Estimating the treatment effect in patients with gastric cancer in the presence of noncompliance.

Aim: In this study, these methods were used to estimate the treatment effect in patients with gastric cancer in the presence of noncompliance.

Background: In medical sciences, simple and advanced methods are used to estimate treatment effects in the presence of noncompliance.

Methods: This historical cohort study surveyed 178 patients with gastric cancer underwent chemotherapy alone (chemotherapy alone group) and 193 patients underwent surgery and chemotherapy (surgery plus chemotherapy group) from 2003 to 2007 at the Cancer Institute of Imam Khomeini Hospital (Tehran). Demographic and clinical characteristics were extracted from patients' hospital records. The survival of patients was calculated as being from diagnosis to death or to the end of the study. The treatment effect was estimated using three methods: treatment as a time-dependent covariate, IPCW, and Structural Nested Models using STATA and R software.

Results: Fifty-six patients (31.5%) who underwent chemotherapy and 69 patients (35.8%) who underwent surgery and chemotherapy died by the end of the study. The hazard ratio in group I compared to group II was estimated between 1.5 to 2.07 times based on the simple analysis method. The modified hazard ratio was estimated to be 1.21 (95% CI: 1.11-1.32) based on the SNM method. Surgery plus chemotherapy is superior to chemotherapy alone, and it improves the overall survival (OS) rate of gastric cancer patients.

Conclusion: Survival was improved in patients undergoing chemotherapy and surgery together compared to those undergoing chemotherapy alone. The results of the current study suggest that treatment effect can be estimated unbiasedly using the appropriate method.