*************************************************************************************************************************************************************** *** SPSS ALGORITHM FOR AQoL-6D UTILITY MODEL FOR ADULTS (WITH MISSING VALUES) (CHE version 12 dated 23 Nov 2015) *************************************************************************************************************************************************************** *** REM THIS ALGORITHM (CHE version 12 dated 23 Nov 2015) IS AN INTERIM RELEASE ***AND MAY BE CHANGED WITHOUT NOTICE. *** Changes between versions are often to do with terminology and generally have slight effect on output. *** REM RESEARCHERS SHOULD CHECK WITH THE AQOL GROUP AT ***MONASH UNIVERSITY FOR ANY MODIFICATION www.aqol.com.au **************************************************************************************************************************************************************** ***REM The transformation programmed in this file have been conceptualised by Prof. Jeff Richardson and *produced by Gang Chen, Centre for Health Economics* ***See the following research paper for more detail (at *http://www.buseco.monash.edu.au/centres/che/pubs/researchpaper66.pdf): *Richardson, J, Chen, G, Iezzi, A & Khan, M. 2011. Transformations between the Assessment of *Quality of Life AQoL Instruments and Test-Retest Reliability * *** REM This file analyses the AQoL-6D instrument and produces dimension disvalue scores and a utility score for the overall instrument for adults. *** Dimension scores are termed disvalue scores. They are not utility values as they have not been evaluated on a life-death scale ***The dimensions are scaled on a "Dimension Worst Health State - Dimension Best Health State" scale *** where DWHS = 0.00 and DBHS = 1.00. ************************************************************* ***Use the AQoL-6D Data Collection copy to collect data. The items are laid out in dimension order, without randomisation** ***Variable names: For this algorithm, the variables in your questionnaire or database should be named * "aqol1, aqol2 ...aqol20". ************************************************************* * The AQoL-6D utility scores are scaled such that the: * "AQoL-6D worst health state" = -0.0444493 ( where Death = 0.00). * "AQoL-6D best health state" = 1.00 ****************************************************************** *Missing Values: Note that missing data are represented by a blank and are handled by imputing values within each dimension. *Dimensions with 3-4 items will allow for 1 missing value to be imputed. *However, if more item responses in the dimensions are missing the observations will be dropped and there will not be a dimension score or an instrument score for the individual. ********************************************************************************** * aqol# are item responses in your data * u and du are utilities and disutilities, v and dv are values and disvalues * Missing values represented by a blank or dot. compute Q1 = aqol1. compute Q2 = aqol2. compute Q3 = aqol3. compute Q4 = aqol4. compute Q5 = aqol5. compute Q6 = aqol6. compute Q7 = aqol7. compute Q8 = aqol8. compute Q9 = aqol9. compute Q10 = aqol10. compute Q11 = aqol11. compute Q12 = aqol12. compute Q13 = aqol13. compute Q14 = aqol14. compute Q15 = aqol15. compute Q16 =aqol16. compute Q17 = aqol17. compute Q18 = aqol18. compute Q19 = aqol19. compute Q20 = aqol20. execute. ********************************************************************* ********* Imputing Missing Values in Database ********* ********************************************************************* ** Independent Living - Dimension 1** Compute ILmiss = Nmiss (Q1,Q2,Q3,Q4). Do if ILmiss < 2. Do repeat A = Q1,Q2,Q3,Q4. If (Missing (A)) A = RND(Mean (Q1,Q2,Q3,Q4)). End repeat. End if. ** Relationships - Dimension 2** Compute RELmiss = Nmiss (Q5,Q6,Q7). Do if RELmiss < 2. Do repeat A = Q5,Q6,Q7. If (Missing (A)) A = RND(Mean (Q5,Q6,Q7)). End repeat. End if. ** Mental Health - Dimension 3** Compute MENmiss = Nmiss (Q8,Q9,Q10,Q11). Do if MENmiss < 3. Do repeat A = Q8,Q9,Q10,Q11. If (Missing (A)) A = RND(Mean (Q8,Q9,Q10,Q11)). End repeat. End if. ** Coping - Dimension 4** Compute COPmiss = Nmiss (Q12,Q13,Q14). Do if COPmiss < 2. Do repeat A = Q12,Q13,Q14. If (Missing (A)) A = RND(Mean (Q12,Q13,Q14)). End repeat. End if. ** Pain - Dimension 5** Compute PAINmiss = Nmiss (Q15,Q16,Q17). Do if PAINmiss < 2. Do repeat A = Q15,Q16,Q17. If (Missing (A)) A = RND(Mean (Q15,Q16,Q17)). End repeat. End if. ** Senses - Dimension 6** Compute SENmiss = Nmiss (Q18,Q19,Q20). Do if SENmiss < 2. Do repeat A = Q18,Q19,Q20. If (Missing (A)) A = RND(Mean (Q18,Q19,Q20)). End repeat. End if. Execute. ******************************************************************************************************** *** ITEM DISUTILITIES.*** ******************************************************************************************************** ***Dimension 1. Independent living* ***1. Household Help if (Q1=1) dvQ1 = 0. if (Q1=2) dvQ1=0.073. if (Q1=3) dvQ1=0.435. if (Q1=4) dvQ1=0.820. if (Q1=5) dvQ1=1. ***2. Getting Around Outside if (Q2=1) dvQ2 = 0. if (Q2=2) dvQ2=0.033. if (Q2=3) dvQ2=0.240. if (Q2=4) dvQ2=0.471. if (Q2=5) dvQ2=0.840. if(Q2=6) dvQ2=1. ***3. Mobility if (Q3=1) dvQ3 = 0. if (Q3=2) dvQ3=0.041. if (Q3=3) dvQ3=0.251. if (Q3=4) dvQ3=0.570. if (Q3=5) dvQ3=0.830. if (Q3=6) dvQ3=1. ***4. Personal Care if (Q4=1) dvQ4 = 0. if (Q4=2) dvQ4=0.040. if (Q4=3) dvQ4=0.297. if (Q4=4) dvQ4=0.797. if (Q4=5) dvQ4=1. *****Dimension 2. Relationships*************** ***5.Intimate if (Q5=1) dvQ5 = 0. if (Q5=2) dvQ5=0.074. if (Q5=3) dvQ5=0.461. if (Q5=4) dvQ5=0.841. if (Q5=5) dvQ5=1. ***6. Family Role if (Q6=1) dvQ6=0. if (Q6=2) dvQ6=0.193. if (Q6=3) dvQ6=0.759. if (Q6=4) dvQ6=1. ***7. Community Role if (Q7=1) dvQ7=0. if (Q7=2) dvQ7=0.197. if (Q7=3) dvQ7=0.648. if (Q7=4) dvQ7=1. ***Dimension 3. Mental Health ***8. Despair if (Q8=1) dvQ8=0. if (Q8=2) dvQ8=0.133. if (Q8=3) dvQ8=0.392. if (Q8=4) dvQ8=0.838. if (Q8=5) dvQ8=1. ***9. Worried if (Q9=1) dvQ9=0. if (Q9=2) dvQ9=0.142. if (Q9=3) dvQ9=0.392. if (Q9=4) dvQ9=0.824. if (Q9=5) dvQ9=1. ***10. Sad if (Q10=1) dvQ10=0. if (Q10=2) dvQ10=0.097. if (Q10=3) dvQ10=0.330. if (Q10=4) dvQ10=0.784. if (Q10=5) dvQ10=1. ***11. Calm if (Q11=1) dvQ11=0. if (Q11=2) dvQ11=0.064. if (Q11=3) dvQ11=0.368. if (Q11=4) dvQ11=0.837. if (Q11=5) dvQ11=1. ***Dimension 4. Coping ***12. Energy if (Q12=1) dvQ12=0. if (Q12=2) dvQ12=0.056. if (Q12=3) dvQ12=0.338. if (Q12=4) dvQ12=0.722. if (Q12=5) dvQ12=1. ***13. Control if (Q13=1) dvQ13=0. if (Q13=2) dvQ13=0.055. if (Q13=3) dvQ13=0.382. if (Q13=4) dvQ13=0.774. if (Q13=5) dvQ13=1. ***14. Coping if (Q14=1) dvQ14=0. if (Q14=2) dvQ14=0.057. if (Q14=3) dvQ14=0.423. if (Q14=4) dvQ14=0.826. if (Q14=5) dvQ14=1. ***Dimension 5. Pain ***15. Serious pain if (Q15=1) dvQ15=0. if (Q15=2) dvQ15=0.133. if (Q15=3) dvQ15=0.642. if (Q15=4) dvQ15=1. ***16. Pain if (Q16=1) dvQ16=0. if (Q16=2) dvQ16=0.200. if (Q16=3) dvQ16=0.758. if (Q16=4) dvQ16=1. ***17. Pain interferes if (Q17=1) dvQ17=0. if (Q17=2) dvQ17=0.072. if (Q17=3) dvQ17=0.338. if (Q17=4) dvQ17=0.752. if (Q17=5) dvQ17=1. ***Dimension 6. Senses ***18. Vision if (Q18=1) dvQ18=0. if (Q18=2) dvQ18=0.033. if (Q18=3) dvQ18=0.223. if (Q18=4) dvQ18=0.621. if (Q18=5) dvQ18=0.843. if (Q18=6) dvQ18=1. ***19. Hearing if (Q19=1) dvQ19=0. if (Q19=2) dvQ19=0.024. if (Q19=3) dvQ19=0.205. if (Q19=4) dvQ19=0.586. if (Q19=5) dvQ19=0.826. if (Q19=6) dvQ19=1. ***20. Communicate if (Q20=1) dvQ20=0. if (Q20=2) dvQ20=0.187. if (Q20=3) dvQ20=0.695. if (Q20=4) dvQ20=1. ********************************************************************************************************* ***2. DIMENSION SCORES *** ********************************************************************************************************* *** dvD1, dvD2, dvD3, dvD4, dvD5, dvD6 are the dimension disvalue scores and have been written this way to not confuse with utility scores. Only the global instrument score is a utility score. ***DIMENSION 1 - IND LIV. ***DIMENSION SCALING CONSTANT kD1=-0.978. ***IND LIV HAS 4 ITEMS. ***ITEM WORST WEIGHTS (Wi). **This model uses w1=0.385412. **This model uses w2=0.593819. **This model uses w3=0.630323. **This model uses w4=0.794888. **4 item formula. **dvD1=(1/kD1)*[(1+(kD1*w1*dvQ1))*(1+(kD1*w2*dvQ2))*(1+(kD1*w3*dvQ3))*(1+(kD1*w4*dvQ4))-1]. Compute dvD1=(1/-0.978)*((1+(-0.978*0.385412*dvQ1))*(1+(-0.978*0.593819*dvQ2))*(1+(-0.978*0.630323*dvQ3))*(1+(-0.978*0.794888*dvQ4))-1). **EXECUTE. ***DIMENSION 2 - REL. ***DIMENSION SCALING CONSTANT kD2 = -0.923. ***REL HAS 3 ITEMS. ***ITEM WORST WEIGHTS (Wi). **This model uses w5=0.64303. **This model uses w6=0.697742. **This model uses w7=0.508658. **3 item formula **dvD2=(1/kD2)*[(1+(kD2*w5*dvQ5))*(1+(kD2*w6*dvQ6))*(1+(kD2*w7*dvQ7))-1]. Compute dvD2=(1/-0.923)*((1+(-0.923*0.64303*dvQ5))*(1+(-0.923*0.697742*dvQ6))*(1+(-0.923*0.508658*dvQ7))-1). **EXECUTE. ***DIMENSION 3 - MEN. *** ***DIMENSION SCALING CONSTANT kD3 = -0.983. *** ***MEN HAS 4 ITEMS. ***ITEM WORST WEIGHTS (Wi). *** **This model uses w8=0.640377. *** **This model uses w9=0.588422. *** **This model uses w10=0.648748. *** ** w11=0.71122. *** ***4 item formula **dvD3=(1/kD3)*[(1+(kD3*w8*dvQ8))*(1+(kd3*w9*dvQ9))*(1+(kD3*w10*dvQ10))*(1+(kD3*w1***1*dvQ11))-1*/ Compute dvD3=(1/-0.983)*((1+(-0.983*0.640377*dvQ8))*(1+(-0.983*0.588422*dvQ9))*(1+(-0.983*0.648748*dvQ10))*(1+(-0.983*0.71122*dvQ11))-1). ***DIMENSION 4 - COPING. ***DIMENSION SCALING CONSTANT kD4 = -0.930. *** ***COPI HAS 3 ITEMS. *** ***ITEM WORST WEIGHTS (Wi). *** **This model uses w12=0.415694. *** **This model uses w13=0.636994. *** **This model uses w14=0.773296. *** **3 item formula **dvD4=(1/kD4)*[(1+(kD4*w12*dvQ12))*(1+(kD4*w13*dvQ13))*(1+(kD4*w14*dvQ14))-1]*/ Compute dvD4=(1/-0.930)*((1+(-0.930*0.415694*dvQ12))*(1+(-0.930*0.636994*dvQ13))*(1+(-0.930*0.773296*dvQ14))-1). ***DIMENSION 5 - PAIN. ***DIMENSION SCALING CONSTANT kD5 = -0.96. ***PAIN HAS 3 ITEMS. ***ITEM WORST WEIGHTS (Wi). ** w15=0.631833. ** w16=0.767573. ** w17=0.652241. Compute dvD5=(1/-0.962)*((1+(-0.962*0.631833*dvQ15))*(1+(-0.962*0.767573*dvQ16))*(1+(-0.962*0.652241*dvQ17))-1). ***DIMENSION 6 - SENSES. ** w18=0.580696. ** w19=0.463022. ** w20=0.604613. Compute dvD6=(1/-0.851)*((1+(-0.851*0.580696*dvQ18))*(1+(-0.851*0.463022*dvQ19))*(1+(-0.851*0.604613*dvQ20))-1). Compute vD1 =1-dvD1. Compute vD2 =1-dvD2. Compute vD3 =1-dvD3. Compute vD4 =1-dvD4. Compute vD5 =1-dvD5. Compute vD6 =1-dvD6. Execute. VARIABLE LABELS vD1 "Score Dimension 1 - Independent Living". VARIABLE LABELS vD2 "Score Dimension 2 - Relationships". VARIABLE LABELS vD3 "Score Dimension 3 - Mental Health". VARIABLE LABELS vD4 "Score Dimension 4 - Coping". VARIABLE LABELS vD5 "Score Dimension 5 - Pain". VARIABLE LABELS vD6 "Score Dimension 6 - Senses". EXECUTE. ******************************************************************************************************** *******Transformation using 8D data to predict 6D - same questions** ******************************************************************************************************** **The optimal model is Model 2A (see Richardson et al (2011, 18-19)** ******************************************************************************************************** **Model 2A: ******************************************************************************************************** Compute uaqol6Dusing8D = 0.0719264*vD1 + 0.1027818*vD2 + 0.2519563*vD3 + 0.3201172*vD4 + 0.1288289*vD5 + 0.2052164*vD6 - 0.0444493. **Note: For those with a predicted utility score higher than 1, adjusted to be 1.** Compute AQoL6D = uaqol6Dusing8D. if (uaqol6Dusing8D>1) AQoL6D=1. EXECUTE. /*Note: _uaqol6Dusing8D!=. covers any missing values*/ VARIABLE LABELS AQoL6D "AQoL-6D Utility score". EXECUTE. DESCRIPTIVES VARIABLES=vD1 vD2 vD3 vD4 vD5 vD6 AQoL6D /STATISTICS=MEAN STDDEV RANGE MIN MAX SEMEAN. EXECUTE. Delete Variables Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 Q16 Q17 Q18 Q19 Q20 ILmiss RELmiss MENmiss COPmiss PAINmiss SENmiss dvQ1 dvQ2 dvQ3 dvQ4 dvQ5 dvQ6 dvQ7 dvQ8 dvQ9 dvQ10 dvQ11 dvQ12 dvQ13 dvQ14 dvQ15 dvQ16 dvQ17 dvQ18 dvQ19 dvQ20 dvD1 dvD2 dvD3 dvD4 dvD5 dvD6 uaqol6Dusing8D. Execute.