The aim of this study was to assess the predictive ability of the frailty phenotype (FP), Groningen Frailty Indicator (GFI), Tilburg Frailty Indicator (TFI) and frailty index (FI) for the outcomes mortality, hospitalization and increase in dependency in (instrumental) activities of daily living ((I)ADL) among older persons. This prospective cohort study with 2-year follow-up included 2420 Dutch community-dwelling older people (65+, mean age 76.3±6.6 years, 39.5% male) who were pre-frail or frail according to the FP. Mortality data were obtained from Statistics Netherlands. All other data were self-reported. Area under the receiver operating characteristic curves (AUC) was calculated for each frailty instrument and outcome measure. The prevalence of frailty, sensitivity and specifcity were calculated using cutoff values proposed by the developers and cutoff values one above and one below the proposed ones (0.05 for FI). All frailty instruments poorly predicted mortality, hospitalization and (I)ADL dependency (AUCs between 0.62–0.65, 0.59–0.63 and 0.60–0.64, respectively). Prevalence estimates of frailty in this population varied between 22.2% (FP) and 64.8% (TFI). The FP and FI showed higher levels of specifcity, whereas sensitivity was higher for the GFI and TFI. Using a different cutoff point considerably changed the prevalence, sensitivity and specifcity. In conclusion, the predictive ability of the FP, GFI, TFI and FI was poor for all outcomes in a population of pre-frail and frail community-dwelling older people. The FP and the FI showed higher values of specifcity, whereas sensitivity was higher for the GFI and TFI.
DOCUMENT
The full potential of predictive maintenance has not yet been utilised. Current solutions focus on individual steps of the predictive maintenance cycle and only work for very specific settings. The overarching challenge of predictive maintenance is to leverage these individual building blocks to obtain a framework that supports optimal maintenance and asset management. The PrimaVera project has identified four obstacles to tackle in order to utilise predictive maintenance at its full potential: lack of orchestration and automation of the predictive maintenance workflow, inaccurate or incomplete data and the role of human and organisational factors in data-driven decision support tools. Furthermore, an intuitive generic applicable predictive maintenance process model is presented in this paper to provide a structured way of deploying predictive maintenance solutions https://doi.org/10.3390/app10238348 LinkedIn: https://www.linkedin.com/in/john-bolte-0856134/
DOCUMENT
Background: Early identification of older cardiac patients at high risk of readmission or mortality facilitates targeted deployment of preventive interventions. In the Netherlands, the frailty tool of the Dutch Safety Management System (DSMS-tool) consists of (the risk of) delirium, falling, functional impairment, and malnutrition and is currently used in all older hospitalised patients. However, its predictive performance in older cardiac patients is unknown. Aim: To estimate the performance of the DSMS-tool alone and combined with other predictors in predicting hospital readmission or mortality within 6 months in acutely hospitalised older cardiac patients. Methods: An individual patient data meta-analysis was performed on 529 acutely hospitalised cardiac patients ≥70 years from four prospective cohorts. Missing values for predictor and outcome variables were multiply imputed. We explored discrimination and calibration of: (1) the DSMS-tool alone; (2) the four components of the DSMS-tool and adding easily obtainable clinical predictors; (3) the four components of the DSMS-tool and more difficult to obtain predictors. Predictors in model 2 and 3 were selected using backward selection using a threshold of p = 0.157. We used shrunk c-statistics, calibration plots, regression slopes and Hosmer-Lemeshow p-values (PHL) to describe predictive performance in terms of discrimination and calibration. Results: The population mean age was 82 years, 52% were males and 51% were admitted for heart failure. DSMS-tool was positive in 45% for delirium, 41% for falling, 37% for functional impairments and 29% for malnutrition. The incidence of hospital readmission or mortality gradually increased from 37 to 60% with increasing DSMS scores. Overall, the DSMS-tool discriminated limited (c-statistic 0.61, 95% 0.56-0.66). The final model included the DSMS-tool, diagnosis at admission and Charlson Comorbidity Index and had a c-statistic of 0.69 (95% 0.63-0.73; PHL was 0.658). Discussion: The DSMS-tool alone has limited capacity to accurately estimate the risk of readmission or mortality in hospitalised older cardiac patients. Adding disease-specific risk factor information to the DSMS-tool resulted in a moderately performing model. To optimise the early identification of older hospitalised cardiac patients at high risk, the combination of geriatric and disease-specific predictors should be further explored.
DOCUMENT
Background: Due to differences in the definition of frailty, many different screening instruments have been developed. However, the predictive validity of these instruments among community-dwelling older people remains uncertain. Objective: To investigate whether combined (i.e. sequential or parallel) use of available frailty instruments improves the predictive power of dependency in (instrumental) activities of daily living ((I)ADL), mortality and hospitalization. Design, setting and participants: A prospective cohort study with two-year followup was conducted among pre-frail and frail community-dwelling older people in the Netherlands. Measurements: Four combinations of two highly specific frailty instruments (Frailty Phenotype, Frailty Index) and two highly sensitive instruments (Tilburg Frailty Indicator, Groningen Frailty Indicator) were investigated. We calculated sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for all single instruments as well as for the four combinations, sequential and parallel. Results: 2,420 individuals participated (mean age 76.3 ± 6.6 years, 60.5% female) in our study. Sequential use increased the levels of specificity, as expected, whereas the PPV hardly increased. Parallel use increased the levels of sensitivity, although the NPV hardly increased. Conclusions: Applying two frailty instruments sequential or parallel might not be a solution for achieving better predictions of frailty in community-dwelling older people. Our results show that the combination of different screening instruments does not improve predictive validity. However, as this is one of the first studies to investigate the combined use of screening instruments, we recommend further exploration of other combinations of instruments among other study populations.
DOCUMENT
The Short-Term Assessment of Risk and Treatability: Adolescent Version (START:AV) is a risk assessment instrument for adolescents that estimates the risk of multiple adverse outcomes. Prior research into its predictive validity is limited to a handful of studies conducted with the START:AV pilot version and often by the instrument’s developers. The present study examines the START:AV’s field validity in a secure youth care sample in the Netherlands. Using a prospective design, we investigated whether the total scores, lifetime history, and the final risk judgments of 106 START:AVs predicted inpatient incidents during a 4-month follow-up. Final risk judgments and lifetime history predicted multiple adverse outcomes, including physical aggression, institutional violations, substance use, self-injury, and victimization. The predictive validity of the total scores was significant only for physical aggression and institutional violations. Hence, the short-term predictive validity of the START:AV for inpatient incidents in a residential youth care setting was partially demonstrated and the START:AV final risk judgments can be used to guide treatment planning and decision-making regarding furlough or discharge in this setting.
DOCUMENT
Since 2016, it is mandatory for all future students at the department of Media, Information and Communication, to participate in a study choice test (SCT), prior their enrollment. However, the outcome is not binding and students are still entitled to enter the first year after receiving negative advice. With the help of a structural model, built for my PhD research, the predictive value of the SCT is tested by comparing the time it takes the students to finish all first year exams, their average grade point and attrition, against the results of the SCT. By using the structural model, various background variables are also measured, such as engagement, effort and commitment are also measured. By using the normed fit index (NFI), the comparative fit index (CFI), the Tucker-Lewis Index (TLI) and the root mean square error of approximation (RMSEA), the fit of the model is established. In addition, a comparison of the direct and indirect influence of the SCT will provide more knowledge about the correlations between the different variables, the SCT and ultimately student success.
DOCUMENT
For my PhD research I build a structural model to predict student success. Initially to show the influence of social media use by first year students in higher education. However, for this research I use the model to investigate the predictive value of a student choice test. This test is mandatory for all students prior to their enrolment at the Amsterdam University of Applied Sciences. In this study two of the Institutes (Communication and Creative Business/Media, Information and Communication) of the Faculty of Digital Media and the Creative Industries participated with a first year enrollment in the year 2017 of 1010 students (respectively of 327 and 683), and in 2018, 1193 students (respectively 225 and 968). This study choice test involved an assignment that the student-to-be had to do at home and bring to the Institute when they took part in the second half of the study choice test. This second half involved an exam in topics central to the curriculum, a Dutch language test and all students had a final meeting with a teacher where they were given a positive or negative advice. Because of the large number of students, a substantial number of teachers and resources were used for this test. In order to see the pros and cons of the test, the predictive value was tested along with other variables which are proven to have a predictive value on student success. The best proven variables from Tinto’s theory were included, based on previous studies. The central variable in Tinto’s study is ‘satisfaction’ (which in other research is revert to as ‘engagement’ of ‘belonging’), consisting originally of a vast number of manifest variables. By using a fraction of those variables, I simplified the model, so it was an easier tool to use for teachers and management and in the meantime, avoiding the capitalization of chance. The smaller latent variable ‘satisfaction’ was tested using principal component analysis to prove the manifest variables where in fact representing one latent variable. Cronbach’s alpha and Guttman’s lambda-2 then provided the internal consistency and reliability of the variable. Along with ‘satisfaction’, the model included different background variables (gender, prior education, ethnicity), commitment and effort, expected progress and of course study success. This was measured by the time it takes a student to finish all first year exams and the average grade point (GPA). SPSS AMOS was used for testing the fit of the model and showed reasonable values for the normed fit index (NFI), the comparative fit index (CFI), the Tucker-Lewis Index (TLI) and the root mean square error of approximation (RMSEA). The advice from the study choice test and the scores were tested in the model to uncover if there was a significant difference. Furthermore, the influence of all variables in the model were compared for their influence on study success.
DOCUMENT
PurposeThis study aims to develop an understanding of how customers of a physical retail store valuate receiving location-based mobile phone messages when they are in proximity of the store. It proposes and tests a model relating two benefits (personalization and location congruency) and two sacrifices (privacy concern and intrusiveness) to message value perceptions and store visit attitudes.Design/methodology/approachThe study uses a vignette-based survey to collect data from a sample of 1,225 customers of a fashion retailer. The postulated research model is estimated using SmartPLS 3.0 with the consistent-PLS algorithm and further validated via a post-hoc test.FindingsThe empirical testing confirms the predictive validity and robustness of the model and reveals that location congruency and intrusiveness are the location-based message characteristics with the strongest effects on message value and store visit attitude.Originality/valueThe paper adds to the underexplored field of store entry research and extends previous location-based messaging studies by integrating personalization, location congruency, privacy concern and intrusiveness into one validated model.
MULTIFILE
BackgroundEarly identification of older cardiac patients at high risk of readmission or mortality facilitates targeted deployment of preventive interventions. In the Netherlands, the frailty tool of the Dutch Safety Management System (DSMS-tool) consists of (the risk of) delirium, falling, functional impairment, and malnutrition and is currently used in all older hospitalised patients. However, its predictive performance in older cardiac patients is unknown.AimTo estimate the performance of the DSMS-tool alone and combined with other predictors in predicting hospital readmission or mortality within 6 months in acutely hospitalised older cardiac patients.MethodsAn individual patient data meta-analysis was performed on 529 acutely hospitalised cardiac patients ≥70 years from four prospective cohorts. Missing values for predictor and outcome variables were multiply imputed. We explored discrimination and calibration of: (1) the DSMS-tool alone; (2) the four components of the DSMS-tool and adding easily obtainable clinical predictors; (3) the four components of the DSMS-tool and more difficult to obtain predictors. Predictors in model 2 and 3 were selected using backward selection using a threshold of p = 0.157. We used shrunk c-statistics, calibration plots, regression slopes and Hosmer-Lemeshow p-values (PHL) to describe predictive performance in terms of discrimination and calibration.ResultsThe population mean age was 82 years, 52% were males and 51% were admitted for heart failure. DSMS-tool was positive in 45% for delirium, 41% for falling, 37% for functional impairments and 29% for malnutrition. The incidence of hospital readmission or mortality gradually increased from 37 to 60% with increasing DSMS scores. Overall, the DSMS-tool discriminated limited (c-statistic 0.61, 95% 0.56–0.66). The final model included the DSMS-tool, diagnosis at admission and Charlson Comorbidity Index and had a c-statistic of 0.69 (95% 0.63–0.73; PHL was 0.658).DiscussionThe DSMS-tool alone has limited capacity to accurately estimate the risk of readmission or mortality in hospitalised older cardiac patients. Adding disease-specific risk factor information to the DSMS-tool resulted in a moderately performing model. To optimise the early identification of older hospitalised cardiac patients at high risk, the combination of geriatric and disease-specific predictors should be further explored.
MULTIFILE
Background: Magnetic resonance imaging (MRI) is being used extensively in the search for pathoanatomical factors contributing to low back pain (LBP) such as Modic changes (MC). However, it remains unclear whether clinical findings can identify patients with MC. The purpose of this explorative study was to assess the predictive value of six clinical tests and three questionnaires commonly used with patients with low-back pain (LBP) on the presence of Modic changes (MC).Methods: A retrospective cohort study was performed using data from Dutch military personnel in the period between April 2013 and July 2016. Questionnaires included the Roland Morris Disability Questionnaire, Numeric Pain Rating Scale, and Pain Self-Efficacy Questionnaire. The clinical examination included (i) range of motion, (ii) presence of pain during flexion and extension, (iii) Prone Instability Test, and (iv) straight leg raise. Backward stepwise regression was used to estimate predictive value for the presence of MC and the type of MC. The exploration of clinical tests was performed by univariable logistic regression models.Results: Two hundred eighty-six patients were allocated for the study, and 112 cases with medical records and MRI scans were available; 60 cases with MC and 52 without MC. Age was significantly higher in the MC group. The univariate regression analysis showed a significantly increased odds ratio for pain during flexion movement (2.57 [95% confidence interval (CI): 1.08-6.08]) in the group with MC. Multivariable logistic regression of all clinical symptoms and signs showed no significant association for any of the variables. The diagnostic value of the clinical tests expressed by sensitivity, specificity, positive predictive, and negative predictive values showed, for all the combinations, a low area under the curve (AUC) score, ranging from 0.41 to 0.53. Single-test sensitivity was the highest for pain in flexion: 60% (95% CI: 48.3-70.4).Conclusion: No model to predict the presence of MC, based on clinical tests, could be demonstrated. It is therefore not likely that LBP patients with MC are very different from other LBP patients and that they form a specific subgroup. However, the study only explored a limited number of clinical findings and it is possible that larger samples allowing for more variables would conclude differently.
DOCUMENT
