Healthcare Resource Utilization 6 Months Before and After Olanzapine/Samidorphan Initiation: Real-World Assessment of Patients with Schizophrenia or Bipolar I Disorder

Abstract

Introduction

A combination of olanzapine and samidorphan (OLZ/SAM), approved by the US Food and Drug Administration (2021) for the treatment of adults with schizophrenia (SZ) or bipolar I disorder (BD-I), mitigates weight gain associated with olanzapine treatment. This study examined changes in healthcare resource utilization (HCRU) among patients with SZ or BD-I after initiating OLZ/SAM.

Methods

This retrospective analysis utilized USA-based administrative claims data from April 19, 2021 to December 31, 2022. Adults aged ≥ 18 years with SZ or BD-I who had continuous enrollment ≥ 6 months before (baseline) and after (follow-up) OLZ/SAM initiation (index) were eligible. HCRU comparisons included inpatient admissions, average number of inpatient days, emergency department (ED) visits, and outpatient visits between 6-month baseline and follow-up periods.

Results

In total, 1546 patients initiated OLZ/SAM (SZ: n = 855, mean age 39.4 years, 52.6% male; BD-I: n = 691, 38.4 years, 31.0% male). Overall, 446 (52.2%) patients with SZ and 302 (43.7%) patients with BD-I remained on OLZ/SAM during the full 6-month follow-up (median days persistent 182 and 106, respectively). In the SZ and BD-I cohorts, proportions of patients with all-cause and mental health (MH)-related inpatient admissions and ED visits significantly decreased between baseline and follow-up (all-cause: all P ≤ 0.026; MH-related: all P < 0.001). Findings were similar for disease-specific outcomes with the exception that reductions in BD-I-related ED visits were not significant. Average number of inpatient days decreased after OLZ/SAM initiation in both cohorts; in patients with BD-I, those reductions were statistically significant (all P ≤ 0.022). The proportions of patients with outpatient visits were similar between baseline and follow-up for both cohorts.

Conclusions

Results from this first real-world study assessing HCRU after OLZ/SAM initiation suggest that OLZ/SAM results in reductions in burden for patients, providers, and the healthcare system, as evidenced by reductions in acute all-cause, MH-related, and disease-specific HCRU among SZ and BD-I cohorts.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12325-025-03211-w.

Plain Language Summary

Olanzapine is an effective medication for people with schizophrenia (SZ) or bipolar I disorder (BD-I), but it causes significant weight gain for many who take it. Bothersome side effects often lead to patients stopping their antipsychotic medications, especially when significant weight gain occurs, and people with SZ or BD-I who do not take their medication regularly are likely to relapse and need hospitalization. A medication that combines olanzapine and samidorphan (OLZ/SAM) is as effective as olanzapine for treating symptoms of SZ and BD-I but causes less weight gain. In this study, we examined insurance records to see if people with either SZ or BD-I would need more or less healthcare services after they started taking OLZ/SAM. We found that fewer people with either SZ or BD-I were hospitalized or had an emergency room visit after they began taking OLZ/SAM compared with before taking OLZ/SAM. Also, people in the BD-I group were hospitalized for fewer days after starting OLZ/SAM than before taking it. The proportions of people with SZ or BD-I who had regular outpatient doctor visits were similar before and after starting the new medication. Our results suggest that taking OLZ/SAM may reduce the burden for people with SZ or BD-I, people providing their care, and the overall healthcare system based on the reductions we saw in their acute healthcare use.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12325-025-03211-w.

Key Summary Points

Why carry out this study?

In clinical studies, treatment with a combination of olanzapine and samidorphan (OLZ/SAM) has consistently mitigated olanzapine-associated weight gain while providing treatment efficacy comparable with that of olanzapine.

This real-world study, using a pre–post analysis of administrative claims data, sought to determine if the initiation of OLZ/SAM affected healthcare resource utilization in the real-world treatment of schizophrenia (SZ) and bipolar I disorder (BD-I).

What was learned from the study?

The proportions of patients with SZ or BD-I who had all-cause and mental health-related inpatient admissions and emergency department visits significantly decreased for patients treated with OLZ/SAM between the baseline and follow-up periods.

The proportions of patients with SZ or BD-I who had outpatient visits were similar 6 months before and 6 months after treatment initiation with OLZ/SAM.

These findings indicate that OLZ/SAM may result in clinically meaningful reductions in patient and provider burden, as evidenced by decreased inpatient admissions and emergency department visits.

Introduction

Serious mental health conditions, including schizophrenia (SZ) and bipolar I disorder (BD-I), are associated with considerable financial cost to the affected individuals, their families, and the healthcare system [1–5]. In the USA, the total economic burdens of SZ and BD-I have been estimated to be $343 billion (2019 US dollars) [5] and $219 billion (2018 US dollars) [1, 3], respectively, and direct healthcare costs are among the largest contributors [5]. Beyond healthcare costs, serious mental health conditions have significant effects on quality of life and burden of care for those patients and their families and caretakers [2, 6–10]. Costs related to caregiving, along with productivity loss due to unemployment and premature mortality, are substantial contributors to the indirect costs associated with patients with a SZ or BD-I diagnosis [3, 5]. The impact of serious mental health conditions on individuals’ lives is further evidenced by the increased risk of hospitalization and incarceration in those with SZ or a bipolar spectrum disorder [11–18].

Among patients with a serious mental health condition, adherence to antipsychotic medication is associated with reduced relapse [19] and hospitalization [13, 20, 21], but adherence is typically low among these patient populations [21, 22]. Adverse effects of atypical antipsychotics contribute to nonadherence and treatment discontinuation [23]; among these effects, weight gain or the expectation of weight gain is frequently cited as a barrier to adherence [23–25]. The atypical antipsychotic olanzapine is highly effective for reducing symptoms of SZ and BD-I in clinical trials [26–29], and treatment with olanzapine is associated with improved health-related quality of life [30–33] and reduced rates of relapse and rehospitalization [27, 34–36]. However, olanzapine is also associated with clinically significant weight gain and increased cardiometabolic risk [26, 27, 37–39], which limit its clinical utility [40–42].

Reducing weight gain and cardiometabolic effects associated with olanzapine could potentially improve adherence, resulting in an effective option for long-term use [43]. Toward that end, a combination product containing olanzapine and the opioid antagonist samidorphan (OLZ/SAM) was evaluated and approved by the US Food and Drug Administration in 2021 for the treatment of adults with SZ and adults with BD-I [44]. In clinical trials, the antipsychotic efficacies of OLZ/SAM and olanzapine were comparable [45–48]. However, treatment with OLZ/SAM resulted in significantly less weight gain compared with olanzapine in clinical studies of 12–24 weeks’ duration [45, 47, 48]. Small changes in weight, and minimal changes in metabolic and glucose parameters, were observed after exposures of up to 5 years of additional treatment in open-label extension studies [49–51]. Importantly, mitigating weight gain associated with olanzapine treatment could increase patients’ willingness to continue their medication [24, 43, 52] and reduce the risks associated with gaps in antipsychotic medication adherence, such as rehospitalization [53]. To date, no real-world studies have examined treatment patterns and healthcare resource utilization (HCRU) among patients with SZ or BD-I who initiate treatment with OLZ/SAM. The current study was designed to examine whether treatment with OLZ/SAM may be associated with reductions in HCRU in a real-world setting. The objective was to assess and compare HCRU 6 months before and after initiating OLZ/SAM in patients treated for SZ or BD-I.

Methods

Data Source

This retrospective study used USA-based administrative claims data from April 19, 2021 through December 31, 2022, obtained from Komodo’s Healthcare Map, the eligibility-controlled (closed-claims) comprehensive data set of Komodo Health, Inc. The Healthcare Map is a fully deidentified USA-based database containing detailed information from approximately 150 million patients covered by a commercial (66%), Medicaid (27%), or Medicare Advantage (7%) insurance plan. The population covered includes patients in all 50 US states and is representative of the geographic distribution of the US census population, based on ZIP-3 code distributions. Patient records contain data from inpatient, outpatient, and pharmacy claims, including submitted, paid, and rejected claims. All database records are fully compliant with US patient confidentiality requirements set forth in the Health Insurance Portability and Accountability Act (HIPAA) of 1996. As this study used only deidentified patient records and did not involve the collection, use, or transmittal of individually identifiable data, it was exempted from institutional review board approval [54]. All study data were obtained using International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) codes.

Study Design and Patient Selection

A schematic of the study design is shown in Fig. 1. The study included an identification period from October 18, 2021 (OLZ/SAM launch date) to June 30, 2022, during which the patients’ first pharmacy or medical claim for OLZ/SAM was identified. The date of the first observed OLZ/SAM claim was defined as the index date. The observation period included the 6 months before the index date (baseline period) and a 6-month period starting on the index date (follow-up period). Patient demographics, clinical characteristics, and medication use were assessed in the baseline period, treatment patterns were assessed during follow-up, and HCRU outcomes were assessed in the baseline and follow-up periods.

Fig. 1.

Study design ^aThe index date could occur at any time within the identification period. OLZ/SAM combination olanzapine and samidorphan

Inclusion Criteria

Patients identified for inclusion were adults (aged ≥ 18 years in the index year) with ≥ 1 medical claim for a primary or secondary ICD-10-CM diagnosis code of SZ (F20.XX) or BD-I (F31.1X, F31.2X, F31.6X, and/or F31.7X) during the study period (April 19, 2021–December 31, 2022) and ≥ 1 pharmacy or medical claim for OLZ/SAM during the identification period. Included patients had continuous enrollment with medical and pharmacy benefits ≥ 6 months before and ≥ 6 months after the index date. Included patients also had ≥ 1 medical claim with an ICD-10-CM diagnosis code for SZ or BD-I (primary or secondary) during the baseline or follow-up period; patients with medical claims for both SZ and BD-I were assigned an indication of SZ. Patients with pharmacy or medical claims for OLZ/SAM during the baseline period were excluded.

Outcomes

Demographic, clinical characteristic, and medication use data included index year (calendar year); age as of the index year and age group (18–24, 25–34, 35–44, 45–54, 55–64, or ≥ 65 years); sex (male or female); insurance type (commercial, Medicaid, Medicare Advantage, multiple, or none/unknown insurance coverage); and US Census Bureau geographic region in which the patient enrolled. The Quan-Charlson comorbidity score [55] was calculated and categorized (0, 1–2, 3–4, or ≥ 5) based on the presence of diagnosis codes on medical claims in the baseline period. General comorbid conditions were defined using Clinical Classifications Software (managed by the Agency for Healthcare Research and Quality) based on ICD-10-CM diagnoses. Patients with ≥ 1 specified behavioral health-related comorbid condition (targeted comorbidities) were identified from medical claim diagnosis codes.

Baseline medication use included immediate previous antipsychotic medication (categorized as first-generation long-acting injectable [LAI], second-generation LAI, first-generation oral antipsychotic, second-generation oral antipsychotic, or none); antipsychotic medication in the baseline period; other behavioral health medication (antidepressant, anticholinergic, anxiolytic, hypnotic or sedative, mood stabilizer, or opioid); and specified other medication (weight-mitigation agent, antiglycemic, antihypertensive, or statin).

Index medication dosage was the dosage on the index date (OLZ/SAM 5/10 mg, 10/10 mg, 15/10 mg, or 20/10 mg). Discontinuation from the index medication was defined as a gap in therapy of ≥ 45 days; the discontinuation date was defined as the last day of the supply of the last prescription filled prior to the first observed gap in therapy, without subsequent fills for the index therapy. Time to discontinuation was the number of days between the index date and discontinuation date. Treatment persistence was defined as the number of days from the index date to the discontinuation date (for patients who discontinued) or from the index date to the end of the follow-up period. The medication possession ratio (MPR) was calculated as the sum of the days’ supply of the index medication in the follow-up period divided by the number of days in the follow-up period. The proportion of days covered (PDC) was calculated as the number of days on which medication was available (based on filled prescriptions) divided by the number of days between the index date through the end of the follow-up period. Adherence to OLZ/SAM treatment was defined based on a MPR of ≥ 0.80 and PDC of ≥ 0.80 in the 6-month period after the index date.

HCRU outcomes assessed in the baseline and follow-up periods were proportion of patients with ≥ 1 inpatient admission and number of inpatient admissions, average number of inpatient days (calculated as the number of inpatient days observed, divided by the total number of patients initiated on treatment), proportion of patients with ≥ 1 emergency department (ED) visit and number of ED visits, and proportion of patients with ≥ 1 outpatient visit and number of outpatient visits (office visits, other visits, and overall number of visits). All-cause, mental health-related, and disease-specific (SZ or BD-I) outcomes were analyzed. ICD-10-CM diagnosis codes for mental, behavioral, and neurodevelopmental disorders (any position on the claim; Supplementary Table S1) were used to identify mental health-related HCRU in the following categories: anxiety, bipolar disorder, delirium, depression, eating disorders, executive functioning, mania, personality disorder, SZ, substance abuse, and other cognitive disorders. HCRU was identified as SZ or BD-I related if the claim had an ICD-10-CM diagnosis code for SZ (F20.XX, F21.XX, and/or F25.XX) or BD-I (F31.XX, F31.1X, F31.2, F31.6X, F31.7X, and/or F34.0), respectively. Diagnosis code lists for identifying disease-specific HCRU were expanded (versus those for patient identification specified in the inclusion criteria) to capture all relevant resource use (Supplementary Table S1).

Statistics

For categorical variables, numbers and percentages are presented; means, medians, and SDs were calculated for continuous variables. Differences between baseline and follow-up utilization measures were analyzed using unadjusted pairwise comparisons (paired t test for normally distributed continuous variables, Wilcoxon signed-rank test for non-normally distributed continuous variables, McNemar test for dichotomous variables). Because the study used a within-patient (pre versus post) design, each patient served as their own control with no adjustment for time-independent covariates. Analyses of HCRU measures were conducted, separately for the SZ and BD-I cohorts, for all patients who initiated index medication; subgroup analyses of HCRU were conducted for patients who did not discontinue OLZ/SAM through the entire 6-month follow-up period and for patients for whom the most recent antipsychotic prior to the OLZ/SAM index date was olanzapine.

Results

Patient Characteristics

A total of 1546 patients met the inclusion criteria for the analysis (Table 1). Of these patients, 855 were in the SZ cohort and 691 were in the BD-I cohort. Mean (SD) ages were 39.4 (13.6) years in the SZ cohort and 38.4 (12.0) years in the BD-I cohort; 52.6% and 31.0% of patients in the SZ and BD-I cohorts, respectively, were male (Table 2). The most common insurance type for the SZ and BD-I cohorts was Medicaid (64.2% and 51.2%, respectively). Comorbid conditions reported by ≥ 10% of patients during the baseline period are listed in Table 3. The three most commonly identified comorbidities were anxiety disorders (42.9%), hypertension (33.2%), and obesity (30.6%) in the SZ cohort and anxiety disorders (65.1%), major depressive disorder (36.8%), and any substance use disorder (32.3%) in the BD-I cohort. Mean (SD) Quan-Charlson comorbidity scores were 0.58 (1.12) and 0.48 (0.93) for the SZ and BD-I cohorts, respectively.

Table 1.

Patient identification and attrition

	Patients, n
Patient criteria
≥ 1 medical claim with an ICD-10-CM diagnosis code for SZ or BD-I (primary or secondary) during the study period	1,458,154
≥ 1 pharmacy or medical claim for OLZ/SAM during the identification period	2599
Aged ≥ 18 years in the index year	2556
Continuously enrolled ≥ 6 months before (baseline period) and ≥ 6 months after (follow-up period) index date	1720
≥ 1 medical claim with an ICD-10-CM diagnosis code for either SZ or BD-I (primary or secondary) in the baseline or follow-up period	1546
Disease cohortsa
SZ	855
BD-I	691

BD-I bipolar I disorder, ICD-10-CM International Classification of Diseases, 10th Revision, Clinical Modification, IP inpatient, OLZ/SAM combination olanzapine and samidorphan, SZ schizophrenia

^aPatients with medical claims for both SZ and BD-I were assigned an indication of SZ (n = 129). Patients with a pharmacy or medical claim for the index medication during the baseline period were excluded (n = 2)

Table 2.

Baseline patient demographics

Characteristics, n (%)	SZ cohort (n = 855)	BD-I cohort (n = 691)
Age, mean (SD), years	39.4 (13.6)	38.4 (12.0)
Age group, n (%)
18–24 years	121 (14.2)	93 (13.5)
25–34 years	240 (28.1)	189 (27.4)
35–44 years	207 (24.2)	197 (28.5)
45–54 years	142 (16.6)	133 (19.2)
55–64 years	108 (12.6)	67 (9.7)
≥ 65 years	37 (4.3)	12 (1.7)
Sex, n (%)
Female	405 (47.4)	477 (69.0)
Male	450 (52.6)	214 (31.0)
Region, n (%)
South	243 (28.4)	212 (30.7)
Midwest	240 (28.1)	224 (32.4)
West	215 (25.1)	124 (17.9)
Northeast	97 (11.3)	97 (14.0)
Othera	60 (7.0)	34 (4.9)
Insurance type, n (%)
Medicaid	549 (64.2)	354 (51.2)
Medicare Advantage	114 (13.3)	48 (6.9)
Commercial	85 (9.9)	213 (30.8)
Multiple	105 (12.3)	73 (10.6)
Unknown	2 (0.2)	3 (0.4)
Index year, n (%)
2021	200 (23.4)	169 (24.5)
2022	655 (76.6)	522 (75.5)

BD-I bipolar I disorder, SZ schizophrenia

^a“Other” includes Armed Forces Americas (except Canada), Armed Forces (Europe, Canada, Middle East, Africa), Armed Forces Pacific, American Samoa, Federated State of Micronesia, Guam, Marshall Islands, Commonwealth of the Northern Mariana Islands, Puerto Rico, Palau, and Virgin Islands

Table 3.

Baseline clinical characteristics

Characteristics	SZ cohort (n = 855)	BD-I cohort (n = 691)
Comorbid conditions,a n (%)
Anxiety disorders	367 (42.9)	450 (65.1)
Hypertension	284 (33.2)	196 (28.4)
Obesity	262 (30.6)	204 (29.5)
Any substance use disorder	261 (30.5)	223 (32.3)
Major depressive disorder	258 (30.2)	254 (36.8)
Hyperlipidemia	255 (29.8)	167 (24.2)
Type 2 DM	172 (20.1)	96 (13.9)
Posttraumatic stress disorder	134 (15.7)	191 (27.6)
Alcohol use disorder	108 (12.6)	101 (14.6)
Last antipsychotic use before index date,b n (%)
Any second-generation oralc	691 (80.8)	578 (83.6)
Any second-generation LAI	70 (8.2)	22 (3.2)
Any first-generation oral	55 (6.4)	19 (2.7)
Any first-generation LAI	15 (1.8)	1 (0.1)
None	47 (5.5)	78 (11.3)
Other common medicationsd with claims during baseline period, n (%)
Antidepressants	553 (64.7)	470 (68.0)
Mood stabilizers	410 (48.0)	490 (70.9)
Antihypertensives	390 (45.6)	321 (46.5)
Anxiolytics	346 (40.5)	356 (51.5)
Anticholinergics	248 (29.0)	105 (15.2)
Statins	204 (23.9)	121 (17.5)
Antiglycemics	169 (19.8)	120 (17.4)
Metformin, overall	144 (16.8)	97 (14.0)
Metformin, patients with type 2 DM	100 (69.4)	49 (50.5)
Metformin, patients without type 2 DM	44 (30.6)	48 (49.5)
Hypnotics and sedatives	91 (10.6)	129 (18.7)
Opioids	70 (8.2)	88 (12.7)

BD-I bipolar I disorder, DM diabetes mellitus, LAI long-acting injectable, SZ schizophrenia

^aComorbidities reported for ≥ 10% of patients in either cohort

^bPercentages sum to > 100 because some patients had multiple antipsychotics with the same claim date, which accounted for > 1 medication

^cOlanzapine was used by 50% of patients with SZ and 45% of those with BD-I; quetiapine was used by 9% of patients with SZ and 11% of patients with BD-I

^dTaken by ≥ 10% in either cohort

In the 6 months before the index OLZ/SAM claim, 94.5% of patients in the SZ cohort and 88.7% of those in the BD-I cohort had used an antipsychotic, most commonly a second-generation oral antipsychotic (SZ, 90.6%; BD-I, 87.0%). The last antipsychotic used by most patients in each cohort before they initiated OLZ/SAM was also a second-generation oral antipsychotic (SZ, 80.8%; BD-I, 83.6%; Table 3). Orally administered olanzapine was the last previous antipsychotic used by 50.4% of patients in the SZ cohort and 45.4% of patients in the BD-I cohort. Other common medications with claims during the baseline period are reported in Table 3.

Treatment Patterns and Persistence

The index OLZ/SAM dosages were most commonly 10/10 mg (31.9%) or 20/10 mg (28.4%) for patients with SZ and 10/10 mg (43.1%) or 5/10 mg (31.8%) for patients with BD-I.

During follow-up, 446 (52.2%) of 855 patients in the SZ cohort and 302 (43.7%) of 691 patients in the BD-I cohort remained on OLZ/SAM treatment for the full 6 months. Mean (SD) times to discontinuation were 56.5 (33.6) and 56.0 (32.7) days for the SZ and BD-I cohorts, respectively (median, 30 days for both cohorts). The average (SD) times patients were persistent with OLZ/SAM treatment were 122.0 (66.9) days for the SZ cohort and 111.1 (67.2) days for the BD-I cohort; median persistence times were 182 and 106 days, respectively. Among patients with SZ, the mean (SD) MPR was 0.88 (0.18), with a median of 0.98; 61.1% of patients had a MPR of ≥ 0.80. Mean (SD) PDC was 0.64 (0.33) in patients with SZ (median, 0.66); 45.7% had a PDC of ≥ 0.80. In the BD-I cohort, the mean (SD) MPR was 0.87 (0.19), with a median of 0.96; 56.0% had a MPR of ≥ 0.80. Mean (SD) PDC was 0.57 (0.33) in patients with BD-I (median, 0.57); 38.2% had a PDC of ≥ 0.80.

HCRU: SZ Cohort

In the SZ cohort, significant reductions from the baseline period in the proportions of patients with ≥ 1 all-cause inpatient admission or ≥ 1 ED visit were observed in the 6 months after OLZ/SAM initiation (Fig. 2a). The proportion of patients with ≥ 1 inpatient admission decreased significantly from 33.2% in the baseline period to 23.9% at follow-up (P < 0.001), a change from baseline of 28.2%. The proportion of patients with ≥ 1 ED visit decreased significantly from 44.7% in the baseline period to 39.5% in the follow-up period (P = 0.009), a reduction from baseline of 11.5%. No significant change in the proportion of patients with ≥ 1 outpatient visit was observed between the baseline and follow-up periods. Average number of all-cause inpatient days was reduced numerically in the SZ cohort during follow-up (mean [SD], 4.8 [17.6] days) versus baseline (mean [SD], 5.9 [15.7] days), but the difference was not statistically significant (P = 0.075; Supplementary Table S2).

Fig. 2.

HCRU in the SZ cohort (n = 855): proportions of patients with ≥ 1 all-cause (a), mental health-related (b), and SZ-related (c) IP admissions and ED and OP visits. *P < 0.05; **P < 0.01; ***P < 0.001. ED emergency department, HCRU healthcare resource utilization, IP inpatient, OLZ/SAM combination olanzapine and samidorphan, OP outpatient, SZ schizophrenia

Similar patterns of change were observed after OLZ/SAM initiation for mental health-related (Fig. 2b) and SZ-related (Fig. 2c) HCRU in the SZ cohort. The proportion of patients with ≥ 1 inpatient admission decreased significantly from baseline to follow-up for both the mental health-related (P < 0.001) and SZ-related (P < 0.001) categories, with reductions from baseline of 34.0% and 34.7%, respectively. The proportion of patients with ≥ 1 ED visit also decreased significantly in each category (mental health-related, P < 0.001; SZ-related, P = 0.018; reductions from baseline, 29.5% and 20.5%, respectively). Mean numbers of mental health-related and SZ-related inpatient admissions (both P ≤ 0.015) and ED visits (both P ≤ 0.003) were also significantly reduced in the follow-up period (Supplementary Table S2) relative to baseline. No significant change in the proportion of patients with ≥ 1 outpatient visit was observed. Mean (SD) mental health-related average numbers of inpatient days were 5.7 (15.2) at baseline and 4.5 (17.3) at follow-up; for SZ-related average number of inpatient days, means (SD) were 4.6 (14.1) and 3.8 (16.6), respectively, at baseline and follow-up. Neither of these differences was statistically significant (Supplementary Table S2).

HCRU: BD-I Cohort

Significant reductions from baseline in the proportions of patients with ≥ 1 all-cause inpatient admission or ≥ 1 ED visit in the 6 months after OLZ/SAM initiation also were observed in the BD-I cohort (Fig. 3a). The proportion of patients with BD-I and ≥ 1 inpatient admission decreased significantly from 27.6% at baseline to 19.1% at follow-up (P < 0.001), a change of 30.9% from baseline. The proportion of those with ≥ 1 ED visit decreased significantly from 42.0% to 36.8% (P = 0.026), a 12.4% change from baseline. No significant change in the proportion of patients with ≥ 1 outpatient visit was observed from baseline to the follow-up period. Mean (SD) all-cause average number of inpatient days decreased significantly from 3.4 (10.7) days in the baseline period to 2.2 (10.4) days in the 6 months after OLZ/SAM initiation (P = 0.011) (Supplementary Table S3).

Fig. 3.

HCRU in the BD-I cohort (n = 691): proportions of patients with ≥ 1 all-cause (a), mental health-related (b), and BD-I-related (c) IP admissions and ED and OP visits. *P < 0.05; ***P < 0.001. BD-I bipolar I disorder, ED emergency department, HCRU healthcare resource utilization, IP inpatient, OLZ/SAM combination olanzapine and samidorphan, OP outpatient

The pattern of change in mental health-related HCRU after OLZ/SAM initiation (Fig. 3b) was similar to that observed for all-cause HCRU in the BD-I cohort. The proportion of patients with ≥ 1 mental health-related inpatient admission decreased significantly from 25.5% at baseline to 16.1% in the 6-month follow-up period (P < 0.001), a change of 36.9% from baseline. The proportion of those with ≥ 1 ED visit decreased significantly from 24.2% to 17.4% (P < 0.001), a 28.1% change between baseline and follow-up periods. Mean number of mental health-related inpatient admissions was also significantly reduced in the follow-up period (P = 0.001; Supplementary Table S3). No significant change in the proportion of patients with ≥ 1 mental health-related outpatient visit was observed. Mental health-related average number of inpatient days decreased significantly from a mean (SD) of 3.3 (10.6) days in the baseline period to 2.0 (10.2) days in the follow-up period (P = 0.007) (Supplementary Table S3).

For BD-I-related HCRU (Fig. 3c), the proportion of patients with ≥ 1 inpatient admission decreased from 20.4% at baseline to 12.2% in the follow-up period (P < 0.001), a change of 40.4% from baseline. Mean number of BD-I-related inpatient admissions was also significantly reduced in the follow-up period (P = 0.001). No significant changes were observed in the proportion of patients with BD-I-related ED visits or in the proportion of those with BD-I-related outpatient visits. BD-I-related average number of inpatient days decreased significantly from a mean (SD) of 2.6 (9.5) days at baseline to 1.6 (9.8) days in the follow-up period (P = 0.022; Supplementary Table S3).

HCRU Subgroup Analyses

Patients Who Remained on OLZ/SAM Through Follow-up

A subgroup analysis was conducted on data from 446 (52.2%) patients in the SZ cohort and 302 (43.7%) patients in the BD-I cohort who remained on OLZ/SAM through the full 6-month follow-up period. Demographics, baseline health comorbidities, and medication use for these “completers” were similar to those for all initiators. As in the initiator population, the most common antipsychotic used just before initiating OLZ/SAM was orally administered olanzapine (SZ cohort, 53.8%; BD-I cohort, 47.0%).

In the SZ cohort, patterns of change in all-cause, mental health-related, and SZ-related HCRU from baseline to the follow-up period were similar to those observed for initiators, with statistically significant reductions in proportions of patients with inpatient admissions and ED visits among completers (all P < 0.001; Supplementary Table S4). Average all-cause and mental health-related inpatient days were statistically significantly reduced by > 1 day in the 6 months after initiation of OLZ/SAM in completers (both P < 0.05).

BD-I cohort completers also had statistically significant reductions from baseline in the proportions of patients with ≥ 1 all-cause, mental health-related, or BD-I-related inpatient admission or ≥ 1 ED visit (all P < 0.05; Supplementary Table S4), and all significant reductions were numerically greater for completers compared with initiators. Average numbers of all-cause and mental health-related inpatient days for patients with BD-I were reduced by > 1 day in the 6 months after initiation of OLZ/SAM, although those reductions were not statistically significant.

Patients Whose Most Recent Antipsychotic Prior to Index Date Was Olanzapine

A subgroup analysis of patients for whom olanzapine was their most recent antipsychotic included 431 (50.4%) patients in the SZ cohort and 314 (45.4%) patients in the BD-I cohort. Baseline patient demographics and clinical characteristics were generally similar to those of the overall patient population, although the proportions of patients with baseline inpatient admissions were slightly higher in this subgroup. In both cohorts, the pattern of changes in HCRU among patients in this subgroup was also similar to that observed in the overall population but with generally numerically greater percentage changes in HCRU measures from baseline to follow-up. In the SZ cohort, the proportions of patients with ≥ 1 all-cause, mental health-related, or SZ-related inpatient admission or ED visit were significantly reduced (all P ≤ 0.028) after initiation of OLZ/SAM. The BD-I cohort had statistically significant reductions in proportions of patients with ≥ 1 all-cause, mental health-related, or BD-I-related inpatient admission and with ≥ 1 mental health-related ED visit (all P ≤ 0.009; Supplementary Table S5).

Discussion

To our knowledge, this is the first real-world study indicating that treatment with OLZ/SAM is associated with reductions in HCRU in patients with SZ or BD-I. Proportions of patients with ≥ 1 all-cause or mental health-related inpatient admission or ED visit decreased significantly in the 6 months after OLZ/SAM initiation in both the SZ and BD-I cohorts. Statistically significant reductions in inpatient admissions and ED visits were also observed in patients who remained on OLZ/SAM through the entire 6-month follow-up period; these reductions were numerically greater compared with all patients who initiated OLZ/SAM treatment. Compared with the overall population, relative reductions in HCRU were also generally numerically greater among those for whom olanzapine was their most recent antipsychotic. Greater changes after OLZ/SAM initiation were observed for outcomes with higher baseline usage in the prior olanzapine subgroup, which could potentially be related to lack of adherence to treatment with olanzapine in the baseline period. Overall, average numbers of inpatient days were reduced by > 1 day between the baseline and follow-up periods in both the SZ and BD-I cohorts; mean decreases in all-cause and mental health-related inpatient days were statistically significant in patients with BD-I. Proportions of patients with all-cause, mental health-related, and disease-related outpatient visits were similar before and after initiation of OLZ/SAM, suggesting that patients maintained their regular outpatient care while reducing ED and inpatient utilization. Overall, these findings indicate that the use of OLZ/SAM may result in clinically meaningful reductions in patient and provider burden, as evidenced by changes in inpatient admissions and ED visits.

The propensity for weight gain and metabolic effects with olanzapine treatment compared with other atypical antipsychotics is well documented [26, 27, 56, 57]. However, olanzapine remains a valuable treatment option because it is highly efficacious [27, 30]. The development of OLZ/SAM shifts the balance further toward the efficacy benefit, as demonstrated by results from a long-term open-label study where patients with SZ or BD-I who were treated with OLZ/SAM for up to 4 years maintained stable symptom control with small changes in body weight and minimal changes in waist circumference and metabolic parameters [51]. Although mitigation of olanzapine-associated weight gain might result in long-term improvements in cardiometabolic risk factors [58, 59], the 6-month follow-up period of the current study would be too short to capture any HCRU effects of limiting sequelae from cardiometabolic risk [48, 60, 61]. Nonetheless, patients report in surveys that expecting weight gain, or experiencing weight gain early in antipsychotic treatment, is likely to negatively affect their adherence to medication [24, 62]. Therefore, limiting weight gain over even the first months of treatment could potentially improve adherence, and reducing gaps in antipsychotic medication could significantly reduce the risk of relapse and rehospitalization [53].

Patient persistence and adherence with a medication may provide some insight into whether patients themselves find benefit in continuing that treatment. Mean (SD) days persistent over the 6-month follow-up period were 122.0 (66.9) for the SZ cohort and 111.1 (67.2) for the BD-I cohort, and more than half of the patients in the SZ cohort and almost 45% of the BD-I cohort remained on OLZ/SAM through the entire 6-month follow-up period. Adherence rates based on MPR ≥ 0.80 and PDC ≥ 0.80 were 61.1% and 45.7%, respectively, for the SZ cohort and 56.0% and 38.2%, respectively, for the BD-I cohort. In a similarly designed, 6-month, pre–post retrospective database study of patients with BD-I, the rates of adherence to other oral antipsychotics, as measured by MPR ≥ 0.80 and PDC ≥ 0.80 during a 6-month follow-up, were observed to be 49% and 44%, respectively, for lurasidone and 41% and 36%, respectively, for cariprazine [63]. The significant reductions in inpatient admissions and ED visits in this study may reflect persistence of treatment and adherence to medication and suggest that patients with SZ or BD-I initiating OLZ/SAM may obtain the benefit of a highly effective antipsychotic medication while reducing the risk of discontinuing medication due to adverse events.

One of the strengths of the current study was use of the pre–post design, which allows for within-patient comparisons without requiring adjustment for between-patient factors. In our study, after initiation of OLZ/SAM, the proportions of patients with all-cause hospitalization decreased significantly from 33.2% in the baseline period to 23.9% at follow-up in the SZ cohort and from 27.6% to 19.1% in the BD-I cohort. Although numerous studies have reported between-group comparisons of HCRU in patients receiving oral antipsychotics [64–67], few have examined within-patient changes in HCRU before versus after initiating a specific oral antipsychotic medication. The above-mentioned similarly designed 6-month pre–post retrospective database study of patients with bipolar disorder also assessed healthcare use after initiation of lurasidone or cariprazine. The proportions of patients with all-cause hospitalization decreased from 21.5% at baseline to 16.2% in the follow-up period for patients initiating lurasidone and from 14.0% to 11.4% for patients who initiated cariprazine [63].

This study used claims data from Komodo’s Healthcare Map, which, because it includes claims from commercial payers, Medicare Advantage, and Medicaid, is able to capture a broad sample of patients with different demographic and clinical characteristics and also is representative of a patient population across multiple insurance coverage types. The study population was nonetheless restricted to patients with insurance coverage; results from this insured population may not be generalizable to uninsured populations. Additionally, results from studies based on single payer types (e.g., commercial only or Medicaid only) might be more representative of those patient groups individually. This study also has several limitations inherent to the analysis of administrative claims. A claim for a filled prescription may not indicate that the medication was consumed or taken as prescribed, and the presence of a diagnosis code may not definitively be indicative of disease presence. Further, the presence of a claim does not indicate disease severity, and neither disease severity nor the reason for initiating OLZ/SAM, each of which could affect HCRU, could be determined from the claims data. The effects of longer-term OLZ/SAM use on HCRU, beyond this study’s 6-month follow-up time, are unknown. Although the pre–post study design allows for within-patient comparisons, conclusions from the current analysis cannot be generalized beyond those receiving OLZ/SAM. Additional studies of real-world data examining treatment patterns and HCRU among patients initiating OLZ/SAM versus comparator antipsychotics are needed to determine the relative benefits of OLZ/SAM. Given that OLZ/SAM provides antipsychotic efficacy comparable with that of olanzapine but mitigates its associated weight gain in clinical trials [45–48], future studies assessing adherence and persistence to treatment with OLZ/SAM versus olanzapine, as well as potential follow-on effects on HCRU, may be of particular interest.

Conclusion

In this first real-world data assessment of HCRU before and after OLZ/SAM initiation, treatment with OLZ/SAM was associated with significant HCRU reductions in patients with SZ or BD-I. In the SZ cohort, proportions of patients with ≥ 1 inpatient admission and with ≥ 1 ED visit in the all-cause, mental health-related, and SZ-related HCRU categories decreased significantly between baseline and follow-up. Among patients with BD-I, proportions of patients with ≥ 1 all-cause, mental health-related, and disease-specific inpatient admissions decreased significantly between baseline and follow-up, as did proportions of patients with ≥ 1 all-cause and ≥ 1 mental health-related ED visits. Average numbers of inpatient days were significantly reduced between the baseline and follow-up periods by > 1 day in the BD-I cohort for all-cause, mental health-related, and BD-I-related admissions. In both cohorts, proportions of patients with all-cause, mental health-related, and disease-specific outpatient visits were similar before and after initiation of OLZ/SAM. Overall, these findings indicate that the use of OLZ/SAM may result in clinically meaningful reductions in patient and provider burden, as evidenced by changes in inpatient admissions and ED visits.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Conceptualization, Hemangi R. Panchmatia, Michael J. Doane; Methodology, Hemangi R. Panchmatia, Michael J. Doane, Alejandro G. Hughes; Data curation, Alejandro G. Hughes, Noah S. Webb; Validation, Alejandro G. Hughes, Hemangi R. Panchmatia; Formal analysis and investigation, Andrew J. Cutler, Michael J. Doane, Hemangi R. Panchmatia, Alejandro G. Hughes, Noah S. Webb, Rakesh Jain; Writing—review and editing, Andrew J. Cutler, Hemangi R. Panchmatia, Alejandro G. Hughes, Noah S. Webb, Michael J. Doane, Rakesh Jain; Supervision, Michael J. Doane, Andrew J. Cutler; Approval of the final manuscript for submission, Andrew J. Cutler, Hemangi R. Panchmatia, Alejandro G. Hughes, Noah S. Webb, Michael J. Doane, Rakesh Jain.

Funding

This study was sponsored by Alkermes, Inc.; Alkermes, Inc., also funded the journal’s Rapid Service fees.

Data Availability

The datasets generated and/or analyzed during the current study are not publicly available due to licensing restrictions. The data supporting the findings of this study were used under license from Komodo Health, Inc., and researchers interested in obtaining those data can access the database through Komodo Health, Inc., under a license agreement.

Declarations

Conflict of Interest

Andrew J. Cutler has acted as consultant for 4M Therapeutics, AbbVie, Acadia, Alfasigma, Alkermes, Anavex Life Sciences, Autobahn Therapeutics, Axsome, Biogen, Biohaven, Boehringer Ingelheim, Brii Biosciences, Bristol Myers Squibb, Cerevel, Corium, Delpor, Evolution Research Group, Intra-Cellular Therapies, J&J Innovative Medicine, Jazz Pharma, Karuna, LivaNova, Lundbeck, Luye Pharma, MapLight Therapeutics, MedAvante-ProPhase, Mentavi, Neumora, Neurocrine, Neuroscience Education Institute, NeuroSigma, Noven, Otsuka, PaxMedica, Relmada, Sage Therapeutics, Supernus, Teva, Thynk, Tris Pharma, Vanda Pharmaceuticals, and VistaGen Therapeutics; has received speaker/promotional honoraria from AbbVie, Alfasigma, Alkermes, Axsome, Bristol Myers Squibb, Corium, Intra-Cellular Therapies, J&J, Lundbeck, Neurocrine, Noven, Otsuka, Supernus, Teva, Tris Pharma, and Vanda Pharmaceuticals; holds stock options in 4M Therapeutics; and has served on a data safety monitoring board for Alar Pharma, COMPASS Pathways, Freedom Biosciences, and Pain Therapeutics. Hemangi R. Panchmatia and Michael J. Doane are or were employees of Alkermes, Inc., and may own stock/options in the company. Alejandro G. Hughes and Noah S. Webb are employees of Optum, Inc., a health services innovation company that received funding from Alkermes, Inc., to conduct this study and analyze the data used for this publication. Rakesh Jain has acted as consultant for AbbVie, Acadia, Adamas, Alfasigma, Alkermes, Almatica, Axsome, Biogen, Boehringer Ingelheim, Cingulate Therapeutics, Corium, Eisai, Evidera, Impel, Janssen, Lilly, Lundbeck, Merck, Neos Therapeutics, Neurocrine Biosciences, Osmotica, Otsuka, Pamlab, Pfizer, Sage Therapeutics, Shire, Sunovion, Supernus, Takeda, Teva, Transcend Therapeutics, and Viatris; has received speaker/promotional honoraria from AbbVie, Alkermes, Almatica, Axsome, Corium, Eisai, Intra-Cellular Therapies, Ironshore Pharmaceuticals, Janssen, Lilly, Lundbeck, Merck, Neos Therapeutics, Otsuka, Pamlab, Pfizer, Shire, Sunovion, Takeda, Tris Pharmaceuticals, and Viatris; has served on an advisory board for Adamas, Alkermes, Corium, Eisai, Janssen, Lilly, Lundbeck, Merck, Neos Therapeutics, Neurocrine Biosciences, Otsuka, Pamlab, Pfizer, Sage Therapeutics, Shire, Sunovion, Supernus, Takeda, and Teva; and has received research funding from AbbVie, Lilly, Lundbeck, Otsuka, Pfizer, Shire, and Takeda.

Ethical Approval

As this study used only deidentified patient records and did not involve the collection, use, or transmittal of individually identifiable data, it was exempted from institutional review board review [54].