ORCID Profile
0000-0002-8020-1587
Current Organisations
Flinders University
,
GenesisCare Radiation Oncology
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
Publisher: JMIR Publications Inc.
Date: 12-2020
Abstract: he Multidimensional Prognostic Index (MPI) is an aggregate, comprehensive, geriatric assessment scoring system derived from eight domains that predict adverse outcomes, including 12-month mortality. However, the prediction accuracy of using the three MPI categories (mild, moderate, and severe risk) was relatively poor in a study of older hospitalized Australian patients. Prediction modeling using the component domains of the MPI together with additional clinical features and machine learning (ML) algorithms might improve prediction accuracy. his study aims to assess whether the accuracy of prediction for 12-month mortality using logistic regression with maximum likelihood estimation (LR-MLE) with the 3-category MPI together with age and gender (feature set 1) can be improved with the addition of 10 clinical features (sodium, hemoglobin, albumin, creatinine, urea, urea-to-creatinine ratio, estimated glomerular filtration rate, C-reactive protein, BMI, and anticholinergic risk score feature set 2) and the replacement of the 3-category MPI in feature sets 1 and 2 with the eight separate MPI domains (feature sets 3 and 4, respectively), and to assess the prediction accuracy of the ML algorithms using the same feature sets. PI and clinical features were collected from patients aged 65 years and above who were admitted to either the general medical or acute care of the elderly wards of a South Australian hospital between September 2015 and February 2017. The diagnostic accuracy of LR-MLE was assessed together with nine ML algorithms: decision trees, random forests, extreme gradient boosting (XGBoost), support-vector machines, naïve Bayes, K-nearest neighbors, ridge regression, logistic regression without regularization, and neural networks. A 70:30 training set:test set split of the data and a grid search of hyper-parameters with 10-fold cross-validation—was used during model training. The area under the curve was used as the primary measure of accuracy. total of 737 patients (female: 370/737, 50.2% male: 367/737, 49.8%) with a median age of 80 (IQR 72-86) years had complete MPI data recorded on admission and had completed the 12-month follow-up. The area under the receiver operating curve for LR-MLE was 0.632, 0.688, 0.738, and 0.757 for feature sets 1 to 4, respectively. The best overall accuracy for the nine ML algorithms was obtained using the XGBoost algorithm (0.635, 0.706, 0.756, and 0.757 for feature sets 1 to 4, respectively). he use of MPI domains with LR-MLE considerably improved the prediction accuracy compared with that obtained using the traditional 3-category MPI. The XGBoost ML algorithm slightly improved accuracy compared with LR-MLE, and adding clinical data improved accuracy. These results build on previous work on the MPI and suggest that implementing risk scores based on MPI domains and clinical data by using ML prediction models can support clinical decision-making with respect to risk stratification for the follow-up care of older hospitalized patients.
Publisher: JMIR Publications Inc.
Date: 21-06-2021
DOI: 10.2196/26139
Abstract: The Multidimensional Prognostic Index (MPI) is an aggregate, comprehensive, geriatric assessment scoring system derived from eight domains that predict adverse outcomes, including 12-month mortality. However, the prediction accuracy of using the three MPI categories (mild, moderate, and severe risk) was relatively poor in a study of older hospitalized Australian patients. Prediction modeling using the component domains of the MPI together with additional clinical features and machine learning (ML) algorithms might improve prediction accuracy. This study aims to assess whether the accuracy of prediction for 12-month mortality using logistic regression with maximum likelihood estimation (LR-MLE) with the 3-category MPI together with age and gender (feature set 1) can be improved with the addition of 10 clinical features (sodium, hemoglobin, albumin, creatinine, urea, urea-to-creatinine ratio, estimated glomerular filtration rate, C-reactive protein, BMI, and anticholinergic risk score feature set 2) and the replacement of the 3-category MPI in feature sets 1 and 2 with the eight separate MPI domains (feature sets 3 and 4, respectively), and to assess the prediction accuracy of the ML algorithms using the same feature sets. MPI and clinical features were collected from patients aged 65 years and above who were admitted to either the general medical or acute care of the elderly wards of a South Australian hospital between September 2015 and February 2017. The diagnostic accuracy of LR-MLE was assessed together with nine ML algorithms: decision trees, random forests, extreme gradient boosting (XGBoost), support-vector machines, naïve Bayes, K-nearest neighbors, ridge regression, logistic regression without regularization, and neural networks. A 70:30 training set:test set split of the data and a grid search of hyper-parameters with 10-fold cross-validation—was used during model training. The area under the curve was used as the primary measure of accuracy. A total of 737 patients (female: 370/737, 50.2% male: 367/737, 49.8%) with a median age of 80 (IQR 72-86) years had complete MPI data recorded on admission and had completed the 12-month follow-up. The area under the receiver operating curve for LR-MLE was 0.632, 0.688, 0.738, and 0.757 for feature sets 1 to 4, respectively. The best overall accuracy for the nine ML algorithms was obtained using the XGBoost algorithm (0.635, 0.706, 0.756, and 0.757 for feature sets 1 to 4, respectively). The use of MPI domains with LR-MLE considerably improved the prediction accuracy compared with that obtained using the traditional 3-category MPI. The XGBoost ML algorithm slightly improved accuracy compared with LR-MLE, and adding clinical data improved accuracy. These results build on previous work on the MPI and suggest that implementing risk scores based on MPI domains and clinical data by using ML prediction models can support clinical decision-making with respect to risk stratification for the follow-up care of older hospitalized patients.
Publisher: Elsevier BV
Date: 05-2020
DOI: 10.1016/J.CLNU.2019.07.011
Abstract: The literature regarding enteral nutrition and mortality in older frail people is limited and still conflicting. Moreover, the potential role of comprehensive geriatric assessment is poorly explored. We therefore aimed to investigate whether the Multidimensional Prognostic Index (MPI), an established tool that assesses measures of frailty and predicts mortality, may help physicians in identifying patients in whom ETF (enteral tube feeding) is effective in terms of reduced mortality. Observational, longitudinal, multicenter study with one year of follow-up. Data regarding ETF were recorded through medical records. A standardized comprehensive geriatric assessment was used to calculate the MPI. Participants were ided in low (MPI-1), moderate (MPI-2) or severe (MPI-3) risk of mortality. Data regarding mortality were recorded through administrative information. 1064 patients were included, with 79 (13 in MPI 1-2 and 66 in MPI-3 class) receiving ETF. In multivariable analysis, patients receiving ETF experienced a higher risk of death (odds ratio, OR = 2.00 95% confidence intervals, CI: 1.19-3.38). However, after stratifying for their MPI at admission, mortality was higher in MPI-3 class patients (OR = 2.03 95%CI: 1.09-3.76), but not in MPI 1-2 class patients (OR = 1.51 95%CI: 0.44-5.25). The use of propensity score confirmed these findings. ETF is associated with a higher risk of death. However, this is limited to more frail patients, suggesting the importance of the MPI in the prognostic evaluation of ETF.
Publisher: MDPI AG
Date: 11-2019
DOI: 10.3390/JCM8111820
Abstract: Background and aims: The Multidimensional Prognostic Index (MPI), an objective and quantifiable tool based on the Comprehensive Geriatric Assessment, has been shown to predict adverse outcomes in European cohorts. We conducted a validation study of the original MPI, and of adapted versions that accounted for the use of specific drugs and cultural ersity in the assessment of cognition, in older Australians. Methods: The capacity of the MPI to predict 12-month mortality was assessed in 697 patients (median age: 80 years interquartile range: 72–86) admitted to a metropolitan teaching hospital between September 2015 and February 2017. Results: In simple logistic regression analysis, the MPI was associated with 12-month mortality (Low risk: OR reference group moderate risk: OR 2.50, 95% CI: 1.67–3.75 high risk: OR 4.24, 95% CI: 2.28–7.88). The area under the receiver operating characteristic curve (AUC) for the unadjusted MPI was 0.61 (0.57–0.65) and 0.64 (95% CI: 0.59–0.68) with age and sex adjusted. The adapted versions of the MPI did not significantly change the AUC of the original MPI. Conclusion: The original and adapted MPI were strongly associated with 12-month mortality in an Australian cohort. However, the discriminatory performance was lower than that reported in European studies.
Publisher: Wiley
Date: 20-05-2015
DOI: 10.1111/BCP.12617
Publisher: PAGEPress Publications
Date: 14-04-2016
DOI: 10.4081/GC.2016.5893
Abstract: Conventional end-points, primarily based on the pharmacodynamic effects of a specific drug, are used to assess the efficacy of pharmacological treatment in clinical trials. However, their application and interpretation in complex frail older patients, a patient group with high inter-in idual variability, multiple coexisting disease states and prescribed medications, is becoming increasingly questionable. National surveys and qualitative studies have convincingly shown that the maintenance of functional independence is key to self-rated health and well being in old age. Therefore, the use of unconventional, patientcentered, end-points focused on functional status and perceived health seems appropriate, in combination with conventional end-points, to comprehensively investigate the impact of pharmacological treatments in this patient group. The recent availability of objective, quantifiable, and robust scoring tools, such as the multidimensional prognostic index, to assess key functional domains and clinical outcomes offers a unique opportunity to adequately characterize patient-centered endpoints in future clinical trials in older patients.
Publisher: Springer Singapore
Date: 2016
No related grants have been discovered for Kimberley Bryant.