DPC system
The DPC database, which has been in place since 2003, is a medical claims database of admissions to acute-care hospitals in Japan. The DPC system was adopted at 1730 hospitals in 2018, and covers approximately 83% of the acute-care beds in Japan 19. There are six distinct categories of diagnosis such as, “main diagnosis,” “main disease triggering admission,” “most resource-consuming diagnosis,” “second most resource-consuming diagnosis,” “comorbidities at admission,” and “complications after admission” in the DPC database. The DPC database also contains patients’ demographics, medical costs, procedures (including stent placement, colectomy, and ileostomy), and condition at discharge.20,21 The physicians input patients’ diagnosis into the DPC database according to International Classification of Diseases, 10th revision (ICD-10). The DPC database has been used for various clinical studies,.16,18 including those for colorectal cancer,17 and its diagnostic validity is widely recognized.20
Patients
This study included patients with obstructive colorectal cancer who were admitted to DPC-participating hospitals from April 2012 through March 2020 (Fig. 1). Colorectal cancer was identified using the ICD-10 code C18-20, which indicates colon cancer or rectal cancer, as the most resource-consuming diagnosis. Entries of colorectal cancer suspicious cases containing the word “suspicious” were excluded. We included patients with the following characteristics, (1) primary colorectal cancer, (2) not scheduled or urgent admissions, (3) containing the phrase “ileus” as main disease triggering admission or comorbidities at admission. We selected conditions with not scheduled or urgent adomissions to exclude patients who were discharged after stent placement and underwent radical surgery after being readmitted. We finally extracted patients who underwent interventions to release obstructions within 3 days after admission.
Data collection
We collected the following data on patients and clinical characteristics, procedures, and condition at discharge from the DPC database: age, sex, body mass index (BMI), smoking history, Charlson comorbidity index (CCI),22 hospital type (academic hospital or not), tumor (T) categories based on TNM classification,23 disease location (including cecum, ascending colon, transverse colon, descending colon, sigmoid colon, and rectum), condition at discharge (in-hospital death), medical costs (available data from 2016 to 2020), length of the hospital stay, and interventions to release obstruction such as SEMS placement and surgery (including colectomy and/or stoma creation). We defined in-hospital death, medical costs and the length of the hospital stay as clinical endpoints to evaluate the efficacy of SEMS.
Data analysis
We classified the eligible patients into five categories according to their age (≤ 49 years, 50‒59 years, 60‒69 years, 70‒79 years, ≥ 80 years) and into three categories according to body mass index (BMI) (underweight: < 18.5 kg/m2, normal range: 18.5‒24.9 kg/m2, overweight: > 25.0 kg/m2) based on the World Health Organization classification.24 The eligible patients were also divided into two groups according to the intervention to release obstruction due to obstructive colorectal cancer, as follows: SEMS group (stent placement) and surgery group (colectomy and/or stoma creation). The DPC database does not include the time of admission; thus, only the date of admission was recorded. Therefore, we extracted interventions which were conducted within 3 days after admission.
We conducted propensity score matching analysis to compare the efficacy of stent placement with that of surgery. We used the following variables for propensity score matching: sex, age categories, and BMI categories as described above, CCI, smoking history, hospital type, T categories and disease location. We subsequently compared them using rates of in-hospital death between the SEMS and surgery groups, using chi-square tests, and the length of hospitalization and medical costs of hospital stay, using Wilcoxon’s signed-rank test. The eligible patients after propensity score matching were divided into two groups based on the disease location, as follows: right-sided colon consisted of cecum, ascending and transverse colon, left-sided colon consisted of descending and sigmoid colon and rectum. We then compared the rate of in-hospital death, length of hospitalization, and medical costs in each group as well. We also performed a multivariate analysis using logistic regression analysis with the data before propensity score matching to identify clinical factors that affect in-hospital death.
Statistics
The threshold for statistical significance was P < 0.05. All analyses were performed using JMP Pro14 (SAS institute, Tokyo, Japan) software. All authors had access to the study data and reviewed and approved the final manuscript.
Ethics
The current study was conducted in accordance with the ethical standards of the Declaration of Helsinki. The study protocol was reviewed and approved by the Ethics Committee of Tohoku University Graduate School of Medicine (2019-1-415). Owing to the anonymous nature of the data, informed consent was waived for the approval of the Ethics Committee of Tohoku University Graduate School of Medicine (2019-1-415).