study, however, involved only 29 patients. Long-term Outcomes A number of negative outcomes are possible, including recurrence of pain, late toxicity, and fractures. The longer the patient survives, the more likely these are to arise. As the research to date has focused on palliation of pain, the patients selected have generally had a relatively short life expectancies. The follow-up period has also been limited, seldom extending beyond 3 months. Patients with different primary tumors have different life expectancies. Patients with bone metastases have a median survival of 0.6 year, an average survival of 1.7 years, and a 5-year survival of 8%. 32 There is a need for research over a longer time frame, evaluating both durability of pain relief and the incidence of long-term complications. Only one prospective study 10 assessed the impact of differing regimens on the incidence of fractures. It found that higher doses were associated with a higher incidence of fractures. None of the single-fraction studies has recorded the frequency of fractures. Long-term morbidity and the possible need for retreatment are major considerations for clinicians, 33 and assessment for longer periods may resolve these uncertainties. Quality of Life There have been a number of studies evaluating quality of life (QOL) at the time of initial diagnosis 34,35 and following systemic treatment. 36,37 Apart from simple global performance measures, however, there has been little documentation of QOL among patients with metastatic disease treated by radiotherapy. Litwin et al. 38 used general, cancer-specific and prostate-specific measures of health-related quality of life (HRQOL) to evaluate alternative treatments for localized prostatic cancer. They identified significantly poorer HRQOL among those men who had received either radiation or surgery, along with a higher rate of complications than hitherto reported. In patients with metastatic prostate cancer, Herr et al. 36 found that hormonal therapy did not enhance QOL. In fact, the reverse was the case. Those who deferred treatment enjoyed better outcomes in terms of physical and sexual functioning. Only two studies have evaluated QOL after palliative radiotherapy. The first assessed QOL and emotional status, and found no difference on these measures between single and fractionated schedules. 9 The second 39 observed that patients rather than the doctors, perception of physical function represented an independent prognostic factor for survival. The latter study also suggested that palliative radiotherapy may improve QOL less than previous reports would indicate; 3 months after radiotherapy only 20% of the surviving patients showed an improved QOL score. This study suggests that QOL measures may bring a fresh perspective to the evaluation of treatment outcomes, and highlights the need to include QOL as an outcome measure in clinical trials, as well as clinical practice. While much work has been done with cancer patients, the reliability of QOL instruments needs further examination in populations undergoing palliative radiotherapy. Among those with bone metastases, other factors related to disease progression and analgesia can impact on QOL, and the sensitivity of QOL measures must be further assessed. Future Directions Future studies of palliative radiotherapy should incorporate sophisticated methodologies. Ideally, these studies should be prospective and randomized, have a large sample size, use precisely defined response criteria, and employ a relevant evaluation schedule (e.g., prior to radiotherapy, weekly for 4 weeks, then at month 2, 4, 6, 9, and 12). There should be follow-up to death to allow assessment of long term outcomes. Both patient and doctor should assess outcomes, and concurrent therapy should be controlled to avoid confounding. A broad assessment of endpoints might include a comprehensive pain assessment using validated pain scales and use of analgesia, com-

8 Vol. 17 No. 3 March 1999 Radiotherapy for Bone Metastases 215 plete response and partial response, performance status, toxicity, quality of life, and satisfaction with outcome. The practical difficulties involved in such a comprehensive evaluation are recognized, particularly with palliative patients. While all of these endpoints may not be possible within the one study, the inclusion of a wider range would enhance the validity and credibility of trial results. A more precise definition of the features of future quantitative research would depend upon interim research to identify which outcomes matter to patients and their relative importance. Variability of Practice Day-to-day practice presents a significant contrast to the world of medical research, with its sophisticated statistical analysis of results. The literature is just one of the many factors contributing to treatment decisions. The variability of practice has been widely reported 5,6 and arises from four main causes. Training and Prejudice Several studies 2,40 comparing treatment plans in different countries highlight the subjectivity of the decision process. To a significant degree, treatment decisions reflect the doctor s ability to assess the prognosis. The ability to prognosticate is an important issue in palliative radiotherapy, as a careful balance must be struck between risks and benefits, in both the short and the long term. Unfortunately, radiation oncologists are reported to be poor at predicting survival of patients with advanced disease. 2 At the more local level, accepted practice varies from center to center, 6,33 reflecting the influence of training and the views of local, senior opinion leaders. These marked regional differences in practice raise questions about both the cost and the quality of care being provided. A health outcomes approach seeks to reduce this variability. Conflicting Clinical Data The results from clinical trials are at times conflicting. Caution is required in interpreting the results because of variation in definition of endpoints, patient selection, sample size, method of evaluation, and follow-up. For example, while most trials have found no conclusive evidence that multiple fractions are superior to single fractions, several have found the reverse. 41,51 The difficulty of assessing treatment effect is illustrated by the RTOG trial 10 and its subsequent reanalysis, which came to an opposite conclusion. 14 Logistical and Resource Issues The availability of resources significantly influences treatment, yet receives little mention in the literature. Compromises are common. If, for example, a patient lives far from a treatment center, or has no caregiver to assist with transport and support, then protracted regimens become difficult to justify. Equally, if there is a shortage of resources (whether staff or equipment) and a heavy workload, then this will operate to curtail treatment plans. 6 Major ethical issues exist as trade-offs are made between competing demands. Just as there are pressures acting to curtail treatment, there may also be economic incentives operating to extend treatment schedules, and this important issue has received little attention. It may be one of the factors contributing to the practice variability in different countries. The influence of such logistical and resource issues on treatment planning may be difficult to evaluate but cannot be underestimated. Diversity of Treatment Aims Any treatment plan must take into consideration the location of the primary, the site of metastases, the likely response to treatment, the history of the disease, the age of the patient, current symptoms, future risks (e.g., fracture), any comorbidity, and the social situation and personal wishes of the patient. Treatment goals are formulated on the basis of this review. Given the number of factors involved, it is hardly surprising that treatment shows variability, and the introduction of practice guidelines will remove some but not all of this. Conclusion This review has examined the shortcomings in the outcome measures that have been used to evaluate radiotherapy to palliate bone metastases. Most important, endpoints have been narrowly defined, with the emphasis on pain

9 216 Dawson et al. Vol. 17 No. 3 March 1999 intensity. The assessment of outcome has rested too much with the physician, rather than the patient. Two additional issues have been overlooked. First, what are the outcomes that matter to patients? Logically the efficacy of radiotherapy should be evaluated in terms of the patients criteria, which requires an understanding of the relative priority that patients attribute to treatment outcomes. For instance, if patients believe that QOL is their prime goal, then measures of QOL should yield key criteria by which future interventions are evaluated. An argument could also be made for the regular inclusion of QOL measures in clinical practice. Numerous instruments exist to measure healthrelated QOL, including several which are specific to cancer, but none specific to bone metastases. The first step in developing such a measure is to conduct comprehensive qualitative research with patients who have had radiotherapy for bone metastases to identify their priorities with regard to palliation. On the basis of this information an instrument can then be developed to assess the patient s utility for treatment outcomes. Second, how do patients perceive their own outcomes from palliative radiotherapy? As far as possible, patients, not doctors, should be responsible for the evaluation of treatment outcomes. Although evaluation of pain is likely to remain one of the key endpoints, there is a need for a more comprehensive definition of pain, and for careful selection of the pain assessment instrument. Again, it is essential that the patient be given a greater role. This shift in orientation may well be more demanding in research terms, but the evaluation of outcomes depends as much on the patient as it does on the physician. Although the published research indicates that single fractions are just as effective as multiple fractions, clinical practice shows little consistency. The continued variability may be due to the influence of local prejudice, financial considerations, reservations about the measures of response selected, and conflicting trial results. A health outcomes approach calls for a significant shift in focus, with greater attention to the value of outcomes as determined by patients. In terms of the palliative treatment of bone metastases, there is clearly a need for a detailed understanding of patient priorities which can then guide the development of appropriate patient-based measures of outcome. Acknowledgment This research was supported by a grant from the New South Wales Health Outcomes Program, 1994/95. References 1. Stevens G, Firth I. Patterns of fractionation for palliation of bone metastases. Australas Radiol 1995; 39: Maher EJ, Dische S, Grosche E, et al. Who gets radiotherapy? Health Trends 1990;2: Bates T. A review of local radiotherapy in the treatment of bone metastases and cord compression. Int J Radiat Oncol Biol Phys 1992;23: McQuay HJ, Carroll D, Moore RA. Radiotherapy for painful bone metastases. Clin Oncol 1997;9: Maher EJ, Coia L, Duncan G, Lawton PA. Treatment strategies in advanced and metastatic cancer: differences in attitude between the USA, Canada and Europe. Int J Radiat Oncol Biol Phys 1992;23: Priestman TJ, Bullimore JA, Godden TP, Deutsch GP. The Royal College of Radiologists Fractionation Survey. Clin Oncol 1989;1: Price P, Hoskin PJ, Easton D, et al. Prospective randomised trial of single and multifraction radiotherapy schedules in the treatment of painful bony metastases. Radiother Oncol 1986;6: Hoskin PJ, Price P, Easton D, et al. A prospective randomised trial of 4 Gy and 8 Gy single doses in the treatment of metastatic bone pain. Radiother Oncol 1992;23: Gaze MN, Kelly CG, Kerr GR, et al. Pain relief and quality of life following radiotherapy for bone metastases: a randomised trial of two fractionation schedules. Radiother Oncol 1997;45: Tong D, Gillick L, Hendrickson FR. The palliation of symptomatic osseous metastases. Final results of the RTOG study. Cancer 1982;50: Poulter CA, Cosmatos D, Rubin P, et al. A report of RTOG 8206: a phase III study of whether the addition of single dose hemibody irradiation to standard fractionated local field irradiation is more effective than local field irradiation alone in the treatment of symptomatic osseous metastases. Int J Radiat Oncol Biol Phys 1992;23: Burmeister BH, Probert JC. Half body irradiation for the palliation of bone metastases. Australas Radiol 1990;34: Madsen EL. Painful bone metastasis: efficacy of

10 Vol. 17 No. 3 March 1999 Radiotherapy for Bone Metastases 217 radiotherapy assessed by the patients. Int J Radiat Oncol Biol Phys 1983;9: Blitzer PH. Reanalysis of the RTOG study of the palliation of symptomatic osseous metastasis. Cancer 1985;55: Gilbert HA, Kagan AR, Nussbaum H, et al. Evaluation of radiation therapy for bone metastases: pain relief and quality of life. Am J Roentgenol 1977; 129: Melzack R, Casey KL. Sensory, motivational and central control determinants of pain: a conceptual model. In: Kenshalo D, ed. The skin senses. C.C. Thomas, New York: 1968: Harrison A. Assessing patient s pain: identifying reasons for error. J Adv Nursing 1991;16: Marvin JA. Pain assessment versus pain measurement. J Burn Care Rehab 1995;16: Holmes S. Preliminary investigations of symptom distress in two cancer patient populations: evaluation of a measurement instrument. J Adv Nursing 1991;16: Hardy JD, Wolff HC, Goodell H. Pain sensations and reactions. Baltimore, Williams & Wilkins, 1952, quoted by Marvin JA. Pain assessment versus measurement. J Burn Care Rehab 1995;16: Melzack R, Torgerson WS. On the language of pain. Anesthesiology 1971;34: Williams DA, Robinson ME, Geisser ME. Pain beliefs: assessment and utility. Pain 1994;59: Walker VA. Pain assessment charts in the management of chronic cancer pain. Palliative Medicine 1987;1: Faries JE, Mills DS, Goldsmith KW, et al. Systematic pain records and their impact on pain control. Cancer Nursing 1991;14: Kahn SB, Houts PS, Harding SP. Quality of life and patients with cancer: a comparative study of patient versus physician perceptions and its implications for cancer education. J Cancer Educat 1992;7: Merkel WT. Physician perception of patient satisfaction: do doctors know which patients are satisfied? Med Care 1984;22: Slevin ML, Plant H, Lynch D, Drinkwater J, Gregory WM. Who should measure quality of life, the doctor or the patient? Br J Cancer 1988;57: Vander Doss AJW. Patients and nurses ratings of pain and anxiety during burn wound care. Pain 1989;39: Salazar OM, Rubin P, Hendrickson FR, et al. Single dose half body irradiation for palliation of multiple bone metastases from solid tumours. Final Radiation Therapy Oncology Group Report. Cancer 1986;58: Okawa T, Kita M, Goto M, et al. Randomized prospective clinical study of small, large and twice a day fraction radiotherapy for painful metastases. Radiother Oncol 1988;13: Cole DJ. A randomized trial of a single treatment versus conventional fractionation in the palliative radiotherapy of painful bone metastases. Clin Oncol 1989;1: Unpublished data based on analysis of records of 3561 radiotherapy patients seen at Westmead Hospital from 1981 to Crellin AM, Marks A, Maher EJ. Why don t British radiotherapists give single fractions of radiotherapy for bone metastases? Clin Oncol 1989;1: Graydon JE. Women with breast cancer: their quality of life following a course of radiation therapy. J Adv Nursing 1994;19: Hughes KK. Psychosocial and functional status of breast cancer patients: the influence of diagnosis and treatment choice. Cancer Nursing 1993;16: Herr HW, Kornblith AB, Ofman U. A comparison of the quality of life of patients with metastatic prostate cancer who received or did not receive hormonal therapy. Cancer 1993;71: Tannock IF, Gospodarowicz M, Panzarella T, et al. Treatment of metastatic prostate cancer with low dose prednisone: evaluation of pain and quality of life as pragmatic indices of response. J Clin Oncol 1989;7: Litwin MS, Hays RD, Fink A, et al. Quality of life outcomes in men treated for localized prostate cancer. JAMA 1995;273: Fossa SD. Quality of life after palliative radiotherapy in patients with hormone-resistant prostate cancer: single institution experience. Br J Urol 1994; 74: Coia LR, Owen JB, Maher EJ, Hanks GE. Factors affecting the treatment patterns of radiation oncologists in the United States in the palliative treatment of cancer. Clin Oncol R Coll Radiol 1992;4: Arcangeli G, Micheli A, Giannarelli D, et al. The responsiveness of bone metastases to radiotherapy: the effect of site, histology and radiation dose on pain relief. Radiother Oncol 1989;14: Zelefsky MJ, Scher HI, Forman JD, et al. Palliative hemoskeletal irradiation for widespread metastatic prostate cancer: a comparison of single dose and fractionated regimes. Int J Radiat Oncol Biol Phys 1989;17: Allen KL, Johnson TW, Hibbs GG. Effective bone palliation as related to various treatment regimes. Cancer 1976;37: Garmatis CJ, Chu FCH. The effectiveness of radiation therapy in the treatment of bone metastases from breast cancer. Radiology 1978;126: Needham PR, Hoskin PJ. Radiotherapy for painful bone metastases. Palliat Med 1994;8:

11 218 Dawson et al. Vol. 17 No. 3 March Vargha ZI, Arvin SG, Boland J. Single dose radiation therapy in the palliation of metastatic disease. Radiology 1969;93: Penn CRH. Single dose and fractionated irradiation for osseous metastases. Clin Radiol 1976;27: Schocker JD, Brady LW. Radiation therapy for bone metastasis. Clin Orthop Rel Res 1982;169: Martin WMC. Multiple daily fractions of radiation in the palliation of pain from bone metastases. Clin Radiol 1983;34: Nag S, Shah V. Once a week lower hemibody irradiation for metastatic cancers. Int J Radiat Oncol Biol Phys 1986;12: Barak F, Werner A, Walach N, Horn Y. The palliative efficacy of a single dose of radiation in treatment of symptomatic osseous metastases. Int J Radiat Oncol Biol Phys 1987;13: Hoskin PJ, Ford HT, Harmer CL. Hemibody irradiation for metastatic bone pain in two histologically distinct groups of patients. Clin Oncol 1989;1: Kuban DA, Delbridge T, el Mahdu AM, et al. Half body irradiation for treatment of widely metastatic carcinoma of the prostate. J Urol 1989;141: Mithal, NP, Needham PR, Hoskin PJ. Retreatment with radiotherapy for painful bone metastases. Int J Radiat Oncol Biol Phys 1994;29: Wilkins MF, Keen CW. Hemi body radiotherapy in the management of metastatic carcinoma. Clin Radiol 1987;38: Price P, Hoskin PJ, Easton D, et al. Low dose single fraction radiotherapy in the treatment of metastatic bone pain: a pilot study. Radiother Oncol 1988;12: Jensen NH, Roesdahl K. Single dose irradiation of bone metastases. Acta Radiol 1976;15: