Regional variation in the use of catheter ablation for patients with arrhythmia in Japan

Abstract Background Regional variation in the use of percutaneous coronary intervention (PCI), especially when performed as an elective procedure, was observed in a previous study. The use of a developing technology, catheter ablation (CA), was compared between regions in Japan. Methods and Results The Diagnostic Procedure Combination data, which are publicly available, were used for the analysis. The number of CAs was summarized and the rates for CA and PCI were calculated based on the prefecture's population aged ≥40 years. A linear regression model was constructed to identify the factors associated with regional variation in the use of CA. The number of CAs performed per hospital consistently increased from 2009 to 2018. The mean rate of CA across Japan was 119 per 100 000 population aged ≥40 years in 2018. The highest CA rate was 166 per 100 000 and the lowest CA rate was 29 per 100 000 in 2018, while the highest and lowest PCI rates for angina per 100 000 were 361 and 88 in 2018, respectively. The significant factor associated with regional variation in the CA rate was the number of specialists. Conclusions A wide regional variation was observed in the use of CA for patients with arrhythmia in Japan. Further research is needed to generate evidence of CA for decision‐making as a treatment option and to appropriately deploy this health service regardless of where patients live.

Research to investigate causes and consequences of regional variation in health care has been actively conducted worldwide since Wennberg and Gittelson published their pivotal paper more than 40 years ago. 3 They addressed unwarranted regional variation in the deployment of resources and utilization of health-care services without clear patterns and reasons. 4 In their subsequent studies of the area of cardiovascular disease, they observed regional variation in the use of coronary artery revascularization and the treatment of patients with myocardial infarction (MI) in the late 1990s. 5,6 From recent studies of the United States, regional variation is still observed in terms of the use of PCI. 7 Furthermore, recent studies reported regional variation in newer technology and medical devices, such as transcatheter aortic valve implantation and cardiac implantable electrical devices. 8,9 As the aging population has grown, the prevalence of arrhythmia, including atrial fibrillation (AF), has increased. 10 An Australian study reported that the provision of catheter ablation (CA) for AF has increased markedly during the past decade. 11 It is important to assess whether access to this technology is secure in Japan. The objective of the present study is to investigate regional variation in the use of CA for patients with arrhythmia in Japan.

| Data source and study population
We obtained the data for the present study from the assessment report on the impact of implementing the Diagnostic Procedure Combination/Per-Diem Payment System (DPC/PDPS) system through the website of the MHLW. 12 DPC/PDPS is used as reimbursement scheme for the acute phase in-patient service in Japan, and the DPC code is defined by the most resource-consuming disease and its medical procedures during hospitalization. The assessment report contains the summary

| Statistical analysis
The numbers of CAs per hospital were summarized from 2009 to 2019; the rates of CA per 100 000 population were calculated for all 47 prefectures, based on the address of the DPC hospitals. To adjust for the age distribution of each prefecture, the number of CAs in each prefecture was divided by the prefecture's population aged ≥40 years because most of the patients receiving this procedure were over 40 years old. The wide regional variation was observed in the use of PCI for angina and acute MI in our previous study. 2 Therefore, as references, the PCI rates for angina and acute MI were calculated in much the same way, and the results were compared with the CA rate via a correlation coefficient. Patient migration was investigated by comparing the patients' address reported in the Patient Summary with the hospitals' address.
A linear regression model was developed to assess factors describing the regional medical supply supporting the use of CA for each prefecture. These factors included the number of arrhythmia specialists reported by the Japanese Heart Rhythm Society, 13 which was surrogate for the source of local supplies of CA and was described as per 100 000 population aged ≥40 years for each prefecture. The percentage of CAs performed in private hospitals was also used as a factor in describing the type of supply in the region.
Additionally, the PCI rates in angina and acute MI were included in the model. A P-value less than .05 was considered to indicate statistical significance. Of them, only 1240 patients (6%) migrated from other prefectures to receive CAs in those hospitals. More than 90% of patients were from same prefecture in 79 of 90 hospitals.

| RE SULTS
As presented in Table 1, the number of arrhythmia specialists per 100 000 population aged ≥40 years and the PCI rates in angina were significant factors for influencing the use of CA. The more these factors increased, the more CA was performed in each prefecture. On the other hand, other factors were not significantly associated with the use of CA for arrhythmia.

| D ISCUSS I ON
A wide regional variation in the use of CA for arrhythmia was observed across Japan. The magnitude of this regional variation was 5.7fold between the lowest and highest regions and 3.4-fold between the highest and second lowest regions across Japan, which was similar to the one in the use with PCI for angina patients, 4.1-fold in the present study. It was also observed that most of the patients received their CAs in the same regions they lived. From the regression model, the supply of health service resources, that is, the number of arrhythmia specialists, was significantly associated with regional variation.
The DPC data showed a nearly 2.4-fold increase in the performance of CA per hospital from 2009 to 2018, especially for patients with AF and atrial flutter ( Figure 1). Randomized controlled trials conducted in the US and Europe suggested that CA was more effective than antiarrhythmic drug therapy for patients with paroxysmal AF as first treatment. [14][15][16] The current Japanese guideline gives class I recommendation for patients with symptomatic paroxysmal AF to undergo CAs when antiarrhythmic drug therapy fails. 17 Corresponding with the evidence and the guideline, the recent result from the Japanese Catheter Ablation Registry for Atrial Fibrillation survey reports a significant increase in the first-line use of CA, as well as application to patients with a low frequency of AF attacks. 18 As indicated, a specialist is thought to be a key factor for regional variation. The more specialists who adopt CA as a treatment option in their clinical practice to follow the recent evidence, the larger the regional variation will be that is generated between regions with few specialists and more specialists. When a region has a larger number of specialists, it is likely that a high-volume center exists in the region, which would also be a large resource for other cardiac procedures, especially for elective procedures. Therefore, the CA rate and Considering our results, reasons for regional variation would be combinations of these inappropriate uses of health care.
Several limitations exist in the present study. According to the survey conducted by the Japanese Circulation Society, the prevalence of AF is different across regions, which also results in regional variation in CA. Thus, the rate of CA was adjusted by population aged ≥40 years to partially mitigate the difference in the prevalence of AF because of aging. Our data were summarized based on the address of the hospital and not the patients, but our data also showed that patients' migration was limited. The assessment reports do not include hospitals with CA performances less than 10 cases per year because of data privacy. In some regions, there exists these small volume hospitals, which may cause or extend the regional variation.
Additionally, the patient clinical characteristics were also not available in the assessment report, which may be associated with the regional variation in the CA rate.
In conclusion, the performance of CA has increased over the past 10 years; wide regional variation was observed in the use of CA for patients with arrhythmia in Japan. The study also suggested that regional variation was observed in elective procedures as similar regional variation was observed in the use of PCI for angina patients. Most of the patients underwent CA in hospitals within the same region they lived.
As patients with arrhythmia will increase as aging population is growing. Further research is needed to generate evidence of the use of CA for decision-making as a treatment option and to appropriately deploy this health service regardless of where patients live.

D I SCLOS U R E
Authors declare no conflict of interests for this article.