Address correspondence to: John Doyle, The Analytica Group, 475 Park Avenue South, 17th Floor, New York, NY 10016. E-mail: email@example.com
Methods: We conducted a multinational pharmacoeconomic evaluation comparing the immediate release form of a new class of serotonin norepinephrine reuptake inhibitor (SNRI), venlafaxine IR to the selective serotonin reuptake inhibitors (SSRIs) and the tricyclic antidepressants (TCAs) in the treatment of acute major depressive disorder (MDD) in 10 countries (Germany, Italy, Netherlands, Poland, Spain, Sweden, Switzerland, United Kingdom, United States, and Venezuela). We designed a decision analytic model assessing the acute phase of MDD treatment within a 6-month time horizon. Six decision tree models were customized with country-specific estimates from a clinical management analysis, meta-analytic rates from two published meta-analyses, and a resource valuation of treatment costs representing the inpatient and outpatient settings within each country. The meta-analyses provided the clinical rates of success defined as a 50% reduction in depression scores on the Hamilton Depression Scale (HAM-D) or the Montgomery-Asberg Depression Rating Scale (MADRS). Treatment regimen costs were determined from standard lists, fee schedules, and communication with local health economists in each country. The meta-analytic rates were applied to the decision analytic model to calculate the expected cost and expected outcomes for each antidepressant comparator. Cost-effectiveness was determined using the expected values for both a successful outcome, and a composite measure of outcome termed symptom-free days. A policy analysis was conducted to examine the health system budget impact in each country of increasing the utilization of the most effective antidepressant found in our study.
Results: Initiating treatment of MDD with venlafaxine IR yielded a lower expected cost compared to the SSRIs and TCAs in all countries except Poland in the inpatient setting, and Italy and Poland within the outpatient settings. The weighted average expected cost per patient varied from US$632 (Poland) to US$5647 (US) in the six-month acute phase treatment of MDD. The estimated total budgetary impact for each 1% of venlafaxine utilization, assuming a population of one million MDD patients, ranged from US$1600 (Italy) to US$29,049 (US).
Conclusions: Within the inpatient and outpatient treatment settings, venlafaxine IR was a more cost-effective treatment of MDD compared to the SSRIs and TCAs. Additionally, the results of this investigation indicate that increased utilization of venlafaxine in most settings across Europe and the Americas will have favorable impact on health care payer budgets.
ADR, adverse drug reaction; CMA, clinical management analysis; ECT, electroconvulsive therapy; HAM-D, Hamilton Depression Scale; MADRS, Montgomery-Asberg depression rating scale; MDD, major depressive disorder; SFD, symptom-free day; SNRI, serotonin-norepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant; WHO, world health organization.
Major depressive disorder (MDD) represents a significant public health problem in the new millennium 2000, posing a substantial burden of illness to health care providers, payers, and patients worldwide. Additionally, it is projected that by the year 2020, depression will rank second in disease burden measured in Disability-Adjusted Life Years (DALYS) .
Intuitively, the economics of diagnosing and treating this disease is daunting. It was estimated in 1993 that the three leading types of depression (manic depression, major depression and dysthymia) cost the United States US$43.7 billion a year , which is approximately equal to the cost of treating heart disease . European countries have also reported significant cost of illness. In the United Kingdom, for example, Kind estimated that in 1993 depression cost some £420 million annually .
As imposing as these estimates of the economic burden of depression seem, it is quite possible that these findings underestimate the true cost of depression, because they represent only the cost of treating diagnosed depression. According to the WHO , depression is estimated to be present in 10% of all patients seeking care at primary health care facilities worldwide, but it is frequently unrecognized by primary care physicians. The DEPRES study attempted to quantify this and showed that only 41% of those suffering from major depression received medication . In terms of indirect costs, the DEPRES study also showed that MDD has a deleterious impact on productivity measured in days-of-work-lost; on average MDD patients lost 13 days over six months vs. three days for nonsufferers . Astonishingly, the rate of suicides for those suffering from major depression has been estimated to be between 10% and 15%, accounting for 60% to 70% of all suicides .
The ratio of health care expenditure to gross domestic product continues to rise among the industrialized nations, driven in part by the aging population, and also by the increasing cost of health care services and medical technology advancements. Of the 10 countries examined in this research, only Italy and Spain have experienced decreases in this ratio from 1990 to 1996 as reported by the Organization for Economic Cooperation and Development (OECD)  and local health economists. Health care policy decision-makers around the world are increasingly instituting measures to control the escalating costs of health care, from the development of guidelines and disease-management initiatives to the capitation of health care expenditures.
Cost-containment initiatives worldwide specific to MDD have traditionally focused on market prices of pharmaceutical interventions to curb treatment costs, particularly with respect to the newer antidepressants such as selective-serotonin reuptake inhibitors (SSRIs). For example in 1993, government funding in Italy was reduced or eliminated for several hundred products, including SSRIs. Although controlling drug expenditures represents a relatively simple and perhaps effective target for short-term policy-making, the cost-effective treatment of MDD for long-term health care policy will depend on evaluation of all resources consumed during the treatment process. The newer antidepressants, generally with higher acquisition costs than the older drugs such as tricyclics (TCAs), generate tangible benefits, such as an improved efficacy and safety profiles. These clinical benefits must be economically quantified and compared to facilitate the effective and efficient allocation of mental health care resources. Toward that end, comprehensive pharmacoeconomic evaluation is warranted to assess total direct costs and cost-effectiveness of competing antidepressants.
In recent years, a number of pharmacoeconomic studies of MDD treatment have been conducted and have produced varied conclusions regarding the patient-level implications of alternative treatment options [9–21]. This research represents the first attempt to investigate patient-level implications of antidepressant treatment in 10 countries, and to quantify those findings in terms of population level impact on health care budgets of primary payers in each market.
We performed a cost-effectiveness comparison between venlafaxine (IR) the immediate release form of the serotonin-norepinephrine reuptake inhibitor (SNRI) class vs. two other classes of antidepressants, the selective serotonin reuptake inhibitors (SSRIs) and the tricyclic antidepressants (TCAs) in the following 10 countries: Germany, Italy, Netherlands, Poland, Spain, Sweden, Switzerland, United Kingdom, United States, and Venezuela. The structural design of our decision-tree model illustrated in Figure 1 and the clinical outcomes data of each comparator shown in Table 1 were based on a previous pharmacoeconomic analysis performed in Canada  and two recently published meta-analyses [22, 23]. A policy-level analysis estimated the budgetary impact that greater venlafaxine utilization would have on the public or private payers responsible for treatment of acute MDD within the respective populations of the 10 countries.
Table 1. Meta-analytic rates
No. of patients
No. of patients
Rates for venlafaxine IR are taken from Einarson et al. (1998).
Rates for the SSRIs and TCAs are adapted from Einarson et al. (1999). SSRIs, selective serotonin reuptake inhibitors; TCAs, tricyclic antidepressants.
We analyzed the health care reimbursement systems of all 10 countries to identify and address the dominant payer perspective [24–39]. With the exception of the United States, the majority of health care funding is public (Table 2). Therefore, for all countries studied but the US, our research was conducted from the perspective of the government body responsible for health care reimbursement. In the United States, a managed care perspective was used in the evaluation.
Table 2. Public expenditures as a percent of total health expenditures
After examining the available clinical outcomes data for acute MDD and the opinion of mental health care providers within each of the countries, we determined that a 6-month time horizon was an appropriate assessment period for treatment in the acute phase of MDD. Specifically, since there is little available clinical trial data assessing relapse rates in patients with acute MDD, our timeline was restricted to 6-months to limit the potential information bias.
Table 1 depicts the clinical outcomes data, representing success rates and dropout rates that were derived from two meta-analyses conducted by Einarson and colleagues in 1998  reporting the rates of the immediate release form of venlafaxine IR and a more recent meta-analysis (1999) reporting the rates for the SSRIs and TCAs . The success rates for each of the three comparators were defined in the meta-analysis as a 50% reduction in depression scores on the Hamilton Rating Scale for Depression (HAM-D) or the Montgomery-Ashberg Depression Rating Scale (MADRS). Dropout rates were defined by the probabilities of failing treatment due to patient perceived lack of improvement (efficacy) or due to adverse drug reactions (ADRs) [20, 21].
Decision Tree Model
Figure 1 illustrates the predictive decision tree model designed by Einarson and colleagues (1997) to represent the possible outcomes of a hypothetical patient treated for acute MDD in the inpatient and/or outpatient setting over a six-month time period. The decision tree probabilities of each branch were determined through the meta-analysis and clinical management analysis. Efficacy and dropout rates obtained from the meta-analysis were assumed to be the same across countries, however, country-specific assumptions were derived from the opinion of local experts to account for differences in the choice of medical treatment options.
Clinical Management Analysis
A clinical management analysis (CMA) was performed to estimate the country-specific treatment options and outcomes within the set structure of the decision tree model  seen in Figure 1. A total of 42 mental health clinicians were consulted to perform this analysis. Surveys were sent to the participating physicians and follow-up interviews were conducted to establish consensus within each of the 10 countries. Six decision trees, representing the three treatment comparators within the inpatient and outpatient setting, were customized within each country to identify the number of physician and psychiatric visits; diagnostic tests, hospitalization days, comparator treatment dosing schedules and additional therapies required at each decision node within each branch in the decision tree. Inpatient and outpatient trees only differed by health care resources identified for each treatment comparator and decision probabilities of branches 1–26 (Fig. 1). Furthermore, the CMA also determined the probabilities of titration, augmentation, combination and switching to second-line therapies within the specific decision node probabilities for each treatment comparator.
Similar to the Canadian model developed by Einarson et al.  results are presented at the drug class level. In all 10 countries, the reference drug, venlafaxine (a serotonin norepinephrine reuptake inhibitor, SNRI) was compared to two classes of antidepressants: the selective serotonin reuptake inhibitors (SSRIs) and the tricyclic antidepressants (TCAs). The CMA Participants in each country also determined the most commonly prescribed antidepressants within the SSRI and TCA classes. The SSRI class was represented by at least two of the following (fluoxetine, fluvoxamine, paroxetine, sertraline and citalopram) while the TCA class was represented by at least three of following (fluoxetine, fluvoxamine, paroxetine, sertraline and citalopram) in each country. The average values of the individual drugs, which constitute each class, were used to calculate aggregated results.
Health Care Resource Valuation
A total of 14 health economists within each country identified the unit costs from standard lists for the antidepressant comparators and medical resources identified by the clinical management analysis. An average weighted cost was calculated for the relevant SSRIs and TCAs to establish a representative cost per class in each country. Furthermore, each inpatient and outpatient decision tree differed by resource valuations. For example, in the United Kingdom, a general practitioner generally manages a patient with acute MDD in the outpatient setting, while a psychiatrist strictly manages these patients in the inpatient setting. Hence, treatment with any of the three comparators will incur a lower cost in the outpatient vs. inpatient settings in all countries. Table 3 compares a selection of the most significant health care resources, which were, along with other relevant resources and costs, incorporated into distinct decision trees to estimate the total expected costs per comparator within the inpatient and outpatient settings in each country. All costs in Table 2 were represented in 1999 US dollars. Considerable variance in resources and their associated costs between countries precluded a more extensive representation of the resource valuation table.
Table 3. Resource valuation of acute MDD treatment
The pharmacoeconomic analysis employed three analytic types: Consequence Analysis, Expected Cost/Outcome Analysis, and Cost-Effectiveness Analysis. Consequence Analysis identifies all of the health outcomes associated with each of the treatment pathways in the decision analytic model. The primary outcome identified for this evaluation was treatment success, calculated by multiplying the success rates, drop-out rates due to lack of efficacy or adverse drug reactions in Table 1.
Expected success rates were calculated by summing the probabilities in each branch (determined in the CMA) of the decision tree for a successful outcome prior to the end of the 6-month time horizon. The secondary outcome measure of this evaluation was symptom-free days (SFDs). SFDs were defined as six months minus the time elapsed before clinical determination of success. Due to the lack of availability of published quality-of-life data, this parameter estimate was a proxy for patient utility. For each branch of the decision tree terminating in a successful treatment, the time after the determination of successful treatment until the end of the 6-month time period was considered symptom-free. It was assumed that there would be zero SFDs for any pathway terminating prior to treatment success.
Unlike cost-consequence analysis, Expected Cost/Outcome Analysis incorporates the probabilities from the decision analytic model. Chance nodes in the model were evaluated based on success and failure rates, which were determined through meta-analysis for each of the treatment comparators. Decision nodes were evaluated through consultation with clinical experts. For example, as shown in Fig. 1, the chance of success after primary treatment with an SSRI, was derived through efficacy rates determined in the meta-analysis . In contrast, the decision to treat with electroconvulsive therapy (ECT) after failure on secondary drug treatment was determined though country-specific physician consensus.
Cost-Effectiveness Analysis was conducted by combining the expected cost with primary and secondary outcome measures. The following cost-effectiveness endpoints were calculated: 1) expected cost per success; and 2) expected cost per SFD. Incremental cost-effectiveness ratios (ICERs) were also calculated to determine the incremental cost per success and the incremental cost per SFD in treatment settings whereby venlafaxine IR did not result in the lowest expected cost of treatment alternative.
Two sensitivity analyses were performed to test assumptions implicit to the model. Univariate or break-even analyses were conducted by varying the success rates and unit costs of venlafaxine independently to determine the break-even point at which venlafaxine's expected cost would equal the expected cost of the SSRIs and the TCAs within each country.
In addition, a series of multivariate analyses were conducted using a Monte Carlo simulation, which generated 10,000 iterations by varying the inputs for unit costs of venlafaxine IR and the other class comparators as well as the success and dropout rates for both primary and secondary therapy for each of the comparators. Success and failure rates were varied randomly within the 95% confidence interval based on a truncated normal distribution. Costs of resources used for services provided throughout therapy were varied randomly within ± 20% of the estimated cost based on a uniform distribution. The Monte Carlo simulation calculated the mean and standard deviation values for the difference in expected cost and expected SFDs between venlafaxine and either the SSRI or TCA class comparators.
To facilitate population-level decision-making among the target audiences for this study, we conducted a policy analysis that extrapolated the patient-level results to provide an estimated budgetary impact of increased utilization of venlafaxine IR. The approach followed for the budget impact analysis is outlined in Figure 2.
Decision probabilities were calculated for each of the 26 possible outcomes or branches for 60 decision trees representing all 10 countries, each assessing treatment patterns and outcomes of the three treatment comparators in two treatment settings. As an example, Table 4 depicts the decision probabilities calculated for the inpatient and outpatient settings for each of the comparators in Germany. As shown in branch 1 of Table 4, the highest rate of success resulted from the first-line treatment with venlafaxine IR at 51% and 57% followed by first-line treatment with the SSRIs at 45% and 46%, and lastly, 41% and 38% with the TCAs in the inpatient and outpatient settings, respectively. Mental health experts in Germany determined that antidepressant combination therapy was not a therapeutic option for outpatients. Outpatient treatment of MDD in the acute phase poses a higher risk of adverse drug interactions due to less medical monitoring compared to the inpatient setting. As such, the decision probabilities for combination therapy were zero for the outpatient decision tree in Germany (branches 11–14). Moreover, the clinical management analysis in Germany also identified the second-line treatment typically prescribed following first-line failure for each comparator. In the event of failure of first-line treatment with venlafaxine or the SSRIs, the TCAs were the second-line or back-up alternative. In contrast, the SSRIs were the second-line or back-up treatment following failure of first-line treatment with the TCAs. This is one reason why the probabilities of success for the second-line treatments (branch 19) were consistently higher with the TCAs using SSRIs as back up vs. venlafaxine IR and the SSRIs, which used TCAs as back-up. The other obvious reason is that the higher initial success rates of venlafaxine IR invariably drove down the probabilities of failure due to titration, augmentation, and adverse drug reactions vs. the comparators.
Table 4. Decision probabilities: example for Germany
Table 5 provides the results of the inpatient and outpatient expected costs associated with each treatment in all 10 countries. Within the inpatient setting, acute MDD treatment with venlafaxine resulted in a lower expected cost vs. treatment with SSRIs or TCAs in all countries with the exception of Poland where treatment with venlafaxine was estimated at $1433 vs. $1314 for the SSRIs and TCAs during the six month period. Within the outpatient settings, treatment with venlafaxine was associated with a lower cost in all countries, other than Poland and Italy where the acquisition costs of the TCAs and even SSRIs were significantly lower than venlafaxine as seen in Table 4.
Table 6 compares the success rates calculated in the country specific decision trees for each treatment comparator based on the efficacy rates following initial treatment in the inpatient or outpatient setting. Initiating treatment of acute MDD with venlafaxine yielded superior outcomes in terms of expected success rates for inpatients and outpatients vs. TCAs and SSRIs in all countries examined (Table 5). Based on the costs and success rates calculated within the decision tree model for each treatment, we determined the expected cost per success as depicted in Table 7. Venlafaxine was more cost-effective than the SSRI and TCA classes in 8 of the 10 countries within the outpatient setting and in 9 of the 10 countries within the inpatient setting. Among the 9 countries in which venlafaxine ranked first in expected cost per success, the difference in expected cost per success for venlafaxine and the second ranking comparator ranged from US$66 (Italy) to US$1617 (US) for outpatients, and from US$524 (UK) to US$3750 (US) for inpatients.
Furthermore, the treatment specific decision trees representing the inpatient and outpatient settings also estimated the number of Symptom-Free Days (SFDs) as shown in Table 8. Since venlafaxine was a more effective treatment, it therefore resulted in a higher number of SFDs vs. SSRIs and TCAs in all countries regardless of inpatient or outpatient setting. Table 9 shows the results of the estimated cost per symptom-free day for each patient treated with either comparator as an inpatient or outpatient. In the countries where venlafaxine ranked first in expected cost per SFD, the difference in expected cost per SFD for venlafaxine and the second ranking comparator (SSRIs) ranged from US$1 (Italy) to US$14 (US) for outpatients, and from US$6 (UK) to US$33 (US) for inpatients.
Expected cost per symptom-free day-outpatient summary
Incremental Cost-effectiveness Ratios (ICERs)
Table 10 presents the incremental cost-effectiveness ratios (ICERs) calculated in Italy and Poland since these were the only treatment settings in which venlafaxine IR was not the lowest treatment cost alternative. Within the outpatient setting in Italy, the TCAs resulted in a lower expected cost despite a higher success rate vs. venlafaxine IR by a difference in cost and success rate of $38 (US) and 5.1%, respectively, As a result, venlafaxine showed a highly favorable ICER measured in terms of the incremental cost per success and the incremental cost per SFD, $1278 and $410, respectively (Table 10). In Poland, greater differences in expected costs between venlafaxine and its comparators resulted in less favorable ICERs, with incremental cost per success and cost per SFD of $2448 and $14, respectively, relative to the TCAs in the outpatient setting. In the inpatient setting, these went as high as $3061 and $20, respectively, relative to the SSRIs (Table 10).
Table 10. Incremental cost-effectiveness ratios (ICERs) for venlafaxine in Italy and Poland
Our analysis revealed that in all settings venlafaxine maintained the highest expected success rate and greatest number of SFDs. Additionally, venlafaxine yielded the lowest expected cost of treating MDD patients in 8 of the 10 health care settings for outpatients and 9 of the 10 for inpatients. Therefore, we evaluated the local budgetary impact of a shift in reimbursement from the payers' current portfolio of antidepressant spending, to one with a greater percentage of spending allocated to venlafaxine.
The weighted average expected cost per patient (WEC, in Analytic Approach, section 2.1) for treating MDD ranged from US$632 (Poland) to US$5647 (US). Table 11 illustrates these results, as well as the total estimated direct cost for MDD treatment and the estimated cost to the primary payer in each country (CDC and CDCP, respectively, in Analytic Approach, section 2.2). To facilitate interpretation of the results, the savings (or loss) is based on the population assumption of one million covered lives (i.e., patients eligible for reimbursement for mental health care services) and a conservative treated prevalence of 0.3% for each country.
Comparison in the estimated cost savings or (loss) with a population assuming a treated population of 1 million with 100% coverage for each 1% of venlafaxine market share. MDD, major depressive disorder.
Inpatient—weighted average cost per patient
Outpatient—weighted average cost per patient
MDD treatment —weighted average expected cost per patient
The results of the univariate and multivariate sensitivity analyses indicate that our patient-level findings are robust with respect to assumptions implicit to the model. For all countries, the results of the break-even analyses illustrated in Tables 12 and 13 showed that venlafaxine IR is equally cost-effective vs. SSRIs and TCAs, even with modest to substantial variations in venlafaxine's success rate and drug price. For example, in Table 12 the break-even analysis for daily cost of treatment within the inpatient setting in Germany suggests that venlafaxine, with a baseline daily treatment cost of $2.54 can increase by almost four-fold to $9.23 and $10.52, respectively, before reaching equivalency in expected cost with the SSRIs and TCAs. Alternatively, Table 13 shows that venlafaxine's baseline efficacy rate of 62.3% in the inpatient setting could drop as much as 53.9% and 52.3%, respectively, to reach cost-effectiveness profiles equivalent to those of SSRIs and TCAs. Similar conclusions can be drawn from results in the other countries.
Table 12. Break-even analysis: daily cost of drug treatment
Although not consistent with the results of our pharmacoeconomic analysis, the Monte Carlo simulation revealed with levels of certainty ranging from 82% to 99% that venlafaxine IR was a significantly lower cost alternative in 9 out of the 10 countries vs. the SSRIs and TCAs in both the inpatient and outpatient settings. Table 14 presents the differences in expected costs of the Monte Carlo simulation expressed in mean differences with upper and lower limits and standard deviations for the distribution curve. The difference in expected cost between venlafaxine and SSRIs in the inpatient setting ranged from $59.92 in Venezuela to $1404 in the Netherlands. Within the outpatient setting, the difference in expected costs between venlafaxine and SSRIs ranged from $73.68 in Venezuela to $488.07 in Germany. Poland was the only country in which the expected costs of venlafaxine were higher than the SSRIs and TCAs resulting in a negative mean value.
Table 14. Monte Carlo analysis: differences in expected cost
Table 15 depicts the results of the Monte Carlo simulation expressed in terms of differences in mean Symptom-Free Days between venlafaxine vs. the SSRIs or TCAs. Consistent with the results of the pharmacoeconomic analysis, venlafaxine was a more effective treatment in terms of SFDs in comparison with either SSRIs or TCAs in both treatment settings for all 10 countries. In the inpatient setting, the greatest mean difference in SFDs was noted in Venezuela, where treatment with venlafaxine resulted in 10.08 more SFDs vs. the TCAs while the lowest difference was seen in UK, where venlafaxine resulted in only 3.39 additional SFDs over the SSRIs.
Table 15. Monte Carlo analysis: differences in symptom-free day
This study confirms results of previous pharmacoeconomic analyses in its conclusion that the use of venlafaxine is a cost-effective strategy for the treatment of MDD for both inpatients and outpatients across diverse health care systems. In addition, initiating treatment with venlafaxine provides the lowest expected cost when all downstream costs are considered in the analysis of alternative antidepressant agents. The superior pharmacoeconomic profile of venlafaxine is attributable to its high success rate and low ADR dropout rate as reported in Table 1 of our analysis.
Treatment success rates and drop-out rates due to either lack of efficacy or adverse drug reactions were the greatest drivers of the decision analytic model in all countries, and therefore had a significant impact on the expected costs of treatment. This observation explains why improvements in treatment success with venlafaxine resulted in lower expected costs in most countries. Since venlafaxine showed a dominant cost-effectiveness profile in terms of expected costs per success over the SSRIs and TCAs in most countries, an incremental cost-effectiveness analysis was not appropriate.
Significant variations were observed among countries in both the magnitude of the difference between the comparators and the confidence of the ranking. These differences were due partly to variations in clinical practice patterns and partly to differences in the cost of health care services and drug prices. For example, in the Netherlands, due to the less restrictive policies regarding the length of inpatient stay for MDD patients, the costs of the failure arms in the model were larger compared to success arms in most of the other countries. As a consequence venlafaxine showed a greater advantage in the Netherlands, in terms of cost-avoidance due to its higher expected success rate. The same effect is evident in Venezuela where the cost of an inpatient stay is high relative to the cost of pharmaceuticals. In Poland, the situation is reversed, since the cost of health care services relative to the cost of antidepressants is small and therefore the avoided cost of failure branches in the model are modest. As a result, the savings in health care resources consumed due to the higher success rate did not offset the higher cost of venlafaxine relative to the other comparators.
The results of this study attempt to reflect the implications of drug selection for the typical MDD patient in the acute phase. However, the significant individual patient variations and differences in practice patterns may compromise the external validity (i.e., generalizability) of the evaluation. Additionally, the deterministic model design and secondary data employed in this evaluation introduce potential information bias to the data analysis. In an effort to mitigate this bias, comprehensive sensitivity analysis was conducted.
The validity of the pharmacoeconomic model was demonstrated through rigorous sensitivity analyses and expert review. Practicing clinicians and local health economists from each country collaborated in the analytic approach and reviewed the methodology against local criteria for health care decision-making. Our cost-effectiveness analysis extracted the efficacy and dropout rates from two published meta-analyses [20, 21]. A statistical analysis demonstrated with 95% confidence that the three comparators were statistically different. Furthermore, both meta-analyses employed strict inclusion/exclusion criteria for selecting MDD patient samples, types of treatment comparators and the measurement of treatment success. All these clinical trial variables are critical elements in reducing the potential for bias and promoting the homogeneity of the input data to justify treatment success in the typical MDD patient. Despite the implementation of rigorous selection criteria, the authors of the meta-analyses did concede that the potential for bias might exist since clinical trials assessing newer agents usually report higher rates of efficacy than the more established treatments. Ironically, one reason may be due to the more homogeneous patient sample selection in these studies assessing newer antidepressant treatments. It is apparent that our research represents the first effort to study the economics of antidepressants on such a large scale. We are therefore confident in the generalizability of the results.
It is important to note that the conclusions of pharmacoeconomic research often encourage the selection of drugs that provide additional benefit in terms of improved health outcome notwithstanding an incremental investment for that improvement (i.e., greater expected cost). However, managing MDD patients with venlafaxine provides less overall financial expenditure in addition to improved health outcome. Viewing health care as an investment, venlafaxine represents excellent value for money. From a policy perspective, configuring payers' antidepressant portfolios to include a greater share of venlafaxine will likely reduce total health care expenditures, as well as improve population health. For each additional 1% of venlafaxine market share, average savings for the 10 countries are estimated at $9,892, based on a population of 1 million, ranging from $1,600 (Italy) to $29,049 (US). Additionally, the utilization of venlafaxine is expected to provide each treated patient with a greater reduction in depressive symptoms compared to other treatments.
When considering the various cost-containment strategies in health care systems across the world, from fixed budgets for doctors and pharmaceutical expenditures to rationing products and services, an optimal strategy to limit the growth of health care costs should focus on the system-wide costs. Optimal health economic efficiency can only be realized when a collective, rather than segmented (e.g., drug acquisition) budget is considered that accounts for total treatment costs and how those treatment costs change based on the uncertainty of health care intervention. In this way, treatment consequences, and not just “up-front” expenditures will figure into the equation of health care budgeting. The results of this international study of system-wide health care costs suggest that increased utilization of venlafaxine for patients with Major Depressive Disorder in most health care settings will favorably impact payer budgets as well as improve patient mental health outcomes.
This study was supported by a grant from Wyeth-Ayerst Inc., St. David's, Pennsylvania. The authors would like to acknowledge the invaluable contribution of the MDD International Study Group: Dr. Antonio Addis, Dr. Richard Allison, Dr. Enrique Baca, Prof. Luc Balant, Dr. Mariano Bassi, Prof. Cesario Bellantuono, Dr. Gilles Bertschy, Dr. Tomas Bocheneck, Prof. Juan M. Cabases, Prof. Dr. J. de Jonghe, Prof. Michael Drummond, Michael J. Einarson, Dr. John Feighner, Dr. Richard Frank, Prof. Hugh Freeman, Dr. F.J. Friesleder, Dr. Alfonso Galán Pajuelo, Dr. Leona Hakkaart-van Roijen, Dr. Chris Hawley, Prof. Alberto Holly, Prof. Edith Holsboer, Prof. Jan Horodnicki, Dr. George Kassianos, Prof. Andrzej Kiejna, Dr. Amanda Kirby, Dr. Charlotte Koczkás, Dr. Sture Lorentzon, Dr. Jose Manchon, Dr. Vittorio Mapelli, Dr. Salvaldor Mata, Dr. G. Matzel, Dr. Royce A. Morrison, Dr. D. Naber, Prof. Rafael Nahmens, Dr. W.A. Nolen, Dr Antonio Pacheco Herndudes, Dr. Paster Oropeza, Dr. Maj-Liz Persson, Mr. Ulf Persson, Prof. Frederico Piccoli, Prof. Staislaw Puzynski, Prof. Joan Rovira, Prof. Dr. Frans Rutten, Prof. Jolanta Rabe-Jablonski, Dr. Raul Rodon, Dr. Salvador Ros Montalban, Prof. Dr. Rychlik, Prof. Janusz Rybakowski, Dr. Hans. J. Salize, Prof. Pierluigi Scappichio, Dr. Manuel Serrano, Dr. David Sheehan, Dr. Camilo Silva Madriz, Dr. David Stasior, Dr. Rosana Tarricone, Mr. H.C.P. Venema, Dr. W.M.A. Verhoeven.