Changes over time in the ADL status of elderly US veterans

The ageing of the US veteran population has greatly out-paced that of the general US population in the last decade and the demographics of this subgroup have changed relative to the US general population. To address the concerns of policy makers within the US Department of Veterans' Affairs (VA), data on elderly US male veterans and non-veterans from the Longitudinal Survey on Aging in 1984, 1986, 1988, and 1990 were used in a prospective study. Health status or well-being was assessed through changes over time in ADL status, according to the following five categories; (I) the presence of no ADL dependence and not developing a dependence between survey periods, (II) no change in ADL dependence between survey periods, (III) an increase in ADL dependence status between survey periods, (IV) a decrease in ADL dependence between survey periods, and (V) the occurrence of death between survey periods. Using a polytomous logistic regression model, these five categories were assessed in relation to a set of variables representing social, health-care utilization, and socioeconomic characteristics. The estimated coefficients from the model indicate that family income, having worked in the last 12 months, having an increase or no change in levels of physical activity in the past 12 months and the presence of Medicare coverage are inversely associated with moving, over time, into a lower state of health status or well-being. The prior existence of an ADL limitation, the number of doctor and hospital visits in the past 12 months, level of education, having been widowed in the past 12 months and veteran status were all proportionally associated with the likelihood of moving into a lower state of health status or well-being. These findings lead to the conclusion that being a veteran in the US does make a difference in regard to moving into various states of health status or well-being. This difference is supported by the estimated coefficient for veteran status being significant at a p value of 0.01 and through an evaluation of the estimated transition likelihood, controlling for veteran status. Veterans are less likely to die than their non-veteran counterparts, but have a greater likelihood of transition into levels of ADL dependence. Similarly, given the existence of a prior ADL limitation, over time, veterans are less likely to die, but more likely to move into a state of increased ADL limitations.