Regional disparity in the use of a financial incentive for device-assisted rehabilitation: an ecological study in Kanagawa, Japan

Abstract

[Purpose] We evaluated the utilization of a financial incentive for functional electrical stimulation and robot-assisted rehabilitation in Kanagawa Prefecture, and assessed geographic variations and related regional indicators. [Participants and Methods] Participants were from nine secondary medical areas in Kanagawa Prefecture. We used fiscal year 2022 standardized claim ratios from the Cabinet Office dataset and summarized the standardized claim ratios for the financial incentive claimable within 2 months post-onset and evaluated associations with selected regional indicators of demand, service provision, and care settings. [Results] The standardized claim ratio for the financial incentive ranged from 3.3 to 223.3 (mean, 64.0; median, 36.8). Only the standardized claim ratios of Yokosuka–Miura and Yokohama exceeded the national average (standardized claim ratio, ≥100). The standardized claim ratio for the financial incentive showed no statistically significant association with standardized claim ratios for cerebral infarction inpatients, convalescent rehabilitation fees, or cerebrovascular rehabilitation volume. [Conclusion] Utilization of the financial incentive program varied widely across secondary medical areas in Kanagawa Prefecture. The eligibility window of 2 months from symptom onset may have contributed to this variation and may not have fully aligned with the current guideline statements for robot-assisted training for upper limb function.

INTRODUCTION

Functional electrical stimulation (FES) and robot-assisted interventions are recommended by clinical practice guidelines for motor impairment after stroke and spinal cord injury. The Japanese Stroke Guidelines recommend FES for foot drop¹⁾. They also recommend robot-assisted training to improve upper limb function, walking ability, and the ability to perform activities of daily living¹⁾. Additionally, physical therapy clinical practice guidelines recommend robot-assisted gait training to improve the walking function of individuals with spinal cord injury²⁾. Therefore, promoting the uptake of these evidence-based interventions in routine practice is important.

In 2020, Japan’s Ministry of Health, Labour and Welfare introduced a reimbursement add-on (financial incentive) for interventions using FES and robotic technologies to address motor impairment after stroke and spinal cord injury. This add-on allows providers to claim 150 points (1,500 Japanese yen) once per month when they conduct an intervention using an approved device within 2 months of onset of motor impairment. A prior descriptive study that used nationwide health insurance claims data reported variations in intervention uptake across prefectures; for example, in fiscal year 2022, the standardized claim ratio ranged from 2.4 to 337.5³⁾. Additionally, in fiscal year 2022, the standardized claim ratios in some prefectures remained below the national average³⁾. These disparities in uptake represent an important challenge from the perspective of equitable access to rehabilitation interventions that include advanced devices.

Kanagawa Prefecture is a prefecture with below-average uptake of the reimbursement add-on; however, the reasons for this are unclear. In particular, Kanagawa has a high demand for rehabilitation services but a shortage of rehabilitation providers, which may limit access to the services needed by its residents. In 2022, Kanagawa’s supply-to-utilization ratio was 1.33, which was lower than the national average of 1.42⁴⁾. Additionally, in Kanagawa, the number of rehabilitation providers per 1,000 population was 0.67, which was the lowest among all prefectures⁴⁾. Clarifying the uptake of the reimbursement add-on in Kanagawa could provide important evidence that could aid in the development of prefecture-level strategies to promote policies that support rehabilitation interventions using advanced devices. Therefore, we aimed to describe the uptake of a reimbursement add-on in Kanagawa Prefecture, characterize the diffusion of this financial incentive, and provide evidence to inform prefecture-level strategies for promoting rehabilitation policies that support the use of advanced devices.

PARTICIPANTS AND METHODS

This study comprised a descriptive ecological design. We used fiscal year 2022 data from the Cabinet Office dataset of regional variations in healthcare provision⁵⁾. This dataset included the standardized claim ratio (SCR), which was calculated based on the data from the National Database.

The SCR is a sex-adjusted and age-adjusted indicator designed to enable comparisons across regions and is defined as the ratio of the observed number of claims based on claim forms to the expected number of claims (observed/expected). An SCR of 100 represents the national average. SCRs higher than 100 indicate more claims than the national average. The SCR reflects claims billed by healthcare institutions located in a given region rather than healthcare utilization of residents. Further details are available from the Cabinet Office website⁵⁾.

Institutional review board approval was not required for this study. Written informed consent was not required because this study used only anonymized publicly available data. This study was not subject to the Ethical Guidelines for Life Science and Medical Research Involving Human Subjects in Japan.

We collected SCRs at the secondary medical area level. We used this unit because prefectural health plans in Japan are developed in secondary medical areas. Financial incentives were identified using the medical service code 180065070. These incentives applied to rehabilitation interventions that used approved devices for upper limb or lower limb motor impairment attributable to neurological conditions such as stroke or spinal cord injury. In principle, providers can claim the incentive once per month for up to 2 months from the date of onset of impairment. The approved devices include FES devices, robot-assisted gait training devices, and robotic upper-limb rehabilitation devices. As of fiscal year 2022, 20 devices have been approved³⁾.

As additional claims-based measures, we used the SCR for the number of hospitalized patients with cerebral infarction as a proxy for regional demand and the SCR for cerebrovascular rehabilitation fees as a proxy for service supply. We also collected SCR-based indicators of healthcare service capacity, including emergency and critical care hospitalization fees, the hyperacute stroke add-on, specific intensive care unit management fees, and convalescent rehabilitation ward hospitalization fees. We selected these items because they approximate the availability of healthcare resources across the care continuum from acute care services such as emergency and intensive care to post-acute care in convalescent rehabilitation wards for patients who may be eligible for the add-on.

First, we calculated summary statistics for the SCR of the financial incentive. Next, we calculated Pearson correlation coefficients between the SCR of the financial incentive and the SCRs of the other claim-based measures as the primary descriptive measure to summarize the observed linear pattern spanning secondary medical areas. We also calculated Spearman’s rank correlation coefficients to account for potential outliers and skewed distributions. As this study was designed as a descriptive ecological analysis of all nine secondary medical areas in Kanagawa Prefecture, the correlation coefficients were interpreted as descriptive and exploratory summaries of the observed regional patterns. We also performed post hoc sensitivity analysis, in which correlation analyses were repeated after excluding Yokosuka–Miura, which had the highest SCR in Kanagawa Prefecture. All analyses were performed using R version 4.5.2 (R Foundation for Statistical Computing, Vienna, Austria). Given the small number of secondary medical areas (n=9), all correlation analyses were interpreted as exploratory and hypothesis-generating. To make the uncertainty around these estimates explicit, we also report 95% confidence intervals (CI) and p-values; p-values are reported on a two-sided basis with α=0.05.

RESULTS

The SCR for Kanagawa Prefecture as a whole was 89.8. Table 1 shows the SCR of the financial incentive according to the secondary medical area in Kanagawa Prefecture. The mean SCR in Kanagawa was 64.0 (median, 36.8; 25th-75th percentile, 7.2–71.0). SCRs across secondary medical areas ranged from 3.3 to 223.3. Only Yokosuka–Miura and Yokohama had SCRs ≥100, which corresponded to the national average (Supplementary Fig 1).

Table 1. Standardized claim ratios for the financial incentive program according to each area in Kanagawa prefecture.

Area	Standardized claim ratio
Yokosuka–Miura	223.3
Yokohama	136.3
Sagamihara	71.0
Western Kanagawa (Ken-sei)	63.5
Eastern Shonan (Shonan-tobu)	36.8
Northern Kawasaki (Kawasaki-hokubu)	30.7
Central Kanagawa (Ken-o)	7.2
Southern Kawasaki (Kawasaki-nanbu)	3.8
Western Shonan (Shonan-seibu)	3.3

Terms in parentheses indicate the Japanese local reading (romanized) for each area name.

Table 2 summarizes the correlation analyses of the SCR of the financial incentive and SCRs of the other claim-based measures. For the SCR for inpatients with a principal diagnosis of cerebral infarction, no statistically significant association was observed with the SCR of the financial incentive (Pearson r=−0.31, 95% CI [−0.81, 0.45], p=0.416; Spearman ρ=−0.30, 95% CI [−0.89, 0.58], p=0.437). For the SCR for cerebrovascular rehabilitation volume, no statistically significant association was observed with the SCR of the financial incentive (Pearson r=0.19, 95% CI [−0.54, 0.76], p=0.628; Spearman ρ=0.25, 95% CI [−0.56, 0.74], p=0.521). Among indicators of the healthcare service capacity, no statistically significant association was observed between the SCR for convalescent rehabilitation ward inpatient fees and the SCR of the financial incentive (Pearson r=−0.26, 95% CI [−0.79, 0.49], p=0.507; Spearman ρ=−0.28, 95% CI [−0.93, 0.58], p=0.463). None of the reported correlations was statistically significant. In the sensitivity analysis excluding Yokosuka–Miura, the directions of the key correlations did not change materially, but their magnitudes weakened to varying degrees; for example, the Pearson correlation attenuated substantially for cerebral infarction inpatients (from r=−0.31 to r=−0.05), whereas the change was more modest for convalescent rehabilitation ward inpatient fees (from r=−0.26 to r=−0.18) (Supplementary Table 1).

Table 2. Pearson correlations between standardized claim ratios for the financial incentive program and related indicators.

Related Indicator (Standardized Claim Ratio)	Pearson			Spearman
	r	95% CI	p-value	rho	95% CI	p-value
Inpatients with a principal diagnosis of cerebral infarction	−0.31	[−0.81, 0.45]	0.416	−0.3	[−0.89, 0.58]	0.437
Cerebrovascular rehabilitation volume	0.19	[−0.54, 0.76]	0.628	0.25	[−0.56, 0.74]	0.521
Emergency admission fees	0.26	[−0.49, 0.79]	0.497	0.12	[−0.68, 0.86]	0.776
Hyperacute stroke add-on	−0.06	[−0.69, 0.63]	0.887	−0.3	[−0.89, 0.74]	0.437
Specified intensive care unit management fees	0.07	[−0.62, 0.7]	0.849	−0.07	[−0.71, 0.89]	0.88
Convalescent rehabilitation ward inpatient fees	−0.26	[−0.79, 0.49]	0.507	−0.28	[−0.93, 0.58]	0.463

All variables are standardized claim ratios (unitless). CI: confidence interval.

DISCUSSION

This study evaluated the uptake of the financial incentive in Kanagawa Prefecture using SCR data calculated based on data from the National Database. Uptake varied substantially across secondary medical areas, with a 67.7-fold difference between the highest and lowest SCRs. In the exploratory correlation analyses, the SCR for the add-on showed a non-significant negative association with the SCR for inpatients with a principal diagnosis of cerebral infarction and the SCR for convalescent rehabilitation ward inpatient fees.

Utilization of rehabilitation interventions using FES and robotic assistive devices varied across regions within Kanagawa Prefecture. A descriptive study that used nationwide health insurance claims data reported SCRs between 2.4 and 337.5³⁾. In our study, SCRs in Kanagawa Prefecture ranged from 3.3 to 223.3. Although the variation in this study was smaller than that reported by the previous study³⁾, it was consistent with the national-level variation reported by the previous study³⁾.

A systematic review of the diffusion of medical technologies reported that reimbursement design elements, including coding, coverage, and payment levels, can influence adoption and access⁶⁾. The observed geographic variations may reflect the eligibility rules of the program. Additionally, the variations may reflect interacting factors such as staff training and rehabilitation professionals’ knowledge of the evidence supporting the intervention⁷⁾. The SCR for the financial incentive showed a non-significant negative association with the SCR for inpatients with a principal diagnosis of cerebral infarction and the SCR for convalescent rehabilitation ward inpatient fees. According to a nationwide descriptive study³⁾, the SCR for inpatients with a principal diagnosis of cerebral infarction showed a non-significant positive association with the SCR for financial incentives. Neither correlation was statistically significant, and the difference in direction should be interpreted cautiously. In the sensitivity analysis excluding Yokosuka–Miura, the directions of the key correlations did not change, but their magnitudes weakened. These findings suggest that the observed coefficients in this small dataset should be interpreted with caution. The varying degrees of attenuation—most pronounced for cerebral infarction inpatients but more modest for convalescent rehabilitation ward inpatient fees—illustrate the inherent instability of ecological analyses based on a small number of units, in which observed correlations can be heavily driven by individual data points. The difference in direction may reflect sampling variability, but a genuine Kanagawa-specific pattern cannot be excluded. This issue warrants further investigation. Furthermore, the non-significant negative correlation with convalescent rehabilitation ward inpatient fees may reflect the program’s eligibility criteria. The financial incentive can be claimed once per month for patients who receive interventions within 2 months of symptom onset. Therefore, patients transferred to a convalescent rehabilitation ward as inpatients may have fewer opportunities to receive rehabilitation interventions using FES and robotic assistive devices. The fiscal year 2025 revision of the Japanese Stroke Treatment Guidelines indicated that the effects of robot-assisted training for upper limb function initiated within 3 months after onset may not differ from those of robot-assisted training for upper limb function initiated more than 3 months after onset⁸⁾. This suggests a potential mismatch between the guidelines and reimbursement rules, which restrict eligibility to within 2 months after onset and allow claims only once per month. Our findings suggest that the utilization of the financial incentive varies across areas and may reflect program design features, particularly the eligibility window and allowance of only one claim per month, in relation to the timing and frequency of device-assisted rehabilitation. Notably, no positive association was observed between the SCR for convalescent rehabilitation ward inpatient fees and program uptake. Although this finding should be interpreted as hypothesis-generating given the limited statistical power from the small number of secondary medical areas (n=9), it suggests that the potential impact of the eligibility window warrants further investigation.

This study had some limitations. First, this was a cross-sectional, descriptive, ecological study that did not include facility-level or individual-level data. Therefore, we could not rule out ecological fallacy or infer causal relationships. Second, because SCRs were calculated based on public health insurance claims, they did not capture care that was paid for through automobile liability insurance or out-of-pocket payments, and we could not account for these services. Third, because our analysis relied on claims data, we could not identify the types of devices used or the details of the intervention (e.g., frequency, intensity, or delivery). Fourth, if facilities provided device-based rehabilitation but did not submit claims for the add-on, then our estimates may have underestimated utilization. Fifth, because the analysis was conducted across nine secondary medical areas, statistical power was limited, estimates were imprecise, and correlation coefficients were sensitive to individual areas. The non-significant findings may therefore reflect a Type II error due to insufficient statistical power (n=9), rather than a definitive absence of association. These findings should be interpreted as exploratory and hypothesis-generating.

This descriptive ecological study found marked within-prefecture variations in claims for the financial incentive for device-assisted rehabilitation across secondary medical areas in Kanagawa Prefecture. The SCRs in several areas were below the national average, indicating uneven uptake under the current eligibility criteria. In Kanagawa Prefecture, areas with higher SCRs for convalescent rehabilitation ward inpatient fees tended to have lower program SCRs. Future reviews of the program should examine whether the current eligibility criteria align with current clinical practice guidelines and, if needed, consider refinements to the program design, including revision of the eligibility window.

Conference presentation

Part of this study was presented at the 42nd Kanagawa Physical Therapy Congress, which was conducted in Kanagawa, Japan (https://pt-kanagawa.or.jp/members/memberspage/abstract/6313/, accessed June 17, 2026).

Funding

The authors received no financial support for the research, authorship, or publication of this article.

Conflict of interest

The authors declare no conflict of interest.