1
|
Lee CC, Su SY, Sung SF. Machine learning-based survival analysis approaches for predicting the risk of pneumonia post-stroke discharge. Int J Med Inform 2024; 186:105422. [PMID: 38518677 DOI: 10.1016/j.ijmedinf.2024.105422] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/14/2024] [Revised: 02/25/2024] [Accepted: 03/19/2024] [Indexed: 03/24/2024]
Abstract
BACKGROUND Post-stroke pneumonia (PSP) is common among stroke patients. PSP occurring after hospital discharge continues to increase the risk of poor functional outcomes and death among stroke survivors. Currently, there is no prediction model specifically designed to predict the occurrence of PSP beyond the acute stage of stroke. This study aimed to explore the use of machine learning (ML) methods in predicting the risk of PSP after hospital discharge. METHODS This study analyzed data from 5,754 hospitalized stroke patients. The dataset was randomly divided into a training set and a holdout test set, with a ratio of 80:20. Several clinical and laboratory variables were utilized as predictors and different ML algorithms were employed to model time-to-event data. The ML model's predictive performance was compared to existing risk-scoring systems. A model-agnostic method based on Shapley additive explanations was utilized to interpret the ML model. RESULTS The study found that 5.7% of the study patients experienced pneumonia within one year after discharge. Based on repeated 5-fold cross-validation on the training set, the random survival forest (RSF) model had the highest C-index among the various ML algorithms and traditional Cox regression analysis. The final RSF model achieved a C-index of 0.787 (95% confidence interval: 0.737-0.840) on the holdout test set, outperforming five existing risk-scoring systems. The top three important predictors were the Glasgow Coma Scale score, age, and length of hospital stay. CONCLUSIONS The RSF model demonstrated superior discriminative ability compared to other ML algorithms and traditional Cox regression analysis, suggesting a non-linear relationship between predictors and outcomes. The developed ML model can be integrated into the hospital information system to provide personalized risk assessments.
Collapse
Affiliation(s)
- Chang-Ching Lee
- Division of Pulmonary Medicine, Department of Internal Medicine, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City, Taiwan
| | - Sheng-You Su
- Clinical Medicine Research Center, Department of Medical Research, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City, Taiwan
| | - Sheng-Feng Sung
- Division of Neurology, Department of Internal Medicine, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City, Taiwan; Department of Beauty & Health Care, Min-Hwei Junior College of Health Care Management, Tainan, Taiwan.
| |
Collapse
|
2
|
Sumiya K, Shogenji M, Ikenaga Y, Ogawa Y, Hirako K, Fujita A, Shimada T, Hashimoto M, Masuda A, Nagamoto T, Tamai I, Ogura H, Toyama T, Wada T, Sai Y. Association between switching prescribed drugs for lower urinary tract symptoms and independence of urination in post-stroke patients: A retrospective cohort study. J Stroke Cerebrovasc Dis 2023; 32:107419. [PMID: 37839304 DOI: 10.1016/j.jstrokecerebrovasdis.2023.107419] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/06/2023] [Revised: 08/24/2023] [Accepted: 10/06/2023] [Indexed: 10/17/2023] Open
Abstract
OBJECTIVES Stroke patients frequently exhibit loss of independence of urination, and their lower urinary tract symptoms change with the phase of stroke. However, it is unclear whether switching prescribed drugs for lower urinary tract symptoms during hospitalization from acute care wards to convalescence rehabilitation wards affects patients' independence of urination at discharge. It is also unclear whether the impact of switching varies by stroke type. This retrospective cohort study aimed to examine these issues. MATERIALS AND METHODS We analyzed 990 patients registered in the Kaga Regional Cooperation Clinical Pathway for Stroke database during 2015-2019. Prescriptions for lower urinary tract symptoms from pre-onset to convalescence rehabilitation were surveyed. Logistic regression analysis was performed to examine the association between switching drugs and independence of urination based on bladder management and voiding location at discharge. Stroke types were also examined in subgroup analyses. RESULTS About 21 % of patients had their lower urinary tract symptoms prescriptions switched during hospitalization. Switching was positively associated with independence of bladder management (odds ratio 1.65, 95 % confidence interval 1.07 to 2.49) and voiding location (odds ratio 2.72, 95 % confidence interval 1.72 to 4.37). Similar associations were observed in different stroke types. CONCLUSIONS Approximately 20 % of patients had their lower urinary tract symptoms medications switched upon transfer from acute to convalescence rehabilitation wards. Switching was significantly associated with improved urinary independence at discharge. Consistent results were observed across different stroke types, suggesting that switching medications contributes to urinary independence after stroke, regardless of the etiology or severity of stroke.
Collapse
Affiliation(s)
- Koyomi Sumiya
- Department of Clinical Pharmacokinetics, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan
| | - Miho Shogenji
- Faculty of Health Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Ishikawa, Japan
| | - Yasunori Ikenaga
- Department of Rehabilitation Medicine, Yawata Medical Center, Ishikawa, Japan; Council of Kaga Local Stroke Network, South Ishikawa, Japan
| | - Yoru Ogawa
- Council of Kaga Local Stroke Network, South Ishikawa, Japan; Department of Pharmacy, Komatsu Municipal Hospital, Ishikawa, Japan
| | - Kohei Hirako
- Frontier Science and Social Co-creation Initiative, Kanazawa University, Ishikawa, Japan; The Establishment Preparation Office for The Faculty of Interdisciplinary Economics, Kinjo University, Ishikawa, Japan
| | - Arimi Fujita
- Department of Clinical Pharmacokinetics, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan; Department of Hospital Pharmacy, University Hospital, Kanazawa University, Ishikawa, Japan
| | - Tsutomu Shimada
- Department of Clinical Pharmacokinetics, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan; Department of Hospital Pharmacy, University Hospital, Kanazawa University, Ishikawa, Japan.
| | | | | | | | - Ikumi Tamai
- Division of Pharmacy, Graduate School of Pharmaceutical Sciences, Kanazawa University, Ishikawa, Japan; AI Hospital/Macro Signal Dynamics Research and Development Center, Kanazawa University, Ishikawa, Japan
| | - Hisayuki Ogura
- AI Hospital/Macro Signal Dynamics Research and Development Center, Kanazawa University, Ishikawa, Japan; Department of Nephrology and Laboratory Medicine, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Ishikawa, Japan
| | - Tadashi Toyama
- AI Hospital/Macro Signal Dynamics Research and Development Center, Kanazawa University, Ishikawa, Japan; Department of Nephrology and Laboratory Medicine, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Ishikawa, Japan
| | - Takashi Wada
- Department of Nephrology and Laboratory Medicine, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Ishikawa, Japan
| | - Yoshimichi Sai
- Department of Clinical Pharmacokinetics, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan; Department of Hospital Pharmacy, University Hospital, Kanazawa University, Ishikawa, Japan; AI Hospital/Macro Signal Dynamics Research and Development Center, Kanazawa University, Ishikawa, Japan
| |
Collapse
|
3
|
Ikenaga Y, Fudeya M, Kusunoki T, Yamaguchi H. Factors Contributing to Complete Oral Intake in Dysphagic Stroke Patients with Enteral Feeding Tubes in Convalescent Rehabilitation Wards. Prog Rehabil Med 2023; 8:20230011. [PMID: 37006382 PMCID: PMC10061229 DOI: 10.2490/prm.20230011] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/19/2022] [Accepted: 03/01/2023] [Indexed: 04/03/2023] Open
Abstract
Objectives: This study investigated the factors contributing to complete oral intake (COI) in dysphagic stroke patients with enteral feeding tubes in the local clinical setting. Methods: Data of patients with percutaneous endoscopic gastrostomy (PEG) or nasogastric tube (NGT) feeding on admission to convalescent rehabilitation wards (CRWs) were extracted from the Kaga Regional Cooperation Clinical Pathway for Stroke database for multiple centers including 19 acute care hospitals and 11 hospitals with CRWs. Patients were divided into two groups based on their status regarding COI or incomplete oral intake (ICOI) at discharge. Logistic regression analysis with forced-entry variables was used to identify factors contributing to COI. Results: On discharge from CRWs, COI and ICOI were observed in 140 and 207 cases, respectively. The COI group was younger, had a higher rate of initial stroke, higher Functional Oral Intake Scale (FOIS) scores, higher Functional Independence Measure (FIM) motor and cognitive scores, higher Body Mass Index (BMI), lower rate of patients with PEG, and shorter stays in acute care wards. Logistic regression analysis with forced entry revealed that younger age; initial stroke; higher FOIS score, FIM cognitive score, and BMI; and shorter stay in the acute care ward contributed to COI. Conclusions: The primary factors contributing to COI in dysphagic stroke patients with enteral feeding tubes were younger age, initial stroke, higher swallowing and cognitive function, good nutritional status, and shorter stay in the acute care ward.
Collapse
Affiliation(s)
- Yasunori Ikenaga
- Department of Rehabilitation Medicine, Yawata Medical Center, Komatsu, Japan
- Council of Kaga Local Stroke Network, Nonoichi, Ishikawa, Japan
| | - Masami Fudeya
- Council of Kaga Local Stroke Network, Nonoichi, Ishikawa, Japan
| | | | - Hiromi Yamaguchi
- Department of Rehabilitation Medicine, Yawata Medical Center, Komatsu, Japan
| |
Collapse
|
4
|
Willman J, Lucke-Wold B. “Timing of percutaneous endoscopic gastrostomy tube placement in post-stroke patients does not impact mortality, complications, or outcomes”: Commentary. World J Gastrointest Pharmacol Ther 2023; 14:1-3. [PMID: 36743219 PMCID: PMC9896468 DOI: 10.4292/wjgpt.v14.i1.1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 11/04/2022] [Revised: 11/24/2022] [Accepted: 01/17/2023] [Indexed: 01/19/2023] Open
Abstract
In this commentary, we summarize some of the key points of the original paper “Timing of percutaneous endoscopic gastrostomy tube placement in post-stroke patients does not impact mortality, complications, or outcomes” and offer support for the proposed results. Specifically, we address how early percutaneous endoscopic gastrostomy (PEG) tube placement may reduce hospital length of stay and costs. We also discuss topics related to the article including PEG weaning and post-stroke nutritional formulation. However, we note that concerns purported by previous studies that early PEG placement may worsen outcomes are not fully addressed, and further research is needed.
Collapse
Affiliation(s)
- Jonathan Willman
- College of Medicine, University of Florida, Gainesville, FL 32610, United States
| | - Brandon Lucke-Wold
- Department of Neurosurgery, University of Florida, Gainesville, FL 32610, United States
| |
Collapse
|
5
|
Shogenji M, Yoshida M, Sumiya K, Shimada T, Ikenaga Y, Ogawa Y, Hirako K, Sai Y. Relationship between Bowel/Bladder Function and Discharge in Older Stroke Patients in Convalescent Rehabilitation Wards: A Retrospective Cohort Study. Prog Rehabil Med 2022; 7:20220028. [PMID: 35663118 PMCID: PMC9126742 DOI: 10.2490/prm.20220028] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/28/2021] [Accepted: 05/10/2022] [Indexed: 11/25/2022] Open
Abstract
Objectives : Many stroke patients experience motor and cognitive dysfunctions that make living at home challenging. We aimed to identify the factors associated with hospital discharge to home in older stroke patients in convalescent rehabilitation wards where intensive and comprehensive inpatient rehabilitation are performed following acute-phase treatment. Methods : A retrospective cohort study was conducted among 1227 older stroke patients registered in the database of the Council of Kaga Local Stroke Network, Japan, between 2015 and 2019. Patients’ basic characteristics, discharge destination, type and severity of stroke, cognitive status, and activities of daily living (ADL) including continence were evaluated. Results : The proportion of subjects discharged to home was 62.3%. The mean hospital stay in the home discharge group was shorter than that in the non-home discharge group (111 days vs. 144.6 days, P <0.001). The following factors were associated with discharge to home: age (adjusted odds ratio [AOR]: 2.801, 95% confidence interval [CI] [1.473, 2.940]; P <0.001), sex (AOR: 1.513, 95% CI [1.112, 2.059]), stroke type (AOR: 1.426, 95% CI [1.013, 2.007]), low cognitive status (AOR: 3.750, 95% CI [2.615, 5.379]), low level of bladder control (AOR: 2.056, 95% CI [1.223, 3.454]), and low level of bowel control (AOR: 2.823, 95% CI [1.688, 4.722]). Conclusions : Age, sex, stroke type, cognitive function, and ADL scores for bladder and bowel control were associated with discharge to home. Improving continence management regarding both voiding and defecation may be a promising care strategy to promote hospital discharge to home in older stroke patients.
Collapse
Affiliation(s)
- Miho Shogenji
- Faculty of Health Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Ishikawa, Japan
| | - Mikako Yoshida
- Department of Women’s Health Nursing & Midwifery, Tohoku University Graduate School of Medicine, Miyagi, Japan
| | - Koyomi Sumiya
- Department of Clinical Pharmacokinetics, Graduate School of Medical Sciences, Kanazawa University, Ishikawa, Japan
| | - Tsutomu Shimada
- Department of Hospital Pharmacy, University Hospital, Kanazawa University, Ishikawa, Japan
| | - Yasunori Ikenaga
- Department of Rehabilitation Medicine, Yawata Medical Center, Ishikawa, Japan
| | - Yoru Ogawa
- Department of Pharmacy, Komatsu Municipal Hospital, Ishikawa, Japan
| | - Kohei Hirako
- Frontier Science and Social Co-creation Initiative, Kanazawa University, Ishikawa, Japan
| | - Yoshimichi Sai
- Department of Hospital Pharmacy, University Hospital, Kanazawa University, Ishikawa, Japan
| |
Collapse
|