Antipsychotic regimen selection and its association with length of stay and hospitalization costs among first-episode schizophrenia patients in Indonesia: a retrospective cross-sectional study

¹Faculty of Pharmacy, Hasanuddin University, Makassar 90245, South Sulawesi, Indonesia, ²Faculty of Health Sciences and Pharmacy, Gunadarma University, Depok 16424, West Java, Indonesia, ³Department of Psychiatry, Dadi Regional Specialized Hospital, Makassar 90113, South Sulawesi, Indonesia, ⁴Department of Pharmacy, Sembilanbelas November University, Kolaka 93561, Indonesia, ⁵Robotics and Artificial Intelligence Study Program, Faculty of Advanced Technology and Multidiscipline, Airlangga University, Surabaya 60115, Indonesia, ⁶Robotics, Mechanics, and Intelligent System Research Group, Universitas Airlangga, Surabaya 60115, Indonesia, ⁷Faculty of Public Health, Tompotika University, Luwuk 94711, Central Sulawesi, Indonesia, ⁸Department of Health Sciences, University of Groningen, University Medical Center Groningen, 9713 AV Groningen, The Netherlands

^&Corresponding author
Bustanul Arifin, Faculty of Pharmacy, Hasanuddin University, Makassar 90245, South Sulawesi, Indonesia

Introduction    

Mental health is a major global public health concern, and schizophrenia represents one of the most severe and disabling psychiatric disorders. It affects approximately 21 million people worldwide and is associated with substantial functional impairment, increased risk of premature mortality, and long-term healthcare needs [1]. Effective management of schizophrenia relies heavily on pharmacological treatment, particularly antipsychotic therapy, which is recommended as first-line management in clinical guidelines [2]. However, treatment response varies considerably, requiring individualized therapeutic approaches based on clinical characteristics and patient needs [3]. In Indonesia, the burden of severe mental disorders has increased substantially, with national data reporting a rise in the prevalence of schizophrenia/psychosis from 1.7‰ in 2013 to 7‰ in 2018 [4]. In South Sulawesi Province, approximately 31,381 individuals are affected, with Makassar City showing the highest distribution [5]. This growing burden places pressure on psychiatric services and highlights the need to optimize both clinical effectiveness and healthcare efficiency. In acute clinical settings, schizophrenia often requires hospitalization for symptom stabilization. Length of stay (LOS) reflects clinical response and resource use, whereas hospitalization costs represent the economic burden of care [6,7]. Although antipsychotics are classified as first- and second-generation agents [8], real-world practice often involves combination regimens, particularly in patients with acute symptoms, inadequate treatment response, or complex clinical presentations [9]. Therefore, regimen selection may influence hospitalization outcomes by affecting symptom stabilization, discharge readiness, adverse event management, and the need for continued inpatient monitoring.

Conceptually, the relationship between antipsychotic regimen selection and cost efficiency can be understood as a clinical-economic pathway. A regimen that supports faster clinical stabilization may shorten length of stay and reduce the cumulative use of inpatient resources, thereby lowering total hospitalization costs. Conversely, regimens associated with delayed stabilization, adverse effects, or prolonged monitoring may extend hospitalization and increase total costs. Therefore, cost efficiency in inpatient schizophrenia care is determined not only by medication costs, but also by the interaction between treatment regimen, length of stay, and overall inpatient resource consumption. This issue is particularly relevant in Indonesia's national health insurance system, BPJS Kesehatan, which applies a standardized case-based reimbursement mechanism [10]. Under this system, hospitals receive fixed payment packages for specific diagnostic groups, while actual treatment costs may vary depending on length of stay, medication use, treatment intensity, and other inpatient resource needs. As a result, prolonged hospitalization or inefficient treatment pathways may create a mismatch between actual hospital expenditure and reimbursement, potentially affecting hospital financial sustainability.

Despite the clinical and economic relevance of this issue, most previous studies have focused on individual antipsychotic effectiveness, medication costs, and general cost-effectiveness [11,12]. Limited evidence has specifically examined how antipsychotic regimen selection is associated with LOS, hospitalization costs, and cost-efficiency indicators in real-world inpatient psychiatric settings, particularly within Indonesia's case-based reimbursement system. In addition, the relative contribution of sociodemographic factors compared with pharmacological factors remains unclear. This study was conducted at Dadi Regional Specialized Hospital (RSKD Dadi) Makassar, a major psychiatric referral center with high inpatient volume and frequent capacity constraints. This setting is therefore relevant for assessing real-world treatment efficiency among hospitalized patients with first-episode schizophrenia. Therefore, this study aimed to examine the association between antipsychotic regimen selection and hospitalization outcomes, including LOS, hospitalization costs, and cost efficiency among patients with first-episode schizophrenia. By integrating clinical and economic outcomes, this study is expected to provide context-specific evidence to support rational pharmacotherapy and more efficient inpatient mental health service delivery within Indonesia's case-based reimbursement setting.

Methods     

Study design and setting: this retrospective cross-sectional study examined the association between antipsychotic therapy regimens and hospitalization outcomes among patients with first-episode schizophrenia. The study was conducted at the Dadi Regional Specialized Hospital (Dadi Hospital), Makassar, the primary psychiatric referral center serving South and West Sulawesi provinces in Indonesia. The hospital manages high inpatient volumes, with more than 590 psychiatric admissions recorded within a typical three-month period, frequently operating near full capacity. This high service demand highlights the importance of optimizing treatment efficiency and resource utilization in psychiatric care. Data were obtained through a systematic review of the electronic medical records of patients hospitalized between January and June 2025. A six-month study period was selected to ensure an adequate sample size while maintaining data consistency and completeness.

Study population: the study population comprised all adult patients hospitalized during the study period with a primary diagnosis of schizophrenia based on the International Classification of Diseases, 10^th Revision (ICD-10: F20.x).

Inclusion and exclusion criteria

Inclusion criteria: eligible participants were adult patients diagnosed with schizophrenia who were undergoing their first documented psychiatric hospitalization, as confirmed by the absence of any recorded previous psychiatric admissions. First psychiatric hospitalization does not necessarily indicate a first episode of schizophrenia. Patients were required to be aged 18 years or older at the time of admission and have a primary diagnosis of schizophrenia ICD-10: F20.x. In addition, patients must have received antipsychotic pharmacotherapy during hospitalization. Only patients with complete medical records containing essential clinical and treatment information were included in the analyses.

Exclusion criteria: patients were excluded if they had incomplete medical records or missing critical data elements. Those with significant medical comorbidities requiring specialized non-psychiatric care that could potentially influence the length of hospital stay or treatment costs were also excluded. In addition, patients with concurrent psychiatric comorbidities requiring different treatment protocols, such as substance abuse or major mood disorders, were not included. Patients who were pregnant or lactating at the time of admission were excluded. Furthermore, patients transferred from another healthcare facility and those admitted through social service agencies or involuntary commitment procedures involving different discharge processes were also excluded from the study.

Sample size and sampling procedure: a total of 937 medical records of patients hospitalized with a primary diagnosis of schizophrenia during the study period were identified. Because all medical records available within the defined study period were screened and all eligible records were included, the study used a total population sampling approach. Therefore, no formal sample size calculation was performed. Records were systematically screened according to predefined inclusion and exclusion criteria. During the first screening stage, 176 records were excluded because the patients had significant medical or psychiatric comorbidities, leaving 761 records for further evaluation. During the second stage, these records were evaluated to determine whether the admission represented the patient's first psychiatric hospitalization with a primary diagnosis of schizophrenia. A total of 454 records did not meet this criterion and were excluded from the study. Following the screening process, 307 medical records fulfilled all eligibility criteria and were included in the final analyses. Each patient contributed only one hospitalization record for the analysis.

Data collection: data were extracted from the hospital's electronic medical record system using a standardized data extraction form developed for this study. The extracted information included sociodemographic characteristics, clinical diagnosis, antipsychotic therapy regimen, admission and discharge dates, length of stay, and total hospitalization costs. Pharmacotherapy data were obtained from medication records, while cost data were obtained from hospital billing records. To improve data accuracy, extracted data were checked by the research team using relevant source documents, including physician notes, medication records, discharge summaries, and billing records. Data validation included checking the consistency of admission and discharge dates, recalculating LOS from the original dates, verifying antipsychotic regimens against medication records, and confirming total hospitalization costs using billing data. Any discrepancies were resolved through discussion and consultation with a senior clinical pharmacist.

Missing data management: medical records were assessed for completeness before inclusion in the final analysis. Critical data elements included age, sex, primary diagnosis, admission date, discharge date, antipsychotic regimen, length of stay, and total hospitalization cost. Records with missing critical data elements were excluded because these variables were essential for defining the exposure and outcome measures. No statistical imputation was performed, and the analysis was conducted using complete cases only. This approach was selected to minimize outcome misclassification and ensure consistency in the calculation of LOS, hospitalization cost, and cost efficiency.

Variables

Antipsychotic therapy regimen (independent variable): the primary independent variable in this study was the antipsychotic therapy regimen administered during hospitalization. The typical antipsychotics used in therapy included trifluoperazine (5 mg), haloperidol tablets (5 mg), fluphenazine decanoate (25 mg/ml) injection, and haloperidol decanoate (50 mg/mL) injection. The atypical antipsychotics used included risperidone (2 mg), quetiapine fumarate (200 mg), olanzapine (5 mg and 10 mg), clozapine (25 mg and 100 mg), aripiprazole (10 mg), Abilify oral solution (1 mg/mL 60 mL), olanzapine injection, and haloperidol injection. Treatment regimens were categorized according to the number of concurrently prescribed antipsychotic agents. Because only eight patients received antipsychotic monotherapy, monotherapy and two-drug regimens were combined into a single category representing regimens involving two or fewer antipsychotic agents (≤2 antipsychotic agents). Regimens involving three or four antipsychotic agents were grouped as regimens involving more than two antipsychotic agents (>2 antipsychotic agents). Concomitant non-antipsychotic psychotropic medications were not used to define the antipsychotic regimen category.

Length of stay (LOS): length of stay (LOS) was defined as the total duration of continuous hospitalization in calendar days, calculated from the date of admission to the date of discharge. For analytical purposes, LOS was categorized into two groups: short LOS (≤21 days) and prolonged LOS (>21 days) [13,14]. The 21-day threshold was used as an operational cutoff based on the maximum hospitalization duration specified in the Indonesian National Clinical Pathway for Schizophrenia and the reimbursement framework applied in psychiatric inpatient care [15]. This threshold was considered clinically and administratively relevant because hospitalization beyond this period may indicate delayed stabilization, extended inpatient monitoring, or greater resource utilization.

Hospitalization cost: hospitalization cost was defined as the total direct medical expenditure incurred during a single inpatient episode. Cost components included pharmaceutical expenses, medical consumables, hospital accommodation charges, professional fees, diagnostic procedures, and ancillary therapeutic services. Costs were initially recorded in Indonesian Rupiah (IDR) and converted to United States Dollars (USD) using the average exchange rate during the study period, namely, 1 USD = IDR 15,504, based on Bank Indonesia exchange rate data [16]. For analytical purposes, hospitalization cost was dichotomized using a predefined threshold of USD 393, equivalent to approximately IDR 6,093,100. This cutoff corresponded to the BPJS INA-CBGs reimbursement rate for schizophrenia hospitalization under the Indonesian INA-CBGs implementation regulations [17]. Costs were categorized as lower cost (≤USD 393) and higher cost (>USD 393). The use of this reimbursement-based threshold allowed the analysis to assess whether actual hospitalization costs were within or above the standardized case-based payment level.

Cost efficiency of hospitalization (LOS and cost): a composite outcome was constructed to assess the cost efficiency of hospitalization by integrating length of stay (LOS) and hospitalization costs, thereby capturing both clinical and economic dimensions of inpatient care. Length of stay (LOS) and treatment costs are commonly used as proxy indicators of hospital efficiency [18], as both reflect the intensity of healthcare resource utilization during hospitalization [15,17]. Length of stay (LOS) represents the duration of clinical stabilization and inpatient care, while hospitalization cost reflects the direct economic burden associated with resource use during the admission episode. For analytical purposes, patients were classified into two categories based on predefined LOS and cost thresholds. Cost-efficient hospitalization was defined as a short LOS (≤21 days) combined with lower hospitalization costs (≤USD 393). In contrast, cost-inefficient hospitalization was defined as prolonged LOS (>21 days) and/or higher hospitalization costs (>USD 393). This composite measure was used to identify hospitalization episodes in which shorter inpatient duration and lower treatment costs occurred concurrently, indicating relatively more efficient inpatient care.

Sociodemographic variables: potential confounders included age, sex, educational attainment, employment status, and marital status. Age was categorized as adult (<60 years) and elderly (≥60 years). Sex was classified as male or female. Educational attainment was grouped into three levels: low (no formal education, kindergarten, elementary school, or junior high school), moderate (senior high school), and high (diploma, bachelor's degree, or master's degree). Employment status was categorized as employed or unemployed. Marital status was classified as single, married, or divorced.

Bias: several potential sources of bias were considered in this study. Selection bias may have occurred because the study was conducted in a single psychiatric referral hospital, and the included patients may not fully represent all patients with first-episode schizophrenia in Indonesia. To reduce this bias, all eligible medical records within the defined study period were systematically screened using predefined inclusion and exclusion criteria. Information bias may also have occurred because the study relied on retrospectively recorded medical record data. To minimize this risk, data were extracted using a standardized form, and key variables were cross-checked across physician notes, medication records, discharge summaries, and billing records. Length of stay (LOS) was recalculated from admission and discharge dates, while hospitalization costs were verified using hospital billing data. Confounding by indication was another important concern, as antipsychotic regimen selection may have been influenced by clinical severity, symptom presentation, treatment response, or physician judgment. Although we excluded patients with major medical or psychiatric comorbidities and used multivariable logistic regression to adjust for available sociodemographic confounders, residual confounding may remain because standardized clinical severity indicators, such as Positive and Negative Syndrome Scale (PANSS) or Clinical Global Impressions (CGI) scores, were not available in the medical records.

Statistical analysis: all statistical analyses were performed using IBM SPSS Statistics for Windows, version 26.0 (IBM Corporation, Armonk, NY, USA). Categorical variables were summarized using frequencies and percentages. Because the exposure and outcome variables were categorical, bivariate associations were assessed using Pearson's chi-square test. Fisher's exact test was used when expected cell counts were small. Three separate binary logistic regression models were constructed to estimate the adjusted associations between antipsychotic regimen and hospitalization outcomes: (i) LOS, categorized as short LOS versus prolonged LOS; (ii) hospitalization cost, categorized as lower cost versus higher cost; and (iii) cost efficiency, categorized as cost-efficient versus cost-inefficient hospitalization. Binary logistic regression was selected because each outcome was dichotomous and the analysis aimed to estimate adjusted odds ratios while adjusting for potential confounders. Variables with a p-value <0.25 in the bivariate analysis were considered for inclusion in the multivariable models. This threshold was used to avoid excluding potentially important confounders at the initial screening stage. Sociodemographic variables and treatment regimen were evaluated as candidate predictors. In all regression models, antipsychotic regimen was entered as a binary predictor, comparing regimens involving two or fewer antipsychotic agents (?2 antipsychotic agents) with regimens involving more than two antipsychotic agents (>2 antipsychotic agents). Multicollinearity among independent variables was assessed before model construction, and model fit was evaluated using the Hosmer-Lemeshow goodness-of-fit test. Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were reported to estimate the strength and precision of the associations. A p-value <0.05 was considered statistically significant.

Ethical considerations: the study protocol was approved by the Research Ethics Committee of the Faculty of Pharmacy, Hasanuddin University (approval number: 771/UN4.17.8/KP.06.05/2025). In addition, administrative permission to access medical records was obtained from the Dadi Regional Specialized Hospital (RSKD Dadi) in Makassar (approval number: 000.9.2/38/P. DLK/III/RSKD DADI; dated March 18, 2025). All patient data were anonymized prior to analysis to ensure confidentiality. This study utilized secondary data extracted from medical records and did not involve direct patient contact.

Results     

Sociodemographic and clinical characteristics: a total of 307 first-episode schizophrenia patients were included in the study (Table 1). Most patients were men (75.9%), with a mean age of 36.73 ± 12.32 years, and most were younger than 60 years (95.4%). Most participants were unemployed (68.1%), had a low educational level (59.6%), and were single (61.2%). Regarding treatment patterns, regimens involving ≤2 antipsychotic agents were the most common category (54.1%). Slightly more than half of the patients had a length of stay ≤21 days (53.4%), and 51.5% incurred hospitalization costs exceeding USD 393. Based on the combined assessment of length of stay and hospitalization costs, 54.7% of cases were classified as cost inefficient. A short length of stay (≤21 days) was observed in 53.4% of patients, whereas 46.6% experienced prolonged hospitalization (>21 days). Hospitalization costs were almost evenly distributed between lower-cost (48.5%) and higher-cost (51.5%) categories. Overall, 45.3% of patients achieved cost-efficient hospitalization outcomes.

Associations between patient characteristics, treatment regimen, and hospitalization outcomes: associations between patient characteristics, treatment regimens, and hospitalization outcomes were examined using the Chi-square test, and the strength of these associations was expressed as odds ratios (OR) with 95% confidence intervals (CI). The results are presented in Table 2. Sex was significantly associated with length of stay (LOS). Female patients had lower odds of prolonged hospitalization compared with male patients (OR = 0.57; 95% CI: 0.33-0.99; p = 0.046). The treatment regimen was significantly associated with all hospitalization outcomes. Patients receiving regimens involving ≤2 antipsychotic agents had higher odds of prolonged hospitalization (OR = 1.86; 95% CI: 1.18-2.93; p = 0.007), higher hospitalization costs (OR = 1.74; 95% CI: 1.11-2.75; p = 0.015), and greater likelihood of cost inefficiency (OR = 1.71; 95% CI: 1.09-2.70; p = 0.019). Other variables, including age, employment status, educational level, and marital status, were not significantly associated with the outcomes (all P > 0.05).

Factors associated with hospitalization outcomes after adjustment: variables that met the inclusion criteria were entered into adjusted logistic regression models to estimate the independent associations with hospitalization outcomes. The results of the adjusted analysis are presented in Table 3. After adjustment for sociodemographic variables, the antipsychotic treatment regimen remained significantly associated with all hospitalization outcomes. Patients receiving regimens involving ≤2 antipsychotic agents had higher odds of prolonged hospitalization than those receiving regimens involving >2 antipsychotic agents (AOR = 1.81; 95% CI: 1.13-2.89; p = 0.013). Similarly, patients treated with regimens involving ≤2 antipsychotic agents were more likely to incur hospitalization costs exceeding the predefined threshold of USD 393 (AOR = 1.72; 95% CI: 1.08-2.72; p = 0.021). In the composite efficiency model, the ≤2 antipsychotic agents' category was also associated with a higher likelihood of cost-inefficient hospitalization based on the combined assessment of length of stay and hospitalization cost (AOR = 1.69; 95% CI: 1.06-2.69; p = 0.027). Overall, these findings should be interpreted in relation to the predefined clinical-administrative and reimbursement-based thresholds used in this study. Additional sensitivity analyses using alternative LOS or cost thresholds were not performed; therefore, the robustness of the observed associations across different operational definitions requires further investigation.

Discussion     

This study examined the association between antipsychotic regimen selection and hospitalization outcomes among patients with first-episode schizophrenia in a high-volume psychiatric referral hospital. The findings demonstrate that the antipsychotic regimen was the only variable consistently and independently associated with length of stay (LOS), hospitalization cost, and a composite measure of cost efficiency. In contrast, sociodemographic variables showed limited or no independent effects after multivariate adjustment. The predominance of male patients and the concentration of cases in the adult age group are consistent with previously reported inpatient demographic patterns among patients hospitalized for schizophrenia [19]. Previous studies have reported that patients with schizophrenia frequently experience socioeconomic disadvantages, including unemployment and lower educational attainment, which are associated with poorer social functioning and quality of life [20]. However, in the present study, these sociodemographic characteristics did not independently predict length of stay, hospitalization costs, or cost efficiency after adjustment, suggesting that therapeutic factors may exert a stronger influence on inpatient outcomes than baseline demographic characteristics [18].

The dominance of atypical antipsychotics observed in this study aligns with current clinical prescribing trends. Risperidone is frequently preferred because of its favorable balance between efficacy and tolerability and its relatively lower risk of extrapyramidal side effects compared with those of typical antipsychotics [2,21,22]. Clozapine remains the gold standard therapy for treatment-resistant schizophrenia because of its superior efficacy in reducing persistent psychotic symptoms and suicide risk [18]. Similar prescribing patterns have been reported in several international and regional studies conducted in Indonesia, China, and Australia [21,23]. Although typical antipsychotics, such as haloperidol, are still prescribed, their use in this study was relatively limited compared with that of atypical agents. Other second-generation antipsychotics, including olanzapine, quetiapine, and aripiprazole, were also used, but at lower frequencies. These findings suggest a gradual shift toward atypical antipsychotic-based regimens in clinical psychiatric practice. The distribution of antipsychotic drugs prescribed in this study is illustrated in Figure 1. In the shortest LOS group (4-7 days), the frequent use of clozapine-risperidone combinations with adjunctive injectable antipsychotics may reflect an intensive treatment strategy used for acute symptom stabilization in this inpatient setting. However, because this study did not measure symptom severity, treatment response trajectories, or adverse events, the clinical mechanisms underlying shorter LOS could not be directly determined.

These findings are supported by large-scale real-world evidence. Tiihonen et al. demonstrated that augmentation of clozapine with selected antipsychotics significantly reduced the hospitalization risk compared with clozapine monotherapy in more than 23,000 patients [24]. Furthermore, a systematic review and meta-analysis by Lin et al. reported that long-acting injectable antipsychotics improved treatment adherence and reduced hospitalization rates without increasing overall healthcare costs [25]. In the bivariate analysis, female sex was associated with shorter LOS, a finding that aligns with prior evidence indicating that male patients may experience longer psychiatric hospitalizations due to behavioral severity, delayed treatment-seeking, or differential social support structures [26]. Nevertheless, this association did not persist in the multivariable analysis, indicating that the initial relationship was likely confounded by other factors, particularly the treatment regimen. This attenuation after adjustment is consistent with established multivariate modeling principles in medical research, in which the effects of sociodemographic variables may diminish when clinical and therapeutic variables are considered simultaneously [27].

The most consistent finding across all models was a significant association between the antipsychotic regimen and hospitalization outcomes. Patients receiving regimens involving ≤2 antipsychotic agents had significantly higher odds of prolonged LOS, higher hospitalization costs, and cost-inefficient outcomes compared with those receiving regimens involving ≤2 antipsychotic agents. However, these findings should not be interpreted as evidence that antipsychotic polypharmacy is broadly superior. Rather, they may reflect real-world clinical decision-making in acute inpatient care, where more intensive regimens are used selectively according to symptom presentation, treatment response, and clinician judgment. Because standardized measures of clinical severity were not available, the possibility of residual confounding by indication cannot be excluded. The high prevalence of antipsychotic polypharmacy observed in this study is consistent with international prescribing patterns, in which combination therapy remains common despite guideline recommendations favoring monotherapy as the first-line treatment [28-30]. Similar prescribing trends have been observed in Asian populations, including those in China and Indonesia, where combination therapy remains common in inpatient psychiatric settings [21,23].

Real-world meta-analytic evidence has further demonstrated that long-acting injectable antipsychotics may offer clinical and economic advantages over oral formulations in selected patient populations [25]. While monotherapy remains the recommended standard in most treatment algorithms, polypharmacy is often employed in cases of severe symptoms, suboptimal response, or complex clinical presentations [31]. Evidence from large cohort analyses suggests that certain forms of targeted antipsychotic polypharmacy may reduce the risk of hospitalization compared with monotherapy, particularly when the pharmacodynamic mechanisms are complementary [24,28,29,32]. Similar findings have been observed in large real-world Medicare cohorts comparing long-acting injectable and oral antipsychotics [33]. From an economic perspective, hospitalization is a major driver of direct schizophrenia-related healthcare expenditures [34]. In the present study, the treatment regimen was independently associated with hospitalization costs, whereas sociodemographic variables were not. This finding reinforces pharmacoeconomic principles, emphasizing that therapeutic decision-making plays a central role in determining cost outcomes within case-based reimbursement systems [31]. Under Indonesia's BPJS INA-CBGs framework, discrepancies between actual costs and standardized reimbursement thresholds directly affect hospital financial sustainability. Therefore, regimen selection has implications not only for clinical stabilization but also for institutional cost containment.

The composite efficiency outcome further highlights the combined clinical and economic implications of regimen selection. Patients receiving regimens involving ≤2 antipsychotic agents had higher odds of cost-inefficient hospitalization than those receiving regimens involving >2 antipsychotic agents. This suggests that, in this specific inpatient context, more intensive regimens may have been associated with shorter hospitalization and lower total inpatient costs. However, these findings should not be interpreted as evidence that antipsychotic polypharmacy is broadly superior. Clinical guidelines continue to emphasize monotherapy as the standard initial approach, while combination therapy should be considered only in selected clinical situations after careful assessment of treatment response, safety, and patient complexity. From a health system perspective, these findings are particularly relevant to Indonesia's BPJS INA-CBGs case-based reimbursement framework. Under this payment model, hospitals receive a fixed reimbursement package, regardless of the actual length of stay or total resource utilization. When hospitalization exceeds the predefined clinical pathway duration or when the total inpatient expenditure surpasses the reimbursement ceiling, the excess operational cost may not be fully covered. In high-volume psychiatric referral hospitals, repeated prolonged admissions may create potential reimbursement gaps and financial pressure on hospital operations. Therefore, treatment regimen selection has implications not only for individual patient stabilization but also for mitigating potential reimbursement gaps within the national health insurance system.

It is also important to note that the present study focused exclusively on direct inpatient medical costs and short-term hospitalization outcomes. Broader economic consequences, including long-term relapse prevention, outpatient expenditures, adverse effect management, and quality-of-life impact, were beyond the scope of this analysis. Although the observed associations suggest potential short-term efficiency advantages of certain multi-drug regimens in acute inpatient settings, a comprehensive pharmacoeconomic evaluation is necessary before translating these findings into broader clinical or policy recommendations. This study has several limitations. First, the retrospective cross-sectional design limits the ability to infer causal relationships between antipsychotic regimen selection and hospitalization outcomes. Second, the study was conducted in a single psychiatric referral hospital, which may limit the generalizability of the findings to other hospitals or healthcare settings. Third, standardized clinical severity measures, such as PANSS or CGI scores, were not available in the medical records. This is an important limitation because antipsychotic regimen selection may have been influenced by symptom severity, clinical complexity, treatment response, or physician judgment. Although we excluded patients with major comorbidities and used multivariable logistic regression to adjust for available sociodemographic factors, residual confounding by indication may remain.

Fourth, this study did not perform additional sensitivity analyses using alternative LOS or cost thresholds. Therefore, the robustness of the findings across different operational definitions of prolonged hospitalization, higher hospitalization cost, and cost efficiency could not be fully assessed. The predefined thresholds used in this study were based on clinical-administrative and reimbursement considerations; however, future studies should examine whether similar associations are observed using alternative cutoff points, continuous outcome measures, or prospective economic evaluation designs. Overall, the findings indicate that within this tertiary psychiatric hospital setting, antipsychotic regimen selection remained consistently associated with both clinical and economic hospitalization outcomes. These results underscore the importance of rational, context-adapted pharmacotherapy strategies that balance effectiveness, safety, and cost considerations within resource-constrained healthcare systems [35].

Conclusion     

This study found that antipsychotic regimen selection was associated with hospitalization outcomes, including LOS, hospitalization cost, and cost efficiency, among patients with first-episode schizophrenia in a psychiatric referral hospital in Indonesia. Regimens involving >2 antipsychotic agents were associated with lower odds of prolonged hospitalization and cost-inefficient outcomes compared with regimens involving ?2 antipsychotic agents. However, these findings should be interpreted cautiously because of the retrospective design, single-center setting, absence of standardized clinical severity measures, and potential residual confounding by indication. Further prospective studies incorporating clinical severity indicators, alternative outcome definitions, and broader hospital settings are needed to confirm these associations and inform rational antipsychotic use within case-based reimbursement systems.

Competing interests     

The authors declare no conflicts of interest.

Authors' contributions     

Nurfadillah Alfiah conceptualized the study, collected the data, and drafted the manuscript. Elly Wahyudin contributed to study supervision and methodological refinement. Mayamariska Sanusi provided clinical input and facilitated access to hospital data. Latifah Rahman and Muhammad Syahrir contributed to data verification and interpretation of findings. Sylmina Dalily Alkaff and Akhmad Ardiansyah supported data analysis and critical revision of the manuscript. Bustanul Arifin supervised the overall study design, performed the statistical analysis, and critically revised the manuscript for important intellectual content. All the authors have read and approved the final version of this manuscript.

Acknowledgments     

The authors express their sincere appreciation to Dadi Regional Specialized Hospital (RSKD Dadi), Makassar, for granting access to the medical record data used in this study. The authors also thank the medical record staff and healthcare personnel at RSKD Dadi Makassar for their assistance in facilitating data retrieval and supporting the research process.

Tables and figure     

Table 1: sociodemographic characteristics, treatment regimens, and hospitalization outcomes (N = 307)

Table 2: association between patient characteristics, antipsychotic regimen, and hospitalization outcomes among first-episode schizophrenia patients (N = 307)

Table 3: multivariable logistic regression analysis of factors associated with hospitalization outcomes among first-episode schizophrenia patients (N = 307)

Figure 1: distribution of antipsychotic prescriptions among hospitalized study participants with schizophrenia

References