cross-over design with longitudinal data

[Bettina Haidich]

Dear list,


I have a cross-over repeated measures design of 40 patients (PATID) with 2 treatment periods (PERIOD) and the variable of interest (IOP) is measured 12 times (TIME). The total length of the dataset is 40x12x2=960. I found a syntax of mixed models for cross-over design in SPSS and I applied it to my data but I think that something is not calculated right, especially in multiple comparisons method. I would like to have Bonferroni adjusted  p-values for the comparisons between treatments in the 12 timepoints. Since this is a cross-over design the comparison in each time point should be paired and not independent. But what I get from the SPSS syntax with EMMEANS seems to be independent t-test comparison and although I use the option BONFERRONI I get unadjusted p-values. I checked that if I use the option LSD or SIDAK I get exactly the same p-values with the BONFERRONI option so I think that something is not calculated right here. Here follows the syntax I
 used:

MIXED
  iop  BY rx period patid time
  /CRITERIA = CIN(95) MXITER(100) MXSTEP(5) SCORING(1)
  SINGULAR(0.000000000001) HCONVERGE(0, ABSOLUTE) LCONVERGE(0, ABSOLUTE)
  PCONVERGE(0.000001, ABSOLUTE)
  /FIXED = rx period time period*rx  | SSTYPE(3)
  /METHOD = REML
  /PRINT = DESCRIPTIVES G SOLUTION TESTCOV
 /RANDOM patid(period)  | COVTYPE(VC)
  /EMMEANS = TABLES(OVERALL)
  /EMMEANS = TABLES(rx) COMPARE ADJ(BONFERRONI)
  /EMMEANS = TABLES(time) COMPARE ADJ(BONFERRONI)
  /EMMEANS = TABLES(rx*time) COMPARE (rx) ADJ(BONFERRONI).

Have someone experienced the same problem? How could paired t-tests with Bonferroni adjusted p-values be performed with the above syntax? I would appreciate for any advice or help.

Thank you in advance,

Bettina


      ____________________________________________________________________________________
Never miss a thing.  Make Yahoo your home page.
http://www.yahoo.com/r/hs

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Zdaniuk, Bozena-2]

I experienced exactly the same problem when running GLM and asking for post-hoc comparisons. the results are exactly the same whether i specify LSD or SIDAK
Bozena
________________________________________
From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Bettina Haidich [[hidden email]]
Sent: Tuesday, November 27, 2007 4:22 PM
To: [hidden email]
Subject: cross-over design with longitudinal data

Dear list,


I have a cross-over repeated measures design of 40 patients (PATID) with 2 treatment periods (PERIOD) and the variable of interest (IOP) is measured 12 times (TIME). The total length of the dataset is 40x12x2=960. I found a syntax of mixed models for cross-over design in SPSS and I applied it to my data but I think that something is not calculated right, especially in multiple comparisons method. I would like to have Bonferroni adjusted  p-values for the comparisons between treatments in the 12 timepoints. Since this is a cross-over design the comparison in each time point should be paired and not independent. But what I get from the SPSS syntax with EMMEANS seems to be independent t-test comparison and although I use the option BONFERRONI I get unadjusted p-values. I checked that if I use the option LSD or SIDAK I get exactly the same p-values with the BONFERRONI option so I think that something is not calculated right here. Here follows the syntax I
 used:

MIXED
  iop  BY rx period patid time
  /CRITERIA = CIN(95) MXITER(100) MXSTEP(5) SCORING(1)
  SINGULAR(0.000000000001) HCONVERGE(0, ABSOLUTE) LCONVERGE(0, ABSOLUTE)
  PCONVERGE(0.000001, ABSOLUTE)
  /FIXED = rx period time period*rx  | SSTYPE(3)
  /METHOD = REML
  /PRINT = DESCRIPTIVES G SOLUTION TESTCOV
 /RANDOM patid(period)  | COVTYPE(VC)
  /EMMEANS = TABLES(OVERALL)
  /EMMEANS = TABLES(rx) COMPARE ADJ(BONFERRONI)
  /EMMEANS = TABLES(time) COMPARE ADJ(BONFERRONI)
  /EMMEANS = TABLES(rx*time) COMPARE (rx) ADJ(BONFERRONI).

Have someone experienced the same problem? How could paired t-tests with Bonferroni adjusted p-values be performed with the above syntax? I would appreciate for any advice or help.

Thank you in advance,

Bettina


      ____________________________________________________________________________________
Never miss a thing.  Make Yahoo your home page.
http://www.yahoo.com/r/hs

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD