Browsing by Subject "Forecasting"
Now showing 1 - 10 of 12
Results Per Page
Sort Options
Item Assessing the utility of deep neural networks in predicting postoperative surgical complications: a retrospective study(Elsevier, 2021) Bonde, Alexander; Varadarajan, Kartik M.; Bonde, Nicholas; Troelsen, Anders; Muratoglu, Orhun K.; Malchau, Henrik; Yang, Anthony D.; Alam, Hasan; Sillesen, Martin; Surgery, School of MedicineBackground: Early detection of postoperative complications, including organ failure, is pivotal in the initiation of targeted treatment strategies aimed at attenuating organ damage. In an era of increasing health-care costs and limited financial resources, identifying surgical patients at a high risk of postoperative complications and providing personalised precision medicine-based treatment strategies provides an obvious pathway for reducing patient morbidity and mortality. We aimed to leverage deep learning to create, through training on structured electronic health-care data, a multilabel deep neural network to predict surgical postoperative complications that would outperform available models in surgical risk prediction. Methods: In this retrospective study, we used data on 58 input features, including demographics, laboratory values, and 30-day postoperative complications, from the American College of Surgeons (ACS) National Surgical Quality Improvement Program database, which collects data from 722 hospitals from around 15 countries. We queried the entire adult (≥18 years) database for patients who had surgery between Jan 1, 2012, and Dec 31, 2018. We then identified all patients who were treated at a large midwestern US academic medical centre, excluded them from the base dataset, and reserved this independent group for final model testing. We then randomly created a training set and a validation set from the remaining cases. We developed three deep neural network models with increasing numbers of input variables and so increasing levels of complexity. Output variables comprised mortality and 18 different postoperative complications. Overall morbidity was defined as any of 16 postoperative complications. Model performance was evaluated on the test set using the area under the receiver operating characteristic curve (AUC) and compared with previous metrics from the ACS-Surgical Risk Calculator (ACS-SRC). We evaluated resistance to changes in the underlying patient population on a subset of the test set, comprising only patients who had emergency surgery. Results were also compared with the Predictive OpTimal Trees in Emergency Surgery Risk (POTTER) calculator. Findings: 5 881 881 surgical patients, with 2941 unique Current Procedural Terminology codes, were included in this study, with 4 694 488 in the training set, 1 173 622 in the validation set, and 13 771 in the test set. The mean AUCs for the validation set were 0·864 (SD 0·053) for model 1, 0·871 (0·055) for model 2, and 0·882 (0·053) for model 3. The mean AUCs for the test set were 0·859 (SD 0·063) for model 1, 0·863 (0·064) for model 2, and 0·874 (0·061) for model 3. The mean AUCs of each model outperformed previously published performance metrics from the ACS-SRC, with a direct correlation between increasing model complexity and performance. Additionally, when tested on a subgroup of patients who had emergency surgery, our models outperformed previously published POTTER metrics. Interpretation: We have developed unified prediction models, based on deep neural networks, for predicting surgical postoperative complications. The models were generally superior to previously published surgical risk prediction tools and appeared robust to changes in the underlying patient population. Deep learning could offer superior approaches to surgical risk prediction in clinical practice.Item The challenges of modeling and forecasting the spread of COVID-19(National Academy of Sciences, 2020-07-02) Bertozzi, Andrea L.; Franco, Elisa; Mohler, George; Short, Martin B.; Sledge, Daniel; Computer and Information Science, School of ScienceThe coronavirus disease 2019 (COVID-19) pandemic has placed epidemic modeling at the forefront of worldwide public policy making. Nonetheless, modeling and forecasting the spread of COVID-19 remains a challenge. Here, we detail three regional-scale models for forecasting and assessing the course of the pandemic. This work demonstrates the utility of parsimonious models for early-time data and provides an accessible framework for generating policy-relevant insights into its course. We show how these models can be connected to each other and to time series data for a particular region. Capable of measuring and forecasting the impacts of social distancing, these models highlight the dangers of relaxing nonpharmaceutical public health interventions in the absence of a vaccine or antiviral therapies.Item Complications at 10 Years of Follow-up in the Infant Aphakia Treatment Study(Elsevier, 2020-11) Plager, David A.; Bothun, Erick D.; Freedman, Sharon F.; Wilson, M. Edward; Lambert, Scott R.; Ophthalmology, School of MedicineItem Dynamic hierarchical state space forecasting(Wiley, 2024) Liu, Ziyue; Guo, Wensheng; Biostatistics and Health Data Science, Richard M. Fairbanks School of Public HealthIn this paper, we aim to both borrow information from existing units and incorporate the target unit's history data in time series forecasting. We consider a situation when we have time series data from multiple units that share similar patterns when aligned in terms of an internal time. The internal time is defined as an index according to evolving features of interest. When mapped back to the calendar time, these time series can span different time intervals that can include the future calendar time of the targeted unit, over which we can borrow the information from other units in forecasting the targeted unit. We first build a hierarchical state space model for the multiple time series data in terms of the internal time, where the shared components capture the similarities among different units while allowing for unit-specific deviations. A conditional state space model is then constructed to incorporate the information of existing units as the prior information in forecasting the targeted unit. By running the Kalman filtering based on the conditional state space model on the targeted unit, we incorporate both the information from the other units and the history of the targeted unit. The forecasts are then transformed from internal time back into calendar time for ease of interpretation. A simulation study is conducted to evaluate the finite sample performance. Forecasting state-level new COVID-19 cases in United States is used for illustration.Item Endoscopic papillectomy: risk factors for incomplete resection and recurrence during long-term follow-up(Elsevier, 2014-02) Ridtitid, Wiriyaporn; Tan, Damien; Schmidt, Suzette E.; Fogel, Evan L.; McHenry, Lee; Watkins, James L.; Lehman, Glen A.; Sherman, Stuart; Coté, Gregory A.; Department of Medicine, IU School of MedicineBackground Endoscopic papillectomy is increasingly used as an alternative to surgery for ampullary adenomas and other noninvasive ampullary lesions. Objective To measure short-term safety and efficacy of endoscopic papillectomy, define patient and lesion characteristics associated with incomplete endoscopic resection, and measure adenoma recurrence rates during long-term follow-up. Design Retrospective cohort study. Setting Tertiary-care academic medical center. Patients All patients who underwent endoscopic papillectomy for ampullary lesions between July 1995 and June 2012. Intervention Endoscopic papillectomy. Main Outcome Measurements Patient and lesion characteristics associated with incomplete endoscopic resection and ampullary adenoma-free survival analysis. Results We identified 182 patients who underwent endoscopic papillectomy, 134 (73.6%) having complete resection. Short-term adverse events occurred in 34 (18.7%). Risk factors for incomplete resection were jaundice at presentation (odds ratio [OR] 0.21, 95% confidence interval [CI] 0.07–0.69; P = .009), occult adenocarcinoma (OR 0.06, 95% CI, 0.01–0.36; P = .002), and intraductal involvement (OR 0.29, 95% CI, 0.11–0.75; P = .011). The en bloc resection technique was strongly associated with a higher rate of complete resection (OR 4.05, 95% CI, 1.71–9.59; P = .001). Among patients with ampullary adenoma who had complete resection (n = 107), 16 patients (15%) developed recurrence up to 65 months after resection. Limitations Retrospective analysis. Conclusion Jaundice at presentation, occult adenocarcinoma in the resected specimen, and intraductal involvement are associated with a lower rate of complete resection, whereas en bloc papillectomy increases the odds of complete endoscopic resection. Despite complete resection, recurrence was observed up to 5 years after papillectomy, confirming the need for long-term surveillance.Item Full report from the first annual Heart Rhythm Society Research Forum: a vision for our research future, "dream, discover, develop, deliver"(Elsevier, 2011) Albert, Christine M.; Chen, Peng-Sheng; Anderson, Mark E.; Cain, Michael E.; Fishman, Glenn I.; Narayan, Sanjiv M.; Olgin, Jeffrey E.; Spooner, Peter M.; Stevenson, William G.; Van Wagonerd, David R.; Packer, Douglas L.; Heart Rhythm Society Research Task Force; Medicine, School of MedicineItem Predictive utility of an adapted Marshall head CT classification scheme after traumatic brain injury(Taylor & Francis, 2019-01-19) Brown, Allen W.; Pretz, Christopher R.; Bell, Kathleen R.; Hammond, Flora M.; Arciniegas, David B.; Bodien, Yelena G.; Dams-O’Connor, Kristen; Giacino, Joseph T.; Hart, Tessa; Johnson-Greene, Douglas; Kowalski, Robert G.; Walker, William C.; Weintraub, Alan; Zafonte, Ross; Physical Medicine and Rehabilitation, School of MedicineObjective: To study the predictive relationship among persons with traumatic brain injury (TBI) between an objective indicator of injury severity (the adapted Marshall computed tomography [CT] classification scheme) and clinical indicators of injury severity in the acute phase, functional outcomes at inpatient rehabilitation discharge, and functional and participation outcomes at 1 year after injury, including death.Participants: The sample involved 4895 individuals who received inpatient rehabilitation following acute hospitalization for TBI and were enrolled in the Traumatic Brain Injury Model Systems National Database between 1989 and 2014.Design: Head CT variables for each person were fit into adapted Marshall CT classification categories I through IV.Main Measures: Prediction models were developed to determine the amount of variability explained by the CT classification categories compared with commonly used predictors, including a clinical indicator of injury severity.Results: The adapted Marshall classification categories aided only in the prediction of craniotomy or craniectomy during acute hospitalization, otherwise making no meaningful contribution to variance in the multivariable models predicting outcomes at any time point after injury.Conclusion: Results suggest that head CT findings classified in this manner do not inform clinical discussions related to functional prognosis or rehabilitation planning after TBI.Item Prostate cancer grading, time to go back to the future(Wiley, 2021-02) Egevad, Lars; Delahunt, Brett; Bostwick, David G.; Cheng, Liang; Evans, Andrew J.; Gianduzzo, Troy; Graefen, Markus; Hugosson, Jonas; Kench, James G.; Leite, Katia R.M.; Oxley, Jon; Sauter, Guido; Srigley, John R.; Stattin, Pär; Tsuzuki, Toyonori; Yaxley, John; Samaratunga, Hemamali; Pathology and Laboratory Medicine, School of MedicineItem Report of the AMIA EHR-2020 Task Force on the status and future direction of EHRs(Oxford University Press, 2015-09) Payne, Thomas H.; Corley, Sarah; Cullen, Theresa A.; Gandhi, Tejal K.; Harrington, Linda; Kuperman, Gilad J.; Mattison, John E.; McCallie, David P.; McDonald, Clement J.; Tang, Paul C.; Tierney, William M.; Weaver, Charlotte; Weir, Charlene R.; Zaroukian, Michael H.; Department of Medicine, IU School of MedicineItem Three-Year Nursing PhD Model Recommendations from the RWJF Future of Nursing Scholars(Slack, 2022) Rosa, William E.; Hartley, Kim; Hassmiller, Susan B.; Frisch, Stephanie O.; Bennett, Stephanie G.; Breen, Katherine; Goldberg, Jessica I.; Koschmann, Kara S.; Missel, Amanda L.; Parekh de Campos, Amisha; Pho, Anthony T.; Rausch, Jamie; Schlak, Amelia E.; Shook, Alic; Tierney, Meghan K.; Umberfield, Elizabeth; Fairman, Julie A.; School of NursingBackground: In response to the 2011 Future of Nursing report, the Robert Wood Johnson Foundation created the Future of Nursing Scholars (FNS) Program in partnership with select schools of nursing to increase the number of PhD-prepared nurses using a 3-year curriculum. Method: A group of scholars and FNS administrative leaders reflect on lessons learned for stakeholders planning to pursue a 3-year PhD model using personal experiences and extant literature. Results: Several factors should be considered prior to engaging in a 3-year PhD timeline, including mentorship, data collection approaches, methodological choices, and the need to balance multiple personal and professional loyalties. Considerations, strategies, and recommendations are provided for schools of nursing, faculty, mentors, and students. Conclusion: The recommendations provided add to a growing body of knowledge that will create a foundation for understanding what factors constitute "success" for both PhD programs and students.