Structural Equation Model And Longitudinal Polytomous Item Response Theory For Assessing Response Shifts In Quality Of Life In Patients With Alzheimer's Disease

Objective:The aims of this study were:(1)to detect different types of response shifts(RS)in quality of life(QOL)measurement in patients with Alzheimer's Disease(AD)and to measure the true change of QOL using structural equation modeling(SEM)and longitudinal polytomous item response theory(LPIRT);(2)to promote the correct understanding of QOL change among dementia patients and to provide methodological guidance for RS assessment in QOL research of other diseases.Methods:The basic principle,model operational theory and procedure of detecting RS using SEM and LPIRT were introduced.The study subjects were patients with AD aging 60 years or older from Department of Neurology and Geriatrics in Taiyuan Central Hospital from May to December 2016.12-item short form health survey(SF-12)was used to measure QOL of AD prior to hospitalization and 1 month following discharge.SEM and LPIRT were fitted based on domain and item level to analyze different types of RS and to measure the true change of QOL using Mplus software.Results:1.The study investigated a total of 238 AD patients with average age 72.17 ± 8.77.The average cognitive level among patients were 23.13 ± 4.50.SF-12 had good reliability and validity.Paired t-tests in SF-12 indicated deterioration in general physical health(PHYS)and physical functioning(PF),improvement in general mental health(MENT),mental health(MH),vitality(VT),and bodily pain(BP).2.Based on domain level,SEM to detect RS showed reprioritization()1(2c =4.128,P<0.05)in the domain of social functioning(SF),reconceptualization()1(2c =17.028,P<0.001)in the domain of role limitations due to emotional problems(RE),uniform recalibration()1(2c =12.241,P < 0.001)in the domain of BP,and non-uniform recalibration()1(2c =4.408,P < 0.05)in the domain of MH,respectively.Considering these RS,the true change of common factors were deteriorated PHYS(-0.098,)1(2c =8.303,P<0.005)and improved MENT(+0.292,)1(2c =20.948,P<0.001).3.Based on item level,LPIRT to detect RS showed that recalibration(G=9.904,P<0.005)was observed in item 2(2),which is one of the items of PF subscale: “Climbing several flights of stairs” with responses: Limited a lot/Limited a little/Not Limited at all.Considering recalibration in item 2(2),the domain of PF did not change statistically between the two measurements(-0.039,)1(2c =0.034,P > 0.05).Reprioritization(G=20.666,P<0.005)and recalibration(G=26.582,P<0.005)were observed in item 6(1),which is one of the items of MH subscale: “Have you felt calm and peaceful?” with responses: All of the time/Most of the time/A good bit of the time/Some of the time/A little of the time/None of the time.Recalibration(G=8.578,P<0.005)was observed in item6(3),which is one of the items of MH subscale: “Have you felt downhearted and blue?”with responses: All of the time/Most of the time/A good bit of the time/Some of the time/A little of the time/None of the time.Considering RS in item 6(1)and 6(3),the true change of the domain of MH deteriorated(-8.824,)1(2c =5236.829,P<0.001).Conclusion:RS should be considered when assessing QOL changes in patients of AD.SEM and LPIRT were found to be useful in detecting RS in QOL measurement.These two methodscan be used not only to detect different types of RS,but also to measure true change.