Dynamic Study On Health-related Quality Of Life In The Progression Of Alzheimer’s Disease Based On Generalized Estimating Equations And Latent Growth Curve Model

Objective The aim of this study was to conduct a more comprehensive quantitative study ofAlzheimer’s disease progression on the basis of preliminary studies and further follow up, withhealth-related quality of life(HRQoL) as the primary outcome measure and appropriate modelsfor longitudinal data statistical analysis. Making quality of life combined with disease status inthe AD process, this study explored the main factors of quality of life and the dynamic changes ofquality of life with state transitions in order to provide theoretical basis for AD prevention, earlyintervention, development of different periods AD countermeasures and quality of lifeimprovement among the elderly in community. In addition, this study discussed the advantagesof longitudinal data statistical analysis models-the Generalized Estimating Equation (GEE) andthe Latent Growth Curve Model(LGM) in the follow-up data analysis.Methods The study subjects were517elderly people aged60and over selected from threecommunities in Taiyuan. Follow-up survey with them were conducted every six months andthree waves data were collected. GEE and LGM were used to focus on the influencing factorsand dynamic changes of HRQoL with the state transitions in the AD process. Cognitive normallyaging, MCI, moderate or severe cognitive impairment, and AD were defined as state1,2,3and4respectively in the research.Results GEE analysis showed that conjugal relation, relation with the children, listening, doinghousework, physical exercise, tea, drinking, first-degree relatives with dementia, hyperlipidemia,MoCA, ADL and GDS were statistically significant factors for elderly’s HRQOL. According tothe comparative analysis results between each outcomes, education, occupation before retire,hobbies, tea, smoking, aluminum cookware, control appetite, brain disorders, ADL and GDSwere statistically significant for transition from state1to state2. Marriage, conjugal relation,number of birth, physical exercise, first-degree relatives with dementia, ADL and GDS werestatistically significant for transition from state2to state3. Education, second career afterretirement, doing housework, physical exercise, tea, first-degree relatives with dementia,hyperlipidemia and GDS were statistically significant for transition from state3to state4.Occupation before retire, marital, recreational and public service activities, smoking and GDSwere statistically significant for transition from state3to state2. LGM analysis showed that the average QoL score was29.865at baseline, a general level,and there were significant differences between individuals. During follow-up periods, there wereincrease tendencies about QoL, but increase speeds did not have significant differences betweenindividuals. Besides, the initial level of QoL was not significantly correlated with the followingchange speed. In each transition group, the initial states of QoL have significant differences. Ingroup2(state2â†'state3) and group3(state3â†'state4), there were significantly increasedtrend about QoL of elderly. Compared with group3, group2had higher increased pace, and hadsignificant differences between individuals. However, group1(state1â†'state2) and group4(state3â†'state2) did not reach significant levels.Conclusion The dynamic changes with the state transitions of HRQOL in AD process weredifferent and were related to a variety of factors. Control and prevention measures in stages canbe taken based on development and influencing factors of different cognitive outcomes. Forlongitudinal data analysis, GEE and LGM can take into account the correlation withinindividuals and the differences between individuals, respectively, and demonstrate their uniqueadvantages in the practical applications.