1. Utilize predictive models to enhance hospice and palliative care services.
2. Understand the interpretations and implications of predictive models' performance parameters.
Several predictive ...models for mortality were developed during the COVID-19 pandemic. However, models become obsolete with clinical advancements. Hence, new predictive models for mortality that consider recent clinical advancements are needed for effective and efficient early-life-saving interventions, hospital resource triaging, and palliative care services.
This study develops a model to identify high-risk patients for mortality, enabling effective and efficient early-life-saving interventions, optimizing triaging of hospital resources, and enhancing palliative care services.
Our study aims to develop a new model that accounts for recent clinical advancements, such as the introduction of COVID-19 vaccination, to predict in-hospital mortality, mortality within 30 days after discharge from hospitalization, or discharge from hospitalization into hospice care.
We conducted a retrospective cohort study by using data from BJC Integrated Health Systems between March 2021 and October 2022 to develop a Long Short-Term Memory (LSTM) neural network model among hospitalized patients. The model predicts a composite outcome of in-hospital mortality, mortality within 30 days after discharge, or discharge into hospice care.
The total sample size used for model development was 94,017, with 75,213 for training data, 9,402 for validation, and 9,402 for test data. Our model achieved an area under the curve (AUC) of 90% and demonstrated high specificity at a threshold of 0.5 (specificity of 99% and sensitivity of 30%). Furthermore, the model exhibited good sensitivity at a threshold of 0.1 (specificity of 86% and sensitivity of 78%). Across different population strata such as race (White and Non-white), COVID-19 status (positive and negative), and Intensive Care Unit (ICU) admissions (positive and negative), the model consistently performed well, with AUC values at or above 85% and sensitivity and specificity mostly exceeding 80% at a threshold of 0.1.
Our developed model effectively identified patients at a high risk of in-hospital mortality, mortality within 30 days after discharge, or discharge into hospice care. Therefore, it can be used to support the identification of patients requiring hospice as well as enhancing palliative care services in the hospital.
Further study is needed in this area to determine how long it takes for predictive models for mortality to become obsolete, allowing for the timely development of new models to support hospice and palliative care services.
Background
Severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) and the coronavirus 19 (COVID‐19) pandemic have had a lasting impact on the care of cancer patients. The impact on patients ...with gastrointestinal (GI) malignancies remains incompletely understood. We aimed to assess the impact of COVID‐19 on mortality, length of stay (LOS), and cost of care among patients with GI malignancies, and identify differences in outcomes based on primary tumor site.
Methods
We analyzed discharge encounters collected from the National Inpatient Sample (NIS) between March 2020 and December 2020 using propensity score matching (PSM) and COVID‐19 as the treatment effect.
Results
Of the 87,684 patient discharges with GI malignancies, 1892 were positive for COVID‐19 (C+) and eligible for matching in the PSM model. Following PSM analysis, C+ with GI tumors demonstrated increased incidence of mortality compared to their COVID‐19‐negative (C‐) counterparts (21.3% vs. 11.9%, p < 0.001). C+ patients with colorectal cancer (CRC) had significantly higher mortality compared to those who were C‐ (40% vs. 24%; p = 0.035). In addition, C+ patients with GI tumors had a longer mean LOS (9.4 days vs. 6.9 days; p < 0.001) and increased cost of care ($26,048.29 vs. $21,625.2; p = 0.001) compared to C‐ patients. C+ patients also had higher odds of mortality secondary to myocardial infarction relative to C‐ patients (OR = 3.54, p = 0.001).
Conclusions
C+ patients with GI tumors face approximately double the odds of mortality, increased LOS, and increased cost of care compared to their C‐ counterparts. Outcome disparities were most pronounced among patients with CRC.
Objective: To group people with type 1/2 diabetes mellitus (T1D/T2D) into clusters reflecting attitudes towards DM-related technology. Methods: Adults with commercial or Medicare Advantage insurance, ...≥1 medical claim for T1D/T2D and treated with insulin MDI or pump during 2/1/19-1/31/20 were identified in the HealthCore Integrated Research Database. Consenting patients completed a survey including 6 patient-reported outcome measures (PROMs) involving DM technology/care (Table 1). The jump method determined the number of clusters; k-means clustering, using PROM scores, segmented respondents into groups. The authors assigned descriptive labels to clusters. Results: Nine clusters were created from 566 patients with complete PROMs. Technology Embracers (clusters 6,4,7) had positive technology/DM care attitudes, low distress, primarily T1D patients with the most pump use. Cautious Technology Adopters (clusters 2,3,9) had positive technology attitudes with mixed DM care attitudes, mix of T1D/T2D patients and primarily MDI users. High Hurdle Technology Adopters (clusters 1,5,8) had lowest technology attitude scores, less positive DM care attitudes, mix of T1D/T2D patients and primarily MDI users. Conclusion: People with DM may be grouped by attitudes towards technology/DM care behaviors with unique clinical characteristics that may inform tailored interventions and technology offerings.
25% of all breast cancer patients have HER-2 overexpression. Breast Cancer patients with HER-2 overexpression are typically treated with HER-2 inhibitors such as Trastuzumab. Trastuzumab is known to ...cause a decrease in left ventricular ejection fraction. The aim of this study is to create a cardiac risk prediction tool among women with Her-2 positive breast cancer to predict cardiotoxicity.
Using a split sample design, we created a risk prediction tool using patient level data from electronic medical records. The study included women 18 years of age and older diagnosed with HER-2 positive breast cancer who received Trastuzumab. Outcome measure was defined as a drop in LVEF by more than 10% to less than 53% at any time in the 1-year study period. Logistic regression was used to test predictors.
The cumulative incidence of cardiac dysfunction in our study was 9.4%. The sensitivity and specificity of the model are 46% and 84%, respectively. Given a cumulative incidence of cardiotoxicity of 9%, the negative predictive value of the test was 94%. This suggests that in a low-risk population, the interval of screening for cardiotoxicity may be performed less frequently.
Cardiac risk prediction tool can be used to identify Her-2 positive breast cancer patients at risk of developing cardiac dysfunction. Also, test characteristics in addition to disease prevalence may inform a rational strategy in performing cardiac ultrasound in Her-2 breast cancer patients. We have developed a cardiac risk prediction model with high NPV in a low-risk population which has an appealing cost-effectiveness profile.
Background
Survival differences between left‐sided colon cancer (LSCC) and right‐sided colon cancer (RSCC) has been previously reported with mixed results, with various study periods not accounting ...for other causes of mortality.
Purpose
We sought to assess the trends in colon cancer cause‐ specific survival (CSS) and overall survival (OS) based on sidedness.
Method
Fine‐Gray competing risk and Cox models were used to analyze Surveillance, Epidemiology, and End Results (SEER) population‐based cohort from 1975 to 2019. Various interval periods were identified based on the timeline of clinical adoption of modern chemotherapy (1975–1989, interval period A; 1990–2004, B; and 2005–2019, C).
Results
Of the 227,637 patients, 50.1% were female and 46.2% were RSCC. RSCC was more common for African Americans (51.5%), older patients (age ≥65; 51.4%), females (50.4%), while LSCC was more common among Whites (53.1%; p < 0.001), younger patients (age 18–49, 64.6%; 50–64, 62.3%; p < 0.001), males (58.1%; p < 0.001). The Median CSS for LSCC and RCC were 19.3 and 16.7 years respectively for interval period A (1975–1989). Median CSS for interval periods B and C were not reached (more than half of the cohort was still living at the end of the follow‐up period). Adjusted CSS was superior for LSCC versus RSCC for the most recent interval period C (HR 0.89; 0.86–0.92; p < 0.001). LSCC consistently showed superior OS for all study periods. Stage stratification showed worse CSS for localized and regional LSCC in the earlier study periods, but the risk attenuated over time. However, left sided distant disease had superior CSS per stage for all interval periods. OS was better for LSCC irrespective of stage, with gradual improvement over time.
Conclusion
LSCC was associated with superior survival compared to right sided tumors. With the adoption of modern chemotherapy regimens, prognosis between LSCC and RSCC became more divergent in favor of LSCC. Colon cancer clinical trials should strongly consider tumor sidedness as an enrollment factor.
Background and Objectives
Colorectal cancer (CRC) sidedness is recognized as a prognostic factor for survival; left‐sided colorectal cancer is associated with better outcomes than right‐sided colon ...cancer (RsCC). We aimed to evaluate the influence of obesity on CRC sidedness and determine how race, age, and sex affect mortality among overweight and obese individuals.
Methods
A survey‐weighted analysis was conducted using data obtained from the National Inpatient Sample between 2016 and 2019.
Results
Of the 24 549 patients with a diagnosis of CRC and a reported body mass index (BMI), 13.6% were overweight and 49.9% were obese. The race distribution was predominantly non‐Hispanic Whites (69.7%), followed by Black (15.6%), Hispanic (8.7%), and other race (6.1%). Overweight (BMI: 25−29.9) and obese (BMI: ≥30) individuals were more likely to have RsCC (adjusted OR aOR = 1.28; 95% CI: 1.17−1.39, p < 0.001 and aOR = 1.45; 95% CI: 1.37−1.54, p < 0.001, respectively). Obese Black individuals were more likely to have RsCC as compared to their White counterparts (aOR = 1.23; 95% CI: 1.09−1.38).
Conclusions
Obesity is associated with an increased risk of RsCC. In addition, racial disparities in CRC sidedness and outcomes are most pronounced among obese patients.
Background
Gastrointestinal extrapulmonary small cell carcinoma (GI EPSCCa) is a rare, aggressive neuroendocrine tumor. Factors affecting survival, including the prognostic significance of primary ...tumor site, remain under investigation.
Methods
Data from the surveillance, epidemiology, and end results (SEER) program were extracted to identify patients diagnosed with GI EPSCCa between 2000 and 2018. Cox proportional hazard models were used to assess prognostic factors based on primary tumor site.
Results
A total of 1687 patients were included in the survival analysis. The distribution of the primary tumor location was as follows: 31.5% colorectum (CRC), 22.1% esophageal, 20.6% pancreatic, 13.3% hepatobiliary (HB), 10.6% stomach, and 1.8% small intestine (SI). Esophagogastric and SI EPSCCa were more common among Black individuals, whereas CRC, HB, and pancreatic EPSCCa were more common among White patients (
p
= 0.012). There were no racial differences in OS for GI EPSCCa. HB EPSCCa was associated with inferior OS compared with esophageal tumors (adjusted hazard ratio aHR 1.21, 95% confidence interval CI 1.00–1.46;
p
= 0.048), and SI EPSCCa was associated with prolonged survival compared with esophageal EPSCCa (aHR 0.76, 95% CI 0.48–1.20;
p
= 0.237) but did not reach statistical significance. Surgical intervention and a treatment period after 2006 were associated with superior OS.
Conclusions
The prognosis for GI ESPCCa varies based on site. Chemotherapy, radiation, and surgical resection are associated with improved outcomes; however, the prognosis for patients with EPSCCa remains dismal. Prospective studies are needed to guide therapy for this aggressive tumor.