The Business Case for Expanded Clozapine Utilization
Abstract
Objective:
In most settings, less than 25% of patients with treatment-resistant schizophrenia receive clozapine, the only medication proven effective for treatment-resistant schizophrenia. Therefore, a business case analysis was conducted to assess whether increasing clozapine utilization for treatment-resistant schizophrenia in a health care system would result in direct health care cost savings.
Methods:
Veterans with treatment-resistant schizophrenia who were treated in the Veterans Health Administration (VHA) were studied. Treatment response, suicides, adverse drug reactions (and associated mortality), and effects on inpatient hospitalization related to clozapine were derived from a systematic review of published studies. A one-factor sensitivity analysis and a probabilistic sensitivity analysis (PSA) with Monte Carlo simulation were conducted to calculate the cost-benefits of increased clozapine utilization.
Results:
Despite monitoring costs, in the base case analysis, the VHA would save $22,444 per veteran with treatment-resistant schizophrenia over the first year of clozapine therapy, primarily from 18.6 fewer inpatient days per patient. If current utilization was doubled, and 50% of those veterans continued clozapine treatment for one year, VHA would save an estimated $80 million. Cost savings were most sensitive to the proportion of treatment-resistant patients who received clozapine, decrease in inpatient days, cost of inpatient stays, clozapine response rate, and number of patients with treatment-resistant schizophrenia. In the PSA, initiation of clozapine for all VHA patients with treatment-resistant schizophrenia who were not currently treated with clozapine would save at least $290 million in 95% of simulations.
Conclusions:
Increased clozapine utilization would result in net cost savings for the VHA.
Clozapine is the only treatment that has been proven effective for treatment-resistant schizophrenia (1–11) and the only treatment approved by the U.S. Food and Drug Administration (FDA) to decrease suicidal behavior associated with schizophrenia (7). However, clozapine is associated with significant adverse drug reactions (ADRs), requiring compliance with an FDA-mandated risk mitigation and evaluation strategy program (7). Given the complexity of the program—which requires patient, prescriber, and pharmacy registration; weekly blood draws; and seven-day fills for the first six months of therapy—clozapine is prescribed for only a minority of patients with treatment-resistant schizophrenia (12–17).
Despite evidence demonstrating clozapine’s benefits, use of clozapine in the United States has steadily declined since the introduction of other second-generation antipsychotics to the U.S. market (12,13). Within the Veterans Health Administration (VHA), 4% of patients with schizophrenia receive clozapine (13). Low clozapine utilization is also reported in non-VHA and international treatment settings (11,14–16). Considering that 20% to 30% of patients have treatment-resistant schizophrenia, the low rate of utilization of clozapine implies that 82% to 88% of patients with treatment-resistant schizophrenia are receiving less effective antipsychotics.
Given clozapine’s unique effectiveness for treatment-resistant schizophrenia and its potential to decrease suicidal behaviors and utilization of more costly forms of care, increasing clozapine utilization could lead to significant cost savings, despite increased monitoring and contact with the health care system. Therefore, we conducted a cost-benefit analysis with data obtained from a medical literature review to simulate potential cost savings associated with increasing clozapine utilization within the VHA. The hypothesis was that potential cost savings, mostly from decreased inpatient hospital days, would constitute a case for expanding clozapine utilization, even if achieving such a goal entailed considerable costs and effort.
Methods
Model Design
We developed a simulation on the basis of data from the medical literature to estimate costs associated with varying degrees of clozapine utilization. The decision model is a tree structure comparing the choice of whether or not to use clozapine for a treatment-resistant patient at a single decision node over a one-year time horizon, calculated from the perspective of the VHA (18). In the clozapine arm, subsequent branch points were event nodes that represent likelihood of response, risk of completed suicide, and risk of serious ADRs resulting in clozapine discontinuation (Figure 1). In the clozapine arm, nonresponse was assumed to lead to clozapine discontinuation. We modeled only serious ADRs (agranulocytosis, seizures, diabetic ketoacidosis, myocarditis, and ileus) that have been attributed to clozapine use (7). [The decision tree for ADRs is available as an online supplement to this article.]
FIGURE 1. Outcomes of a decision model involving whether to initiate clozapine for a patient with treatment-resistant schizophreniaa
aADR, adverse drug reaction
In the nonclozapine arm, suicide risk was modeled at a single branch point, and we assumed that there would be no ADRs, even though other medications cause ADRs. This is a simplifying assumption, but a conservative model was chosen to decrease the likelihood of overestimating the benefits of clozapine treatment. Costs of inpatient and outpatient psychiatric care, laboratory monitoring, and health care expenditures related to ADRs were incorporated into the clozapine arm. Differential mortality risks from completed suicides between the clozapine and nonclozapine arms and aggregate risks of mortality from clozapine-related ADRs were calculated.
This study was exempted from review by the Edith Nourse Rogers Memorial Veterans Hospital’s Institutional Review Board.
Model Inputs
Response to clozapine.
We conducted a systematic literature search of MEDLINE articles indexed between January 1, 1985, and June 30, 2015, for studies using the terms “clozapine” AND “refractory schizophrenia” OR “treatment resistant” OR “resistant schizophrenia.” Thirty-nine articles were evaluated for number of persons given clozapine and clinical response rate (4,19–56). If these values could not be ascertained from the text or if the subjects were under 18 years of age, the article was excluded. Response was most commonly defined as a 20% improvement in baseline score on the Brief Psychiatric Rating Scale, but inclusion of an article was not conditional on this definition. Of the 39 identified articles, 31 articles comprising 2,571 clozapine patients were included (4,19–45,47–49) [see online supplement]. We calculated a pooled clozapine response rate of 51%.
Suicide rate.
Completed suicide rate was calculated as the product of the annual suicide rate for the U.S. general population ages 35–64, the standardized mortality ratio from suicide for schizophrenia, and a published hazard ratio of suicide for male veterans with schizophrenia (57–60). The effect of clozapine on suicide risk was determined on the basis of a meta-analysis of five studies that calculated the relative risk of completed suicide associated with use of clozapine. The relative risk of completed suicide was 2.90 among patients who did not use clozapine compared with those who used clozapine (61). For this model, reduction in the risk of suicide was applied only among clozapine responders; nonresponders were assumed to have a similar risk of suicide as patients receiving nonclozapine antipsychotics.
Clozapine-related ADRs.
Pooled risks and mortality of the clozapine-related ADRs were calculated from studies identified through the literature review and the clozapine-prescribing information (7,62–64). A pooled risk of clozapine-related seizures and subsequent risk of discontinuation were calculated from an analysis of the clozapine patient management system database (64).
Costs of treatment.
Costs for the first year of clozapine treatment were determined to be the cost of laboratory monitoring and follow-up visits plus the VHA cost of clozapine tablets. On the basis of current guidelines for monitoring, there are a total of 39 visits over the course of one year of treatment (weekly for six months and biweekly for the next six months) (7). We used the pricing schedule used by the Centers for Medicare and Medicaid Services for the 2015 CPT code for a complete blood cell count with differential cell count and for the evaluation and management code for a level II visit with a psychiatrist (65). Patients were assumed to be reimbursed $20 for travel costs by VHA for each follow-up visit. Clozapine costs in the base case were for a total daily dose of 600 mg, based on 2015 VHA pricing of 41 cents for a 100-mg tablet.
Costs of ADRs.
The estimated cost of the ADRs was the sum of annual direct health care costs (prescriptions, inpatient treatment, emergency services, outpatient treatment, and office-based medical visits). Adults with relevant three-digit ICD-9-CM codes and Clinical Classification Software codes were identified through examination of the Medical Expenditure Panel Survey from 2002 to 2011 for condition-related annual health care expenditures (66). For diabetic ketoacidosis, costs were for a single inpatient hospitalization rather than yearly costs (67).
Inpatient psychiatric stays and costs.
The length of inpatient stay was calculated from a pooled, random-effects meta-analysis of seven studies identified in the systematic review (33,34,36,38,39,68,69) [see online supplement]. These studies evaluated the impact of clozapine therapy on hospital days, using either a prepost comparison or a comparison between hospital days for intervention and control patients. The cost of an inpatient stay at a VA hospital was estimated at $1,414 per day (13).
Treatment-eligible population.
In 2009, a total of 87,000 veterans had schizophrenia or schizoaffective disorder (13). Although 20% to 30% of patients are reported to have treatment-resistant schizophrenia, we assumed that 20% of patients, an estimated 18,000 veterans, had treatment-resistant schizophrenia (1). Assuming that the 3.6% of veterans with schizophrenia already treated with clozapine had treatment-resistant schizophrenia, we estimated an additional 14,400 veterans with treatment-resistant schizophrenia would be eligible for a clozapine trial.
Cost Reporting
For each parameter, costs that were not already in 2015 U.S. dollars were inflated to 2015 U.S. dollars by calculating the rate of medical inflation from the Consumer Price Index (CPI) between the year in which the parameter was reported and 2014 (last available year of the CPI) and then adjusting again by the average yearly rate of medical inflation from 2010 to 2014 (70).
Analysis
This analysis was a cost-benefit model, comparing costs and monetizing benefits associated with clozapine use and nonuse in treatment-resistant schizophrenia. The deterministic base case model used the parameter estimates for the model inputs detailed above to determine the difference in costs for treatment-resistant schizophrenia between clozapine treatment and nonuse of clozapine. In the clozapine arm, the costs of treatment, ADRs, and inpatient stays were summed. In the nonclozapine arm, it was assumed that ADRs did not occur and medications had no costs, so the costs reflected inpatient stays. Mortality was estimated by the effect of clozapine on suicide among veterans and anticipated mortality from serious ADRs (61–63). The business case for clozapine was broadened by multiplying the results over the treatment-eligible population within the VHA.
Several assumptions were made for model variability. Costs of expenditures directly attributable to clozapine use (prescriptions, laboratory monitoring, office visits, and treatment of serious ADRs) were not included in the control arms. Health care expenditures unrelated to clozapine effects were assumed to be identical in the two arms to prevent needlessly complicating the model; these expenditures included metabolic monitoring, which is recommended for all patients on antipsychotics. Benefits of reduced inpatient stay and diminished incidence of completed suicide were applied only for clozapine responders. Costs of clozapine treatment were incurred for the entire year for clozapine responders and were prorated until day of discontinuation for nonresponders. All costs for persons who died by suicide were curtailed at the day of suicide. ADR costs were prorated from the day on which the ADR occurred until the end of the year, with the exception of diabetic ketoacidosis, which was assumed to be a single event. In the base case deterministic model, suicide, clozapine discontinuation due to nonresponse, and ADRs occurred at six months, the midpoint of the year (day 183). Among clozapine responders, the ADRs were assumed to lead to discontinuation, except seizures, in which case we used published rates of discontinuation of clozapine following a seizure to estimate the proportion of patients who would continue clozapine (64). For clozapine nonresponders, all ADRs were assumed to result in drug discontinuation.
To incorporate uncertainty, one-factor sensitivity analyses and a probabilistic sensitivity analysis (PSA) with Monte Carlo simulation were calculated (71). In the factor analysis, we varied the range of input parameters by 20% in each direction from the base case, except the proportion of treatment-resistant schizophrenia patients initiated on clozapine, which were varied from 20% to 80%. We then evaluated the relative influence of high and low values for each input on cost savings for the health care system. For the PSA, method of moments were used to fit probability parameters on beta distributions, relative risk and hazard ratios on log normal distributions, and event days (ADRs, clozapine discontinuation, and suicide). Fixed costs (lab monitoring, veteran travel, and psychiatrist visits) were varied on uniform distributions. Gamma distributions were used to fit aggregate ADR costs, and Poisson distributions were used to fit inpatient days (Table 1). Random numbers generated output values within distributional assumptions. Simulated cost outcomes were determined for the entire health care system, with 1,000 replications each for initiation of clozapine among 20%, 40%, 60%, and 80% of patients with treatment-resistant schizophrenia. The last of these proportions would equate to 100% clozapine use because it would be added to the current rate of use of clozapine for treatment-resistant schizophrenia, which is approximately 20%. Simulations and analysis were performed by using Excel 2010 and Stata, version 13, software packages.
| Variable | Base case | SD | Distribution | Notes |
|---|---|---|---|---|
| Veterans with treatment-resistant schizophrenia | 18,000 | Uniform | 2009 VHA estimate of ∼90,000 patients with schizophrenia × 20% rate of treatment-resistant schizophrenia (13) | |
| Proportion initiated clozapine treatment | .20 | Uniform | Assumption | |
| Probability of clozapine response | .510 | .01 | β | Pooled meta-analysis of 31 studies (4,19–45,47–49) [see online supplement] |
| Probability of completed suicideb | .003 | β | U.S. suicide rate for ages 35–64 × SMR for suicides among persons with schizophrenia × HR for suicides among male patients with schizophrenia (57–60) | |
| Completed suicide rate given clozapine response | .34 | .8 | Log normal | Prevention of suicides with clozapine (61)c |
| Probability of ADR (clozapine arm only) | ||||
| Agranulocytosis | .005 | .0002 | β | Calculated from Cohen et al., 2012 (63) |
| Myocarditis | .0001 | .000 | β | Calculated from Cohen et al., 2012 (63) |
| Ileus | .004 | .0004 | β | Calculated from Cohen et al., 2012 (63) |
| Seizures | .032 | .003 | β | Calculated from Cohen et al., 2012 (63) |
| Diabetic ketoacidosis | .0001 | .000 | β | Calculated from Cohen et al., 2012 (63) |
| Probability of discontinuation of clozapine due to seizures | .352 | .046 | β | Calculated from Pacia and Devinsky (64) |
| Probability of mortality due to ADR | .28 | .28 | β | Calculated from Cohen et al., 2012 (63) |
| Myocarditis | ||||
| Agranulocytosis | .03 | .007 | β | Calculated from Cohen et al., 2012 (63) |
| Ileus | .20 | .036 | β | Calculated from Cohen et al., 2012 (63) |
| Diabetic ketoacidosis | .03 | .04 | β | Calculated from Cohen et al., 2012 (63) |
| Fixed costs ($)d | ||||
| Lab monitoring (event) | 10.58 | Uniform | WBC; CPT G0306 (CMS 2015 Lab Diagnostic Fee Schedule, global, nonfacility) (64) | |
| Psychiatrist visit | 43.98 | Uniform | CPT 99212 (CMS 2015 Physician Fee Schedule, global, nonfacility) (64) | |
| Veteran travel (event) | 20.00 | Uniform | Assumption | |
| Clozapine treatment (day) | 2.46 | Uniform | 100-mg tablet, 600 mg/day (VHA VISN 1 costs) | |
| Inpatient stay (day) | 1,413.70 | Uniform | 2009 VHA estimate (2015 $) | |
| Cost for ADR (2015 $) | ||||
| Seizures (year) | 1,621.62 | 245.75 | γ | MEPS 2002–2011 for adults (ICD-9 code 345) (66) |
| Myocarditis (year) | 3,695.89 | 934.70 | γ | MEPS 2002–2011 for adults (ICD-9 code 422) (66) |
| Agranulocytosis (year) | 2,428.31 | 1,641.52 | γ | MEPS 2002–2011 for adults (ICD-9 code 288) (66) |
| Ileus (year) | 6,796.33 | 1,234.79 | γ | MEPS 2002–2011 for adults >18 (ICD-9 code 560) (66) |
| Diabetic ketoacidosis (single hospitalization) | 20,141.34 | 20,415.43 | γ | (67) |
| Day of suicide | 183 | Uniform | Assumption | |
| Day of discontinuation of clozapine | 183 | Uniform | Assumption | |
| Day of ADR (clozapine arm) | 183 | Uniform | Assumption | |
| Inpatient days | 138 | Poisson | Pooled analysis of 7 studies (33,34,36,38,39,68,69) [see online supplement] | |
| Reduction in inpatient days among clozapine responders | –37e | Poisson | Unstandardized mean difference in inpatient days of 7 studies (33,34,36,38,39,68,69) [see online supplement] |
TABLE 1. Adjustment of base case values during sensitivity analyses of the cost-benefits of increased clozapine utilization among veterans with treatment resistant schizophreniaa
Results
Cost Savings
If 20% of veterans with treatment-resistant schizophrenia initiated clozapine, the one-year costs to the VHA would decrease by a mean of $22,444 per patient treated with clozapine. Savings were driven primarily through an average reduction of 18.6 inpatient hospital days. If current utilization was doubled and 50% of those veterans continued clozapine treatment for one year, the VHA would accrue an estimated cost savings of $80 million (Table 2).
| Outcome | Base case | Simulated outcome (percentile) | ||
|---|---|---|---|---|
| Mean | 5th | 95th | ||
| Aggregate cost and utilization | ||||
| Cost ($) | ||||
| Control arm | 700,410,000 | 695,541,000 | 633,082,000 | 757,851,000 |
| Clozapine arm | 619,613,000 | 615,697,000 | 560,652,000 | 671,316,000 |
| Cost savings ($) | 80,797,000 | 79,845,000 | 72,406,000 | 87,685,000 |
| Inpatient days | ||||
| Control arm | 495,000 | 492,000 | 447,920 | 536,000 |
| Clozapine arm | 429,000 | 426,000 | 388,000 | 464,000 |
| Inpatient days saved | 67,000 | 66,000 | 60,000 | 72,000 |
| Clozapine responders | 1,787 | 1,794 | 1,629 | 1,964 |
| Adverse drug reaction (ADR)b | ||||
| Seizures | 116 | 115 | 94 | 137 |
| Agranulocytosis | 16 | 16 | 10 | 23 |
| Myocarditis | <1 | <1 | 0 | 1 |
| Ileus | 16 | 16 | 9 | 23 |
| Diabetic ketoacidosis | <1 | <1 | 0 | 2 |
| Mortality | ||||
| Suicides (control arm) | 11 | 11 | 6 | 17 |
| Suicides (clozapine arm) | 8 | 8 | 3 | 12 |
| Deaths due to ADRs | ||||
| Agranulocytosis | <1 | <1 | 0 | 2 |
| Myocarditis | <1 | <1 | 0 | 1 |
| Ileus | 3 | 3 | 1 | 6 |
| Diabetic ketoacidosis | <1 | <1 | 0 | 1 |
TABLE 2. Simulated outcomes from doubling baseline rates of clozapine utilization at the VHA for one yeara
ADRs
An additional 743 serious ADRs would occur in year 1 if all veterans with treatment-resistant schizophrenia who had previously not been treated with clozapine initiated clozapine treatment. If only 20% of those veterans began treatment with clozapine, an additional 149 serious ADRs would occur (Table 2).
Deaths
If all veterans with treatment-resistant schizophrenia who had not been treated with clozapine later initiated clozapine, 19 suicides would be averted annually, and there would be a total of 18 additional deaths, three due to clozapine-related agranulocytosis and 15 due to ileus. If only 20% of veterans with treatment-resistant schizophrenia who had not previously initiated clozapine began treatment with clozapine, there would be three deaths due to clozapine-related ADRs and three fewer suicides in the clozapine group versus the nonclozapine group. Greater increases in clozapine use would lead to more lives saved by suicide prevention than lives lost to clozapine-associated ADRs (Figure 2).
FIGURE 2. Estimated one-year VHA-wide savings and mortality associated with incremental gains in clozapine utilization among veterans with treatment-resistant schizophreniaa
aBars indicate mean, 5th, and 95th percentiles from 1,000 simulations for each incremental increase of 20%.
Sensitivity Analysis
One-way sensitivity analysis was performed, varying each of the model inputs across a plausible range to evaluate the input’s influence on our findings. Cost savings were most sensitive to changes in the proportion of patients with treatment-resistant schizophrenia who received clozapine (92% of the swing in range of costs), reduction in inpatient days (2%), cost of inpatient stay (2%), rate of response to clozapine (2%), and number of patients with treatment-resistant schizophrenia (2%) (Figure 3). Costs associated with laboratory monitoring, ADRs, outpatient visits, and travel had negligible impact on costs for the payer. Probabilistic sensitivity analysis confirmed that increasing the number of patients with treatment-resistant schizophrenia receiving clozapine would result in incremental cost savings. Assuming a baseline rate of 20% for clozapine utilization, cost savings increased to $323,188,000 when 100% of patients with treatment-resistant schizophrenia were prescribed clozapine.
FIGURE 3. Effects of varying the model inputs on the cost of clozapine carea
aInput parameters were varied from the base case by 20% in each direction, except proportion of patients with treatment-resistant schizophrenia who were treated with clozapine, which varied from 20% to 80%. Reduced inpatient days were prorated for the clozapine response rate probability. Rates of adverse drug reactions (ADRs), suicide, clozapine discontinuation, costs of ADRs, and laboratory monitoring had smaller impacts.
Figure 2 shows the cost implications of initiating clozapine for 20% to 80% of patients with treatment-resistant schizophrenia (representing 100% utilization at beginning of year 1, including current users). A relatively modest change, doubling utilization from 20% to 40% of eligible patients, would lead to cost savings of more than $80 million in the first year of treatment. In 95% of simulations, increasing clozapine utilization from 20% to 40% of eligible patients saved a minimum of $72 million in the first year. Initiation of clozapine for all VHA patients with treatment-resistant schizophrenia who were not currently treated with clozapine would save at least $290 million in 95% of simulations.
Discussion
Clozapine has repeatedly demonstrated superiority to other antipsychotics for treatment-resistant schizophrenia, but it is underutilized. Therefore, we simulated the effects of increasing clozapine utilization for patients with treatment-resistant schizophrenia. Doubling the proportion of patients with treatment-resistant schizophrenia who received clozapine from 20% to 40% would save $80 million for the VA, without considering any value accruing to patients from improved outcomes. In addition, despite the much-feared adverse effects of clozapine, the increase in ADR-related mortality associated with clozapine would be more than offset by suicide prevention.
The benefits of increased clozapine utilization may have been underestimated because of several assumptions. First the model assumed that 20% of patients with schizophrenia have treatment-resistant schizophrenia, although many studies have estimated a rate as high as 30% (1). Second, the costs of clozapine treatment are substantially higher in the first year compared with later years because of a need for more office visits and laboratory monitoring. After the first year, the continued costs of clozapine treatment are more similar to those of other antipsychotics, but the benefits of clozapine treatment continue to accrue. Third, this model assumed that treatment with other antipsychotics did not result in serious ADRs or include prescription or other treatment costs. Thus it is possible savings could be higher than reported in this study.
Our analyses had several important uncertainties. First, it was difficult to estimate the incidence and costs of obesity and diabetes risk and the expected increased risk in cardiovascular mortality and morbidity associated with use of clozapine. Some comparator agents have similar risks, whereas most have substantially lower risks than clozapine (72). In addition, studies have reported mixed effects of clozapine on mortality (73–75). As such, cardiometabolic ADRs were not included in the analyses. Second, it should be remembered that extrapolation of cost savings to subsequent years and to use outside the VHA is uncertain. Third, some of the studies providing evidence for clozapine benefits were conducted years ago and thus may be less relevant to current practice. Finally, the base model assumed a reduction of 30 inpatient days in the first year. Although that may seem extreme in an era of shortened hospitalizations, the average length of stay for schizophrenia both within and outside the VHA in 2009 and 2011 was approximately 21 days, and patients with treatment-resistant schizophrenia averaged three hospitalizations per year (13,76). In addition, studies indicate that more severe illness, as would be expected with treatment-resistant schizophrenia, further increases the length of stay (76). Thus it is reasonable to assume that clozapine would be associated with 30 days of inpatient hospitalization. Varying the analyses by limiting the number of days of hospitalization avoided to seven still resulted in cost savings [see online supplement].
This study had important strengths. A number of underlying assumptions suggest that, if anything, benefits of increased clozapine utilization have been underestimated. In addition, much is known about the improved outcomes, adherence, and persistence rates associated with clozapine treatment (75,76). However, the study also had some limitations. First, all the data were not from the VHA or specifically related to treatment-resistant schizophrenia. For example, the suicide rate was not based on rates for schizophrenia populations with treatment-resistant disease, given the lack of data on suicide among patients with treatment-resistant schizophrenia. Second, rates of response, ADRs, and costs associated with use of clozapine were based on the published literature and may not accurately reflect real-world systems. Third, dispensing costs were not included, given that we found no estimates that included the added cost of maintenance and oversight of the VHA clozapine registry.
Clozapine monitoring guidelines have recently changed. It is unlikely that these changes will have a significant impact on cost of monitoring, but the possibility cannot be excluded. In addition, monitoring of clozapine serum concentrations is recommended by some guidelines. Because clozapine serum concentrations are not routinely monitored, we excluded these costs from our analyses. If completed, such monitoring would be expected to increase clozapine costs.
It is important to note the circumscribed nature of the study. This study was not a cost-utility analysis and, therefore, did not formally model patient-related outcomes, such as health-related quality of life. Instead, we opted to conduct a simpler business case analysis from the VHA perspective, including direct health care costs and veterans’ travel costs but excluding indirect costs, such as caregiver time and missed work. It should be noted that the value to society from increased clozapine utilization for treatment-resistant schizophrenia would greatly exceed the cost savings we reported here because patients and, presumably, their families prefer that the symptoms of schizophrenia be well controlled (77). However, payers are likely most responsive to a business case analysis, which is why we chose this approach.
In a separate study, we found that some relatively straightforward strategies were associated with higher clozapine utilization within the VHA (78). Among these are having a dedicated clozapine clinic; ensuring that the clinic has sufficient capacity to accommodate all patients with treatment-resistant schizophrenia and adequate staffing levels by nonphysicians, including registered nurses and pharmacists; and providing transportation to appointments for patients using clozapine. Although these strategies are allowable under VHA rules and regulations and are currently used at some sites, they are not without cost.
It is hoped that this study provides an impetus to make these relatively straightforward, if not inexpensive, changes. Extrapolating from our findings, an average-sized VHA site (managing 700 schizophrenia patients) could expect to have 140 patients with treatment-resistant schizophrenia. Increasing the proportion of patients with treatment-resistant schizophrenia who are treated with clozapine from 20% to 40% would save the facility more than $625,000 in the first year alone. It is difficult to imagine that the cost of implementing our recommendations could approach this figure. Thus, even a modest increase in clozapine utilization, achieved at a very high cost, could still result in a net cost savings for the VHA.
Conclusions
This study found that modest increases in clozapine utilization could yield significant cost savings. It is very unusual to have the opportunity to save money while improving patient outcomes, particularly in such a vulnerable population. Our results suggest that the VHA should strongly consider initiatives, possibly based on practices already used by high-performing sites, to substantially increase clozapine utilization for treatment-resistant schizophrenia.
1 : World Federation of Societies of Biological Psychiatry (WFSBP) Guidelines for Biological Treatment of Schizophrenia: part 1. update 2012 on the acute treatment of schizophrenia and the management of treatment resistance. World Journal of Biological Psychiatry 13:318–378, 2012Crossref, Medline, Google Scholar
2 : The costs of schizophrenia. Journal of Clinical Psychiatry 68(suppl 14):4–7, 2007Crossref, Medline, Google Scholar
3 Villa K, Sclar D, Schmidt J, et al: Cost of treatment-resistant schizophrenia: an assessment of state Medicaid programs. Presented at the U.S. Psychiatric and Mental Health Congress, Orlando, Fla, Sept 20–23, 2014. http://www.psychcongress.com/posters/costs-treatment-resistant-schizophrenia-assessment-state-medicaid-programsGoogle Scholar
4 : Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Archives of General Psychiatry 45:789–796, 1988Crossref, Medline, Google Scholar
5 : Randomized controlled trial of effect of prescription of clozapine versus other second-generation antipsychotic drugs in resistant schizophrenia. Schizophrenia Bulletin 32:715–723, 2006Crossref, Medline, Google Scholar
6 : Practice guideline for the treatment of patients with schizophrenia, 2nd ed. American Journal of Psychiatry 161(suppl 2):1–56, 2004Link, Google Scholar
7 Clozaril [package insert]. East Hanover, NJ, Novartis Pharmaceuticals Corp, Dec 2014Google Scholar
8 : Clozapine treatment for suicidality in schizophrenia: International Suicide Prevention Trial (InterSePT). Archives of General Psychiatry 60:82–91, 2003Crossref, Medline, Google Scholar
9 : Overt aggression and psychotic symptoms in patients with schizophrenia treated with clozapine, olanzapine, risperidone, or haloperidol. Journal of Clinical Psychopharmacology 24:225–228, 2004Crossref, Medline, Google Scholar
10 : Effectiveness of clozapine versus olanzapine, quetiapine, and risperidone in patients with chronic schizophrenia who did not respond to prior atypical antipsychotic treatment. American Journal of Psychiatry 163:600–610, 2006Link, Google Scholar
11 : Underprescribing of clozapine and unexplained variation in use across hospitals and regions in the Canadian province of Québec. Clinical Schizophrenia Related Psychoses 7:33–41, 2013Crossref, Medline, Google Scholar
12 : Current status of clozapine in the United States. Shanghai Archives of Psychiatry 24:110–113, 2012Medline, Google Scholar
13 Care for Veterans With Psychosis in the Veterans Health Administration, FY09, 11th Annual National Psychosis Registry Report. Ann Arbor, Mich, VA HSR&D Serious Mental Illness Treatment Resource and Evaluation Center, 2009Google Scholar
14 : Comparison of clozapine use in Maryland and in Victoria, Australia. Psychiatric Services 56:320–323, 2005Link, Google Scholar
15 : Geographic and clinical variation in clozapine use in the United States. Psychiatric Services 65:186–192, 2014Link, Google Scholar
16 : Clozapine utilization and outcomes by race in a public mental health system: 1994–2000. Journal of Clinical Psychiatry 67:1404–1411, 2006Crossref, Medline, Google Scholar
17 : Antipsychotic prescribing pathways, polypharmacy, and clozapine use in treatment of schizophrenia. Psychiatric Services 64:527–533, 2013Link, Google Scholar
18 : Primer on medical decision analysis: part 2. building a tree. Medical Decision Making 17:126–135, 1997Crossref, Medline, Google Scholar
19 : Treatment outcome with clozapine in tardive dyskinesia, neuroleptic sensitivity, and treatment-resistant psychosis. Journal of Clinical Psychiatry 48:263–267, 1987Medline, Google Scholar
20 : Clozapine for refractory schizophrenia: an open study of 14 patients treated up to 2 years. Journal of Clinical Psychiatry 50:389–391, 1989Medline, Google Scholar
21 : Treatment with clozapine and its effect on plasma homovanillic acid and norepinephrine concentrations in schizophrenia. Psychiatry Research 46:151–163, 1993Crossref, Medline, Google Scholar
22 : The use of clozapine in treatment-refractory schizophrenia. Schizophrenia Research 10:29–32, 1993Crossref, Medline, Google Scholar
23 : Clozapine treatment of outpatients with schizophrenia: outcome and long-term response patterns. Hospital and Community Psychiatry 44:1145–1149, 1993Abstract, Google Scholar
24 : Clozapine therapy and clinical outcomes in Kentucky psychiatric hospitals. Southern Medical Journal 87:1092–1096, 1994Crossref, Medline, Google Scholar
25 : Clinical effects of clozapine in chronic schizophrenia: response to treatment and predictors of outcome. American Journal of Psychiatry 151:1744–1752, 1994Link, Google Scholar
26 : Clozapine response in treatment-refractory first-episode schizophrenia. Biological Psychiatry 35:278–280, 1994Crossref, Medline, Google Scholar
27 : A clinical study of clozapine treatment and predictors of response in a Canadian sample. Canadian Journal of Psychiatry 40:208–211, 1995Crossref, Medline, Google Scholar
28 : Long-term impact of clozapine and psychosocial treatment on psychiatric symptoms and cognitive functioning. Psychiatric Services 47:41–45, 1996Link, Google Scholar
29 : The effects of clozapine on symptom reduction, neurocognitive function, and clinical management in treatment-refractory state hospital schizophrenic inpatients. Neuropsychopharmacology 15:361–369, 1996Crossref, Medline, Google Scholar
30 : Time required for initial improvement during clozapine treatment of refractory schizophrenia. American Journal of Psychiatry 153:951–952, 1996Link, Google Scholar
31 : Neuroleptic-resistant schizophrenic patients treated by clozapine: clinical evolution, plasma and red blood cell clozapine and desmethylclozapine levels. Thérapie 52:227–232, 1997Medline, Google Scholar
32 : Time to clozapine response in a standardized trial. American Journal of Psychiatry 154:1243–1247, 1997Link, Google Scholar
33 : Cost-effectiveness of clozapine: a UK clinic-based study. British Journal of Psychiatry 171:125–130, 1997Crossref, Medline, Google Scholar
34 : A comparison of clozapine and haloperidol in hospitalized patients with refractory schizophrenia. New England Journal of Medicine 337:809–815, 1997Crossref, Medline, Google Scholar
35 : Health status and health care costs for publicly funded patients with schizophrenia started on clozapine. Psychiatric Services 49:1590–1593, 1998Link, Google Scholar
36 : Initiating clozapine treatment in the outpatient clinic: service utilization and cost trends. Psychiatric Services 49:1034–1038, 1998Link, Google Scholar
37 : Clozapine therapy for older veterans. Psychiatric Services 49:340–344, 1998Link, Google Scholar
38 : Clozapine for refractory schizophrenia: the Illinois experience. Journal of Clinical Psychiatry 60(suppl 1):18–22, 1999Medline, Google Scholar
39 : Health care costs of therapy-refractory schizophrenic patients treated with clozapine: a study in a community psychiatric service in Italy. Acta Psychiatrica Scandinavica 99:274–280, 1999Crossref, Medline, Google Scholar
40 : How long to wait for a response to clozapine: a comparison of time course of response to clozapine and conventional antipsychotic medication in refractory schizophrenia. Schizophrenia Bulletin 25:709–719, 1999Crossref, Medline, Google Scholar
41 : Double-blind study of clozapine dose response in chronic schizophrenia. American Journal of Psychiatry 156:1744–1750, 1999Abstract, Google Scholar
42 : Long-term therapeutic drug monitoring of clozapine and metabolites in psychiatric in- and outpatients. Psychopharmacology 152:80–86, 2000Crossref, Medline, Google Scholar
43 : An assessment of clinical practice of clozapine therapy for veterans. Psychiatric Services 51:669–671, 2000Link, Google Scholar
44 : Risperidone versus clozapine in treatment-resistant schizophrenia: a randomized pilot study. Progress in Neuropsychopharmacology and Biological Psychiatry 24:911–922, 2000Crossref, Medline, Google Scholar
45 : Clozapine and haloperidol in moderately refractory schizophrenia: a 6-month randomized and double-blind comparison. Archives of General Psychiatry 58:965–972, 2001Crossref, Medline, Google Scholar
46 : Clinical predictors of response to clozapine treatment in ambulatory patients with schizophrenia. Journal of Clinical Psychiatry 63:420–424, 2002Crossref, Medline, Google Scholar
47 : Efficacy of clozapine in the treatment of atypical antipsychotic refractory schizophrenia: a pilot study. Journal of Clinical Psychopharmacology 23:103–104, 2003Crossref, Medline, Google Scholar
48 : Clozapine alone versus clozapine and risperidone with refractory schizophrenia. New England Journal of Medicine 354:472–482, 2006Crossref, Medline, Google Scholar
49 : Clinical predictors of therapeutic response to clozapine in a sample of Turkish patients with treatment-resistant schizophrenia. Progress in Neuropsychopharmacology and Biological Psychiatry 31:1330–1336, 2007Crossref, Medline, Google Scholar
50 : Serotonin function and treatment response to clozapine in schizophrenic patients. American Journal of Psychiatry 150:1337–1342, 1993Link, Google Scholar
51 : Cost-effectiveness of clozapine treatment in therapy-refractory schizophrenia. Acta Psychiatrica Scandinavica 92:199–201, 1995Crossref, Medline, Google Scholar
52 : Plasma clozapine levels and clinical response for treatment-refractory schizophrenic patients. American Journal of Psychiatry 152:179–182, 1995Link, Google Scholar
53 : Response of patients with treatment-refractory schizophrenia to clozapine within three serum level ranges. American Journal of Psychiatry 153:1579–1584, 1996Link, Google Scholar
54 : Clinical and neurocognitive effects of clozapine and risperidone in treatment-refractory schizophrenic patients: a prospective study. Journal of Clinical Psychiatry 59:521–527, 1998Crossref, Medline, Google Scholar
55 : Impact of clozapine prescription on inpatient resource utilization. Journal of Nervous and Mental Disease 189:766–773, 2001Crossref, Medline, Google Scholar
56 : Ziprasidone vs clozapine in schizophrenia patients refractory to multiple antipsychotic treatments: the MOZART study. Schizophrenia Research 113:112–121, 2009Crossref, Medline, Google Scholar
57 Suicide among adults aged 35–64 years—United States, 1999–2010. Morbidity and Mortality Weekly Report 62:321–325, 2013Medline, Google Scholar
58 : Psychiatric diagnoses and risk of suicide in veterans. Archives of General Psychiatry 67:1152–1158, 2010Crossref, Medline, Google Scholar
59 : Reducing clozapine-related morbidity and mortality: 5 years of experience with the Clozaril National Registry. Journal of Clinical Psychiatry 59(suppl 3):3–7, 1998Crossref, Medline, Google Scholar
60 : Suicide as an outcome for mental disorders: a meta-analysis. British Journal of Psychiatry 170:205–228, 1997Crossref, Medline, Google Scholar
61 : Suicidal risk during treatment with clozapine: a meta-analysis. Schizophrenia Research 73:139–145, 2005Crossref, Medline, Google Scholar
62 : Rare and very rare adverse effects of clozapine. Neuropsychiatric Disease and Treatment 11:1995–2003, 2015Crossref, Medline, Google Scholar
63 : Beyond white blood cell monitoring: screening in the initial phase of clozapine therapy. Journal of Clinical Psychiatry 73:1307–1312, 2012Crossref, Medline, Google Scholar
64 : Clozapine-related seizures: experience with 5,629 patients. Neurology 44:2247–2249, 1994Crossref, Medline, Google Scholar
65 Medicare Provider Utilization and Payment Data. Baltimore, Centers for Medicare and Medicaid Services, 2015. https://www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/medicare-provider-charge-data/Google Scholar
66 Medical Expenditure Panel Survey. Rockville, Md, Agency for Healthcare Research and Quality, 2015. http://meps.ahrq.gov/Google Scholar
67 : Economic impact of diabetic ketoacidosis in a multiethnic indigent population: analysis of costs based on the precipitating cause. Diabetes Care 26:1265–1269, 2003Crossref, Medline, Google Scholar
68 : Cost-effectiveness of clozapine therapy for severe psychosis. Psychiatric Services 49:829–831, 1998Link, Google Scholar
69 : Outcome following clozapine discontinuation: a retrospective analysis. Journal of Clinical Psychiatry 68:1027–1030, 2007Crossref, Medline, Google Scholar
70 Consumer Price Index. Washington, DC, United States Department of Labor, Bureau of Labor Statistics, Dec 2014. http://www.bls.gov/cpi/Google Scholar
71 : Probabilistic sensitivity analysis using Monte Carlo simulation: a practical approach. Medical Decision Making 5:157–177, 1985Crossref, Medline, Google Scholar
72 : Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. New England Journal of Medicine 353:1209–1223, 2005Crossref, Medline, Google Scholar
73 : 11-year follow-up of mortality in patients with schizophrenia: a population-based cohort study (FIN11 study). Lancet 374:620–627, 2009Crossref, Medline, Google Scholar
74 : Serious adverse drug events reported to the Food and Drug Administration, 1998–2005. Archives of Internal Medicine 167:1752–1759, 2007Crossref, Medline, Google Scholar
75 : Factors associated with length of psychiatric hospitalization. Comprehensive Psychiatry 55:681–687, 2014Crossref, Medline, Google Scholar
76 : Comparative effectiveness of clozapine and standard antipsychotic treatment in adults with schizophrenia. American Journal of Psychiatry 173:166–173, 2016Link, Google Scholar
77 : Side effects and treatment with clozapine: a comparison between the views of consumers and their clinicians. International Journal of Mental Health Nursing 17:2–8, 2008Crossref, Medline, Google Scholar
78 : Organizational characteristics of Veterans Health Administration clinics with high and low utilization of clozapine. Psychiatric Services (Epub ahead of print, June 15, 2016)Medline, Google Scholar
