Towards the systematic simplification of mechanistic models

Mechanistic models used for prediction should be parsimonious, as models which are over-parameterised may have poor predictive performance. Determining whether a model is parsimonious requires comparisons with alternative model formulations with differing levels of complexity. However, creating alternative formulations for large mechanistic models is often problematic, and usually time-consuming. Consequently, few are ever investigated. In this paper, we present an approach which rapidly generates reduced model formulations by replacing a model’s variables with constants. These reduced alternatives can be compared to the original model, using data based model selection criteria, to assist in the identification of potentially unnecessary model complexity, and thereby inform reformulation of the model. To illustrate the approach, we present its application to a published radiocaesium plant-uptake model, which predicts uptake on the basis of soil characteristics (e.g. pH, organic matter content, clay content). A total of 1024 reduced model formulations were generated, and ranked according to five model selection criteria: Residual Sum of Squares (RSS), AICc, BIC, MDL and ICOMP. The lowest scores for RSS and AICc occurred for the same reduced model in which pH dependent model components were replaced. The lowest scores for BIC, MDL and ICOMP occurred for a further reduced model in which model components related to the distinction between adsorption on clay and organic surfaces were replaced. Both these reduced models had a lower RSS for the parameterisation dataset than the original model. As a test of their predictive performance, the original model and the two reduced models outlined above were used to predict an independent dataset. The reduced models have lower prediction sums of squares than the original model, suggesting that the latter may be overfitted. The approach presented has the potential to inform model development by rapidly creating a class of alternative model formulations, which can be compared.

Interventions for dysphagia in acute stroke (Review)

Background It is unclear how dysphagic patients should be fed and treated after acute stroke. Objectives The objective of this review was to assess the effect of different management strategies for dysphagic stroke patients, in particular how and when to feed, whether to supplement nutritional intake, and how and whether to treat dysphagia. Search strategy We searched the Cochrane Stroke Group trials register, Medline, Embase, ISI, and existing review articles.We contacted researchers in the field and equipment manufacturers. Date of the most recent searches: March 1999. Selection criteria Unconfounded truly or quasi randomised controlled trials in dysphagic patients with acute/subacute (within 3 months) stroke. Data collection and analysis Three reviewers independently applied the trial inclusion criteria. Two reviewers assessed trial quality and extracted the data. Main results Percutaneous endoscopic gastrostomy (PEG) versus nasogastric tube (NGT) feeding: two trials (49 patients) suggest that PEG reduces end-of-trial case fatality (Peto Odds Ratio, OR 0.28, 95% CI 0.09 to 0.89) and treatment failures (OR 0.10, 95% CI 0.02 to 0.52), and improves nutritional status, assessed as weight (Weighted Men Difference, WMD +4.1 kg, 95% CI -4.3 to +12.5), mid-arm circumference (WMD +2.2 cm, 95% CI -0.5 to +4.9) or serum albumin (WMD + 7.0 g/l, 95% CI +4.9 to +9.1) as compared with NGT feeding; two larger studies are ongoing. Timing of feeding: no completed trials; one large study is ongoing. Swallowing therapy for dysphagia: two trials (85 patients) suggest that formal swallowing therapy does not significantly reduce end-of-trial dysphagia rates (OR 0.55, 95%CI 0.18 to 1.66). Drug therapy for dysphagia: one trial (17 patients); nifedipine did not alter end-of-trial case fatality or the frequency of dysphagia. Nutritional supplementation: one trial (42 patients) found a non-significant trend to a lower case fatality, and significantly increased energy and protein intake; one large trial is ongoing and data is awaited from two other studies. Fluid supplementation: one trial (20 patients) found that supplementation did not alter the time to resolution of dysphagia. Authors’ conclusions Too few studies have been performed, and these have involved too few patients. PEG feeding may improve outcome and nutrition as compared with NGT feeding. Further research is required to assess how and when patients are fed, and the effect of swallowing or drug therapy on dysphagia.