Comparing the accuracy of survival prediction with two different prognostic scoring systems, in Indian Cancer Patients receiving Palliative Care/Palliative Radiotherapy

Abraham, Samuel (2010) Comparing the accuracy of survival prediction with two different prognostic scoring systems, in Indian Cancer Patients receiving Palliative Care/Palliative Radiotherapy. Masters thesis, Christian Medical College, Vellore.


Download (1MB) | Preview


AIM OF THE STUDY: To evaluate the Palliative Prognostic Score (Pap)and the Palliative Prognostic Index (PPI) in patients receiving palliative radiotherapy and /or palliative care. Primary Objective: To compare the sensitivity, specificity and positive predictive values of the two scales in prognosticating 1 week, one month, and three month survivals. The primary outcome was sensitivity and specificity for estimating 1 month survival. Secondary objectives: 1. 1.To document the median survivals of patients classified into good, intermediate, and better prognostic categories. 2. To assess the sensitivity and specificity of Score B when clinician’s prediction of survival is added to the score (modified score B). (Is modified score B better than the original score B?). 3. To assess the sensitivity and specificity of score A when laboratory tests are excluded from the scoring system (How accurate is a simpler score A?). MATERIALS AND METHODS: This was an observational cohort study done in patients receiving palliative radiotherapy. Concurrently a second cohort was followed up by the palliative care team, in the palliative care service. Inclusion Criteria: 1. All patients receiving palliative radiotherapy. Exclusion Criteria: 1. Patients less than eighteen years of age. 2. Patients or families who decline to participate. 3. Cases where a suitable informant could not be identified . 4. Cases where language or other barriers prevent adequate monthly telephone/postal/ clinic follow up. Pap Versus PPI Patients with incurable cancer who are receiving palliative radiotherapy were clinically assessed and scored at baseline using both the Palliative Prognostic Scale (PaP) and the Palliative Prognostic Index (PPI). The scoring systems classified patients into poor, intermediate or better prognostic groups at baseline. These scoring systems were based on the presence of symptoms such as dyspnoea, anorexia, oedema, the general performance status. A family member or contact person was contacted once a month to enquire about the patient’s condition. Survival or date of death was noted based on this information. This enabled us to know which patients have survived less than 1 week, 1 month or three months. The survival data thus obtained, was compared with the baseline categorization to assess and compare the sensitivity, specificity and positive predictive values of the PaP and the PPI. PaP(Palliative Prognostic Score): PaP is a scoring system developed to estimate the survival of terminally ill cancer patients. It was constructed by adding six variables which were individually predictors of survival. These variables were clinician prediction of survival, Karnofsky performance status, anorexia, dyspnoea, total white blood count and lymphocyte percentage. Patients who have not had a WBC count in the past two weeks will have a blood test done The PaP score is generated by applying a ‘weighted’ scoring system to each of these variables. Total scores range from 0 to 17.5 and patients were divided into three prognostic categories. The total scores for the three groups A,B,C were 0-5.5,5.6 -11,11.1-17.5 and the 30 days survival probabilities were >70%,30–70%,<30% respectively. CONCLUSION: 1) All scores were highly statistically significant prediction of the survivals in the population studied. These tests were less specific for individual patients, except those who are very sick. 2) Clinician estimate was as good as other scores for less than 1 month and more 3 month categories. It had a better sensitivity compared to the PPI score. Hence even with availability of these scores its value should not be negligible. 3) No score was superior to another as seen in the ROC curves. There was no loss of AUC by removing laboratory investigations from the Pap Score for predicting on month survival. 4) PPI was a simple score and useful especially in very sick patients, to predict one month survival. 5) Three month survival prediction is possibly more accurate by incorporating lab tests and clinician estimates than by PPI scores alone. The shortest median survivals were for PPI category C. The longest were for patients where the clinician estimated more than three months. 6) Bone secondaries had the longest survival by metastatic site and many patients could be considered for further systemic therapy to improve survival, especially if they have good Pap or PPI scores. 7) Patients with local recurrence had survivals in the range, median survival 11.3 weeks with confidence intervals from 5.25 to 17.35. Many of these patients are not suitable for further radiotherapy, and it is necessary to improve the quality of life (QOL) through good palliative care. 8) Patients with brain metastases, lung cancer and stomach cancer have shorter survivals, and decisions regarding long or expensive courses of treatment should be made judiciously.

Item Type: Thesis (Masters)
Uncontrolled Keywords: survival prediction ; two different prognostic scoring systems ; Indian Cancer Patients ; Palliative Care ; Palliative Radiotherapy.
Subjects: MEDICAL > Radio Therapy
Depositing User: Subramani R
Date Deposited: 22 Jun 2017 11:02
Last Modified: 23 May 2018 16:33

Actions (login required)

View Item View Item