Improving palliative care with deep learning

Avati, Anand; Jung, Kenneth; Harman, Stephanie; Downing, Lance; Ng, Andrew Y.; Shah, Nigam H.

doi:10.1186/s12911-018-0677-8

Cited by 248 publications

(215 citation statements)

References 46 publications

Supporting

Mentioning

206

Contrasting

Unclassified

Order By: Relevance

“…Out of the numerous potential predictors of outcome, 3,[8][9][10][11][12][13] our findings suggest that there are four very important ones, all of which are readily available. Our findings also highlight the importance of accurately recording a complete set of vital signs and assessing mobility on all ED attendees.…”

Section: Ta B L E 1 Differences Between Cohortsmentioning

confidence: 99%

“…A modification of the CriSTAL prognostic model has recently been suggested for the assessment of frail older ED patients. 6 However, it is clearly not practical to use complex or inconvenient scores, [7][8][9] or those that require laboratory information, 10,11 or access to large amounts of administrative data 12,13 on every patient attending an ED. The question "Would you be surprised if this patient died within the next 6 to 12 months?"…”

Section: Importancementioning

confidence: 99%

See 1 more Smart Citation

A simple prognostic score predicts one‐year mortality of alert and calm emergency department patients: A prospective two‐center observational study

et al. 2020

View full text Add to dashboard Cite

Study Objective To derive and validate a prognostic score to predict 1‐year mortality using vital signs, mobility and other variables that are readily available at the bedside at no additional cost. Methods Post hoc analysis of two independent prospective observational studies in two emergency departments, one in Denmark and the other in Switzerland. Participants Alert and calm emergency department patients. Measurements The prediction of mortality from presentation to 365 days by vital signs, mobility and other variables that are readily available at the bedside at no additional cost. Results One thousand six hundred and eighteen alert and calm patients were in the Danish cohort and 1331 in the Swiss cohort. Logistic regression identified age >68 years, abnormal vital signs, impaired mobility and the decision to admit as significant predictors of 365‐day mortality. A simple prognostic score awarded one point to each of these predictors. Less than two of these predictors were present in 45.6% of patients, and only 0.4% of these patients died within a year. If two or more of these predictors were present, 365‐day mortality increased exponentially. Conclusion Age >68 years, the decision for hospital admission, any vital sign abnormality at presentation and impaired mobility at presentation are equally powerful predictors of 1‐year mortality in alert and calm emergency department patients. If validated by others these predictors could be used to discharge patients with confidence since nearly half of these patients had less than two predictors and none of them died within 30 days. However, when two or more predictors were present 365‐day mortality increased exponentially.

show abstract

Section: Ta B L E 1 Differences Between Cohortsmentioning

confidence: 99%

Section: Importancementioning

confidence: 99%

A simple prognostic score predicts one‐year mortality of alert and calm emergency department patients: A prospective two‐center observational study

et al. 2020

View full text Add to dashboard Cite

show abstract

“…To explain a patient's risk, statistics of these discretized features can be used such as the odds ratio or the Rothman index [12]. Other lines of work have studied latent Dirichlet allocation [44], convolutional neural networks with feature ablation [45], RNNs with an attention mechanism [46,47], and co-distillation [48,49].…”

Section: Related Workmentioning

confidence: 99%

Explaining an increase in predicted risk for clinical alerts

Hardt

Rajkomar

Flores

et al. 2020

Proceedings of the ACM Conference on Health, Inference, and Learning

View full text Add to dashboard Cite

Much work aims to explain a model's prediction on a static input. We consider explanations in a temporal setting where a stateful dynamical model produces a sequence of risk estimates given an input at each time step. When the estimated risk increases, the goal of the explanation is to attribute the increase to a few relevant inputs from the past.While our formal setup and techniques are general, we carry out an in-depth case study in a clinical setting. The goal here is to alert a clinician when a patient's risk of deterioration rises. The clinician then has to decide whether to intervene and adjust the treatment. Given a potentially long sequence of new events since she last saw the patient, a concise explanation helps her to quickly triage the alert.We develop methods to lift static attribution techniques to the dynamical setting, where we identify and address challenges specific to dynamics. We then experimentally assess the utility of different explanations of clinical alerts through expert evaluation. * Google

show abstract

“…. , θ (k+1) ) be a partitioning of the full parameter vector, and assume a similar partitioning of the momenta (p (1) , p (2) , . .…”

Section: Partitioned Discretization Algorithms For Deep Neural Networkmentioning

confidence: 99%

“…For h → 0, due to the h 2 dependency in (19), we can think of θ (2) n as approximately fixed at sayθ (2) in (17), (18), (20). This is then a symplectic (leapfrog) discretization of a Hamiltonian system with Hamiltonian H(θ (1) , p (1) ) = p (1) 2 /2 + L(θ (1) ,θ (2) ), meaning that it is not at all dissipative as h → 0. Even if, for small h, the system would seem to make incremental loss-reducing steps via the slow gradient descent component, it would seem unwise to rely on this weak mechanism of dissipation.…”

Section: Langevin-overdamped Langevin (Lol) and Dissipated Hamiltoniamentioning

confidence: 99%

Partitioned integrators for thermodynamic parameterization of neural networks

Leimkuhler¹,

Matthews²,

Vlaar³

2019

Foundations of Data Science

View full text Add to dashboard Cite

Traditionally, neural networks are parameterized using optimization procedures such as stochastic gradient descent, RMSProp and ADAM. These procedures tend to drive the parameters of the network toward a local minimum. Various algorithmic devices have been introduced to enhance (or, in some cases, delay) the convergence of the optimizer, to increase the exploration of low loss states. These methods ultimately provide a crude sampling for a probability distribution on parameter space. As an alternative, one may derive "sampling" algorithms (referred to here as "thermodynamic parameterization methods") for a defined target distribution on parameter space. By using state-of-theart discretizations of stochastic differential equations with a known invariant probability distribution, we argue that it is much easier to control the properties of the distribution. Stochastic Gradient Langevin Dynamics, the "unadjusted Langevin algorithm", and Adaptive Langevin Dynamics (also known as Stochastic Gradient Nosé-Hoover dynamics) are examples of existing thermodynamic parameterization methods in use for machine learning, but these can be substantially improved. We find that by partitioning the parameters based on natural layer structure we obtain schemes with rapid convergence for data sets with complicated loss landscapes. We describe easy-to-implement hybrid partitioned numerical algorithms, based on discretized stochastic differential equations, which are adapted to feed-forward neural networks, including LaLa (a multi-layer Langevin algorithm), AdLaLa (combining the adaptive Langevin and Langevin algorithms) and LOL (combining Langevin and Overdamped Langevin); we examine the convergence of these methods using numerical studies and compare their performance among themselves and in relation to standard alternatives such as stochastic gradient descent and ADAM. We present evidence that thermodynamic parameterization methods can be (i) faster, (ii) more accurate, and (iii) more robust than standard algorithms incorporated into machine learning frameworks, in particular for data sets with complicated loss landscapes. Moreover, we show in numerical studies that methods based on sampling excite many degrees of freedom. The equipartition property, which is a consequence of their ergodicity, means that these methods keep in play an ensemble of low-loss states during the training process. By drawing parameter states from a sufficiently rich distribution of nearby candidate states, we show that the thermodynamic schemes produce smoother classifiers, improve generalization and reduce overfitting compared to traditional optimizers.

show abstract

Improving palliative care with deep learning

Cited by 248 publications

References 46 publications

A simple prognostic score predicts one‐year mortality of alert and calm emergency department patients: A prospective two‐center observational study

A simple prognostic score predicts one‐year mortality of alert and calm emergency department patients: A prospective two‐center observational study

Explaining an increase in predicted risk for clinical alerts

Partitioned integrators for thermodynamic parameterization of neural networks

Contact Info

Product

Resources

About