| Breast cancer is currently one of the most common cancer diseases worldwide,and the number of patients with the disease is increasing every year.Surgery is the most common treatment for breast cancer,but there is no uniformity in the choice of post-operative treatment,and how to assign post-operative treatment to patients is a challenge that clinicians are eager to address.This paper uses data from 2,211 breast cancer patients provided by a hospital in Shanghai to investigate the following two aspects.Firstly,it examines the benefits of different post-operative treatments(with or without chemotherapy)for patients? secondly,it examines personalised medical decision making,i.e.how to assign a reasonable post-operative treatment to each patient in order to achieve optimal treatment outcomes.To address these two issues,this paper combines medical advice from clinicians to construct causal diagrams between confounders,post-operative treatments and prognostic variables,and calculates causal effects based on survival outcome variables to measure the benefits of different treatments for patients? furthermore,this paper seeks to personalise healthcare decisions within a causal inference framework,using the method of maximising t-year survival IPSWKME was used to determine the optimal post-operative treatment plan,i.e.to maximise the patient’s t-year survival rate,and to solve the optimal personalised medicine guideline to provide some basis for clinical care.The final conclusion is that: overall,chemotherapy has a negative benefit for all patients? when patients are split by population,the causal effect of chemotherapy is negative for pre-menopausal patients,but the difference between chemotherapy and no chemotherapy for post-menopausal patients is small? the optimal personalised medical treatment criterion is solved from the perspective of maximising the ten-year survival rate of patients after surgery,at which point the ten-year survival rate after model selection is 96.25%. |
PDF Full Text Request