PDF
Citation
Print
INTRODUCTION
Aneurysmal subarachnoid hemorrhage (aSAH), which accounts for approximately 5% of all strokes in the United States, is a life-threatening condition with profound clinical and economic challenges [1,2]. The clinical severity of aSAH is measured using the Hunt and Hess grade, which is a key indicator of overall survival and hospitalization costs. Hunt and Hess grades 1–4 incur hospital costs as high as $213,606, $264,425, $271,879, and $421,942, respectively [3]. In the United States, the financial burden of aSAH increases over time with advances in treatment, prolonged hospital stays, and inflationary pressures on hospitalization costs [4,5]. These costs are further compounded by frequent readmissions and the complexity of managing the complications of aSAH, such as hydrocephalus, vasospasm, and delayed cerebral ischemia (DCI) [4].
Immediate interventions after aSAH, including endovascular or microsurgical treatment of the ruptured aneurysm, administration of medications to control blood pressure, and placement of a surgical drain to treat hydrocephalus, primarily focus on preventing rebleeding and managing early complications [2]. However, these interventions are resource-intensive and do not directly address the inflammatory processes that contribute to poor outcomes or extended hospital stays [6]. This is important given the pivotal role of inflammation in the pathophysiology of aSAH and in exacerbating complications, such as vasospasm and DCI, which are strongly associated with increased hospital costs [7-9]. Novel therapies such as transcutaneous auricular vagus nerve stimulation (taVNS) offer a promising adjunct to standard care by modulating this deleterious inflammatory response [10-12]. Emerging evidence indicates that taVNS can reduce systemic inflammation and mitigate radiographic vasospasms, thereby improving clinical outcomes [13,14]. In particular, taVNS was associated with reduced levels of tumor necrosis factor-alpha (TNF-α) and interleukin-6 (IL-6), lower vasospasm incidence, and improved discharge outcomes [14]. Although clinically promising, little is known about the potential cost implications of taVNS for health systems and patients.
This study evaluated the potential cost reductions of integrating taVNS into standard aSAH management using data from the Non-invasive Auricular Vagus Nerve Stimulation for Subarachnoid Hemorrhage (NAVSaH) trial, a prospective, triple-blinded, randomized controlled study [15]. By analyzing direct hospital costs and resource utilization, we aimed to determine whether taVNS could decrease the economic burden of aSAH on healthcare systems. Our investigation provides critical insights into the potential of innovative neuromodulation techniques in advancing aSAH care and supporting economically sustainable healthcare practices.
METHODS
Clinical trial design
The NAVSaH trial (NCT04557618) is a prospective, triple-blinded, randomized controlled trial (Fig. 1). This study investigated the effects of taVNS in patients with spontaneous subarachnoid hemorrhage (SAH). The primary outcome measures were the impact of taVNS on inflammatory markers, such as TNF-α in plasma and cerebrospinal fluid, mitigation of radiographic vasospasm, and a decrease in the incidence of chronic hydrocephalus. Eligible participants were adult intensive care unit (ICU) patients with spontaneous SAH who were randomized 1:1 into either the taVNS or sham stimulation group. Blinding extended to the patients, medical providers, and outcome assessors. Each participant underwent twice-daily 20-minute sessions of either taVNS or sham stimulation targeting the auricular vagus nerve branch. We report the clinical and health economic outcomes of 27 patients who completed their index hospitalization and at least one follow-up visit. This reflects the 27 patients in the NAVSaH trial for whom a larger-than-expected effect size at the primary endpoint of radiographic vasospasm was previously reported following the interim analysis [14].
Health economics review
A literature review was conducted following PICOS (Problem, Intervention, Comparison, Outcomes and Study) criteria to evaluate the economic burden of SAH. Searches were performed in PubMed, Embase, Web of Science, EconLit, National Health Service Economic Evaluation Database, and the Health Care Utilization Project (HCUP). The focus was on studies reporting the direct and indirect costs associated with SAH care, covering radiology, operating rooms, ICU stay, neurosurgical interventions, complications, and readmissions.
Hospital cost data
Cost data were obtained from the institutional cost accounting system of the Neurosciences Clinical Service Line and reflected direct and indirect costs rather than charges or reimbursements. Costs were analyzed per patient and stratified according to Hunt-Hess grade to account for variations in disease severity. Outliers were retained in the analysis to preserve real-world variability; however, sensitivity analyses were performed to assess the impact of extreme length of stay (LOS) cases on cumulative cost differences (Supplementary Fig. 1).
Statistical analysis
Statistical analyses were performed using Python version 3.9.13 and its libraries (pandas, NumPy, SciPy, statsmodels, and matplotlib). To assess baseline comparability between the treatment arms, we compared clinical severity at admission using both the modified Rankin Scale (mRS) and the Hunt-Hess grading system (Supplementary Fig. 2). Group comparisons were performed using two-tailed Mann-Whitney U-tests, with visual inspection supported by overlaid histograms and boxplots. These analyses confirmed that there were no statistically significant differences in baseline severity between the sham and taVNS groups (P=0.980 for mRS; P=0.640 for Hunt-Hess), indicating well-balanced randomization and supporting an unbiased downstream comparison of outcomes.
Statistical significance was defined as either a two-tailed P-value <0.05 or a one-tailed P-value <0.05, depending on the underlying null hypothesis. Descriptive statistics summarized key variables, such as hospital costs, LOS, and Hunt-Hess grades. Chi-square statistics were computed for statistical analysis of categorical variables. Group comparisons were conducted using the Mann-Whitney U-test. ICU LOS was measured as the total number of days spent in the ICU, and group comparisons were performed using the Mann-Whitney U-test. All visualizations were generated using matplotlib.
Cost minimization analysis
To isolate modifiable costs from the total hospitalization expenses, we first identified and excluded non-modifiable costs associated with initial patient assessment, primary aneurysm treatment, and patient stabilization (Fig. 2). Non-modifiable costs included those for initial computed tomography imaging, diagnostic angiography, procedural interventions (microsurgical clipping or endovascular aneurysm repair), and immediate postoperative care within the first 24 hours. All remaining costs were categorized as modifiable and included expenses related to ongoing critical care management, respiratory therapy, rehabilitation services, pharmaceutical interventions, routine imaging studies, laboratory testing, and rooms and boards. Finally, 30-day costs represented modifiable hospital costs plus the cost of any patient readmission within 30 days of discharge. Thirty-day costs were calculated for each patient and analyzed both in aggregate and stratified by the Hunt-Hess grade to account for severity-based variations in resource utilization. Importantly, these modifiable costs were for patient care after the initiation of the taVNS or sham therapy protocols. The analysis focused on comparing these modifiable costs between the taVNS and sham groups to assess the economic impact of the intervention on potentially modifiable expenses during hospitalization.
Resource utilization analysis
We utilized the service area allocations of charges to perform a detailed analysis of clinical resource utilization patterns. A charge is a standard health-billing unit that reflects a single unit of service provided to a patient. Unlike cost analysis, which evaluates actual financial expenditure, charge analysis provides a relative measure of resource consumption across different hospital services. To evaluate the daily resource utilization trends, we computed the cumulative sums of the daily average charges for the sham and taVNS groups. Differences in cumulative distributions were assessed using the Kolmogorov-Smirnov test with adjustment for multiple comparisons using the standard false discovery rate correction.
RESULTS
Patient and group characteristics
The patient groups in this study, consisting of 14 individuals in the sham group and 13 individuals in the taVNS group, had similar demographic and clinical characteristics (Table 1). The average age was 57 years (standard deviation [SD], 12.3) in the sham group and 63 years (SD, 16.1) in the taVNS group, with no significant difference (P=0.34). Both groups were predominantly women, making up 86% of the sham group and 69% of the taVNS group (P=0.57). The distribution of treatment modalities showed a higher frequency of endovascular intervention in both groups (10 in the sham group and 11 in the taVNS group). Disease severity, measured by Hunt-Hess grade and the mRS score on admission, was similar between groups according to the unpaired Mann-Whitney U-test (P=0.17 and P=0.554, respectively). The taVNS group had fewer hospital-acquired infections, but the difference was not statistically significant (30.7% vs. 57%, P=0.084). Vasospasm severity was significantly reduced in the taVNS group, with the sham group showing a higher incidence of moderate and severe vasospasms (P=0.018). Furthermore, the taVNS group had a higher rate of favorable dispositions, defined as discharge to an inpatient rehabilitation center or home (92% vs. 64%, P=0.047) [14].
Thirty-day hospitalization costs
In terms of 30-day hospitalization costs, significant reductions were noted in the taVNS treatment group than the sham group during both initial hospitalization and readmission (Fig. 3). The mean 30-day cost in the sham group was $103,998 (SD, $67,013), whereas the taVNS group had a markedly lower mean cost of $82,632 (SD, $44,436) (P=0.003). This represents a $21,366 (20%) decrease in the average modifiable costs in the taVNS group. Readmissions were more frequent in the sham group, with four patients (28.6%) readmitted within 30 days of discharge, than in the 30-day readmission (7.7%) in the taVNS group (P=0.08). Individual patients with readmissions are represented in Fig. 3 based on the cost of readmission, represented in green on the stacked bar plot. The average readmission costs for the sham and taVNS groups were significantly different, with mean readmission costs in the sham group of $13,085 (SD, $16,512) compared to $2,716 for sole readmission in the taVNS group (Table 1).
Service area utilization
The service area resource volume and utilization analysis revealed statistically significant differences between the sham and taVNS groups in several clinical areas, as determined by the Kolmogorov-Smirnov test for cumulative distributions (Table 2). Notably, respiratory service use was significantly reduced in the taVNS group (mean, 17.0 charges per patient; SD, 17.85) compared to the sham group (mean, 64.60 charges per patient; SD, 41.76; P=0.003) (Fig. 4). Laboratory utilization was significantly lower in the taVNS group (mean, 112.23 charges per patient; SD, 66.67) than in the sham group (mean, 140.36 charges per patient; SD, 101.35; P<0.001). The pharmacy resource volume was markedly lower in the taVNS group (mean, 266.54 charges per patient; SD, 153.97) than in the sham group (mean, 313.29 charges per patient; SD, 215.56; P=0.004). Additionally, room and board resource usage was reduced in the taVNS group (mean, 5.08 charges per patient; SD, 4.42) than the sham group (mean, 8.64 charges per patient; SD, 7.49; P=0.008) (Table 3, ).
Impact of severity of subarachnoid hemorrhage on costs
Hunt-Hess cost analysis indicated a trend of reduced 30-day hospitalization costs in the taVNS group than the sham group across all grades (Fig. 5). For Hunt-Hess grade 1 and 2 patients, the mean cost was $62,497 (SD, $28,883) in the sham group and $57,807 (SD, $37,042) in the taVNS group, representing an 8% reduction. Similarly, for Grade 3 patients, there was a 15% reduction, with average costs of $134,441 (SD, $58,412) in the sham group and $114,315 (SD, $35,430) in the taVNS group. The greatest cost reduction (42%) was observed in grade 4 patients with a mean cost of $184,224 (SD, $70,106) in the sham group and $106,153 (SD, $46,112) in the taVNS group.
The analysis of hospital charges, expressed as group mean charge count (±SD), revealed lower resource utilization in patients receiving taVNS compared to the sham group, with greater reductions observed in patients with higher Hunt-Hess grades (Supplementary Fig. 3, Supplementary Table 1). Specifically, the mean number of hospital charges for grades 1 and 2 was similar (3% lower in taVNS), whereas grades 3 and 4 showed substantial reductions of 24% and 26%, respectively.
LOS and impact of severity of subarachnoid hemorrhage
We examined the difference in the mean LOS between the sham and taVNS groups. The mean LOS for the sham group was 22.9 days, while the mean LOS for the taVNS group was 19.15 days. This represents a difference of approximately 3.45 days, which was not statistically significant (P=0.232) (Table 1). Although LOS did not differ significantly across Hunt-Hess grades, notable trends emerged (Table 2). For Hunt-Hess grade 3, the sham group had a mean LOS of 31.33 days (SD, 18.01) compared with 25.5 days (SD, 4.65) in the taVNS group, reflecting a 19% shorter stay with taVNS. Lastly, for Hunt-Hess grade 4, the sham group had a mean LOS of 37 days (SD, 11) compared with 24.5 days (SD, 4.95) in the taVNS group, which is a 34% difference. However, none of these differences reached statistical significance, likely because of the small sample size in each Hunt-Hess category.
DISCUSSION
In this secondary analysis of the previously reported NAVSaH trial [14,15], we demonstrated that taVNS significantly reduced the economic burden associated with aSAH management in healthcare systems. The taVNS group showed markedly lower mean hospitalization costs ($82,632.62 vs. $103,998.60, P=0.003) (Table 1) and substantial decreases in resource utilization across multiple hospital services, including respiratory service utilization (17.00 vs. 64.60 charges per patient, P=0.003) (Table 3). These findings are particularly noteworthy, given that aSAH-related healthcare costs have risen dramatically over the past decade [4]. The reduction in both 30-day hospital costs and resource utilization suggests that taVNS could represent an important approach to improving the financial sustainability of aSAH management while potentially enhancing patient outcomes.
Length of stay
Studies have suggested that each additional day of hospitalization for aSAH treatment incurs an average cost increase of $3,815, highlighting LOS as a significant driver of total expenses [16]. We observed this trend in our data, with a longer LOS correlating with increased total hospital charges and costs (Supplementary Fig. 4). The use of taVNS therapy shows promise in decreasing this cost driver, with the intervention group demonstrating not only lower costs but also a 3.45-day difference in the mean LOS (19.15 vs. 22.9 days) (Table 1). Our results are in accordance with those of other studies that showed that the LOS is higher in patients with higher Hunt-Hess grades. Despite the lack of statistical significance, our results suggest that taVNS intervention may be associated with decreased LOS, and that this decrease in LOS may scale with the Hunt-Hess grade. For example, Hunt-Hess grade 4 patients in the taVNS arm showed a 42% difference in cost and a 34% difference in LOS compared to their sham-treated counterparts (Fig. 5, Table 2).
The need for prolonged hospital care is correlated with complications including cerebral vasospasm, hydrocephalus, and DCI, each of which requires intensive monitoring and intervention [3,5,16-18]. Avoiding these complications is associated with decreased hospital LOS, 30-day readmission rates, and healthcare costs [16,17]. Although not statistically significant, the differences observed in LOS between the taVNS and sham groups across the Hunt-Hess grades may be explained by our previous clinical findings of decreased inflammation, lower vasospasm incidence, and improved discharge outcomes in patients receiving taVNS therapy [14].
Thirty-day readmission for aSAH
The management of aSAH readmissions can be resource-intensive and requires intensive care, specialized nursing, and extensive diagnostic imaging [4,19]. Similar to the LOS, the difference in 30-day readmission rates observed with taVNS therapy (7.7% vs. 28.6% in the sham group) (Table 1) represents a potential economic benefit for healthcare systems. Other studies have reported 30-day readmission rates ranging from 11% to 14.6% [20,21]. Although limited by the small sample size of our current study, if this 20.9-percentage point reduction in readmissions were re-demonstrated in larger studies, these findings would have significant financial implications. Based on the mean readmission cost observed in our study ($13,085 for the sham group) (Table 1), preventing a single readmission could represent savings by avoiding additional costs from specialty consultations, interventions, and rehabilitation services.
These calculated cost savings from avoiding a single readmission are likely to be underestimated, given the opportunity to utilize ICU beds and specialized resources for new admissions, potentially generating additional revenue. Furthermore, although aSAH is not included in the Centers for Medicare & Medicaid Services Readmissions Reduction Program, reduction in all-cause readmissions with taVNS may improve hospital readmission metrics that support value-based payments and improve competitiveness in the quality-driven healthcare market [22,23]. Effectively, the reduction in readmission and readmission costs, combined with lower initial hospitalization costs, positions taVNS as a financially advantageous intervention that aligns with both clinical quality improvement goals and healthcare system economic objectives.
Resource utilization and capacity
Patients in the taVNS arm incurred lower modifiable hospital costs during index admission ($82,423 vs. $101,194, P=0.003) (Table 1). We demonstrate that these cost reductions are largely driven by the decreased utilization across several service lines. For example, respiratory therapy charges were reduced by over 70% (17.00 vs. 64.6 mean charges per patient, P<0.001) (Table 3, Fig. 5), and laboratory, pharmacy, and room and board services showed statistically significant differences between the treatment groups (Table 3). Similar to hospital costs and LOS, these decreases in resource utilization between the taVNS and sham groups were found to scale with the Hunt-Hess grade, suggesting a greater benefit for patients with severe aSAH (Supplementary Fig. 3, Supplementary Table 1).
From an operational standpoint, the reduced demand for inpatient services and reduced readmissions suggest a capacity benefit: by lowering utilization intensity, taVNS frees up staff, equipment, and procedural bandwidth, particularly in departments with fixed throughput, such as respiratory therapy and laboratory medicine [24]. These findings suggest that taVNS may function as a capacity multiplier, reducing the average resource burden per patient across index hospitalizations and readmissions, thereby expanding the hospital’s ability to accommodate additional admissions, reducing procedural delays, and repurposing staff and services more efficiently.
Cost of the taVNS intervention
Currently, there is no Food and Drug Administration (FDA)-approved taVNS device for aSAH; however, we can estimate potential hospital cost savings using data from analogous research-use-only (RUO) neuromodulation devices. For example, the Soterix taVNS system used in this research context has a fixed cost of $4,950 [25]. Historically, when neuromodulation technologies transition from RUO to FDA-cleared commercial products, device costs typically increase by a factor of 1.5 to 3× due to regulatory compliance, clinical validation, and commercialization infrastructure. This pattern is observed in devices such as transcranial direct current stimulation systems (e.g., Soterix vs. Neuroelectrics), external trigeminal nerve stimulators (e.g., Cefaly), and non-invasive vagus nerve stimulators (e.g., gammaCore), all of which experienced cost increases upon achieving FDA clearance. Applying this precedent, a reasonable future cost estimate for a commercially available FDA-approved taVNS system and therapy would range from $7,500 to $12,000. Even under this more conservative pricing scenario, the modeled hospital savings remain substantial. This underscores the economic value of taVNS in neurocritical care settings, even when accounting for realistic commercial pricing assumptions.
The cost of performing this clinical trial was not factored into the hospital cost analysis for several reasons. First, the research nature of the clinical trial, which includes research personnel and research-grade equipment, is not representative of clinical deployment and clinical practice cost savings from the use of this method. Additionally, a substantial portion of the resources for the NAVSaH trial supported basic scientific analysis (e.g., cytokine panels, biostatistical support, and specimen storage), which would not be relevant to the hospital cost analysis. The sources of funding for the NAVSaH trial include grants from the NIH/NINDS R21 NS128307, American Association of Neurological Surgeons, and The Aneurysm and AVM Foundation.
Indirect patient costs
Although our study did not directly measure patient-incurred costs, taVNS may provide financial benefits for patients. This intervention is, first and foremost, clinically beneficial [14]. Unmeasured patient costs included out-of-pocket costs of hospitalization and recovery. Based on our results, we believe that taVNS can reduce these costs. Out-of-pocket costs included cash pay, copay, and coinsurance for medical services. The observed decrease in 30-day costs and reduced resource utilization imply a decreased bill for patients from the hospital and/or private insurers. For example, among patients with similar insurance coverage, Hunt-Hess grade, and co-insurance terms, our findings suggest a potential 20% reduction in out-of-pocket expenses for those receiving taVNS.
Studies estimate that only 32.4% of aSAH survivors return to their previous employment level, with many facing a permanent disability that affects their earning capacity during their productive years [26]. Loss of wages during extended recovery periods, reduced future earnings potential owing to cognitive or physical disabilities, and early exit from the workforce are significant unmeasured costs. Studies suggest that patients discharged home or to an inpatient rehabilitation center have a higher rate of return to work than those going to a skilled nursing facility [27]. Our findings showed that the taVNS treatment arm had a higher rate of favorable disposition, defined as discharge to an inpatient rehabilitation center or home (92% vs. 64%) [14]. Therefore, the improved discharge outcomes in the taVNS arm suggest the potential to increase the return to employment, which may substantially decrease the unmeasured costs associated with aSAH.
Limitations
This study was a secondary analysis of a small single-center randomized controlled trial. Several studies have indicated significant variability in hospitalization costs with geographic and institutional factors, such as hospital size and teaching status [5,28-31]. These contextual differences highlight the need for larger multicenter studies to ensure generalizability of the results across various healthcare systems and practices. Quality of life implications, although not directly measured in economic terms in this study, involve substantial societal costs [32]. Due to a lack of quality-of-life data, this analysis did not calculate quality-adjusted life years or incremental cost-effectiveness ratios. Therefore, this study should be interpreted as a hospital cost-offset or cost-minimization analysis.
Future directions
Our findings suggest an economic benefit of taVNS in addition to a clinical benefit during initial hospitalization. Future multicenter trials should evaluate the long-term cost-effectiveness of taVNS for aSAH across diverse healthcare systems, focusing on both direct and indirect costs, including workforce reintegration, caregiver burden, and post-discharge care. To quantify value, future studies should have a priori defined cost-utility and cost-effectiveness endpoints in accordance with Consolidated Health Economic Evaluation Reporting Standards (CHEERS) guidelines [33]. Furthermore, practical implementation studies should address the logistical and financial barriers to clinical adoption, including training needs and reimbursement policies. Future comprehensive economic evaluations are crucial to establish taVNS as a clinically and economically viable treatment for aSAH.
XML Download