The SEQTEST Procedure

This example compares two binomial proportions by using a log odds ratio statistic in a five-stage group sequential test. A clinic is studying the effect of vitamin C supplements in treating flu symptoms. The study consists of patients in the clinic who exhibit the first sign of flu symptoms within the last 24 hours. These patients are randomly assigned to either the control group (which receives placebo pills) or the treatment group (which receives large doses of vitamin C supplements). At the end of a five-day period, the flu symptoms of each patient are recorded.

Suppose that you know from past experience that flu symptoms disappear in five days for 60% of patients who experience flu symptoms. The clinic would like to detect a 70% symptom disappearance with a high probability. A test that compares the proportions directly specifies the null hypothesis with a one-sided alternative and a power of 0.90 at , where and are the proportions of symptom disappearance in the treatment group and control group, respectively. An alternative trial tests an equivalent hypothesis by using the log odds ratio statistics:

Then the null hypothesis is and the alternative hypothesis is

The following statements invoke the SEQDESIGN procedure and request a five-stage group sequential design by using an error spending function method for normally distributed statistics. The design uses a two-sided alternative hypothesis with early stopping to reject the null hypothesis .

ods graphics on;
proc seqdesign altref=0.441833
               boundaryscale=mle
               ;
   OneSidedErrorSpending: design method=errfuncpow
                                 nstages=5
                                 alt=upper
                                 stop=accept
                                 alpha=0.025;
   samplesize model=twosamplefreq( nullprop=0.6 test=logor);
   ods output Boundary=Bnd_CSup;
run;
ods graphics off;

The ODS OUTPUT statement with the BOUNDARY=BND_CSUP option creates an output data set named BND_CSUP which contains the resulting boundary information for the subsequent sequential tests.

The “Design Information” table in Output 84.5.1 displays design specifications and derived statistics. With the specified alternative reference, the maximum information 56.30934 is derived.

The “Boundary Information” table in Output 84.5.2 displays information level, alternative reference, and boundary values at each stage. With the specified BOUNDARYSCALE=MLE option, the procedure displays the output boundaries in terms of the MLE scale.

With ODS Graphics enabled, a detailed boundary plot with the rejection and acceptance regions is displayed, as shown in Output 84.5.3.

With the SAMPLESIZE statement, the “Sample Size Summary” table in Output 84.5.4 displays the parameters for the sample size computation.

The “Sample Sizes” table in Output 84.5.5 displays the required sample sizes for the group sequential clinical trial.

Thus, 101 new patients are needed in each group at stages 1, 2, and 4, and 100 new patients are needed in each group at stages 3 and 5. Suppose that 101 patients are available in each group at stage 1. Output 84.5.6 lists the 10 observations in the data set count_1.

The TrtGrp variable is a grouping variable with the value Control for a patient in the placebo control group and the value C_Sup for a patient in the treatment group who receives vitamin C supplements. The Resp variable is an indicator variable with the value 1 for a patient without flu symptoms after five days and the value 0 for a patient with flu symptoms after five days.

The following statements use the LOGISTIC procedure to compute the log odds ratio statistic and its associated standard error at stage 1:

proc logistic data=CSup_1 descending;
   class TrtGrp / param=ref;
   model Resp= TrtGrp;
   ods output ParameterEstimates=Parms_CSup1;
run;

The DESCENDING option is used to reverse the order for the response levels, so the LOGISTIC procedure is modeling the probability that Resp = 1.

The following statements create and display (in Output 84.5.7) the data set for the log odds ratio statistic and its associated standard error:

data Parms_CSup1;
   set Parms_CSup1;
   if Variable='TrtGrp' and ClassVal0='C_Sup';
   _Scale_='MLE';
   _Stage_= 1;
   keep _Scale_ _Stage_ Variable Estimate StdErr;
run;

proc print data=Parms_CSup1;
   title 'Statistics Computed at Stage 1';
run;

The following statements invoke the SEQTEST procedure to test for early stopping at stage 1:

ods graphics on;
proc seqtest Boundary=Bnd_CSup
             Parms(Testvar=TrtGrp)=Parms_CSup1
             infoadj=prop
             errspendadj=errfuncpow
             boundarykey=both
             boundaryscale=mle
             ;
   ods output test=Test_CSup1; 
run;
ods graphics off;

The BOUNDARY= option specifies the input data set that provides the boundary information for the trial at stage 1, which was generated in the SEQDESIGN procedure. The PARMS=PARMS_CSUP1 option specifies the input data set PARMS_CSUP1 that contains the test statistic and its associated standard error at stage 1, and the TESTVAR=TRTGRP option identifies the test variable TRTGRP in the data set.

If the computed information level for stage 1 is not the same as the value provided in the BOUNDARY= data set, the INFOADJ=PROP option (which is the default) proportionally adjusts the information levels at future interim stages from the levels provided in the BOUNDARY= data set. The ERRSPENDADJ=ERRFUNCPOW option adjusts the boundaries with the updated error spending values generated from the power error spending function. The BOUNDARYKEY=BOTH option maintains both the and levels. The BOUNDARYSCALE=MLE option displays the output boundaries in terms of the MLE scale.

The ODS OUTPUT statement with the TEST=TEST_CSUP1 option creates an output data set named TEST_CSUP1 which contains the updated boundary information for the test at stage 1. The data set also provides the boundary information that is needed for the group sequential test at the next stage.

The “Design Information” table in Output 84.5.8 displays design specifications. With the specified BOUNDARYKEY=BOTH option, the information levels and boundary values at future stages are modified to maintain both the and levels.

The “Test Information” table in Output 84.5.9 displays the boundary values for the test statistic with the specified MLE scale. With the INFOADJ=PROP option (which is the default), the information levels at future interim stages are derived proportionally from the observed information at stage 1 and the information levels in the BOUNDARY= data set.

Since the information level at stage 1 is derived from the PARMS= data set and other information levels are not specified, equal increments are used at remaining stages. At stage 1, the MLE statistic 0.32474 is greater than the corresponding upper boundary value –0.29906, so the sequential test continues to the next stage.

With ODS Graphics enabled, a boundary plot with the boundary values and test statistics is displayed, as shown in Output 84.5.10. As expected, the test statistic is in the continuation region.

The following statements use the LOGISTIC procedure to compute the log odds ratio statistic and its associated standard error at stage 2:

proc logistic data=CSup_2 descending;
   class TrtGrp / param=ref;
   model Resp= TrtGrp;
   ods output ParameterEstimates=Parms_CSup2;
run;

The following statements create and display (in Output 84.5.11) the data set for the mean positive response and its associated standard error at stage 2:

data Parms_CSup2;
   set Parms_CSup2;
   if Variable='TrtGrp' and ClassVal0='C_Sup';
   _Scale_='MLE';
   _Stage_= 2;
   keep _Scale_ _Stage_ Variable Estimate StdErr;
run;
proc print data=Parms_CSup2;
   title 'Statistics Computed at Stage 2';
run;

The following statements invoke the SEQTEST procedure to test for early stopping at stage 2:

proc seqtest Boundary=Test_CSup1
             Parms( testvar=TrtGrp)=Parms_CSup2
             infoadj=prop
             errspendadj=errfuncpow
             boundarykey=both
             boundaryscale=mle
             ;
   ods output Test=Test_CSup2;
run;

The BOUNDARY= option specifies the input data set that provides the boundary information for the trial at stage 2, which was generated by the SEQTEST procedure at the previous stage. The PARMS= option specifies the input data set that contains the test statistic and its associated standard error at stage 2, and the TESTVAR= option identifies the test variable in the data set.

The ODS OUTPUT statement with the TEST=CSUP_LDL2 option creates an output data set named CSUP_LDL2 which contains the updated boundary information for the test at stage 2. The data set also provides the boundary information that is needed for the group sequential test at the next stage.

The “Test Information” table in Output 84.5.12 displays the boundary values for the test statistic with the specified MLE scale. The test statistic 0.2356 is greater than the corresponding upper boundary value –0.01068, so the sequential test continues to the next stage.

Similar results are found at stages 3 and stage 4, so the trial continues to the final stage. The following statements use the LOGISTIC procedure to compute the log odds ratio statistic and its associated standard error at stage 5:

proc logistic data=CSup_5 descending;
   class TrtGrp / param=ref;
   model Resp= TrtGrp;
   ods output ParameterEstimates=Parms_CSup5;
run;

The following statements create and display (in Output 84.5.13) the data set for the log odds ratio statistic and its associated standard error at stage 5:

data Parms_CSup5;
   set Parms_CSup5;
   if Variable='TrtGrp' and ClassVal0='C_Sup';
   _Scale_='MLE';
   _Stage_= 5;
   keep _Scale_ _Stage_ Variable Estimate StdErr;
run;

proc print data=Parms_CSup5;
   title 'Statistics Computed at Stage 5';
run;

The following statements invoke the SEQTEST procedure to test for the hypothesis at stage 5:

ods graphics on;
proc seqtest Boundary=Test_CSup4
             Parms( testvar=TrtGrp)=Parms_CSup5
             errspendadj=errfuncpow
             boundaryscale=mle
             cialpha=.025
             rci
             plots=rci
             ;
run;
ods graphics off;

The BOUNDARY= option specifies the input data set that provides the boundary information for the trial at stage 5, which was generated by the SEQTEST procedure at the previous stage. The PARMS= option specifies the input data set that contains the test statistic and its associated standard error at stage 5, and the TESTVAR= option identifies the test variable in the data set. By default (or equivalently if you specify BOUNDARYKEY=ALPHA), the boundary value at stage 5 is derived to maintain the level.

The “Test Information” table in Output 84.5.14 displays the boundary values for the test statistic with the specified MLE scale. The test statistic 0.2043 is less than the corresponding upper boundary 0.25375, so the sequential test stops to accept the null hypothesis. That is, there is no reduction in duration of symptoms for the group receiving vitamin C supplements.

The “Test Plot” displays boundary values of the design and the test statistics, as shown in Output 84.5.15. It also shows that the test statistic is in the “Acceptance Region” at the final stage.

After a trial is stopped, the “Parameter Estimates” table in Output 84.5.16 displays the stopping stage, parameter estimate, unbiased median estimate, confidence limits, and the p-value under the null hypothesis . As expected, the p-value 0.0456 is not significant at level and the lower 97.5% confidence limit is less than the value . The p-value, unbiased median estimate, and confidence limits depend on the ordering of the sample space , where k is the stage number and z is the standardized Z statistic.

Since the test is accepted at stage 5, the p-value computed by using the default stagewise ordering can be expressed as

where is the test statistic at stage 5, is a standardized normal variate at stage k, and is the upper boundary value in the standardized Z scale at stage .

With the RCI option, the “Repeated Confidence Intervals” table in Output 84.5.17 displays repeated confidence intervals for the parameter. For a one-sided test with an upper alternative hypothesis, since the upper acceptance repeated confidence limit 0.3924 at the final stage is less than the alternative reference 0.441833, the null hypothesis is accepted.

With the PLOTS=RCI option, the “Repeated Confidence Intervals Plot” displays repeated confidence intervals for the parameter, as shown in Output 84.5.18. It shows that the upper acceptance repeated confidence limit at the final stage is less than the alternative reference 0.441833. This implies that the study accepts the null hypothesis at the final stage.