We conducted a national cohort study retrospectively from Taiwan’s National Health Insurance Research Database (NHIRD), which is regulated and maintained by the Data Science Centre of the Ministry of Health and Welfare (MOHW) of Taiwan. The NHIRD is the database of National Health Insurance (NHI) program in Taiwan, which included over 23 million Taiwanese, almost all of Taiwan’s residents. Thus, the database could be regarded as the medical record of entire Taiwan population. In this database, the clinical diagnosis of the patients were determined according to International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM). This study was approved by the Institutional Review Board of E-Da Hospital (EMRP-106-063) and Kaohsiung Medical University Hospital (KMUHIRB-E(I)-20180308) conducting in accordance with the Declaration of Helsinki. As this work is a retrospective, case control study gathering data from NHIRD, informed consent for study participation was not needed, approving by the Institutional Review Board.
2.2 Inclusion and Exclusion criteria
The subjects of this study were selected by using NHIRD data for the period from January 1, 2003 to December 31, 2013. All patients with the diagnosis of congenital collagen disease as Collagen group. It was established by ICD-9-CM code including 756.51, 756.59, 756.83, 756.89, 756.4 which represented for Osteogenesis imperfecta, Albright syndrome, Ehlers-Danlos syndrome, Amyotrophia congenita, Chondrodystrophy respectively. The index day was 2003/01/01. We excluded female patients, patients who was born after index day, patients who was died before index day, and patients receive hernia repair before 2003/01/01.
On the other hand, we randomly sampled 50000 active civilians as the general population. After excluding those included in our experimental cohort and those who had hernia repair before 2003/01/01, we used propensity score-matched (1:4) analysis by age and comorbidities including Charlson Comorbidity Index (CCI) score, Chronic Obstruction Pulmonary Disease (COPD), prostate disease and obese.
All the included patients were followed until their withdrawal record presented in the NHI or the end of our study period, December 31, 2013. The flow chart of selection criteria is shown in Fig. 1.
2.3 Study outcomes and covariates
The primary outcome of our study is accepting inguinal hernia repair. The diagnosis of hernia was confirmed by both ICD-9 CM code of a hernia (550.xx to 553.xx) and the surgical procedure code that included for inguinal hernia (53.00 to 53.05)[12, 24, 25]. In Taiwan, the cost of hernia repair, including both traditional open approach and laparoscopic assist, were fully covered by National Health Insurance. Thus, all the medical records were subjected to a detailed evaluation to ensure the diagnosis and the treatment were appropriated. Follow up time was defined as the time from beginning of inclusion, 2003/01/01, to receiving hernia repair or the end of the research.
In our study, patients' demographic data, including age and baseline comorbidities, were recorded. These baseline comorbidities that may be the potential risk factors of h in prior studies and possibly have affected our result were examined. During analysis, we assessed several independent variables as comorbidities, including prostate disease (ICD-9-CM code 600.x, 601.x, 602.x) ; obesity (ICD-9-CM code: 278.00, 278.01) and Chronic obstructive pulmonary disease (COPD) (ICD-9-CM code: 491.x-496.x, 501.x-504.x), which was reported as a risk factor for hernia repair[12, 27]. Comorbidities identified by an ICD-9 code within the NHIRD database before admission were included as comorbidities.
2.4 Statistical analysis
The baseline characteristics of the two groups (congenital collangen disease cohort and general population) were analyzed using descriptive statistics. The Kaplan − Meier (KM) curve was used to estimate the cumulative incidences of receiving hernia repair for the two groups, and the difference between two groups was estimated with log-rank test. Hazard ratios (HRs) with 95% confidence intervals (CI) were calculated using chi-squared, and multivariable-adjusted Cox proportional hazards models were used to test the association between these two groups. SPSS version 25 software (IBM, Chicago, IL, USA) was used for the statistical analysis. A P value less than 0.05 was considered statistically significant.