A predictive model to identify hospitalized cancer patients at risk for 30-day mortality based on admission criteria via the electronic medical record

Artificial intelligence-based risk stratification, accurate diagnosis and treatment prediction in gynecologic oncology

As data-driven science, artificial intelligence (AI) has paved a promising path toward an evolving health system teeming with thrilling opportunities for precision oncology. Notwithstanding the tremendous success of oncological AI in such fields as lung carcinoma, breast tumor and brain malignancy, less attention has been devoted to investigating the influence of AI on gynecologic oncology. Hereby, this review sheds light on the ever-increasing contribution of state-of-the-art AI techniques to the refined risk stratification and whole-course management of patients with gynecologic tumors, in particular, cervical, ovarian and endometrial cancer, centering on information and features extracted from clinical data (electronic health records), cancer imaging including radiological imaging, colposcopic images, cytological and histopathological digital images, and molecular profiling (genomics, transcriptomics, metabolomics and so forth). However, there are still noteworthy challenges beyond performance validation. Thus, this work further describes the limitations and challenges faced in the real-word implementation of AI models, as well as potential solutions to address these issues.

Vital Signs in Palliative Care: A Scoping Review

Vital signs are routinely measured in secondary healthcare settings and can be used to detect clinical problems, guide treatment, and monitor response to treatment. Vital signs are less frequently measured in palliative care settings. Reasons for this are unclear. This scoping review aimed to assess the generic use of vital signs in palliative care, and its role in prognostication for adult patients with cancer receiving palliative care. Medline, Embase, and CINAHL were searched for articles involving adult patients with advanced cancer receiving palliative care who had their vital signs measured. Twenty-six articles were identified in which one or a combination of vital signs, with or without other clinical parameters, was used to prognosticate for patients. An additional three articles investigated the generic use of vital signs in patients with advanced cancer. There was significant heterogeneity between identified studies, with some indication that changes in vital signs may indicate that a patient is close to death. However, other studies suggested that patients may maintain normal vital signs until the time of death. Further studies are needed to explore whether abnormal vital signs may be used as a prognostic indicator for patients with cancer receiving palliative care.

Development of a Machine Learning Model Using Limited Features to Predict 6-Month Mortality at Treatment Decision Points for Patients With Advanced Solid Tumors

Patients with advanced solid tumors may receive intensive treatments near the end of life. This study aimed to create a machine learning (ML) model using limited features to predict 6-month mortality at treatment decision points (TDPs).

Predicting 6-month mortality at treatment decisions for patients with advanced solid tumors.

Development and Validation of a 30-Day In-hospital Mortality Model Among Seriously Ill Transferred Patients: a Retrospective Cohort Study

BACKGROUND

Predicting the risk of in-hospital mortality on admission is challenging but essential for risk stratification of patient outcomes and designing an appropriate plan-of-care, especially among transferred patients.

OBJECTIVE

Develop a model that uses administrative and clinical data within 24 h of transfer to predict 30-day in-hospital mortality at an Academic Health Center (AHC).

DESIGN

Retrospective cohort study. We used 30 putative variables in a multiple logistic regression model in the full data set (n = 10,389) to identify 20 candidate variables obtained from the electronic medical record (EMR) within 24 h of admission that were associated with 30-day in-hospital mortality (p < 0.05). These 20 variables were tested using multiple logistic regression and area under the curve (AUC)-receiver operating characteristics (ROC) analysis to identify an optimal risk threshold score in a randomly split derivation sample (n = 5194) which was then examined in the validation sample (n = 5195).

PARTICIPANTS

Ten thousand three hundred eighty-nine patients greater than 18 years transferred to the Indiana University (IU)-Adult Academic Health Center (AHC) between 1/1/2016 and 12/31/2017.

MAIN MEASURES

Sensitivity, specificity, positive predictive value, C-statistic, and risk threshold score of the model.

KEY RESULTS

The final model was strongly discriminative (C-statistic = 0.90) and had a good fit (Hosmer-Lemeshow goodness-of-fit test [X² (8) =6.26, p = 0.62]). The positive predictive value for 30-day in-hospital death was 68%; AUC-ROC was 0.90 (95% confidence interval 0.89-0.92, p < 0.0001). We identified a risk threshold score of -2.19 that had a maximum sensitivity (79.87%) and specificity (85.24%) in the derivation and validation sample (sensitivity: 75.00%, specificity: 85.71%). In the validation sample, 34.40% (354/1029) of the patients above this threshold died compared to only 2.83% (118/4166) deaths below this threshold.

CONCLUSION

This model can use EMR and administrative data within 24 h of transfer to predict the risk of 30-day in-hospital mortality with reasonable accuracy among seriously ill transferred patients.

Artificial intelligence for clinical oncology

Clinical oncology is experiencing rapid growth in data that are collected to enhance cancer care. With recent advances in the field of artificial intelligence (AI), there is now a computational basis to integrate and synthesize this growing body of multi-dimensional data, deduce patterns, and predict outcomes to improve shared patient and clinician decision making. While there is high potential, significant challenges remain. In this perspective, we propose a pathway of clinical cancer care touchpoints for narrow-task AI applications and review a selection of applications. We describe the challenges faced in the clinical translation of AI and propose solutions. We also suggest paths forward in weaving AI into individualized patient care, with an emphasis on clinical validity, utility, and usability. By illuminating these issues in the context of current AI applications for clinical oncology, we hope to help advance meaningful investigations that will ultimately translate to real-world clinical use.

Oxygen use and survival in patients with advanced cancer and low oxygen saturation in home care: a preliminary retrospective cohort study

Background

The role of oxygen therapy in end-of-life care for patients with advanced cancer is incompletely understood. We aimed to evaluate the association between oxygen use and survival in patients with advanced cancer and low oxygen saturation in home care.

Methods

We conducted a retrospective cohort study at a primary care practice in suburban Tokyo. Adult patients in home care with advanced cancer demonstrating first low oxygen saturation (less than 90%) detected in home visits were consecutively included in the study. Cox proportional hazards regression was used to investigate the effect of oxygen use on overall survival and survival at home, adjusted for systolic blood pressure, decreased level of consciousness, dyspnea, oral intake, performance status, and cardiopulmonary comorbidity.

Results

Of 433 identified patients with advanced cancer, we enrolled 137 patients (oxygen use, n = 35; no oxygen use, n = 102) who developed low oxygen saturation. In multivariable analysis, the adjusted hazard ratio (HR) of oxygen use was 0.68 (95% confidence interval 0.39–1.17) for death and 0.70 (0.38–1.27) for death at home. In patients with dyspnea, the HR was 0.35 (0.13–0.89) for death and 0.33 (0.11–0.96) for death at home; without dyspnea, it was 1.03 (0.49–2.17) for death and 0.84 (0.36–1.96) for death at home.

Conclusions

Oxygen use was not significantly associated with survival in patients with advanced cancer and low oxygen saturation, after adjusting for potential confounders. It may not be necessary to use oxygen for prolongation of survival in such patients, particularly in those without dyspnea.

Leveraging hybrid biomarkers in clinical endpoint prediction

Background

Clinical endpoint prediction remains challenging for health providers. Although predictors such as age, gender, and disease staging are of considerable predictive value, the accuracy often ranges between 60 and 80%. An accurate prognosis assessment is required for making effective clinical decisions.

Methods

We proposed an extended prognostic model based on clinical covariates with adjustment for additional variables that were radio-graphically induced, termed imaging biomarkers. Eight imaging biomarkers were introduced and investigated in a cohort of 68 non-small cell lung cancer subjects with tumor internal characteristic. The subjects comprised of 40 males and 28 females with mean age at 68.7 years. The imaging biomarkers used to quantify the solid component and non-solid component of a tumor. The extended model comprises of additional frameworks that correlate these markers to the survival ends through uni- and multi-variable analysis to determine the most informative predictors, before combining them with existing clinical predictors. Performance was compared between traditional and extended approaches using Receiver Operating Characteristic (ROC) curves, Area under the ROC curves (AUC), Kaplan-Meier (KM) curves, Cox Proportional Hazard, and log-rank tests (p-value).

Results

The proposed hybrid model exhibited an impressive boosting pattern over the traditional approach of prognostic modelling in the survival prediction (AUC ranging from 77 to 97%). Four developed imaging markers were found to be significant in distinguishing between subjects having more and less dense components: (P = 0.002–0.006). The correlation to survival analysis revealed that patients with denser composition of tumor (solid dominant) lived 1.6–2.2 years longer (mean survival) and 0.5–2.0 years longer (median survival), than those with less dense composition (non-solid dominant).

Conclusion

The present study provides crucial evidence that there is an added value for incorporating additional image-based predictors while predicting clinical endpoints. Though the hypotheses were confirmed in a customized case study, we believe the proposed model is easily adapted to various clinical cases, such as predictions of complications, treatment response, and disease evolution.

Telehealth for Cancer Care in Veterans: Opportunities and Challenges Revealed by COVID

The Veterans Health Administration system is one of the largest integrated health care providers in the United States, delivering medical care to > 9 million veterans. Barriers to delivering efficient health care include geographical limitations as well as long wait times. Telehealth has been used as a solution by many different health care services. However, it has not been as widely used in cancer care. In 2018, the US Department of Veterans Affairs (VA) Pittsburgh Healthcare System expanded the use of telehealth to provide antineoplastic therapies to rural patients by creating a clinical video telehealth clinic of the Virtual Cancer Care Network. This allows oncologists located at the tertiary center to virtually deliver care to remote sites. The recent COVID-19 pandemic forced oncologists across the VA system to adopt telehealth to provide continuity of care. On the basis of our review and personal experience, we have outlined opportunities for telehealth to play a role in every step of the cancer care journey from diagnosis to therapy to surveillance to clinical trials for medical, surgical, and radiation oncology. There are many advantages, such as decreased travel time and potential cost savings; however, there continues to be challenges with veterans having access to devices and the Internet as well as understanding how to use telehealth equipment. The lessons learned from this assessment of the VA telehealth system for cancer care can be adopted and integrated into other health systems. In the future, there needs to be evaluation of how telehealth can be further incorporated into oncology, satisfaction of veterans using telehealth services, overcoming telehealth barriers, and defining metrics of success.

A retrospective analysis of emergency surgery for cases of acute abdomen during cancer chemotherapy. Case series

Background

Treatment for acute abdomen during chemotherapy is frequently difficult because of the complicated status of the patients, and there have been only a few case series summarizing the outcomes of emergent surgery during chemotherapy. The aim of this study was to clarify the clinical outcomes of emergency surgery for acute abdomen during chemotherapy and identify predictive factors associated with mortality.

Methods

We retrospectively analyzed the records of patients who underwent emergency surgery for acute abdomen within 30-days after anti-cancer drugs administration between 2009 and 2020.

Results

Thirty patients were identified. The primary malignancies were hematological (n = 7), colorectal (n = 4), lung (n = 4), stomach (n = 2), breast (n = 2), prostate (n = 2) and others (n = 5). Fifteen patients were treated with the regimen, including molecular-targeted anti-cancer drugs (Bevacizumab: 8 cases, Rituximab: 4, Ramucirumab: 2, and Gefitinib: 1). Indications for emergency surgery were perforation of the gastrointestinal tract (n = 24), appendicitis (n = 3), bowel obstruction (n = 2), and gallbladder perforation (n = 1). Severe morbidity (Clavien-Dindo IIIa or more) occurred in 8 cases (27%), and there were 6 in-hospital deaths (20%). Significant factors related to in-hospital death were age >70 years old (P = 0.029), poor performance status (ECOG score 1 or 2) (P = 0.0088), and serum albumin level <2.6 g/dl (P = 0.026). The incidence of acute abdomen (odds ratio 5.31, P = 0.00017) was significantly higher in the patients receiving anti-VEGF drugs than in those without anti-VEGF drugs.

Conclusion

This study identified three predictive factors associated with in-hospital death after emergency surgery during chemotherapy: an older age, poor performance status, and low serum albumin level.

Comparison of Artificial Neural Networks and Logistic Regression for 30-days Survival Prediction of Cancer Patients

Introduction

A machine learning technique that imitates neural system and brain can provide better than traditional methods like logistic regression for survival prediction and create an algorithm by determining influential factors.

Aim

To determine the influential factors on survival time of palliative care cancer patients and to compare two statistical methods for better prediction of survival.

Methods

One-year data is gathered from the patients that we followed in the palliative care clinic of our hospital (2017-2018) (n = 189). All data were retrospectively evaluated. After descriptive statistics, we used Pearson and Spearman correlations for parametric and non-parametric variables. The Artificial Neural Networks (ANN) and logistic regression model were applied to parameters which have a significant correlation with short survival.

Results

Significantly correlated variables with short survival were Palliative Performance Scale (PPS), Edmonton Symptom Assessment System (ESAS), Karnofsky Performance Scale (KPS), brain, liver, and distant metastasis, hemogram parameters, cero-reactive protein (CRP) and albumin (ALB). ANN model showed 89.3% prediction accuracy while the logistic regression model showed 73.0%. ANN model achieved a better AUC value of 0.86 than logistic regression model (0.76).

Discussion

There are several prognostic evaluation tools such as PPS, KPS, CRP, albumin, leukocytes, neutrophil were reported several studies as survival-related parameters in logistic regression models, also. Many studies compare ANN with logistic regression. When we evaluated these parameters totally, we observed the same relations with survival then we used the same parameters in the ANN model. The effectivity of the survival prediction models can be improved with the use of ANN.

Conclusion

ANN provides a more accurate estimation than logistic regression. ANN model is an important statistical method for survival prediction of cancer patients.

Utilizing time series data embedded in electronic health records to develop continuous mortality risk prediction models using hidden Markov models: A sepsis case study

Continuous mortality risk monitoring is instrumental to manage a patient's care and to efficiently utilize the limited hospital resources. Due to incompleteness and irregularities of electronic health records (EHR), developing continuous mortality risk prediction using EHR data is a challenge. In this study, we propose a framework to continuously monitor mortality risk, and apply it to the real-world EHR data. The proposed method employs hidden Markov models (temporal technique) that take account of both the previous state of patient's health and the current value of clinical signs. Following the Sepsis-3 definition, we selected 3898 encounters of patients with suspected infection to compare the performance of temporal and non-temporal methods (Decision Tree (DT), Logistic Regression (LR), Naive Bayes (NB), Random Forest (RF), and Support Vector Machine (SVM)). The area under receiver operating characteristics (AUROC) curve, sensitivity, specificity and G-mean were used as performance measures. On the selected data, the AUROC of the proposed temporal framework (0.87) is 9-12% greater than the nontemporal methods (DT: 0.78, NB: 0.79, SVM: 0.79, LR: 0.80 and RF: 0.80). The results also show that our model (G-mean=0.78) provides a better balance between sensitivity and specificity compared to clinically acceptable bed-side criteria (G-mean=0.71). The proposed framework leverages the longitudinal data available in EHR and performs better than the non-temporal methods. The proposed method facilitates information related to the time of change of the patient's health that may help practitioners to plan early and develop effective treatment strategies.

Automated Survival Prediction in Metastatic Cancer Patients Using High-Dimensional Electronic Medical Record Data

BACKGROUND

Oncologists use patients' life expectancy to guide decisions and may benefit from a tool that accurately predicts prognosis. Existing prognostic models generally use only a few predictor variables. We used an electronic medical record dataset to train a prognostic model for patients with metastatic cancer.

METHODS

The model was trained and tested using 12 588 patients treated for metastatic cancer in the Stanford Health Care system from 2008 to 2017. Data sources included provider note text, labs, vital signs, procedures, medication orders, and diagnosis codes. Patients were divided randomly into a training set used to fit the model coefficients and a test set used to evaluate model performance (80%/20% split). A regularized Cox model with 4126 predictor variables was used. A landmarking approach was used due to the multiple observations per patient, with t0 set to the time of metastatic cancer diagnosis. Performance was also evaluated using 399 palliative radiation courses in test set patients.

RESULTS

The C-index for overall survival was 0.786 in the test set (averaged across landmark times). For palliative radiation courses, the C-index was 0.745 (95% confidence interval [CI] = 0.715 to 0.775) compared with 0.635 (95% CI = 0.601 to 0.669) for a published model using performance status, primary tumor site, and treated site (two-sided P < .001). Our model's predictions were well-calibrated.

CONCLUSIONS

The model showed high predictive performance, which will need to be validated using external data. Because it is fully automated, the model can be used to examine providers' practice patterns and could be deployed in a decision support tool to help improve quality of care.

Improving palliative care with deep learning

BACKGROUND

Access to palliative care is a key quality metric which most healthcare organizations strive to improve. The primary challenges to increasing palliative care access are a combination of physicians over-estimating patient prognoses, and a shortage of palliative staff in general. This, in combination with treatment inertia can result in a mismatch between patient wishes, and their actual care towards the end of life.

METHODS

In this work, we address this problem, with Institutional Review Board approval, using machine learning and Electronic Health Record (EHR) data of patients. We train a Deep Neural Network model on the EHR data of patients from previous years, to predict mortality of patients within the next 3-12 month period. This prediction is used as a proxy decision for identifying patients who could benefit from palliative care.

RESULTS

The EHR data of all admitted patients are evaluated every night by this algorithm, and the palliative care team is automatically notified of the list of patients with a positive prediction. In addition, we present a novel technique for decision interpretation, using which we provide explanations for the model's predictions.

CONCLUSION

The automatic screening and notification saves the palliative care team the burden of time consuming chart reviews of all patients, and allows them to take a proactive approach in reaching out to such patients rather then relying on referrals from the treating physicians.

Development of Imminent Mortality Predictor for Advanced Cancer (IMPAC), a Tool to Predict Short-Term Mortality in Hospitalized Patients With Advanced Cancer

PURPOSE

End-of-life care for patients with advanced cancer is aggressive and costly. Oncologists inconsistently estimate life expectancy and address goals of care. Currently available prognostication tools are based on subjective clinical assessment. An objective prognostic tool could help oncologists and patients decide on a realistic plan for end-of-life care. We developed a predictive model (Imminent Mortality Predictor in Advanced Cancer [IMPAC]) for short-term mortality in hospitalized patients with advanced cancer.

METHODS

Electronic health record data from 669 patients with advanced cancer who were discharged from Yale Cancer Center/Smilow Cancer Hospital were extracted. Statistical learning techniques were used to develop a tool to estimate survival probabilities. Patients were randomly split into training (70%) and validation (30%) sets 20 times. We tested the predictive properties of IMPAC for mortality at 30, 60, 90, and 180 days past the day of admission.

RESULTS

For mortality within 90 days at a 40% sensitivity level, IMPAC has close to 60% positive predictive value. Patients estimated to have a greater than 50% chance of death within 90 days had a median survival time of 47 days. Patients estimated to have a less than 50% chance of death had a median survival of 290 days. Area under the receiver operating characteristic curve for IMPAC averaged greater than .70 for all time horizons tested. Estimated potential cost savings per patient was $15,413 (95% CI, $9,162 to $21,665) in 2014 constant dollars.

CONCLUSION

IMPAC, a novel prognostic tool, can generate life expectancy probabilities in real time and support oncologists in counseling patients about end-of-life care. Potentially avoidable costs are significant.

Retrospective study of unplanned hospital admission for metastatic cancer patients visiting the emergency department

PURPOSE

The purpose of the study was to identify factors that predict unplanned admission for metastatic cancer patients visiting the emergency department (ED).

METHODS

Patients visiting the ED of a general hospital from April 2012 to March 2013 were investigated retrospectively. Data including demographics, vital signs, and laboratory measurements were collected from a chart review for each patient. Factors related to emergency admission were identified by univariate and multivariate analyses.

RESULTS

A total of 15,716 individuals visiting the ED during the study period included 1244 (7.9%) patients with cancer. Among the 491 cancer patients with metastasis, univariate analysis revealed that emergency admission was significantly associated with an age of ≥76 years; an altered mental status; fever (≥38 °C); a blood oxygen saturation of <90%; a white blood cell (WBC) count of ≤2000 or ≥10,000/μL; hypoalbuminemia (≤2.5 g/dL); and elevated levels of aspartate aminotransferase (≥100 IU/L), blood urea nitrogen (≥25 mg/dL), and C-reactive protein (CRP, ≥10 mg/dL). Multivariate analysis identified age, an altered mental status, hypoxemia, an abnormal WBC count, and elevated CRP as putative independent predictive factors for emergency admission. The number of these five factors present was also correlated with 30-day mortality (c-statistic = 0.72).

CONCLUSIONS

Age, unconsciousness, hypoxemia, an abnormal WBC count, and elevated CRP were found to be associated with emergency admission and 30-day mortality for metastatic cancer patients. Prospective validation of a predictive scoring system based on these findings is warranted.

Predicting Risk of Suicide Attempt Using History of Physical Illnesses From Electronic Medical Records

BACKGROUND

Although physical illnesses, routinely documented in electronic medical records (EMR), have been found to be a contributing factor to suicides, no automated systems use this information to predict suicide risk.

OBJECTIVE

The aim of this study is to quantify the impact of physical illnesses on suicide risk, and develop a predictive model that captures this relationship using EMR data.

METHODS

We used history of physical illnesses (except chapter V: Mental and behavioral disorders) from EMR data over different time-periods to build a lookup table that contains the probability of suicide risk for each chapter of the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) codes. The lookup table was then used to predict the probability of suicide risk for any new assessment. Based on the different lengths of history of physical illnesses, we developed six different models to predict suicide risk. We tested the performance of developed models to predict 90-day risk using historical data over differing time-periods ranging from 3 to 48 months. A total of 16,858 assessments from 7399 mental health patients with at least one risk assessment was used for the validation of the developed model. The performance was measured using area under the receiver operating characteristic curve (AUC).

RESULTS

The best predictive results were derived (AUC=0.71) using combined data across all time-periods, which significantly outperformed the clinical baseline derived from routine risk assessment (AUC=0.56). The proposed approach thus shows potential to be incorporated in the broader risk assessment processes used by clinicians.

CONCLUSIONS

This study provides a novel approach to exploit the history of physical illnesses extracted from EMR (ICD-10 codes without chapter V-mental and behavioral disorders) to predict suicide risk, and this model outperforms existing clinical assessments of suicide risk.

Opportunities and challenges in developing risk prediction models with electronic health records data: a systematic review

OBJECTIVE

Electronic health records (EHRs) are an increasingly common data source for clinical risk prediction, presenting both unique analytic opportunities and challenges. We sought to evaluate the current state of EHR based risk prediction modeling through a systematic review of clinical prediction studies using EHR data.

METHODS

We searched PubMed for articles that reported on the use of an EHR to develop a risk prediction model from 2009 to 2014. Articles were extracted by two reviewers, and we abstracted information on study design, use of EHR data, model building, and performance from each publication and supplementary documentation.

RESULTS

We identified 107 articles from 15 different countries. Studies were generally very large (median sample size = 26 100) and utilized a diverse array of predictors. Most used validation techniques (n = 94 of 107) and reported model coefficients for reproducibility (n = 83). However, studies did not fully leverage the breadth of EHR data, as they uncommonly used longitudinal information (n = 37) and employed relatively few predictor variables (median = 27 variables). Less than half of the studies were multicenter (n = 50) and only 26 performed validation across sites. Many studies did not fully address biases of EHR data such as missing data or loss to follow-up. Average c-statistics for different outcomes were: mortality (0.84), clinical prediction (0.83), hospitalization (0.71), and service utilization (0.71).

CONCLUSIONS

EHR data present both opportunities and challenges for clinical risk prediction. There is room for improvement in designing such studies.

Factors Associated with Short-Term Mortality After Surgical Oncologic Emergencies

BACKGROUND

The clinical outcome of patients with oncologic emergencies is often poor and mortality is high. It is important to determine which patients may benefit from invasive treatment, and for whom conservative treatment and/or palliative care would be appropriate. In this study, prognostic factors for clinical outcome are identified in order to facilitate the decision-making process for patients with surgical oncologic emergencies.

METHODS

This was a prospective registration study for patients over 18 years of age, who were consulted for surgical oncologic emergencies between November 2013 and April 2014. Multiple variables were registered upon emergency consultation, and the follow-up period was 90 days. Multivariate logistic regression analysis was performed to identify factors associated with 30- and 90-day mortality.

RESULTS

During the study period, 207 patients experienced surgical oncologic emergencies-101 (48.8 %) men and 106 (51.2 %) women, with a median age of 64 years (range 19-92). The 30-day mortality was 12.6 % and 90-day mortality was 21.7 %. Factors significantly associated with 30-day mortality were palliative intent of cancer treatment prior to emergency consultation (p = 0.006), Eastern Cooperative Oncology Group performance score (ECOG-PS) >0 (p for trend: p = 0.03), and raised lactate dehydrogenase (LDH) (p < 0.001). Additional factors associated with 90-day mortality were low handgrip strength (HGS) (p = 0.01) and low albumin (p = 0.002).

CONCLUSIONS

Defining the intent of prior cancer treatment and the ECOG-PS are of prognostic value when deciding on treatment for patients with surgical oncologic emergencies. Additional measurements of HGS, LDH, and albumin levels can serve as objective parameters to support the clinical assessment of individual prognosis.

Palliative radiotherapy during the last month of life: Predictability for referring physicians and radiation oncologists

Oncologists commonly overestimate the survival time of patients receiving palliative therapy, which may result in the administration of treatments that are too aggressive for patients near the end of their lives. Previous studies have discussed the negative implications of palliative radiotherapy if administered during the last month of life. Models predicting a limited survival time may improve the ability of the oncologists to tailor the treatment according to the needs of each individual patient. In the present study, prognostic factors for survival time, and the use of palliative radiotherapy during the last month of life, were analyzed in 873 patients. Models predicting the likelihood of administering such therapy were examined, and the risk of receiving radiotherapy during the last month of life was observed to be lower in patients with non-metastatic cancer than in those with metastatic cancer (7 vs. 13%, respectively; P=0.12). On multivariate analysis, 11 factors that significantly influenced the survival time were identified. These findings emphasize the complexity of potential prediction models. The most important risk factor regarding the prediction of extremely short survival times was observed to be an Eastern Cooperative Oncology Group performance status (ECOG PS) of 4, followed by an ECOG PS of 3 (median survival times, 14 and 64 days, respectively). A limited number of patients who received palliative radiotherapy during their last month of life died unexpectedly. Disease-specific prediction models were developed; however, the small number of events available for analysis limited their immediate clinical impact. Furthermore, these prediction models identified a minority of patients who received radiotherapy during the last month of life. In conclusion, the majority of the palliative radiotherapy courses administered to patients with advanced cancer during their last month of life may be preventable if accurate decision models for the clinic are developed. However, due to the complexity associated with the prediction of survival times in patients receiving palliative radiotherapy, large databases are required to allow accurate models to be established. The present study also discusses the recommendations of the Department of Oncology and Palliative Medicine of Nordland Hospital (Bodø, Nordland, Norway) with regard to the use of palliative radiotherapy during the last month of life of patients with terminal cancer.

Current management of surgical oncologic emergencies

Objectives

For some oncologic emergencies, surgical interventions are necessary for dissolution or temporary relieve. In the absence of guidelines, the most optimal method for decision making would be in a multidisciplinary cancer conference (MCC). In an acute setting, the opportunity for multidisciplinary discussion is often not available. In this study, the management and short term outcome of patients after surgical oncologic emergency consultation was analyzed.

Method

A prospective registration and follow up of adult patients with surgical oncologic emergencies between 01-11-2013 and 30-04-2014. The follow up period was 30 days.

Results

In total, 207 patients with surgical oncologic emergencies were included. Postoperative wound infections, malignant obstruction, and clinical deterioration due to progressive disease were the most frequent conditions for surgical oncologic emergency consultation. During the follow up period, 40% of patients underwent surgery. The median number of involved medical specialties was two. Only 30% of all patients were discussed in a MCC within 30 days after emergency consultation, and only 41% of the patients who underwent surgery were discussed in a MCC. For 79% of these patients, the surgical procedure was performed before the MCC. Mortality within 30 days was 13%.

Conclusion

In most cases, surgery occurred without discussing the patient in a MCC, regardless of the fact that multiple medical specialties were involved in the treatment process. There is a need for prognostic aids and acute oncology pathways with structural multidisciplinary management. These will provide in faster institution of the most appropriate personalized cancer care, and prevent unnecessary investigations or invasive therapy.

Impact of Implementing an Electronic Health Record on Surgical Resident Work Flow, Duty Hours, and Operative Experience

Our objective was to assess the effect of implementing an electronic health record (EHR) on surgical resident work flow, duty hours, and operative experience at a large teaching hospital. In May 2012, an EHR was put into effect at our institution replacing paper documentation and orders. Resident time to complete patient documentation, average duty hours, and operative experience before EHR and afterward (at 1, 4, 6, 8, and 24 weeks) were surveyed. We obtained 100 per cent response rate from 15 surgical residents at all time intervals. The average time spent documenting before EHR was 9 ± 2 minutes per patient document and at Weeks 1, 4, 6, 8, and 24 after EHR implementation was 22 ± 10, 15 ± 7, 15 ± 7, 14 ± 8, and 12 ± 4 minutes, respectively. Repeated measures analysis of variance demonstrated a difference among the means ( P < 0.0001). Discharge summary and operative note remained significantly longer to complete at Week 24 compared with paper documentation ( P < 0.05). Average resident work hours and operative cases per week before EHR were 77 ± 5 hours and 12 ± 5 cases, respectively, which were similar at all time points after EHR implementation ( P > 0.05). At 24 weeks after EHR, 74 per cent of residents felt their risk of performing a medical error using electronic documentation and order entry was higher compared with paper charting and orders. Transition to EHR led to a significant doubling in resident time spent performing documentation for each patient. It improved over 6 months after implementation but never reached the pre-EHR baseline for operative notes and discharge summaries. Average resident work hours and case logs remained similar during this transition.

Change in physiological variables in the last 2 weeks of life: an observational study of hospital in-patients with cancer

BACKGROUND

Recognising dying remains a difficult clinical skill which has gained increasing importance in the United Kingdom since the Neuberger review. Clinical and research methods exist to aid recognition of dying but do not exhibit the level of accuracy required for such an important decision.

AIM

To explore change in key clinical parameters as cancer patients near the end of life.

DESIGN

This is a retrospective cohort study of terminally ill patients. Data were collected from hospital case-notes. Case-note data were analysed using multilevel modelling to explore absolute values and rates of change of given variables.

SETTING/PARTICIPANTS

Hospital in-patients who died from solid-tumour malignancies within a 3-month period in 2009 formed the cohort. The setting was an acute hospital trust in the North of England.

RESULTS

A total of 15,337 data points from the case-notes of 102 patients were analysed. There was a clinically and statistically significant deterioration in respiratory function and renal function over the last 2 weeks of life. Heart rate and serum sodium also changed but did not vary greatly from normal limits. White cell parameters, haemoglobin and albumin showed evidence for change over longer periods.

CONCLUSION

Results demonstrate statistically and clinically significant change in routinely measured respiratory and renal function variables during the final 2 weeks of life in people dying with cancer. Although useful in acute early warning scores, in a terminally ill patient, relative haemodynamic stability should not be interpreted as reassuring. Further work is needed to understand how these findings apply to the individual or inform other prognostic work.

Pre- and in-therapy predictive score models of adult OSAS patients with poor adherence pattern on nCPAP therapy

OBJECTIVES

To identify patterns of adherence to nasal continuous positive airway pressure (nCPAP) use in the first 3 months of therapy among newly diagnosed adult patients with obstructive sleep apnea/hypopnea syndrome (OSAS) and their predictors. To develop pretherapy and in-therapy scores to predict adherence pattern.

METHODS

Newly diagnosed adult OSAS patients were consecutively recruited from March to August 2013. Baseline clinical information and measures such as Epworth Sleepiness Scale (ESS), Fatigue Severity Scale (FSS), Zung's Self-Rating Depression Scale (SDS), and The Pittsburgh Sleep Quality Index (PSQI) at baseline and at the end of 3rd-week therapy were collected. Twelve weeks' adherence data were collected from the nCPAP memory card, and K-means cluster analysis was used to explore adherence patterns. Predictive scores were developed from the coefficients of cumulative logit models of adherence patterns using variables available at baseline and after 3 weeks of therapy. Performance of the score was validated using 500 bootstrap resamples.

RESULTS

Seventy six patients completed a 12-week follow-up. Three patterns were revealed. Patients were identified as developing an adherence pattern that was poor (n=14, mean ± SD, 2.3±0.9 hours per night), moderate (n=19, 5.3±0.6 hours per night), or good (n=43, 6.8±0.3 hours per night). Cumulative logit regression models (good → moderate → poor) revealed independent baseline predictors to be ESS (per unit increase) (OR [95% CI], 0.763 [0.651, 0.893]), SDS (1.461 [1.238, 1.724]), and PSQI (2.261 [1.427, 3.584]); and 3-week therapy predictors to be ESS (0.554 [0.331, 0.926]), PSQI (2.548 [1.454, 4.465]), and the changes (3rd week-baseline data) in ESS (0.459 [0.243, 0.868]), FSS (3.556 [1.788, 7.070]), and PSQI (2.937 [1.273, 6.773]). Two predictive score formulas for poor adherence were developed. The area under the curve (AUC) of the receiver operating characteristics (ROC) curves for baseline and 3-week formulas were 0.989 and 0.999, respectively. Bootstrap analysis indicated positive predictive values of baseline and 3-week predictive scores in our patient population of 0.82 (95% CI [0.82, 0.83]) and 0.94 (95% CI [0.93, 0.94]), respectively.

CONCLUSION

A high level of prediction of poor adherence pattern is possible both before and at the first 3 weeks of therapy. The predictive scores should be further evaluated for external validity.

Variations in vital signs in the last days of life in patients with advanced cancer

CONTEXT

Few studies have examined variation in vital signs in the last days of life.

OBJECTIVES

We determined the variation of vital signs in the final two weeks of life in patients with advanced cancer and examined their association with impending death in three days.

METHODS

In this prospective, longitudinal, observational study, we enrolled consecutive patients admitted to two acute palliative care units and documented their vital signs (heart rate, blood pressure, respiratory rate, oxygen saturation, and temperature) twice a day serially from admission to death or discharge.

RESULTS

Of 357 patients, 203 (57%) died in hospital. Systolic blood pressure (P < 0.001), diastolic blood pressure (P < 0.001), and oxygen saturation (P < 0.001) decreased significantly in the final three days of life, and temperature increased slightly (P < 0.04). Heart rate (P = 0.22) and respiratory rate (P = 0.24) remained similar in the last three days. Impending death in three days was significantly associated with increased heart rate (odds ratio [OR] = 2; P = 0.01), decreased systolic blood pressure (OR = 2.5; P = 0.004), decreased diastolic blood pressure (OR = 2.3; P = 0.002), and decreased oxygen saturation (OR = 3.7; P = 0.003) from baseline readings on admission. These changes had high specificity (≥ 80%), low sensitivity (≤ 35%), and modest positive likelihood ratios (≤ 5) for impending death within three days. A large proportion of patients had normal vital signs in the last days of life.

CONCLUSION

Blood pressure and oxygen saturation decreased in the last days of life. Clinicians and families cannot rely on vital sign changes alone to rule in or rule out impending death. Our findings do not support routine vital signs monitoring of patients who are imminently dying.

Optimal screening for geriatric assessment in older allogeneic hematopoietic cell transplantation candidates

OBJECTIVE

Older patients who receive hematopoietic cell transplantation (HCT) may be at risk for adverse outcomes due to age-related conditions or frailty. Geriatric assessment (GA) has been used to evaluate HCT candidates but can be time-consuming. We therefore sought to determine the predictive ability of two screening tools, the Vulnerable Elders Survey (VES-13) and the G8, for abnormal GA or frailty.

MATERIALS AND METHODS

We enrolled 50 allogeneic HCT candidates age ≥60 years. The GA included measures of medical, physical, functional, and social health. Frailty was defined as 3 or more abnormalities on grip strength, gait speed, weight loss, exhaustion, and activity. We associated baseline characteristics and abnormal GA or frailty. We determined the sensitivity and predictive ability of the VES-13 and G8 for GA and frailty.

RESULTS

Overall, 33 (66%) patients (mean age 65.4 years) had an abnormal GA, and 11 patients (22%) were frail. The G8 screening tool had a higher sensitivity for an abnormal GA (69.7%), and the VES-13 had a higher specificity (100%). Both tools had similar discriminatory ability.

CONCLUSIONS

Older HCT candidates had a significant number of deficits on baseline GA and a high prevalence of frailty. Existing screening tools may not be able to replace a full GA.