|
Economic model of first-line drug strategies to achieve recommended glycaemic control in newly diagnosed type 2 diabetes mellitus |
Ramsdell J W, Braunstein S N, Stephens J M, Bell C E, Botteman M F, Devine S T |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology The use of first-line therapies aimed at achieving the glycaemic control recommended by the American Diabetes Association (ADA) in patients with Type 2 diabetes. The six therapies examined were:
glipizide gastrointestinal therapeutic system (GLI-GITS),
metformin immediate release (MET-IR),
metformin extended release (MET-XR),
glibenclamide (glyburide)-metformin (GLI-MET),
rosiglitazone (ROS), and
repaglinide (REP).
Economic study type Cost-effectiveness analysis.
Study population The study population comprised a hypothetical cohort of newly diagnosed Type 2 diabetes patients.
Setting The setting is likely to have been primary care. The economic study was carried out in the USA.
Dates to which data relate The effectiveness and resource use data were derived from studies published between 1993 and 2001. The costs were reported as 2001/2002 values.
Source of effectiveness data The effectiveness evidence was derived from a review of completed studies and experts' assumptions.
Modelling A decision model based on a Markov approach was used to estimate the expected total costs and number of patients achieving ADA glycaemic control over 3 years. ADA glycaemic control was defined as glycosylated haemoglobin values (HbA1c) of at least 7%). The length of the Markov cycle was 3 months, which reflected the frequency of physician office visits. In the case of primary failure or adverse events, the patients were switched to second-line therapies before moving to an insulin-containing regimen. Primary failure was defined as little or no change in blood glucose or HbA1c during the first 3 months of therapy. The model did not consider diabetic complications because the model was short-term and all the patients were treated to reach the recommended glycaemic control.
Outcomes assessed in the review The outcomes assessed in the review were the probabilities values for:
the rate of ADA control after first-line therapy,
the rate of primary failure after first-line therapy,
the rate of discontinuation due to adverse events,
the rate of control with the addition of a second or third agent if patients did not reach control with first-line therapy, and
the rate of secondary failure (HbA1c value drifts to 7% or more after being in good control).
Study designs and other criteria for inclusion in the review The design of the primary studies was not reported, but the authors stated that randomised trials were used whenever possible.
Sources searched to identify primary studies MEDLINE was searched from 1966 to 2001 for relevant studies. A manual search of bibliographic references was also conducted.
Criteria used to ensure the validity of primary studies Methods used to judge relevance and validity, and for extracting data Number of primary studies included Twenty-four primary studies were included in the review.
Methods of combining primary studies The authors did not state how the primary estimates were combined. When a range of values was retrieved from the literature, a specific value was selected. However, the approach used was not reported.
Investigation of differences between primary studies Results of the review The rate of ADA control after first-line therapy was 0.51 with GLI-GITS, MET-IR and MET-XR, 0.717 with GLI-MET, 0.335 with ROS and 0.32 with REP.
The rate of primary failure after first-line therapy was 0.105 with GLI-GITS, 0.1 with MET-IR and MET-XR, 0.017 with GLI-MET, 0.25 with ROS and 0.08 with REP.
The rate of discontinuation due to adverse events was 0.021 with GLI-GITS, 0.06 with MET-IR, 0.006 with MET-XR, 0.02 with GLI-MET, 0.05 with ROS and 0.01 with REP.
The rate of control with an additional (second or third) agent if patients did not reach control with first-line therapy was 0.58 with GLI-GITS plus MET-IR, 0.34 with GLI-GITS plus ROS, 0.281 with MET-IR plus ROS, 0.59 with MET-IR plus REP, 0.50 with GLI-MET plus ROS, 0.50 with GLI-GITS plus MET-IR and ROS, and 0.50 with REP plus MET-IR and ROS.
The rate of secondary failure was 0.07 with all possible drug combinations.
The ranges of values used in the base-case model, the ranges observed in the literature, and ranges based on experts' opinions were also reported.
Methods used to derive estimates of effectiveness The authors used opinions from a panel of endocrinologists, and internal medicine physicians to supplement literature rates when no data were available.
Estimates of effectiveness and key assumptions Data from expert opinions were used only to define the ranges of values, because the final estimate was derived using data coming from the literature (reported above). Expert opinion mainly contributed to the estimation of the probability of second failure with combination therapies.
Measure of benefits used in the economic analysis The summary benefit measure was the number of patients controlled (HbA1c of less than 7%) on oral agents (monotherapy, combination, or triple therapies) within each first-line tree at the end of year 3. This measure was obtained using the decision model.
Direct costs The costs were discounted at an annual rate of 3% since the time horizon of the model was 3 years. The unit costs were reported separately from the quantities of resources used. The categories of costs included in the analysis were drugs, continuing medical care, and the treatment of adverse events. Continuing medical care included physician office visits, patient education, laboratory tests, preventive vaccination, podiatry and eye exam visits, and home glucose monitoring. The adverse effects were hypoglycaemia, severe gastrointestinal discomfort, lactic acidosis, elevated liver enzymes, and oedema. The perspective of the payer was applied.
The resource use data were estimated from the ADA Clinical Practice Recommendations and probability values obtained from the literature. Average wholesale prices of drug and supplies were based on the Drug Topics Red Book and Drug Topics Red Book Updates. Physician services were estimated from the Medicode's National Fee Analyzer. The Medicare's Clinical Diagnostic Laboratory Fee Schedule was used to estimate the laboratory costs. Finally, the hospital costs related to the treatment of adverse events came from the St. Anthony' DRG Guidebook. All the costs were presented in 2001/2002 values.
Statistical analysis of costs Statistical tests of the costs were not conducted in the base-case.
Indirect Costs The indirect costs were not included.
Sensitivity analysis Univariate and multivariate sensitivity analyses, varying all model inputs by +/- 20%, were conducted to assess the robustness of the model results. Whenever possible, the ranges estimated in the literature (supplemented with expert opinions) were used. A probabilistic second-order Monte Carlo simulation was carried out in which 10,000 trials of the decision model were run. Triangular distributions were used for all model variables. The costs were varied by +/-50% and the discount rates were varied from 0 to 6% annually.
Estimated benefits used in the economic analysis The number of patients controlled in the cohort of 1,000 patients after 3 years' treatment was 823 with GLI-GITS, 822 with MET-IR, 828 with MET-XR, 793 with GLI-MET, 750 with ROS and 798 with REP. Thus, the effectiveness results were similar across the six treatments. The number of patients controlled did not vary in the sensitivity analysis. When the probabilistic analysis was run, there was no statistically significant difference between the treatments.
Cost results The cumulative discounted overall costs of treatment per patient in the cohort of 1,000 patients after 3 years' treatment were $6,106 with GLI-GITS, $6,727 with MET-IR, $6,826 with MET-XR, $7,141 with GLI-MET, $7,759 with ROS and $9,298 with REP. Thus, the GLI-DITS strategy led to cost-savings of $621 to $3,192 in comparison with the other therapies.
The estimated costs were sensitive to variations in the control rate on first-line therapy, costs of routine medical care, and drug costs. However, GLI-GITS remained the least expensive strategy in 86 to 99% of all model simulations.
Synthesis of costs and benefits A synthesis of the costs and benefits was not carried out because the authors concluded that the benefit measure (number of patients controlled) was similar between the treatments, which were considered equally effective. Therefore, it appears that a cost-minimisation analysis was conducted.
Authors' conclusions The oral agents included in the economic evaluation were all similarly effective in achieving glycaemic control in newly diagnosed Type 2 diabetes patients. However, GLI-GITS treatment provided early cost-savings when compared with the other first-line strategies.
CRD COMMENTARY - Selection of comparators The authors stated that all the comparators selected in the analysis were the most commonly prescribed individual agents in the USA, or were new technologies or entrants to the market. The authors stated that they did not consider acarbose, used to reach glycaemic control, as a comparator because it was not commonly prescribed in the USA as first-line therapy. You should decide whether the treatments considered in the analysis represent widely used comparators in your own setting.
Validity of estimate of measure of effectiveness The effectiveness evidence was derived from a review of the literature. The methods and conduct of the review were reported, but details of the primary studies (design, patient population, duration of treatment) were not provided. The authors did not state how the baseline point estimates were selected from the ranges of values observed in the literature. The internal validity of the primary studies was not discussed. Literature estimates were supported by expert opinion, but the method used to derive the estimates (e.g. a Delphi approach) was not reported. Extensive (deterministic and probabilistic) sensitivity analyses were carried out to deal with the uncertainty of the estimates used in the model.
Validity of estimate of measure of benefit No summary benefit measure was used in the economic analysis and the health outcomes turned out to be equivalent. The study was therefore categorised as a cost-minimisation analysis.
Validity of estimate of costs The authors explicitly stated the perspective adopted in the study and reported many details of the economic analysis. In particular, the price year, quantities of resources used, unit costs, source of data, and discount rate. Statistical tests of the costs were not conducted in the base-case, but wide ranges of values were tested in the sensitivity analysis to take variability in the data into account. The costs were estimated over 3 three years, and the authors justified their choice of the time horizon adopted in the decision model. A breakdown of the costs was provided.
Other issues The authors compared their findings with those from other studies, although it was stated that few comparable articles were available in the literature. The findings of the present study were consistent with those observed in other economic evaluations, mainly in relation to the key role played by drug acquisition costs. In terms of the issue of the generalisability of the study results, the authors stated that US-specific data were used and the drug prices were generally lower in Europe. However, the conclusions of the study were shown to have been robust to variations in almost all of the model inputs. The authors noted some limitations to the validity of their analysis. First, the use of data referring to a definition of glycaemic control slightly different from that used in the decision model. Second, the price variability and fluctuations. Third, the new availability of cheaper generic drugs. Finally, medication compliance, diabetic complications, and beneficial lipid effects were not directly included. However, all of these issues were addressed in the sensitivity analyses.
Implications of the study The main implication of the analysis was that treatment with GLI-GITS represented an effective and efficient treatment for newly diagnosed Type 2 diabetes patients. The authors stressed that patient-specific characteristics such as metabolic profile, organ function and co-morbidities, have to be considered when choosing any therapy for the treatment of newly diagnosed Type 2 diabetes patients. In addition, the economic implications of oral therapy represent a key issue for the payer, who may be interested in selecting the least expensive strategy given the comparable effectiveness and safety profile of all the therapies evaluated.
Source of funding Sponsored in part by Pfizer Inc.
Bibliographic details Ramsdell J W, Braunstein S N, Stephens J M, Bell C E, Botteman M F, Devine S T. Economic model of first-line drug strategies to achieve recommended glycaemic control in newly diagnosed type 2 diabetes mellitus. PharmacoEconomics 2003; 21(11): 819-837 Indexing Status Subject indexing assigned by NLM MeSH Carbamates /economics /therapeutic use; Cohort Studies; Decision Trees; Diabetes Mellitus, Type 2 /drug therapy /economics /prevention & Direct Service Costs /statistics & Drug Costs /statistics & Drug Therapy, Combination; Glipizide /economics /therapeutic use; Glyburide /economics /therapeutic use; Hemoglobin A, Glycosylated /drug effects; Humans; Hypoglycemic Agents /classification /economics /therapeutic use; Managed Care Programs /economics /utilization; Markov Chains; Metformin /economics /therapeutic use; Models, Economic; Piperidines /economics /therapeutic use; Thiazoles /economics /therapeutic use; Thiazolidinediones; United States; control; numerical data; numerical data AccessionNumber 22003008251 Date bibliographic record published 30/04/2004 Date abstract record published 30/04/2004 |
|
|
|