Background: Modern modeling techniques may potentially provide more accurate predictions of dichotomous outcomes than classical techniques. Objective: In this study, we aimed to examine the predictive performance of eight modeling techniques to predict mortality by frailty. Methods: We performed a longitudinal study with a 7-year follow-up. The sample consisted of 479 Dutch community-dwelling people, aged 75 years and older. Frailty was assessed with the Tilburg Frailty Indicator (TFI), a self-report questionnaire. This questionnaire consists of eight physical, four psychological, and three social frailty components. The municipality of Roosendaal, a city in the Netherlands, provided the mortality dates. We compared modeling techniques, such as support vector machine (SVM), neural network (NN), random forest, and least absolute shrinkage and selection operator, as well as classical techniques, such as logistic regression, two Bayesian networks, and recursive partitioning (RP). The area under the receiver operating characteristic curve (AUROC) indicated the performance of the models. The models were validated using bootstrapping. Results: We found that the NN model had the best validated performance (AUROC=0.812), followed by the SVM model (AUROC=0.705). The other models had validated AUROC values below 0.700. The RP model had the lowest validated AUROC (0.605). The NN model had the highest optimism (0.156). The predictor variable “difficulty in walking” was important for all models. Conclusions: Because of the high optimism of the NN model, we prefer the SVM model for predicting mortality among community-dwelling older people using the TFI, with the addition of “gender” and “age” variables. External validation is a necessary step before applying the prediction models in a new setting.
Achtergrond: De Two-Minute Step Test (TMST) is een meetinstrument gericht op het beoordelen van uithoudingsvermogen. Verscheidene psychometrische eigenschappen van de TMST-NL (Nederlands vertaalde versie) zijn onderzocht bij intramuraal wonende ouderen. De gevoeligheid voor verandering en de responsiviteit is bij deze patiëntenpopulatie nog niet vastgesteld. Doel: Het vaststellen van de gevoeligheid voor verandering en de responsiviteit (Minimal Clinical Important Difference) van de TMST-NL bij intramuraal wonende ouderen. Design: Prospectief responsiviteitsonderzoek.Methode: De onderzoekspopulatie bestond uit intramuraal wonende ouderen. Deelnemers hebben twee meetmomenten (T0 en T1) ondergaan waartussen ze drie maanden fysiotherapie gericht op uithoudingsvermogen ontvingen. Om de gevoeligheid van verandering te meten werd de distributie methode gebruikt waarbij de correlatie met de 6-minuten wandeltest (6MWT) werd getoetst. Via de anker methode met de Receiver Operating Characteristic (ROC) curve werd de MCID bepaald.Metingen voor het aerobe uithoudingsvermogen werden verricht met de TMST-NL en de 6-minuten wandeltest (6MWT). De Global Rating of Change (GRC) en de Borg Category-Ratio10 (BORG-CR10) werden gebruikt als subjectieve vragenlijsten om verandering van de gezondheidssituatie en vermoeidheid te meten.Resultaten: Intramurale ouderen (N=50) met een gemiddelde (SD) leeftijd van 83,96 jaar (6,96) zijn geïncludeerd. De correlatie tussen de verschilscores van de TMST-NL en de 6MWT over de deelnemerspopulatie die T1 ook hebben afgerond (N= 36) kwam uit op r=0.51 (P <0.05). Vanuit de ROC curve werd een MCID van 8,50 stappen berekend. De AUC-waarde was 0,74 (95% CI 0,54-0,94; P =0.02). Conclusie: De TMST-NL is gevoelig voor verandering en responsief bij intramuraal wonende ouderen. Echter doordat de MCID binnen de minimale meetfout (MDC) valt moeten de resultaten voor individuele evaluatie bij deze doelgroep met voorzichtigheid worden geïnterpreteerd.
BACKGROUND: Visceral obesity is associated with the metabolic syndrome. The metabolic risk differs per ethnicity, but reference values for visceral obesity for body composition analyses using Computed Tomography (CT) scans in the Caucasian population are lacking. Therefore, the aim of this study was to define gender specific reference values for visceral obesity in a Caucasian cohort based upon the association between the amount of visceral adipose tissue (VAT) and markers of increased metabolic risk.METHODS: Visceral Adipose Tissue Area Index (VATI cm 2/m 2) at the level of vertebra L3 was analyzed using CT scans of 416 healthy living kidney donor candidates. The use of antihypertensive drugs and/or statins was used as an indicator for increased metabolic risk. Gender specific cut-off values for VATI with a sensitivity ≥80% were calculated using receiver operating characteristic (ROC) curves. RESULTS: In both men and women who used antihypertensive drugs, statins or both, VATI was higher than in those who did not use these drugs (p ≤ 0.013). In males and females respectively, a value of VATI of ≥38.7 cm 2/m 2 and ≥24.9 cm 2/m 2 was associated with increased metabolic risk with a sensitivity of 80%. ROC analysis showed that VATI was a better predictor of increased metabolic risk than BMI (area under ROC curve (AUC) = 0.702 vs AUC = 0.556 in males and AUC = 0.757 vs AUC = 0.630 in females). CONCLUSION: Gender and ethnicity specific cut-off values for visceral obesity are important in body composition research, although further validation is needed. This study also showed that quantification of VATI is a better predictor for metabolic risk than BMI.
