Economic and clinical comparison of atypical depot antipsychotic drugs for treatment of chronic schizophrenia in the Czech Republic

Abstract

Purpose:

The Czech Republic is faced with making choices between pharmaceutical products, including depot injectable antipsychotics. A pharmacoeconomic analysis was conducted to determine the cost-effectiveness of atypical depots.

Methods:

An existing 1-year decision-analytic framework was adapted to model drug use in this healthcare system. The average direct costs to the General Insurance Company of the Czech Republic of using paliperidone palmitate (Xeplion®), risperidone (Risperdal Consta®), and olanzapine pamoate (Zypadhera®) were determined. Literature-derived clinical rates populated the model, with costs adjusted to 2012 Euros using the consumer price index. Outcomes included quality-adjusted life-years (QALYs), days in remission, and proportions hospitalized or visiting emergency rooms. One-way sensitivity analyses were calculated for all important inputs. A multivariate probability analysis was used to examine the stability of results using 10,000 iterations of simulated input over reasonable ranges of all included variables.

Results:

Expected average costs/per patient treated were €5377 for PP-LAI, €6118 for RIS-LAI, and €6537 for OLZ-LAI. Respective QALYs were 0.817, 0.809, and 0.811; ER visits were 0.127, 0.134, and 0.141; hospitalizations were 0.252, 0.298, and 0.289. Results were generally robust in sensitivity analyses. PP-LAI dominated RIS-LAI and OLZ-LAI in 90.2% and 92.1% of simulations, respectively. Results were insensitive to drug prices but sensitive to adherence and hospitalization rates.

Conclusions:

PP-LAI dominated the other two drugs, as it had a lower overall cost and superior clinical outcomes, making it the preferred choice. Using PP-LAI in place of RIS-LAI for chronic relapsing schizophrenia would reduce the overall costs of care for the healthcare system.

Keywords::

• Czech Republic
• Depot
• Antipsychotic
• Cost-effectiveness
• Paliperidone

Introduction

The Czech Republic has an estimated population of just over 10 million1. The country has adopted a compulsory national health insurance model with essentially universal coverage2,3. The total healthcare budget in 2007 was €8.5 billion, of which 51.2% was for hospital care, 24.0% for ambulatory care, 18.6% for drugs, and 6.2% other2. The majority of these costs (∼86%) are managed by the General Insurance Company of the Czech Republic2.

Pharmaceuticals are included as benefits and are managed by the State Institute for Drug Control (SÚKL). When setting allowable prices for new drugs, they use international price comparisons; however, they do not officially require the use of pharmacoeconomic guidelines2. They do require a wide range of evidence be presented, including expected costs to the healthcare system.

A search of Medline and Embase found a single full peer reviewed pharmacoeconomic analysis addressing schizophrenia in this country, which was published more than a decade ago4. Hosák and Bahbouh4 compared costs and outcomes between 67 patients treated with oral risperidone and 67 with traditional antipsychotics with data derived from chart reviews. Clinical outcomes were similar between groups, but hospitalization costs were 40% lower in the risperidone group. The increased acquisition cost of risperidone was not offset by this reduction, which was attributed to the very low labor costs in the healthcare field at that time. Two other pharmacoeconomic analyses from the Czech Republic have appeared in abstract form5,6. Both studies examined risperidone long-acting injectable (RIS-LAI) in a mirror image model and found that it was cost-effective compared to the previous treatment for schizophrenia and schizoaffective disorder.

In all three of these studies4–6, the main cost driver was hospitalization, which comprised 60–70% of the total costs of care. A major causative factor in rehospitalization among persons with chronic schizophrenia is poor adherence to prescribed drug regimens7,8. This problem has been identified in the Czech Republic, and is more problematic than in persons with somatic disorders9,10. Even partial adherence is associated with large increases in hospitalizations and costs10. Therefore, if adherence can be improved, patients can have better management of their disease and thereby avoid relapses. As well, their quality-of-life (QoL) would be improved11.

To that end, depot medications have been developed. Because their clinical effect is prolonged, they have the potential to preclude both intentional and unintentional non-adherence by the patient12. In a meta-analysis of data from 1700 patients, Leucht et al.13 found that depot antipsychotics were superior to oral drugs in reducing hospitalizations due to relapse. Some of the older depot drugs required weekly or biweekly injections14; however, improved formulations have extended their effect to the point that some of them can be administered monthly. Such is the case with paliperidone palmitate15 (PP-LAI) and olanzapine pamoate16 (OLZ-LAI).

The economics of these newer drugs in the Czech Republic healthcare system is currently unknown. We therefore undertook this research to determine the cost-effectiveness of the newer atypical depot antipsychotics from the viewpoint of the General Insurance Company of the Czech Republic.

Methods

A decision tree model with a 1-year time horizon that was previously developed for used in another country was adapted for the Czech healthcare system. Drugs of interest included paliperidone palmitate (Xeplion®), risperidone (Risperdal Consta®), and olanzapine pamoate (Zypadhera®). Figure 1 depicts the model and clinical events. Patients began with stable disease and were managed as described in Table 115,17–35. Data for that table were extracted from the published literature and represent standard care. Local variations were implemented based on the input from experts. Patients would move along the branches of the decision tree until they arrived at a decision node, which occurred every 3 months. They would either remain stable or deteriorate according to rates found in the literature (Table 2)17,23,32,36–39. Their symptoms could relapse, prompting a visit to the emergency room (ER) where they would be treated. Those who could be managed as outpatients would be treated and discharged with follow-up in the community. The others would be admitted to hospital and managed as inpatients. Their course of illness would continue until they reached the 1 year time horizon.

Figure 1. Decision tree used for the economic analysis.

Table 1. Clinical management of schizophrenia and sources of information.

Drug	Treatment	Regimen	Data source
PP-LAI	Treatment initiation for relapse requiring hospitalization	150 mg week 1, 100 mg week 2, then 82.8 mg every 4 weeks maintenance	Xeplion® product summary15, Hough et al.17
	Relapse not requiring hospitalization	84.9 mg every 4 weeks	Average of Gopal et al.18, Pandina et al.19, Hough et al.20, Nasrallah et al.21, Pandina et al.22
	Maintenance	69.3 mg every 4 weeks	Gopal et al.23, Fleischhacker et al.24
		Adjunct: 6.8 mg oral risperidone for 30.5 days	Gopal et al.23
RIS-LAI	Treatment initiation for relapse requiring hospitalization	50 mg every 2 weeks	Maintenance dose pro-rated as per PP-LAI (equals the maximum dose in product summary)25
	Relapse not requiring hospitalization	50 mg every 2 weeks	Maintenance dose pro-rated as per PP-LAI (also equals the maximum dose in Risperdal Consta® product summary)25
	Maintenance	40.3 mg every 2 weeks	Average of Fleischhacker et al.24, Kissling et al.26, Lee et al.27, Lindenmayer et al.28, Olivares et al.29
		Adjunct: oral 3.2 mg/day for 43 days	Möller et al.30
OLZ-LAI	Treatment initiation for relapse requiring hospitalization	300 mg every 2 weeks ×3, then 473 mg every 4 weeks	Zypadhera® product summary16, Lauriello et al.31
	Relapse not requiring hospitalization	473 mg every 4 weeks	Lauriello et al.31
	Maintenance	432 mg every 4 weeks	Kane et al.32
		Adjunct: oral 10.8 mg/day for 31 days	Ascher-Svanum et al.33
clozapine	Two failed LAIs	500 mg daily	Average dose in 9 trials reported by Chakos et al.34
	Incomplete response	750 mg daily	Average of dose range in European Medicines Agency recommendations35

LAI, long acting antipsychotic; OLZ-LAI, olanzapine (Zypadhera®); PP, paliperidone palmitate (Xeplion®), RIS, risperidone (Risperdal Consta®).

Table 2. Clinical rates used as inputs for calculations in the model.

Drug	Rate	Value input (probability)	Source
PP-LAI	Stable disease	0.803	Calculation: 1 – (Relapses + hospitalizations)
	Relapse not requiring hospitalization	0.059	Calculation: Average hospitalization rate from Gopal et al.23 and Hough et al.17× relapse:hospitalization ratio from Ascher-Svanum et al.36
	Relapse requiring hospitalization	0.138	Gopal et al.23, Hough et al.17
	Re-hospitalization	0.108	Nicholl et al.37
	Adherence	0.872	Olivares et al.38 adjusted via Mehnert and Diels39
RIS-LAI	Stable disease	0.763	Calculation: 1 – (Relapses + hospitalizations)
	Relapse not requiring hospitalization	0.071	Olivares et al.38
	Relapse requiring hospitalization	0.166	Olivares et al.38
	Re-hospitalization	0.131	Nicholl et al.37
	Adherence	0.823	Olivares et al.38
OLZ-LAI	Stable disease	0.793	Kane et al.32
	Relapse not requiring hospitalization	0.062	Calculation: 1 – (Stable + hospitalizations)
	Relapse requiring hospitalization	0.145	Kane et al.32
	Re-hospitalization	0.114	Nicholl et al.37
	Adherence	0.803	Ascher-Svanum et al.36

Patients who had relapsed because they had discontinued their primary drug for reasons other than intolerance or ineffectiveness would be restarted on the same drug using recommended dosing regimens (Table 2). Those who stopped due to intolerance or lack of efficacy would be switched to an alternative drug. Those failing PP-LAI or RIS-LAI would be switched to OLZ-LAI, and those who failed OLZ-LAI would be switched to RIS-LAI. After two failures, patients would receive clozapine.

The outcomes incorporated into the analyses were those related to patient care such as number of days with stable disease, number of relapses requiring treatment at the ER, number of hospitalizations, and quality-adjusted life years. The utility weights were those previously calculated for these patients, based on data from the literature (0.890 for stable disease)40–44, 0.659 for a relapse treated in the ER40–42, and 0.490 for hospitalization43,44.

Since the perspective was that of the General Insurance Company, the analysis incorporated all direct costs of care such as drug costs, medical care, and hospital care. Unit prices for cost inputs are listed in Table 345–48. Each year, the company issues a list of drugs and services that will be reimbursed for the following year. If generic drugs are available and are considered acceptable, they are put on the list and only that price is allowed. Otherwise, the brand name drug is used. Such is the case in this analysis; olanzapine and risperidone tablets are both available as generics and those prices have been used. We used a simple average cost per mg across all available strengths of each drug. Indirect costs were not considered, nor were costs accrued by other agencies such as the justice system, police, or welfare.

Table 3. Units costs for resources.

Resource	Item	Unit	Cost/unit	Source
Drugs	Clozapine tablets	mg	€0.005/mg	SÚKL45
	Olanzapine tablets	mg	€0.148/mg	SÚKL45
	Olanzapine LAI	mg	€0.945/mg	SÚKL45
	Paliperidone LAI	mg	€4.440/mg	SÚKL45
	Risperidone tablets	mg	€0.125/mg	SÚKL45
	Risperidone LAI	mg	€4.977/mg	SÚKL45
Hospital care	Emergency room	1 visit	€50.33	UZIS46,47
	Psychiatric ICU	1 day	€167.50	Expert opinion
	Acute psychiatric ward	1 day	€43.35	UZIS46,47
	Medical ward	1 day	€39.41	UZIS46,47
	ICU	1 day	€788.25	UZIS46,47
Medical care	Psychiatrist (outpatient)	1 visit	€14.27	Act 411/201148
	Psychiatric nurse	per month	€301.74	Act 411/201148
	Staff nurse	1 visit	€6.15	Act 411/201148
	Psychologist	1 visit	€10.35	Act 411/201148

ICU, intensive care unit; LAI, long acting injectable.

The primary pharmacoeconomic outcome was the incremental cost-effectiveness ratio based on the additional cost that would be required to pay for one additional quality-adjusted life year (QALY). Model robustness was tested using a series of one-way sensitivity analyses on all major cost and clinical inputs. As well, a multivariate probability-based analysis was conducted for all model inputs. Clinical values and costs were varied over plausible ranges using 10,000 iterations in a Monte Carlo synthesis.

Results

Table 4 presents the results from the pharmacoeconomic analyses. PP-LAI had the lowest overall expected cost per patient treated. At the same time, it was associated with the greatest number of QALYs and days with stable disease as well as the lowest numbers of ER visits and hospitalizations for relapses. Thus, PP-LAI dominated the other alternatives in the base case analysis.

Table 4. Pharmacoeconomic results.

Type	Outcome	PP-LAI	RIS-LAI	OLZ-LAI
Clinical	Stable days	329.5	324.7	324.9
	ER visits*	0.127	0.134	0.141
	Hospitalizations*	0.252	0.298	0.289
	QALYs	0.817	0.809	0.811
Economic	Cost	€5377	€6118	€6537

LAI, long acting antipsychotic; OLZ-LAI, olanzapine (Zypadhera®); PP, paliperidone palmitate (Xeplion®), QALY, quality-adjusted life year; RIS, risperidone (Risperdal Consta®).

*Proportions of patients experiencing the outcome.

In the one-way sensitivity analyses, PP-LAI was insensitive to alterations in acquisition costs, since it would require a 37% increase in the price of PP-LAI to lose its dominance over OLZ-LAI and 24% for RIS-LAI. On the other hand, dominance would be lost if the price of OLZ-LAI were to decrease by 41% or RIS-LAI by 21%. In addition, modest changes in adherence rates or hospitalization rates could result in a loss of dominance.

In the 10,000 iteration probability sensitivity analyses, PP-LAI dominated OLZ-LAI in 92.1% and RIS-LAI in 90.2%, and was itself dominated 1.3% of the time. Figures 2a and b illustrate those findings. Using the cost-effectiveness threshold of ∼€30,000 suggested by NICE49 (∼750,000 Kč), PP-LAI was cost-effective in 98.8% of simulations when compared with RIS-LAI and in 93.8% when compared with OLZ-LAI. Thus, we conclude that the results were quite stable over usual variations that would be encountered in the majority of situations.

Figure 2. Incremental cost-utility ratio plotting cost in euros on the y-axis against quality adjusted life-years on the x-axis of PP-LAI vs (a) OLZ-LAI, and (b) RIS-LAI—results from a probability sensitivity analysis with 10,000 iterations.

Discussion

Since PP-LAI is the drug with the lowest cost per patient treated and the highest number of QALYs, it dominates the other two atypical antipsychotic depots. Sensitivity analyses generally confirmed the robustness of those results. Unless dramatic changes were made, the likelihood that PP-LAI would not be the cost-effective alternative is low. Therefore, it should be considered the drug of choice in many situations. The lower costs appear to be associated with the higher efficacy of PP-LAI, which decreases hospitalizations and ER visits. Similar findings in the Czech Republic have been reported previously. In a review of the use of clozapine, Zvolský and Hulínský50 concluded that the drug’s efficacy reduced the number of hospital re-admissions and length of stay, which offset the added costs of monitoring.

QoL is the standard outcome measured in pharmacoeconomic evaluations as it captures many aspects of the patient’s being. Sidlova et al.51 found that patients with chronic schizophrenia have much lower quality-of-life than do persons not suffering from the disease. Therefore, it is an especially important outcome to measure in these people and drugs that can improve QoL are welcome additions as therapeutic choices.

Among the limitations in this study, not all costs were considered in the calculations. They did not include the costs of adverse events, despite the fact that they can be quite common with antipsychotic drugs. A previous analysis in the Czech Republic reported that adverse events accounted for only ∼1.2% of the total cost of care52. Other government agencies have made similar statements53,54. Therefore, it was assumed that results would not be affected to any great extent due to their omission.

In addition, indirect costs, such as loss of work days, were not included in the analysis. One reason is that few of these people regularly participate in the work force. A multi-country study in Europe indicated that <10% from a sample of almost 4000 persons with chronic schizophrenia worked full-time and ∼12% worked part-time55. Therefore, the overall cost burden may have been a little under-estimated; however, the difference between drugs would have been minimal. Similarly, violence does occur with some patients56, which would also add to the under-estimation, but not bias.

It must be kept in mind that a model such as this represents the average patient from the defined population treated with average doses for an average length of time. There will be much variation in real life situations, as has been described in the literature. There has especially been so with Risperdal Consta®, which has few blinded trials57. However, the literature must be reviewed with caution, as there are issues of different forms of bias such as selection bias, channeling bias, and sponsorship bias, among others.

Conclusions

Among the atypical depot antipsychtoics examined in this analysis, paliperidone palmitate (Xeplion®) emerged as the drug having the best pharmacoeconomic profile. It had the lowest cost per patient treated, which would result in cost savings for the system. At the same time, it had the best clinical outcomes, which would result in better patient care and improved quality-of-life. If feasible, paliperidone palmitate should possibly be considered the first choice when selecting from among these alternatives for persons with chronic relapsing schizophrenia in the Czech Republic.

Transparency

Declaration of funding

This study was funded by Janssen Cilag.

Declaration of financial/other interest

TTE and RZ have disclosed they are consultants to Janssen Cilag and received funding from Janssen to help conduct this study and prepare it for publication. JS has no relevent financial relationships to disclose. SV, MG and MEHH have disclosed that they are employees of Janssen Cilag. JME Peer reviewers on this manuscript have no relevant financial relationships to disclose.