A simple and accurate prediction model to estimate the intrahospital mortality risk of hospitalised cancer patients

Development and validation of a stacking ensemble model for death prediction in the Chinese Longitudinal Healthy Longevity Survey (CLHLS)

OBJECTIVE

This study aimed to develop and validate a mortality risk prediction model for older people based on the Chinese Longitudinal Healthy Longevity Survey using the stacking ensemble strategy.

MATERIAL AND METHODS

A total of 12,769 participants aged 65 or more at baseline were included. Ensemble machine learning models were applied to develop a mortality prediction model. We selected three base learners, including logistic regression, eXtreme Gradient Boosting, and Categorical + Boosting, and used logistic regression as the meta-learner. The primary outcome was five-year survival. Variable importance was evaluated by the SHapley Additive exPlanations method.

RESULTS

The mean age at baseline was 88, and 57.8 % of participants were women. The CatBoost model performed the best among the three base learners, the area under the receiver operating characteristics curve (AUC) reached 0.8469 (95%CI: 0.8345-0.8593), and the stacking ensemble model further improved the discrimination ability (AUC = 0.8486, 95%CI: 0.8367-0.8612, P = 0.046). Conventional logistic regression had comparable performance (AUC = 0.8470, 95 % CI: 0.8346-0.8595). Older age, higher scores for self-care activities of daily living, being male, higher objective physical performance capacity scores, not undertaking housework, and lower scores on the Mini-Mental State Examination contributed to higher risk.

CONCLUSIONS

We successfully constructed and validated a few death risk prediction models for a Chinese population of older adults. While the stacking ensemble approach had the best prediction performance, the improvement over conventional logistic regression was insubstantial.

Prevalence and predictors for 72-h mortality after transfer to acute palliative care unit

Purpose

Accurate prediction of survival is important to facilitate clinical decision-making and improve quality of care at the end of life. While it is well documented that survival prediction poses a challenge for treating physicians, the need for clinically valuable predictive factors has not been met. This study aims to quantify the prevalence of patient transfer 72 h before death onto the acute palliative care unit in a tertiary care center in Switzerland, and to identify factors predictive of 72-h mortality.

Methods

All patients hospitalized between January and December 2020 on the acute palliative care unit of the Competence Center Palliative Care of the Department of Radiation Oncology at the University Hospital Zurich were assessed. Variables were retrieved from the electronic medical records. Univariable and multivariable logistic regressions were used to identify predictors of mortality.

Results

A total of 398 patients were screened, of which 188 were assessed. Every fifth patient spent less than 72 h on the acute palliative care unit before death. In multivariable logistic regression analysis, predictors for 72-h mortality after transfer were no prior palliative care consult (p = 0.011), no advance care directive (p = 0.044), lower performance status (p = 0.035), lower self-care index (p = 0.003), and lower blood albumin level (p = 0.026).

Conclusion

Late transfer to the acute palliative care unit is not uncommon, which can cause additional distress to patients and caretakers. Though clinically practical short-term survival predictors remain largely unidentified, early integration of palliative care should be practiced more regularly in patients with life-limiting illness.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00520-022-07075-6.

A catalogue of tools and variables from crisis and routine care to support decision-making about allocation of intensive care beds and ventilator treatment during pandemics: Scoping review

PURPOSE

This scoping review sought to identify objective factors to assist clinicians and policy-makers in making consistent, objective and ethically sound decisions about resource allocation when healthcare rationing is inevitable.

MATERIALS AND METHODS

Review of guidelines and tools used in ICUs, hospital wards and emergency departments on how to best allocate intensive care beds and ventilators either during routine care or developed during previous epidemics, and association with patient outcomes during and after hospitalisation.

RESULTS

Eighty publications from 20 countries reporting accuracy or validity of prognostic tools/algorithms, or significant correlation between prognostic variables and clinical outcomes met our eligibility criteria: twelve pandemic guidelines/triage protocols/consensus statements, twenty-two pandemic algorithms, and 46 prognostic tools/variables from non-crisis situations. Prognostic indicators presented here can be combined to create locally-relevant triage algorithms for clinicians and policy makers deciding about allocation of ICU beds and ventilators during a pandemic. No consensus was found on the ethical issues to incorporate in the decision to admit or triage out of intensive care.

CONCLUSIONS

This review provides a unique reference intended as a discussion starter for clinicians and policy makers to consider formalising an objective a locally-relevant triage consensus document that enhances confidence in decision-making during healthcare rationing of critical care and ventilator resources.

Applied Informatics Decision Support Tool for Mortality Predictions in Patients With Cancer

PURPOSE

With rapidly evolving treatment options in cancer, the complexity in the clinical decision-making process for oncologists represents a growing challenge magnified by oncologists' disposition of intuition-based assessment of treatment risks and overall mortality. Given the unmet need for accurate prognostication with meaningful clinical rationale, we developed a highly interpretable prediction tool to identify patients with high mortality risk before the start of treatment regimens.

METHODS

We obtained electronic health record data between 2004 and 2014 from a large national cancer center and extracted 401 predictors, including demographics, diagnosis, gene mutations, treatment history, comorbidities, resource utilization, vital signs, and laboratory test results. We built an actionable tool using novel developments in modern machine learning to predict 60-, 90- and 180-day mortality from the start of an anticancer regimen. The model was validated in unseen data against benchmark models.

RESULTS

We identified 23,983 patients who initiated 46,646 anticancer treatment lines, with a median survival of 514 days. Our proposed prediction models achieved significantly higher estimation quality in unseen data (area under the curve, 0.83 to 0.86) compared with benchmark models. We identified key predictors of mortality, such as change in weight and albumin levels. The results are presented in an interactive and interpretable tool ( www.oncomortality.com ).

CONCLUSION

Our fully transparent prediction model was able to distinguish with high precision between highest- and lowest-risk patients. Given the rich data available in electronic health records and advances in machine learning methods, this tool can have significant implications for value-based shared decision making at the point of care and personalized goals-of-care management to catalyze practice reforms.

Febrile neutropenia (FN) occurrence outside of clinical trials: occurrence and predictive factors in adult patients treated with chemotherapy and an expected moderate FN risk. Rationale and design of a real-world prospective, observational, multinational study

Background

Febrile neutropenia (FN) is a common occurrence during chemotherapy. Granulocyte colony-stimulating factors (G-CSFs) can significantly reduce the risk of FN. International guidelines recommend G-CSF for patients receiving chemotherapy with FN risk of ≥20% or 10% to 20% with defined risk factors. Prophylaxis is not typically recommended for FN risk of < 10%; however, few studies have investigated FN incidence in lower-risk patients in real-world settings and tried to identify higher-risk subgroups.

Methods

This real-world prospective, observational, multinational study aims to estimate the rate of development of FN with a chemotherapy line expected to be associated with a 10% to 20% risk of FN. Eligible patients (> 18 years of age) will have a solid tumour or Hodgkin/non-Hodgkin lymphoma and a planned chemotherapy regimen with expected risk of FN of 10% to 20% (according to published guidelines). Patients will be observed for the duration of the chemotherapy line (first cycle administered without FN prophylaxis). Primary endpoint is incidence of FN after the first chemotherapy cycle. Secondary outcomes include: FN-associated morbidity and mortality; time to first FN occurrence; other FN risk factors and impact of FN on quality of life. A risk model using occurrence of FN as a binary outcome will be developed. Data will be stratified by age, comorbidities and other risk factors.

Discussion

This study will provide insight into the real FN risk for common chemotherapy regimens and predictive factors for FN, including patients generally excluded from randomised clinical trials, from which reported FN rates have been variable. This study builds on knowledge of predictive factors from other research and will provide information on patients with 10% to 20% FN risk.

Development of a User-Friendly Graphic Tool to Estimate Individualized Survival Curves for Advanced Cancer Patients in Hospice Care

This prospective study aimed to develop an individualized prognostic tool for predicting the survival probability at any given point for a hospice patient with advanced cancer. A total of 286 patients with advanced cancer were included in the study. Median observational time was 18 days (range: 1 to 60 days). Cox proportional hazards regression analysis revealed that faster heart rate (hazard ratio [HR] = 1.01), jaundice (HR = 2.32), poorer performance status (HR = 2.01), and antifungal treatment (HR = 1.62) were independent predictors of shorter survival time. Patients with infections who received aminoglycoside treatments (HR = 0.45) were associated with longer survival times. Based on this model, we could construct a covariate-adjusted individualized survival curve for a given patient according to his or her clinical condition. This user-friendly tool for estimating the survival probability of patients with advanced cancer in hospice settings could facilitate clinical decision making and medical care planning.

Objective palliative prognostic score among patients with advanced cancer

CONTEXT

The accurate prediction of survival is one of the key factors in the decision-making process for patients with advanced illnesses.

OBJECTIVES

This study aimed to develop a short-term prognostic prediction method that included such objective factors as medical history, vital signs, and blood tests for use with patients with advanced cancer.

METHODS

Medical records gathered at the admission of patients with advanced cancer to the hospice palliative care unit at a tertiary hospital in Taiwan were reviewed retrospectively. The records included demographics, history of cancer treatments, performance status, vital signs, and biological parameters, Multivariate logistic regression analyses and receiver operating characteristic curves were used for model development.

RESULTS

The Objective Palliative Prognostic Score was determined by using six objective predictors identified by multivariate logistic regression analysis. The predictors include heart rate >120/min, white blood cells >11,000/mm(3), platelets <130,000/mm(3), serum creatinine level >1.3 mg/dL, serum potassium level >5 mg/dL, and no history of chemotherapy. The area under the receiver operating characteristic curve used to predict seven-day survival was 82.0% (95% confidence interval 75.2%-88.8%). If any three predictors of the six were reached, death within seven days was predicted with 68.8% sensitivity, 86.0% specificity, 55.9% positive predictive value, and 91.4% negative predictive value.

CONCLUSION

The Objective Palliative Prognostic Score consists of six objective predictors for the estimation of seven-day survival among patients with advanced cancer and showed a relatively high accuracy, specificity, and negative predictive value. Objective signs, such as vital signs and blood test results, may help clinicians make decisions at the end of life.

A brief review of survival prediction of advanced cancer patients

Survival prediction of advanced cancer patients remains an important task for palliative clinicians. It has transformed from an art form into a more scientific branch of the discipline with the evolution of palliative medicine and use of statistical estimates of survival. Both clinician predicted survival and actuarial estimation of survival have their uses and drawbacks. This article gives a practical and quick summary of the pros and cons of clinician survival prediction and actuarial-based prognostic tools used at the bedside.

Predicting survival time in noncurative patients with advanced cancer: a prospective study in China

BACKGROUND

Accurate prediction of prognosis for cancer patients is important for good clinical decision making in therapeutic and care strategies. The application of prognostic tools and indicators could improve prediction accuracy.

OBJECTIVE

This study aimed to develop a new prognostic scale to predict survival time of advanced cancer patients in China.

METHODS

We prospectively collected items that we anticipated might influence survival time of advanced cancer patients. Participants were recruited from 12 hospitals in Shanghai, China. We collected data including demographic information, clinical symptoms and signs, and biochemical test results. Log-rank tests, Cox regression, and linear regression were performed to develop a prognostic scale.

RESULTS

Three hundred twenty patients with advanced cancer were recruited. Fourteen prognostic factors were included in the prognostic scale: Karnofsky Performance Scale (KPS) score, pain, ascites, hydrothorax, edema, delirium, cachexia, white blood cell (WBC) count, hemoglobin, sodium, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), and alkaline phosphatase (ALP) values. The score was calculated by summing the partial scores, ranging from 0 to 30. When using the cutoff points of 7-day, 30-day, 90-day, and 180-day survival time, the scores were calculated as 12, 10, 8, and 6, respectively.

CONCLUSIONS

We propose a new prognostic scale including KPS, pain, ascites, hydrothorax, edema, delirium, cachexia, WBC count, hemoglobin, sodium, total bilirubin, direct bilirubin, AST, and ALP values, which may help guide physicians in predicting the likely survival time of cancer patients more accurately. More studies are needed to validate this scale in the future.

The median informs the message: accuracy of individualized scenarios for survival time based on oncologists' estimates

PURPOSE

To determine the accuracy and usefulness of oncologists' estimates of survival time in individual patients with advanced cancer.

PATIENTS AND METHODS

Twenty-one oncologists estimated the "median survival of a group of identical patients" for each of 114 patients with advanced cancer. Accuracy was defined by the proportions of patients with an observed survival time bounded by prespecified multiples of their estimated survival time. We expected 50% to live longer (or shorter) than their oncologist's estimate (calibration), 50% to live from half to double their estimate (typical scenario), 5% to 10% to live ≤ one quarter of their estimate (worst-case scenario), and 5% to 10% to live three or more times their estimate (best-case scenario). Estimates within 0.67 to 1.33 times observed survival were deemed precise. Discriminative value was assessed with Harrell's C-statistic and prognostic significance with proportional hazards regression.

RESULTS

Median survival time was 11 months. Oncologists' estimates were relatively well-calibrated (61% shorter than observed), imprecise (29% from 0.67 to 1.33 times observed), and moderately discriminative (Harrell C-statistic 0.63; P = .001). The proportion of patients with an observed survival half to double their oncologist's estimate was 63%, ≤ one quarter of their oncologist's estimate was 6%, and three or more times their oncologist's estimate was 14%. Independent predictors of observed survival were oncologist's estimate (hazard ratio [HR] = 0.92; P = .004), dry mouth (HR = 5.1; P < .0001), alkaline phosphatase more than 101 U/L (HR = 2.8; P = .0002), Karnofsky performance status ≤ 70 (HR = 2.3; P = .007), prostate primary (HR = 0.23; P = .002), and steroid use (HR = 2.4; P = .02).

CONCLUSION

Oncologists' estimates of survival time were relatively well-calibrated, moderately discriminative, independently associated with observed survival, and a reasonable basis for estimating worst-case, typical, and best-case scenarios for survival.

Prediction of In-Hospital Mortality of Patients With Advanced Cancer Using the Chuang Prognostic Score

The prediction of in-hospital mortality may help in improving end-of-life care for patients dying of cancer. The Chuang Prognostic Score (CPS) was developed to predict survival of terminally ill patients with cancer. The CPS was assessed in 61 hospitalized adult patients with advanced cancer. Using a CPS cutoff point of ≥6, in-hospital mortality was predicted with 71% positive predictive value, 91% negative predictive value, 75% sensitivity, 89% specificity, and 85% overall accuracy. The patients were divided according to the CPS score into 3 groups (Group 1: CPS < 3.5, Group 2: CPS ≥ 3.5-<6, and Group 3: CPS ≥ 6) with a median survival of not reached, 118 days, and 16 days, respectively ( P < .001). The CPS may be useful in predicting in-hospital mortality of hospitalized patients with advanced cancer.

Development and cross-validation of the in-hospital mortality prediction in advanced cancer patients score: a preliminary study

PURPOSE

Acute palliative care units (APCUs) provide intensive symptom support and transition of care for advanced cancer patients. Better understanding of the predictors of in-hospital mortality is needed to facilitate program planning and patient care. In this prospective study, we identified predictors of APCU mortality, and developed a four-item In-hospital Mortality Prediction in Advanced Cancer Patients (IMPACT) predictive model.

METHODS

Between April and July 2010, we documented baseline demographics, the Edmonton Symptom Assessment Scale (ESAS), 80 clinical signs including known prognostic factors, and 26 acute complications on admission in consecutive APCU patients. Multivariate logistic regression analysis was used to identify factors for inclusion in a nomogram, which was cross-validated with bootstrap analysis.

RESULTS

Among 151 consecutive patients, the median age was 58, 13 (9%) had hematologic malignancies, and 52 (34%) died in the hospital. In multivariate analysis, factors associated with in-hospital mortality were advanced education (odds ration [OR]=11.8, p=0.002), hematologic malignancies (OR=8.6, p=0.02), delirium (OR=4.3, p=0.02), and high ESAS global distress score (OR=20.8, p=0.01). In a nomogram based on these four factors, total scores of 6, 10, 14, 17, and 21 corresponded to a risk of death of 10%, 25%, 50%, 75%, and 90%, respectively. The model has 92% sensitivity and 88% specificity for predicting patients at low/high risk of dying in the hospital, and a receiver-operator characteristic curve concordance index of 83%.

CONCLUSIONS

Higher education was associated with increased utilization of the interdisciplinary palliative care unit until at the end of life. Patients with higher symptom burden, delirium, and hematologic malignancies were also more likely to require APCU care until death.

A community-based approach to translational research addressing breast cancer disparities

Disparities in breast cancer survival rates suggest that biological processes contribute. Translational research addressing health disparities would benefit from using a community-based participatory approach (CBPR) to examine biological processes commonly seen as the proximal causes of illness as well as behavioral and social-ecological "causes of the causes" within an integrated conceptual framework. This paper describes a CBPR study that explored perceptions regarding breast cancer relevant behaviors, and the application of the study's results to develop translational research. Data from eight focus groups of African American (n = 29) and Caucasian women (n = 27) were analyzed, using the framework of the social-ecological model. Nutrition and physical activity were valued over screening and research participation. Treatment of illness was emphasized over prevention. Women's perspectives are presented within a framework that facilitated the collaborative development of translational research to examine associations among biological, behavioral, and societal processes contributing to disparities.

Validation of prognostic scores for survival in cancer patients beyond first-line therapy

Background

We aimed to validate prognostic scores for survival in patients undergoing chemotherapy for advanced or metastatic cancer after first-line treatment.

Methods

We previously described two models with good prognostic value based on a combination of Performance Status (PS) and either lactate dehydrogenase (LDH) level or lymphocyte count. These factors were evaluated for their ability to predict overall survival (OS) in a prospective cohort of 299 patients. Clinical and blood parameters were prospectively recorded. Candidate prognostic factors for OS with 0.05 significance level in univariate analysis were included in a multivariate Cox model.

Results

Median age was 59 years (range: 26-85). Primary tumor sites were breast (45%), lung (15%), ovaries (11%) and others (29%). The number of metastatic sites was 1 (29%), 2 (48%), >2 (23%). Median follow-up and median OS were 12 and 6 months, respectively. Multiple regression analysis confirmed that PS >1, lymphocyte count ≤700/μL and LDH >600 UI/L were independent predictors of short OS, as well as interleukin 6 (IL-6) level, serum albumin concentration and platelet count.

Conclusions

Prognostic scores using PS plus LDH level or PS plus lymphocyte count were validated for predicting survival in metastatic cancer patients in relapse beyond first-line treatment. A score combining PS, LDH, lymphocyte and platelet count, serum albumin and IL-6 level was superior in determining patients' prognosis.

Discharge outcomes and survival of patients with advanced cancer admitted to an acute palliative care unit at a comprehensive cancer center

BACKGROUND

Acute palliative care units (APCUs) are new programs aimed at integrating palliative and oncology care. Few outcome studies from APCUs are available.

OBJECTIVES

We examined the frequency, survival, and predictors associated with home discharge and death in our APCU.

METHODS

All patients discharged from the APCU between September 1, 2003 and August 31, 2008 were included. Demographics, cancer diagnosis, discharge outcomes, and overall survival from discharge were retrieved retrospectively.

RESULTS

The 2568 patients admitted to APCU had the following characteristics: median age, 59 years (range, 18-101); male, 51%; median hospital stay, 11 days; median APCU stay, 7 days; and median survival 21 days (95% confidence interval [CI] 19-23 days). Five hundred ninety-two (20%), 89 (3%), and 1259 (43%) patients were discharged to home, health care facilities, and hospice, respectively, with a median survival of 60, 29, and 14 days, respectively (p < 0.001). Nine hundred fifty-eight (33%) patients died during admission (median stay, 11 days). Compared to hospice transfers, home discharge (hazard ratio = 0.35, 95% CI 0.30-0.41, p < 0.001) was associated with longer survival in multivariate analysis, with a 6-month survival of 22%. Multivariate logistic regression revealed that male gender, specific cancer primaries, and admissions from oncology units were associated with death in the APCU, while younger age and direct admissions to the APCU were associated with home discharge.

CONCLUSIONS

Our APCU serves patients with advanced cancer with diverse clinical characteristics and survival, and discharged home a significant proportion with survival greater than 6 months. RESULTS from this simultaneous care program suggest a pattern of care different from that of traditional hospice and palliative care services.

Association between cardiovascular autonomic functions and time to death in patients with terminal hepatocellular carcinoma

CONTEXT

A better time-to-death (TTD) prediction can facilitate decision-making processes related to plans for providing effective end-of-life care for patients in hospice wards.

OBJECTIVE

To explore the association of cardiovascular autonomic functions with TTD in patients with terminal hepatocellular carcinoma.

METHODS

A prospective study was conducted with 33 patients with hepatocellular carcinoma recruited from the hospice ward of a regional hospital in Chiayi county, Taiwan. Serum creatinine, serum glutamate oxaloacetate transaminase, serum glutamate pyruvate transaminase, blood urea nitrogen (BUN), and serum albumin were measured on the admission day. Cardiovascular autonomic functions were evaluated by frequency-domain measures of heart rate variability (HRV) on admission.

RESULTS

TTD was significantly associated with total spectrum power (TP) (r=0.55, P=0.001) and high frequency (HF power) (r=0.44, P=0.010) of HRV measurement. The accuracy of within-one-week TTD prediction was 67% for TP and HF power. The accuracy of within-two-week TTD prediction was 82% for TP and 73% for HF. In addition, TTD of the patients was also significantly associated with serum creatinine (r=-0.42, P=0.015), serum albumin (r=-0.46, P=0.007), and BUN (r=-0.44, P=0.010).

CONCLUSION

This is the first study to evaluate the association between cardiovascular autonomic functions and TTD in patients with terminal hepatocellular carcinoma. The inclusion of HRV measurement in prognostic models may improve accuracy in TTD prediction and, hence, facilitate medical decision making in hospice care.

A computer-assisted model for predicting probability of dying within 7 days of hospice admission in patients with terminal cancer

OBJECTIVE

The aim of the present study is to compare the accuracy in using laboratory data or clinical factors, or both, in predicting probability of dying within 7 days of hospice admission in terminal cancer patients.

METHODS

We conducted a prospective cohort study of 727 patients with terminal cancer. Three models for predicting the probability of dying within 7 days of hospice admission were developed: (i) demographic data and laboratory data (Model 1); (ii) demographic data and clinical symptoms (Model 2); and (iii) combination of demographic data, laboratory data and clinical symptoms (Model 3). We compared the models by using the area under the receiver operator curve using stepwise multiple logistic regression.

RESULTS

We estimated the probability dying within 7 days of hospice admission using the logistic function, P = Exp(betax)/[1 + Exp(betax)]. The highest prediction accuracy was observed in Model 3 (82.3%), followed by Model 2 (77.8%) and Model 1 (75.5%). The log[probability of dying within 7 days/(1 - probability of dying within 7 days)] = -6.52 + 0.77 x (male = 1, female = 0) + 0.59 x (cancer, liver = 1, others = 0) + 0.82 x (ECOG score) + 0.59 x (jaundice, yes = 1, no = 0) + 0.54 x (Grade 3 edema = 1, others = 0) + 0.95 x (fever, yes = 1, no = 0) + 0.07 x (respiratory rate, as per minute) + 0.01 x (heart rate, as per minute) - 0.92 x (intervention tube = 1, no = 0) - 0.37 x (mean muscle power).

CONCLUSIONS

We proposed a computer-assisted estimated probability formula for predicting dying within 7 days of hospice admission in terminal cancer patients.

Prediction of survival time in advanced cancer: a prognostic scale for Chinese patients

This study reports the development of a simple Chinese Prognostic Scale (ChPS) for predicting survival in advanced cancer patients. Data relating to 1,019 advanced cancer patients referred to a palliative home care service were retrospectively analyzed. The records were divided into two sets using stratified random sampling: 80% as a "training set" for developing the scale and 20% as a "testing set" for validating it. Demographic data, symptoms/signs, Karnofsky Performance Status (KPS), quality of life (QOL), and survival time were statistically analyzed to create the scale. In the training set, a total of 10 prognostic factors were determined: weight loss, nausea, dysphagia, dyspnea, edema, cachexia, dehydration, gender, KPS, and QOL. The ChPS score was calculated for each case by summing the partial scores of prognostic factors, ranging from 0 (no altered variables) to 124 (maximal altered variables). The score for a cutoff point of three months' survival was 28 (95% confidence interval: 26.6, 28.9). When scores were more than 28, survival appeared to be usually less than three months. The accuracy rate was 69.4% in the training set and 65.4% in the testing set. In conclusion, it is possible with this prognostic scale to guide physicians in predicting more accurately the likely survival time of Chinese cancer patients, and to help policy makers in establishing appropriate referral for hospice care.

A proposed prognostic 7-day survival formula for patients with terminal cancer

BACKGROUND

The ability to identify patients for hospice care results in better end-of-life care. To develop a validated prognostic scale for 7-day survival prediction, a prospective observational cohort study was made of patients with terminal cancer.

METHODS

Patient data gathered within 24 hours of hospital admission included demographics, clinical signs and symptoms and their severity, laboratory test results, and subsequent survival data. Of 727 patients enrolled, data from 374 (training group) was used to develop a prognostic tool, with the other 353 serving as the validation group.

RESULTS

Five predictors identified by multivariate logistic regression analysis included patient's cognitive status, edema, ECOG performance status, BUN and respiratory rate. A formula of the predictor model based on those five predictors was constructed. When probability was >0.2, death within 7 days was predicted in the training group and validation group, with sensitivity of 80.9% and 71.0%, specificity of 65.9% and 57.7%, positive predictive value of 42.6% and 26.8%, and negative predictive value (NPV) of 91.7% and 90.1%, respectively.

CONCLUSION

This predictor model showed a relatively high sensitivity and NPV for predicting 7-day survival among terminal cancer patients, and could increase patient satisfaction by improving end-of-life care.

Construction of a new, objective prognostic score for terminally ill cancer patients: a multicenter study

GOALS OF WORK

The goal of this study was to develop a new, objective prognostic score (OPS) for terminally ill cancer patients based on an integrated model that includes novel objective prognostic factors.

MATERIALS AND METHODS

A multicenter study of 209 terminally ill cancer patients from six training hospitals in Korea were prospectively followed until death. The Cox proportional hazard model was used to adjust for the influence of clinical and laboratory variables on survival time. The OPS was calculated from the sum of partial scores obtained from seven significant predictors determined by the final model. The partial score was based on the hazard ratio of each predictor. The accuracy of the OPS was evaluated.

MAIN RESULTS

The overall median survival was 26 days. On the multivariate analysis, reduced oral intake, resting dyspnea, low performance status, leukocytosis, elevated bilirubin, elevated creatinine, and elevated lactate dehydrogenase (LDH) were identified as poor prognostic factors. The range of OPS was from 0.0 to 7.0. For the above cutoff point of 3.0, the 3-week prediction sensitivity was 74.7%, the specificity was 76.5%, and the overall accuracy was 75.5%.

CONCLUSIONS

We developed the new OPS, without clinician's survival estimates but including a new prognostic factor (LDH). This new instrument demonstrated accurate prediction of the 3-week survival. The OPS had acceptable accuracy in this study population (training set). Further validation is required on an independent population (testing set).

Palliative medicine review: prognostication

Prognostication, along with diagnosis and treatment, is a traditional core clinical skill of the physician. Many patients and families receiving palliative care want information about life expectancy to help plan realistically for their futures. Although underappreciated, prognosis is, or at least should be, part of every clinical decision. Despite this crucial role, expertise in the art and science of prognostication diminished during the twentieth century, due largely to the ascendancy of accurate diagnostic tests and effective therapies. Consequently, "Doctor, how long do I have?" is a question most physicians find unprepared to answer effectively. As we focus on palliative care in the twenty-first century, prognostication will need to be restored as a core clinical proficiency. The discipline of palliative medicine can provide leadership in this direction. This paper begins by discussing a framework for understanding prognosis and how its different domains might be applied to all patients with life limiting illness, although the main focus of the paper is predicting survival in patients with cancer. Examples of prognostic tools are provided, although the subjective assessment of prognosis remains important in the terminally ill. Other issues addressed include: the importance of prognostication in terms of clinical decision-making, discharge planning, and care planning; the impact of prognosis on hospice referrals and patient/family satisfaction; and physicians' willingness to prognosticate.

Predicting survival in patients with advanced disease

Prognostication is an important clinical skill for all clinicians, particularly those clinicians working with patients with advanced cancer. However, doctors can be hesitant about prognosticating without a fundamental understanding of how to formulate a prognosis more accurately and how to communicate the information with honesty and compassion. Irrespective of the underlying type of malignancy, most patients with advanced cancer experience a prolonged period of gradual decline (months/years) before a short phase of accelerated decline in the last month or two. The main indicators of this final phase are poor performance status, weight loss, symptoms such as anorexia, breathlessness or confusion and abnormalities on laboratory parameters (e.g. high white cell count, lymphopaenia, hyopalbuminaemia, elevated lactate dehydrogenase or C-reactive protein). The clinical estimate of survival remains a powerful independent prognostic indicator, often enhanced by experience, but research has only begun to understand the different biases affecting clinicians' estimates. More recent research has shown probabilistic predictions to be more accurate than temporal predictions. Simple, reliable and valid prognostic tools have been developed in recent years that can be used readily at the bedside of terminally ill cancer patients. The greatest accuracy occurs with the use of a combination of subjective prognostic judgements and objective validated tools. Communicating survival predictions is an important part of cancer care and guidelines exist for improving delivery of such information. Important cultural differences may influence communication strategies and should be recognised in clinical encounters. More well-designed studies of prognosis and its impact on decision making are needed. The benefits and limitations of prognostication should be considered in many clinical decisions.

Uric acid as a prognostic factor for survival time: a prospective cohort study of terminally ill cancer patients

The aim of this prospective cohort study was to determine whether serum uric acid level is useful as a predictor of survival in terminally ill cancer patients. One hundred eighteen terminally ill cancer patients, including 63 (53.4%) males, were categorized into four groups by serum uric acid levels and followed up until death or to the end of the study. Cox's proportional hazard model was adopted to evaluate the joint effect of some clinicobiological variables on survival. From an initial model containing 51 variables, a final parsimonious model was obtained by means of a stepwise method. Repetitive dispersion analysis was performed for serum uric acid level in 39 subjects for 3 weeks until death. During the study period, 113 (95.76%) subjects expired, and the median survival time was 14 days. In univariate analysis, survival time of the fourth highest group (> or =7.2mg/dL) was significantly shorter than that of the others (hazard ratio (HR)=2.784, P<0.001). After adjustment for low performance status, moderate to severe pain, prolonged prothrombin time, hypocholesterolemia, and high lactate dehydrogenase (LDH) level, high serum uric acid level (> or =7.2mg/dL) was significantly and independently associated with short survival time (HR=2.637, P=0.001). Serum uric acid levels were also significantly increased between the first and the second week before death. These findings suggest that serum uric acid level can be useful in predicting life expectancy in terminally ill cancer patients.