Abstracts 151 METHODS: A 3-state Markov model was developed. Transition probabilities, utilities, and costs were assumed. Costs and effects were discounted at 3%. Decision analytic CL and matrix CL were calculated as the interquartile range (IQR) from 10,000 simulations. For each simulated value, over or under was defined relative to the exact value. Bias was defined as the ratio of (over - under)/ (over + under). DATA software was used for decision analysis; S Plus was used for matrix programming. RESULTS: Estimated life-years, quality-adjusted life-years (QALY), and costs ($) are summarized in the table below. CONCLUSIONS: Decision analytic software may yield bi- ased estimates of costs and effects. The implications of this must be considered. Analysts and policy-makers should carefully validate all decision models prior to using them to determine health policy. TPDM4 MODELING LIFETIME TREATMENT COSTS OF HIV/AIDS PATIENTS Pleil AM 1 , Kempel A 2 , Willke RJ 3 1 Pharmacia & Upjohn, Stockholm, Sweden; 2 Pharmametrics GmbH, Freiburg, Germany; 3 Pharmacia & Upjohn, Kalamazoo, MI, USA OBJECTIVES: A model was developed to evaluate the cost-effectiveness of a non-nucleoside reverse transcriptase inhibitor (RTI), delavirdine (DLV) in combination with two nucleoside RTIs (AZT + 3TC) versus a control arm of AZT + 3TC in the management of HIV/AIDS pa- tients. METHODS: A Markov chain model is used to describe the clinical progression of HIV/AIDS patients based on discrete CD4 cell count categories. The first year transi- tion probabilities are estimated from a head-to-head clini- cal trial comparing these regimens and subsequent annual probabilities are derived from previously reported trials describing the natural progression of the disease. Re- source use and cost data are based on information col- lected from clinical experts and include medical resource use for routine care and the prophylaxis and treatment of opportunistic infections. The economic and clinical effects of antiviral treatment are examined. Sensitivity analysis was performed to determine the robustness of the cost- effectiveness results. The effect of time preference is in- cluded by discounting future costs and quality-adjusted life-years (QALYs) in the sensitivity analysis. RESULTS: The lifetime treatment cost for a cohort of pa- tients at beginning CD4 cell counts of 350 to 500 is $105,880 for the DLV arm and $101,962 for the control arm. The DLV incremental cost per QALY gained was $13,262 for the undiscounted base case. Applying a dis- count rate of 5% for costs alone resulted in US $12,637 per QALY gained; discounting costs and benefits at 5% resulted in US $6854 per QALY gained. CONCLUSIONS: The results indicate that DLV in com- bination is a cost-effective treatment for this cohort of patients as compared to a standard combination therapy. Future application of the model to other treatment pat- terns with DLV are forthcoming. TPDM5 COST-BENEFIT ANALYSIS OF AN INTRA- UTERINE LEVONORGESTREL-RELEASING DEVICE MIRENA VERSUS HYSTERECTOMY FOR WOMEN WITH MENORRHAGIA Milne RJ 1,2 , Farquhar CF 3 1 Health Outcomes Associates Ltd., Auckland, New Zealand; 2 University of Auckland, Auckland, New Zealand OBJECTIVE: The study purpose was to evaluate the overall cost of hysterectomy versus a levonorgestrel- releasing intra-uterine device Mirena backed up by hys- terectomy for women with heavy menstrual bleeding (menorrhagia). METHODS: A decision tree was based on a clinical trial of Mirena versus hysterectomy. The model considers pharmaceuticals, GP and specialist consultations, hyster- ectomy, complications of surgery and surgical cancella- tion rates. Future costs are discounted to present value at 5% per annum. RESULTS: In the base case, over 20 years from a health- care payer perspective, first-line treatment with Mirena is predicted to cost $3047 versus $3800 per individual for first-line treatment with hysterectomy. From a healthcare funding agency (HFA) perspective, the costs are $2289 for Mirena versus $2867 for hysterectomy. From a phar- maceuticals budget perspective, Mirena costs $639 versus $284 for hysterectomy. In the base case, first-line treat- ment with Mirena costs 70% to 80% of the cost of hys- terectomy over periods of 5 to 20 years. Switching a can- didate for hysterectomy to Mirena has the potential to avert costs in the range $753 to $1076 per woman from a healthcare payer perspective and $578 to $807 from an HFA perspective. Threshold analysis shows that therapy with Mirena will be less costly than hysterectomy pro- vided that more than 16% of women with menorrhagia who accept Mirena subsequently cancel hysterectomy in Data analytic Matrix inversion Median (IQR) Bias Median Bias Life-years 5.187 0.086 5.186 0.068 (4.812; 5.611) (4.803; 5.605) QALY 0.074 0.011 3.960 3.927 $ (3.669; 4.286) 0.006 (3.639; 4.251) -0.014 $ per 134,900 -0.035 133,800 -0.041 life-year (110,900; 160,800) (109,700; 160,000) -0.038 -0.017 $ per 25,900 25,900 QALY (20,900; 31,500) (21,700; 30,300) 33,800 34,100 (27,400; 41,300) (28,200; 40,600)