s7 2 nusselder physicalactivitycomp ression

PHYSICAL ACTIVITY AND COMPRESSION OF MORBIDITY : PHYSICAL ACTIVITY AND COMPRESSION OF MORBIDITY Wilma Nusselder Caspar Looman Anna Peeters Oscar Franco-Durán Johan Mackenbach Department of Public Health


Background : Background Lower levels of physical activity (PA) are associated with higher all-cause mortality and higher incidence of major chronic diseases and disability Health policies aim at increasing the level of PA in the population Reducing the proportion of people with lower PA levels in the population is likely to extend total life expectancy and disability-free life expectancy However, the effects of reducing the proportion of people with lower PA levels on the life expectancy with disability are largely unknown


Objective : Objective AIM: To explore the effects on life expectancy with(out) disability of reducing the proportion of people with lower levels of physical activity in the population MAIN QUESTIONS: What will be the effect of increasing the level of physical activity in the population on the number of years lived with(out) disability? Will reduction of the lowest levels of PA in the population result in a compression of morbidity?


General approach : General approach Combines information on the effect of PA on different transitions (e.g. incidence, recovery and mortality) to assess the population disability burden in terms of years lived with(out) disability associated with different PA levels Uses transition rates by PA level, based on the combination of data on: - population transition rates - population prevalence of PA levels - RRs of each transition associated with each PA level Simulates a change in PA levels - change of the PA level , but NOT of other characteristics Uses multi-state life tables as integrative tool


Data : Data GLOBE: Eindhoven and surroundings, the Netherlands Persons aged 15-74 in 1991 Follow-up in 1993, 1995, 1997 N=5669 Transition rates: information on disability and death by age and sex Prevalence of PA: information on disability by age, sex and disability status RRs: information disability and death, PA and baseline characteristics Longitudinal Study on Aging (LSOA), USA Persons aged 70+ in 1984 Follow-up in 1986, 1988, 1990 N=7527 (1986: 5151) Transition rates: information on disability and death by age and sex


Population transition rates : Population transition rates


2. Prevalence of PA : 2. Prevalence of PA Basal 11% Light 54% Moderate/High 35% PA=LTPA = Leisure Time Physical Activity Walking/biking to work or shops Walking/biking/gardening in leisure time Doing sports Basal no sports and less than 20 min. per day walking/biking/gardening Moderate/high more than 2 hours per week sports, OR 1-2 hours per week sports and more than 20 min. per day walking/biking/gardening Light intermediate levels e.g., no sports and more than 20 min. per day walking/biking/gardening


3. RR of physical activity (PA) for each transition (1) : 3. RR of physical activity (PA) for each transition (1)


3. RR of physical activity (PA) for each transition (2) : 3. RR of physical activity (PA) for each transition (2)


3. RR of physical activity (PA) for each transition (3) : 3. RR of physical activity (PA) for each transition (3)


3. RR of physical activity (PA) for each transition (4) : 3. RR of physical activity (PA) for each transition (4)


3. RR of physical activity (PA) for each transition (all) : 3. RR of physical activity (PA) for each transition (all)


Health expectancy by level of PAunadjusted : Health expectancy by level of PA unadjusted


Effects of increasing PA levels on health expectancy Total men, adjusted : Effects of increasing PA levels on health expectancy Total men, adjusted


Effects of increasing PA levels on health expectancy Total women, adjusted : Effects of increasing PA levels on health expectancy Total women, adjusted


Further work : Further work - Refinements of RRs - More realistic definitions of intervention strategies - Confidence intervals for health expectancies - Sensitivity analyses


Limitations (1) : Limitations (1) Cohort study introduces risk of reverse causation + confounding Use habitual physical activity (not exercise) Adjusted for baseline disease, education, smoking, locus of control Modeling approach could not be based on review of literature Limited number of studies about the effect of PA on transitions between disability states and death Huge variation in PA measures Huge variation in disability measures


Limitations (2) : Limitations (2) Simulations are based on pooled datasets (GLOBE+LSOA) For age 80+ (LSOA): no comparable PA measurement is available Limited power of the study: Non-significant effects of PA after correction for confounding for most transitions We used RR adjusted for sex We used RR constant by age Strong effects of PA on mortality from non-disabled


Conclusions : Conclusions Conclusions are based on preliminary analyses Increasing the level of physical activity in the population extends total life expectancy and disability-free life expectancy, but does NOT reduce life expectancy with disability Increasing the level of physical activity in the population will NOT lead to a compression of morbidity


Slide20 : Thank you for your attention


Slide21 :


Published cohort studies + GLOBE to quantify RRs : Published cohort studies + GLOBE to quantify RRs


Disability items : Disability items Eating Getting in and out bed/chair Dressing and undressing Bathing or showering Getting outside Walking 10 steps without rest Lifting or carrying 10 pound bag Walking quarter of a mile See enough to read newspaper print Living in an institution


2. Prevalence of PA : 2. Prevalence of PA


Slide25 :