© 2010 Amgen Inc. Presented at: 2013 European Cancer Congress; September 27-October 1, 2013; Amsterdam, the Netherlands BACKGROUND • Bone is the third most common site for metastasis in advanced cancer • More than 100,000 cases of bone metastases are reported each year in Europe • Patients with bone metastases from solid tumors often experience skeletal-related events (SREs), clinically deined as pathologic fracture, radiation to bone, surgery to bone, and spinal cord compression • Several bone-targeted agents are approved to prevent SREs, but patients’ preferences among available options has not been reported previously OBJECTIVE • To assess patients’ preferences for eficacy, safety, and mode of administration in relation to currently available treatment for the prevention of SREs in metastatic patients in Europe MATERIAL AND METHODS Study Population • Inclusion criteria: – Aged 18 years or older – Patients’ self-reported physician diagnosis of bone metastases from a solid tumor • Patients were recruited from existing online panels • Participants were from France, Germany, and the United Kingdom (UK) • All participants provided informed consent Survey Instrument • An online discrete-choice experiment survey instrument was developed for this study using best practices 1 – Discrete-choice experiments, also known as choice-format conjoint analyses, are a valid and reliable approach for assessing trade-off preferences 2-4 – Survey instrument validation included a pretest of open-ended interviews with 15 patients in the United States – Additional open-ended pretest interviews were conducted in Europe (four in France, four in Germany, and three in the UK) to test the local language translations • After reviewing prescribing information for the currently approved products and consulting with clinical experts, ive attributes and their ranges were selected to describe the available bone metastases treatments (Table 1) • Each respondent was required to answer 10 choice questions between pairs of hypothetical medication proiles (Figure 1) • The experimental design included: – Combinations of attribute levels in each choice question – Main-effects D-eficient criteria obtained using Statistical Analysis System (SAS) Version 9.3 5,6 – Four survey versions each with 9 choice questions (36 choice questions in total), which were randomly assigned to respondents • The third choice question shown was repeated as the seventh, eighth, or ninth choice question as an internal validity test • Patient demographic information (e.g., age, gender) and items about the patients’ experiences with bone metastases treatments were collected Patients’ Preferences for Attributes of Bone Metastases Treatments in France, Germany, and the United Kingdom Guy Hechmati, 1 Jorge Arellano, 2 Ateesha F Mohamed, 3 A Brett Hauber, 3 Yi Qian, 2 Francesca Gatta, 1 Ian Haynes, 4 Amit Bahl 5 1 Amgen (Europe) GmBH, Zug, Switzerland; 2 Amgen Ltd., Thousand Oaks, CA, United States; 3 RTI Health Solutions, Research Triangle Park, NC, United States; 4 Amgen Ltd., Uxbridge, United Kingdom; 5 University Hospitals Bristol, United Kingdom Table 3. Patient and Disease Characteristics Category France (n = 159) Germany (n = 166) UK (n = 159) Gender Female 53.2% 38.0% 43.4% Male 46.8% 62.0% 56.6% Age < 46 years 37.2% 58% 42.8% 46-65 years 44.2% 32.9% 37.0% ≥ 66 years 18.6% 9.1% 20.2% Type of cancer that led to bone metastases Breast 36.1% 23.2% 36.5% Prostate 14.6% 17.7% 21.4% Lung 20.9% 20.7% 13.2% Other 28.4% 38.4% 28.9% Less than 2 years since diagnosed with cancer 69.1% 79.5% 67.8% Less than 1 year since diagnosed with bone metastases 73.0% 67.2% 65.4% Currently taking treatment to delay complications of bone metastases 68.4% 75.8% 70.9% Losing ability to move around affected the patient the most in the past 2 weeks 30.6% 46.7% 53.5% Had a complication because of bone metastases 50.9% 57.4% 59.5% Most worrisome complication of bone metastases A bone fracture 39.7% 45.5% 30.8% Radiation therapy to prevent a fracture and/or to treat bone pain 17.9% 18.8% 22.6% Surgery to the bone to prevent a fracture 12.8% 18.2% 6.9% Increased pressure on the spinal cord 29.5% 17.6% 39.6% Severity of worst pain in the past week for any reason No pain 1.9% 1.8% 5.7% Mild 9.4% 22.4% 28.9% Moderate 45.3% 50.3% 44.7% Severe 43.4% 25.5% 20.8% Severity of average pain in the past week for any reason No pain 2.5% 3.0% 6.3% Mild 19.0% 29.9% 32.3% Moderate 64.6% 51.2% 53.2% Severe 13.9% 15.9% 18.2% Figure 2. Preference Weights for French Patients Preference weights -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 French Patients (N = 159) Months until first SRE Months until worsening of pain Risk of ONJ Risk of renal impairment Mode of administration 28 months 18 months 10 months 10 months 6 months 3 months None 1 out of 100 (1%) 5 out of 100 (5%) None 4 out of 100 (4%) 10 out of 100 (10%) Daily oral tablet Injection every 4 weeks 15-minute infusion every 4 weeks 120-minute infusion every 4 weeks Note: The vertical bars surrounding each mean preference weight denote the 95% CI about the point estimate. If the CIs do not overlap for adjacent levels in a particular attribute, the mean estimates are statistically different from each other at the 5% level of signiicance. Figure 3. Preference Weights for German Patients Preference weights -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.4 0.2 0.8 0.6 1.0 German Patients (N = 166) Months until first SRE Months until worsening of pain Risk of ONJ Risk of renal impairment Mode of administration 28 months 18 months 10 months 10 months 6 months 3 months None 1 out of 100 (1%) 5 out of 100 (5%) None 4 out of 100 (4%) 10 out of 100 (10%) Daily oral tablet Injection every 4 weeks 15-minute infusion every 4 weeks 120-minute infusion every 4 weeks Note: The vertical bars surrounding each mean preference weight denote the 95% CI about the point estimate. If the CIs do not overlap for adjacent levels in a particular attribute, the mean estimates are statistically different from each other at the 5% level of signiicance. Figure 4. Preference Weights for United Kingdom Patients Preference weights -1.5 -1.0 -0.5 0.0 0.5 1.5 1.0 UK Patients (N = 159) Months until first SRE Months until worsening of pain Risk of ONJ Risk of renal impairment Mode of administration 28 months 18 months 10 months 10 months 6 months 3 months None 1 out of 100 (1%) 5 out of 100 (5%) None 4 out of 100 (4%) 10 out of 100 (10%) Daily oral tablet Injection every 4 weeks 15-minute infusion every 4 weeks 120-minute infusion every 4 weeks Note: The vertical bars surrounding each mean preference weight denote the 95% CI about the point estimate. If the CIs do not overlap for adjacent levels in a particular attribute, the mean estimates are statistically different from each other at the 5% level of signiicance. Figure 1. Example Choice Question Medication Feature Medication A Medication B Months until you have a complication of bone metastases 28 months 18 months Months until your pain gets worse 3 months 6 months Chance of a problem with your teeth and/or jaw bone each year because of the medicine 5 out of 100 (5%) None Chance of kidney dam- age because of the medi- cine each year 4 out of 100 (4%) 10 out of 100 (10%) How you take the medicine 120-minute infusion every 4 weeks Injection every 4 weeks Which would you choose? • • Analyses • Random-parameters logit models were used to quantify trade-off preferences among patients in each country – Parameters represented the relative preference of each attribute level 3-4 – The vertical bars around each preference weight indicated the 95% conidence interval (CI) around the mean estimate • If the CIs between adjacent levels of a single attribute did not overlap, the mean estimates were statistically different from each other at the 5% level of signiicance • Better clinical outcomes logically should be preferred to worse clinical outcomes – If better clinical outcomes were preferred to worse clinical outcomes, then all attributes were consistent with the natural ordering of the categories • Relative importance weight of attributes – The distance between the preference weight for the best and worst levels of an attribute was interpreted as the overall relative importance of the attribute over the speciic ranges presented in the survey 4 • Predicted choice probabilities were estimated using the model results for a product with characteristics similar to denosumab, zoledronic acid, clodronate, and pamidronate (Table 2). LIMITATIONS • Patients were a self-selected group: members of online panels • Respondents were asked to evaluate hypothetical treatment attributes • Differences can arise between stated and actual choices • Diagnosis was a patient self-reported physician diagnosis CONCLUSIONS • When considering treatment choices for preventing skeletal complications associated with bone metastases, patients focused mainly on delaying SREs, avoiding renal impairment, and delaying worsening of pain • This study adds to the medical information available and is useful data to aid clinicians in decision making by understanding the factors that are important to patients REFERENCES 1. Bridges JFP, Hauber AB, Marshall D, Lloyd A, Prosser LA, Regier DA, et al. Value Health 2011;14:403-11. 2. Marshall D, Bridges JFP, Hauber AB, Cameron R, Donnalley L, Fyie K, et al. Patient. 2010 Dec 1;3(4):249-56. 3. Hauber AB, Arden NK, Mohamed AF, Johnson FR, Peloso PM, Watson DJ, et al. Osteoarthritis Cartilage. 2013 Feb;21(2):289-97. 4. Mohamed AF, Hauber AB, Neary MP. Pharmacoeconomics. 2011 Nov;29(11): 977-88. 5. Johnson FR, Lancsar E, Marshall D, Kilambi V, Mühlbacher A, Regier DA, et al. Value Health. 2013 Jan;16(1):3-13. 6. Kuhfeld W, Tobias F, Garratt M. J Mark Res. 1994;31:545-57. DISCLOSURE Funding for this study was obtained from Amgen, Thousand Oaks, California, United States. Medical writing and editorial assistance was provided by Emma Thomas of Amgen (Europe) GmbH. Table 1. Attributes and Levels for the Choice Questions Attribute Levels Months until you have a complication of bone metastases (Months until irst SRE) 28 months 18 months 10 months Months until worsening of pain 10 months 6 months 3 months Chance of a problem with your teeth and/or jaw bone (Annual risk of ONJ) None 1 out of 100 (1%) 5 out of 100 (5%) Chance of kidney damage (Annual risk of renal impairment) None 4 out of 100 (4%) 10 out of 100 (10%) How you take the medicine (Mode of administration) Daily oral tablet Injection every 4 weeks 15-minute infusion every 4 weeks 120-minute infusion every 4 weeks ONJ = osteonecrosis of the jaw. Table 2. Proiles Attribute Characteristics Similar to Denosumab Characteristics Similar to Zoledronic Acid Characteristics Similar to Clodronate Characteristics Similar to Pamidronate Months until worsening of pain 6.6 months 4.7 months 3 months From 1 day to several months (assume 3 months) Risk of ONJ 1.8% 1.3% Yes, but value not stated (assume 1%) Yes, but value not stated (assume 1%) Risk of renal impairment 0% 9.3% Yes, but value not stated (assume 5%) 8.1% Mode of administration Injection every 4 weeks 15-minute infusion every 4 weeks Daily oral tablet 120-minute infusion every 4 weeks Predicted Choice Probabilities • Table 4 presents predicted choice probabilities. RESULTS Patients • Of those eligible patients who consented to participate, 166 patients from France, 175 patients from Germany, and 165 patients from the UK answered at least one choice question • Among all the eligible patients who agreed to participate and answered at least one choice question, 7 patients from France, 9 patients from Germany, and 6 patients from the UK were excluded because they always chose the same answer (Medication A or Medication B) in the choice questions (i.e., showing no understanding of the exercise) • The inal sample sizes were 159 patients from France, 166 from Germany, and 159 from the UK • Baseline patient and disease characteristics are reported in Table 3. Preference Weights • Preference weights for all the attribute levels were consistent with natural ordering for all countries (Figures 2-4) • For all countries, the preference weights for all levels were statistically signiicantly different (P < 0.05) for the attributes: – Months until irst SRE – Months until worsening of pain – Risk of renal impairment • For France and Germany, preference weights for the irst two levels were not statistically different (P > 0.05) for risk of ONJ; preference weights for the irst three levels were not statistically signiicantly different (P > 0.05) for mode of administration • In the UK, the preference weights for adjacent levels were not statistically signiicantly different (P > 0.05) for risk of ONJ and mode of administration • For French and German patients risk of renal impairment, months until irst SRE, and months until worsening of pain were the top three relatively important attributes (in decreasing order) – For French patients this was followed by risk of ONJ and then mode of administration, with a 120-minute infusion every 4 weeks as the least preferred mode of administration – In Germany, the relative importance of the last two attributes was reversed, mode of administration (120-minute infusion every 4 weeks was again the least preferred) and the risk of ONJ • UK patients determined the relative importance of attributes in decreasing order to be slightly different: months until irst SRE, months until worsening of pain, risk of renal impairment, mode of administration (120-minute infusion every 4 weeks was the least preferred mode of administration), and risk of ONJ Table 4. Predicted Choice Probabilities Sample Characteristics Similar to Denosumab Characteristics Similar to Zoledronic Acid Characteristics Similar to Clodronate Characteristics Similar to Pamidronate French patients 94.8% (90.3%, 97.4%) 2.5% (1.0%, 4.7%) 2.2% (1.1%, 4.1%) 0.6% (0.3%, 1.5%) German patients 75.6% (65.3%, 83.3%) 9.1% (5.7%, 13.3%) 12.3% (8.3%, 17.7%) 3.1% (1.5%, 5.7%) UK patients 83.5% (74.1%, 89.9%) 7.1% (3.9%, 11.5%) 8.0% (4.6%, 13.0%) 1.6% (0.7%, 3.2%)