Biostatistics in healthcare
Using the Survival Curve dataset tab located in the ((Framingham Heart Study dataset))
perform a Cox Proportional Hazards Regression Analysis to determine the survival time (risk of dying) for the Survival Curve data where the patients are divided into treatment groups. Upload the Excel spreadsheet into R Studio or perform the Cox Proportional Hazards Regression Analysis in Excel.
H0 The risk of dying is not related to the patient treatment group. (Null Hypothesis)
H1 The risk of dying is related to the patient treatment group. (Alternative Hypothesis)
Present your findings as a Survival Time chart in a Word document, with a title page,
-Introduction explaining why you would conduct a survival analysis,
-Discussion where you interpret the meaning of the survival analysis,
-Conclusion should be included.
Provide support for your statements with in-text citations from scholarly, peer-reviewed articles.
Cox Proportional Hazards Regression Analysis
The Cox proportional hazards regression analysis is a regression modelling method for survival analysis used in medical research to investigate association between predictor variables and patients’ survival time. The importance of survival analysis in the medical field has become progressively enormous with the multiplication of diseases and evolution of healthcare, as there is high clinical presentation of known characteristics or quantities which act as potential covariates influencing disease prognosis and outcome for patients, hence the phenomenon has incessantly evoked expeditions in medical research especially as immune diseases are on escalation (Park, et al., 2018). For example, the pre-COVID-19 rise in cancer infection and mortality scaled research in the field and it was found that survival analysis is unequivocally important in cancer patients’ care due to the unidirectional trajectory of the infection (Makar, et al., 2020). Wu, et al (2020), in an exemplary research with appended meaning in relation to the pandemic, used survival analysis to found that while the effects of age, gender, among other characteristics on prognosis of COVID-19 patients had been established by research previously, the contingency of blood glucose levels was assessed by this research study carried out in Wuhan, the epicentre of the current global pandemic – multivariate cox proportional hazards regression was used to assess the association between these blood glucose indices and patient prognosis – a sensitivity analysis that evaluated the association of admission blood glucose level with the risk of critical case of mortality of non-critical patient who does not suffer from diabetes at the point of admission, and a significant relation was established with a recommendation that could enhance clinical handling of COVID-19 through hierarchical management if implemented. This paper is a presentation of utilization of survival analysis, Cox proportional hazards regression analysis, to examine the length of time patients take before the event of death and assess the association with treatment measure provided. Intrinsically, the cox proportional-hazards model is used to explore the association between prognosis factors and death risk, as it is advantageous over the Kaplan-Meier method in that it allows for analysis of risk effects of the factors (Lee, & Lee, 2017). This paper presents a comparison of survival between patients on chemotherapy and those on placebo treatments over a period of 3.5 years.
The survival time charts are presented below.
Fig 1: Kaplan-Meier survival time chart.
Fig 2: Survival analysis comparison plot between Chemo and Placebo treatments.
The cox proportional regression analysis results are as in the outputs below.
The tests that check the overall model significance, the C-test shows us that two-thirds of the model is concordant, while the log-rank test shows that there is no significant difference in survival between the two treatment groups (chemotherapy and placebo) as the null hypothesis of no difference was not rejected.
The tabular summary is as follows:
The analysis shows that patients on placebo treatment have a hazard ratio of 3.42 – this means that someone placed on placebo treatment is 242% higher likelihood of dying than one placed on chemo. The exponentiated negative coefficient of 0.292 indicates that patients on chemo have a 0.3 times dying chance as compared to those on placebo treatment.
However, there is no significance, at 95% confidence level, of association between treatment and mortality (p-value>0.05) in the univariate cox proportional regression model.
Based on the data, the null hypothesis is not rejected at 95% confidence level. This essentially means that the data did not provide sufficient evidence to prove that there is indeed a difference between the two treatment groups in terms of survival, at 95% confidence level.
Lee, T., & Lee, M. (2017). Analysis of stage III proximal colon cancer using the Cox proportional hazards model. Journal of the Korean Data and Information Science Society, 28(2), 349-359.
Makar, G. S., Makar, M., Obinero, C., Davis, W., Gaughan, J. P., & Kwiatt, M. (2020). Refusal of cancer-directed surgery in patients with colon cancer: risk factors of refusal and survival data. Annals of surgical oncology, 1-11.
Park, J., Blackburn, B. E., Ganz, P. A., Rowe, K., Snyder, J., Wan, Y., … & Herget, K. (2018). Risk factors for cardiovascular disease among thyroid cancer survivors: findings from the Utah Cancer Survivors Study. The Journal of Clinical Endocrinology & Metabolism, 103(7), 2468-2477.
Wu, J., Huang, J., Zhu, G., Wang, Q., Lv, Q., Huang, Y., … & Liu, Y. (2020). Elevation of blood glucose level predicts worse outcomes in hospitalized patients with COVID-19: a retrospective cohort study. BMJ Open Diabetes Research and Care, 8(1), e001476.