1
|
Roy S, Mincu D, Loreaux E, Mottram A, Protsyuk I, Harris N, Xue Y, Schrouff J, Montgomery H, Connell A, Tomasev N, Karthikesalingam A, Seneviratne M. Multitask prediction of organ dysfunction in the intensive care unit using sequential subnetwork routing. J Am Med Inform Assoc 2021; 28:1936-1946. [PMID: 34151965 PMCID: PMC8363803 DOI: 10.1093/jamia/ocab101] [Citation(s) in RCA: 6] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/23/2020] [Revised: 03/07/2021] [Accepted: 05/14/2021] [Indexed: 12/29/2022] Open
Abstract
Objective Multitask learning (MTL) using electronic health records allows concurrent prediction of multiple endpoints. MTL has shown promise in improving model performance and training efficiency; however, it often suffers from negative transfer – impaired learning if tasks are not appropriately selected. We introduce a sequential subnetwork routing (SeqSNR) architecture that uses soft parameter sharing to find related tasks and encourage cross-learning between them. Materials and Methods Using the MIMIC-III (Medical Information Mart for Intensive Care-III) dataset, we train deep neural network models to predict the onset of 6 endpoints including specific organ dysfunctions and general clinical outcomes: acute kidney injury, continuous renal replacement therapy, mechanical ventilation, vasoactive medications, mortality, and length of stay. We compare single-task (ST) models with naive multitask and SeqSNR in terms of discriminative performance and label efficiency. Results SeqSNR showed a modest yet statistically significant performance boost across 4 of 6 tasks compared with ST and naive multitasking. When the size of the training dataset was reduced for a given task (label efficiency), SeqSNR outperformed ST for all cases showing an average area under the precision-recall curve boost of 2.1%, 2.9%, and 2.1% for tasks using 1%, 5%, and 10% of labels, respectively. Conclusions The SeqSNR architecture shows superior label efficiency compared with ST and naive multitasking, suggesting utility in scenarios in which endpoint labels are difficult to ascertain.
Collapse
Affiliation(s)
| | | | | | | | | | | | - Yuan Xue
- Google Health, Mountain View, California, USA
| | | | - Hugh Montgomery
- Centre for Human Health and Performance, University College London, London, United Kingdom
| | | | | | | | | |
Collapse
|
2
|
Tomašev N, Glorot X, Rae JW, Zielinski M, Askham H, Saraiva A, Mottram A, Meyer C, Ravuri S, Protsyuk I, Connell A, Hughes CO, Karthikesalingam A, Cornebise J, Montgomery H, Rees G, Laing C, Baker CR, Peterson K, Reeves R, Hassabis D, King D, Suleyman M, Back T, Nielson C, Ledsam JR, Mohamed S. A clinically applicable approach to continuous prediction of future acute kidney injury. Nature 2019; 572:116-119. [PMID: 31367026 PMCID: PMC6722431 DOI: 10.1038/s41586-019-1390-1] [Citation(s) in RCA: 482] [Impact Index Per Article: 96.4] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/13/2018] [Accepted: 06/18/2019] [Indexed: 12/31/2022]
Abstract
The early prediction of deterioration could have an important role in supporting healthcare professionals, as an estimated 11% of deaths in hospital follow a failure to promptly recognize and treat deteriorating patients1. To achieve this goal requires predictions of patient risk that are continuously updated and accurate, and delivered at an individual level with sufficient context and enough time to act. Here we develop a deep learning approach for the continuous risk prediction of future deterioration in patients, building on recent work that models adverse events from electronic health records2-17 and using acute kidney injury-a common and potentially life-threatening condition18-as an exemplar. Our model was developed on a large, longitudinal dataset of electronic health records that cover diverse clinical environments, comprising 703,782 adult patients across 172 inpatient and 1,062 outpatient sites. Our model predicts 55.8% of all inpatient episodes of acute kidney injury, and 90.2% of all acute kidney injuries that required subsequent administration of dialysis, with a lead time of up to 48 h and a ratio of 2 false alerts for every true alert. In addition to predicting future acute kidney injury, our model provides confidence assessments and a list of the clinical features that are most salient to each prediction, alongside predicted future trajectories for clinically relevant blood tests9. Although the recognition and prompt treatment of acute kidney injury is known to be challenging, our approach may offer opportunities for identifying patients at risk within a time window that enables early treatment.
Collapse
Affiliation(s)
| | | | - Jack W Rae
- DeepMind, London, UK
- CoMPLEX, Computer Science, University College London, London, UK
| | | | | | | | | | | | | | | | | | | | | | | | - Hugh Montgomery
- Institute for Human Health and Performance, University College London, London, UK
| | - Geraint Rees
- Institute of Cognitive Neuroscience, University College London, London, UK
| | - Chris Laing
- University College London Hospitals, London, UK
| | | | - Kelly Peterson
- VA Salt Lake City Healthcare System, Salt Lake City, UT, USA
- Division of Epidemiology, University of Utah, Salt Lake City, UT, USA
| | - Ruth Reeves
- Department of Veterans Affairs, Nashville, TN, USA
| | | | | | | | | | - Christopher Nielson
- University of Nevada School of Medicine, Reno, NV, USA
- Department of Veterans Affairs, Salt Lake City, UT, USA
| | | | | |
Collapse
|
7
|
Mottram A. "Like a trip to McDonalds": a grounded theory study of patient experiences of day surgery. Int J Nurs Stud 2010; 48:165-74. [PMID: 20678770 DOI: 10.1016/j.ijnurstu.2010.07.007] [Citation(s) in RCA: 33] [Impact Index Per Article: 2.4] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/28/2010] [Revised: 07/05/2010] [Accepted: 07/07/2010] [Indexed: 10/19/2022]
Abstract
BACKGROUND The amount and complexity of (ambulatory) day surgery is rapidly expanding internationally. Nurses have a responsibility to provide quality care for day surgery patients. To do this they must understand all aspects of the patient experience. There is dearth of research into day surgery using a sociological frame of reference. OBJECTIVE The study investigated patients' experiences of day surgery using a sociological frame of reference. DESIGN A qualitative study using the grounded theory approach was used. SETTING The study was based in two day surgery units in two urban public hospitals in the United Kingdom. PARTICIPANTS 145 patients aged 18-70 years and 100 carers were purposely selected from the orthopaedic, ear nose and throat and general surgical lists. They were all English speaking and were of varied socio-economic background. METHODS The data was collected from 2004 to 2006. Semi-structured interviews were conducted on three occasions: before surgery, 48 h following surgery and one month following discharge. Permission was received from the Local Research Ethics Committee. Analysis of the data involved line-by-line analysis, compilation of key words and phrases (codes) and constant comparison of the codes until categories emerged. FINDINGS Patients liked day surgery and placed it within the wider societal context of efficiency and speed. Time was a major issue for them. They wished surgery, like all other aspects of their life to be a speedy process. They likened it to a McDonald's experience with its emphasis on speed, predictability and control. CONCLUSION This study throws new light on patient experiences and offers an understanding of day surgery against a western culture which emphasises the importance of speed and efficiency. It is a popular choice for patients but at times it can be seen to be a mechanistic way of providing care. The implications for nurses to provide education and information to add to the quality of the patient experience are discussed.
Collapse
Affiliation(s)
- Anne Mottram
- School of Nursing and Health Visiting, Mary Seacole Building, University of Salford, Frederick Road, Salford M6 6PU, United Kingdom.
| |
Collapse
|
11
|
Mottram A. The hospital revolution John Riddington Young The hospital revolution Metro £9.99 243 978184454595 [Formula: see text]. Nurs Manag (Harrow) 2009; 16:9. [PMID: 27753332 DOI: 10.7748/nm.16.2.9.s14] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 06/06/2023]
Abstract
THIS BOOK appears to have been born out of the sense of frustration of its authors, who, as doctors, must cope with the current management ethos of the NHS.
Collapse
|
12
|
Abstract
An increasing number of complex surgical interventions are now taking place on a day case basis with some surgical specialities able to perform 80% of their elective surgery as day surgery. It is important that student nurses are exposed to clinical practice within the day surgery unit. Some students, for a variety of reasons, exhibit a reluctance to experience a day-surgery placement. The writer describes a programme of study which takes place before the students take up their placements to demonstrate that day surgery offers many opportunities for the delivery of highly skilled, specialised nursing care. Day Surgery Nursing is emerging as a speciality in its own right. The clinical skills of the nurse are required alongside the interpersonal, informational and psychological care-giving skills to ensure safety and comfort for the patient whilst in the unit and transfer home where responsibility for care, normally performed by nurses, now lies with the patients and their carers.
Collapse
Affiliation(s)
- A Mottram
- University of Salford, Peel House, Albert Street, M30 ONN, Eccles, UK
| |
Collapse
|