End of Life Care in Gynaecological Cancer

  • Sonali Kaushik

Student thesis: Doctoral ThesisDoctor of Medicine (MD)


End of life care is a basic human right. Most patients wish to die at home or in familiar surroundings. This is the largest study worldwide to explore aspects of the quality and equity of end of life care for gynaecological cancer patients in England using routine data sources.

The two hypotheses tested were:

‘Can routine data be used to describe aspects of end of life care for gynaecological cancer patients in England?’


‘the null hypothesis that there is no difference in the end of life care of gynaecological cancer patients in England from a socio-demographic or geographical perspective as measured by routine data’.

A quantitative approach was adopted for this study using ONS - HES linked data recorded for 71, 269 gynaecological cancer patients. Significant variables predicting death in hospital were, year of death, age group, deprivation status, frequency and length of elective and emergency admissions. Each subsequent emergency admission in the last month of life increased the odds of death in hospital by 2.4 times. A validated model to predict death in hospital was developed which correctly identified 73% of hospital deaths with a sensitivity of 75% and a specificity of 72%, and area under ROC curve of 0.77.

Following emergency admissions it is more likely that interventions take place. In the last month of life 56% of patients had at least one invasive procedure as inpatient or day case and of those admitted to hospital through elective or emergency routes, 72% had an intervention in the last month of life. Patients who had emergency ascites drainage had an increased risk of death in hospital of 11%. This thesis has shown significant geographical variations in the number of hospital deaths, emergency admissions and interventions in the last month of life for patients with gynaecological cancer.

This work demonstrates the usefulness of routine data to explore variations in end of life care for gynaecological cancer patients, to predict death in hospital following an emergency admission and the potential for monitoring service improvement. This thesis also raises some important policy implications. It highlights variation in ‘proxy’ outcomes as described by place of death and patient experience represented by emergency admissions, length of stay and procedures in the last month of life.
Date of Award25 Sept 2018
Original languageEnglish
Awarding Institution
  • The University of Bristol
SupervisorKathleen M Gillespie (Supervisor), Susan Glew (Supervisor), Robert Gornall (Supervisor) & Julia Verne (Supervisor)

Cite this