Comparison of methods for modelling a count outcome with excess zeros: an application to Activities of Daily Living (ADL-s)
Résumé
Background: count outcomes are commonly encountered in many epidemiology applications, and are often characterized by a large proportion of zeros. While linear or logistic regression models have often been used to analyse count outcomes, the resulting estimates are likely to be inefficient, inconsistent or biased. Methods: data are from wave1 of the English Longitudinal Study of Ageing (ELSA). The main outcome measure is number of difficulties (ranging from 0 to 6) with ‘Activities of Daily Living (ADL-s)' such as dressing, walk across a room, bathing, eating, getting in and out of bed and using the toilet. We fitted four regressions models specifically developed for count outcomes: Poisson, NB, ZIP and ZINB. We then compared these models using the Likelihood Ratio test of overdispersion, the Vuong test, and graphical methods. Results: The plots of predictions showed that overall, the ZINB model fit best. Although the ZINB and the ZIP models showed similar fit the LR test provided strong evidence that the ZINB improves the fit over the ZIP. Increasing number of difficulties with ADL-s was associated with fair/poor self-reported health, limiting longstanding illness and physical inactivity. The probability of not having any difficulty with ADL-s decreases with a limiting longstanding illness, increasing age, no education, fair/poor self-reported health and with not living with the partner. Conclusion: models specifically developed for count outcomes with excess zeros such as ZINB can provide better insights into the investigation of the factors associated with the number of difficulties with Activities of Daily Living.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...