SAS Support Communities

kc · ‎08-02-2018

I wonder, Why do you want a piecewise linear regression? How many breakpoints do you need? Do you know where the breakpoint(s) for the pieces is(are) (for example, at 6 months), or do you need to estimate the location(s) of the breakpoint(s)? Well, from prior clinical knowledge, there is no significant effect of treatment on summary score beyond the 6 month timepoint. Therefore, there is need for only 2 breakpoints, one each at 1 month and 6 month. I included only one breakpoint at 6 month in my code as an example. Is your actual dataset larger than the one you posted? (In the future, please post as a CSV file, rather than Excel.) Yes, much larger. Few subjects in your posted dataset have values for all 7 times (5 out of 27 subjects), and many have only 1 value (8 out of 27). Only 14 subjects have data at both 1 month and 6 months. Correct - some subjects have summary score data at all times, and some don't. The data, although dummy, mimics the data from an actual trial. So, missing data is quite common. Can you sensibly fit a model (with or without random slopes) for multiple linear pieces using a data set that is so incomplete? I'd say, probably not. I also would be concerned about potential bias in either ANOVA or regression models due to missing data and why data are missing. Multiple methods (paired t-tests, ANCOVA, mixed effects models) are employed in analyzing these data. Also, an underlying assumption of longitudinal growth curve models is that the missing data is missing at random. So, any help with the syntax in running a piecewise regression to fill the table in my original post would be great. Let me know if more data (in CSV format this time) would be helpful in carrying out this task!

kc · ‎08-02-2018

I am trying to run piece wise linear regression on a longitudinal dataset (because growth curve modeling is giving results which are clinically not plausible at certain time points). The sample dataset is attached in excel format and has the following columns/variables 1. Subject ID 2. Clinically planned event name (total 7 time points possible for a subject - 1 month, 6 months, 1 year, 2 years, 3 years, 4 years, 5 years) 3. Time (time in years) 4. Summary Score (the dependent variable in the model) 5. m6 (dummy variable for the first 6 months) 6. post6m (dummy variable for post - 6 months) 7. Group (treatment group) The model I currently have is as follows: proc glimmix data=PLR; class subjectid; model summaryscore = m6 post6m/solution; random intercept m6 post6m/ subject=subjectid type=chol; run; Here is the question I have: How do I get mean summary score by treatment group and difference between treatment groups in mean summary score (along with 95% CI and p-values) at the 7 different time points in the study? In other words, can someone help me with the syntax to add treatment group and time, and the interaction between treatment and time as covariates in the model, to get the mean and mean difference in summary scores between treatment groups, at various time points? Here is what I need: Predicted Mean Values (95% CI) Predicted Mean Difference (TRT1-TRT2), 95%CI P-value TRT1 TRT2 Summary Score 1 Month xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) 0.xxxx 6 Months xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) 0.xxxx 1 Year xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) 0.xxxx 2 Years xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) 0.xxxx 3 Years xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) 0.xxxx 4 Years xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) 0.xxxx 5 Years xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) xx.x (xx.x, xx.x) 0.xxxx Any suggestions are greatly appreciated!

kc · ‎05-03-2017

I have a health status scale measured at 2 different timepoints (baseline and 1 month) with 5 different response levels (0-4). I want to see if there is a statistically significant difference in within group change from baseline to the 1 month time point for each of these response levels. My questions are as follows: 1. Am I using the right code to accomplish this task? 2. If so, which statistics do I report to highlight if the difference is significant? 3. Finally, I think the procedure below gives an overall p-value and not a level-by-level p-value and statistics. How do I test a level by level difference between baseline and 1 month response? Like the difference between response level 0 between baseline and 1 month. I am running the following proc freq code (analogous to paired t-test for continous variables): proc freq data=health; tables hs_BASELINE*hs_1_MONTH / agree norow nocol nopercent; run; The result is as follows: SAS Output Frequency Table of hs_BASELINE by hs_1_MONTH hs_BASELINE hs_1_MONTH 0 1 2 3 4 Total 0 327 55 25 12 17 436 1 236 100 37 17 14 404 2 140 66 42 21 22 291 3 106 39 36 14 21 216 4 62 38 33 30 67 230 Total 871 298 173 94 141 1577 Frequency Missing = 325 Statistics for Table of hs_BASELINE by hs_1_MONTH Test of Symmetry Statistic (S) 328.8670 DF 10 Pr > S <.0001 Kappa Statistics Statistic Value ASE 95% Confidence Limits Simple Kappa 0.1402 0.0139 0.1130 0.1674 Weighted Kappa 0.2379 0.0171 0.2044 0.2714 Effective Sample Size = 1577 Frequency Missing = 325 WARNING: 17% of the data are missing.

kc · ‎03-22-2017

Here is a contrast statement that works: (you can write the statement for other time points in a similar fashion) proc mixed data=growth noclprint order=formatted covtest noitprint method=reml; class group pt strata time(ref="1"); model score = baseline_score group time time*group strata group*strata time*strata time*group*strata/solution ddfm=bw notest; random pt /vcorr; contrast 'pvalue at 1 month' group*strata 1 -1 -1 1 time*group*strata 1 0 0 0 0 0 -1 0 0 0 0 0 -1 0 0 0 0 0 1 0 0 0 0 0; lsmeans time*group*strata/pdiff cl ; ods output Contrasts=cont_(keep=label probf) SolutionF=est_(keep=effect estimate stderr probt) LSMeans=lsm_(keep=effect group strata time estimate lower upper) Diffs=diff_(keep=effect group strata _group _strata time _time estimate lower upper probt); run;

kc · ‎03-16-2017

I am running the following model to run a repeated measures ancova. the data has 2 treatment groups, 2 stratification levels and 5 time points (baseline, 1m, 6m, 12m, 24m). The dependent variable is the score at post-baseline time points (1,6,12,24 months). My question is, how do I get an overall p-value for the 3-way interaction at each post-baseline timepoint? I am not sure of how to write a contrast statement to accomplish this? Please help! proc mixed data=growth noclprint order=formatted covtest noitprint method=reml; class group pt strata time(ref="1"); model score = baseline_score group time time*group strata group*strata time*strata time*group*strata/solution ddfm=bw notest; random pt /vcorr; lsmeans time*group*strata/pdiff cl ; ods output SolutionF=est_(keep=effect estimate stderr probt) LSMeans=lsm_(keep=effect group strata time estimate lower upper) Diffs=diff_(keep=effect group strata _group _strata time _time estimate lower upper probt); run;

kc · ‎07-12-2016

Thanks Dan! I should have probably mentioned that along with the code.Will try and see if the code works after upgrading to TS1M3 release.

kc · ‎07-12-2016

I am using SAS 9.4 (TS1M1) - Do I need TS1M3 for the code to work (I see that on some blogs)?

kc · ‎07-12-2016

Hi Dan, I tried using the default escape sequence, but with no luck. The symbol does show up as its supposed to in the title and footnote statements no matter which escape sequence I use, but just not as an axis value for some reason.

kc · ‎07-12-2016

Here is the sgplot code and the proc format code (no titles or footnote statements) proc format library=work; value yscale 200 = 'Treatment' 160 = 'Age ^{unicode 2265} 75' 140 = 'Age < 75' 100 = 'US' 80 = 'Outside US' other = '' ; run; ods options orientation=portrait; ODS ESCAPECHAR='^'; ods graphics on / border=off; ods rtf file='C:\Users\Z1000\Desktop\newplot.rtf' sge=on; proc sgplot data=combined sganno=anno; scatter x=meanestimate y=Order / xerrorlower=MeanLowerCL xerrorupper=MeanUpperCL markerattrs=OR (symbol=DiamondFilled size=6); refline 1/axis=x; format Order yscale.; xaxis label="Relative Rate" LABELATTRS=(Family='Arial' ) VALUEATTRS=(Family=Arial SIZE=10 ) type=LOG LOGBASE= 10 min=0.5 max=2 values = (0.5 1 2); yaxis label="SubGroups" display=(nolabel) tickvalueformat=yscale. VALUEATTRS=(Family=Arial SIZE=10 ) offsetmin=0.1 offsetmax=0.1 VALUES = (80 100 140 160 200); run; ods rtf close;

kc · ‎07-09-2016

I am trying to display the ≥ symbol (greater than or equal to) on y-axis in a plot using the sgplot procedure. But, the ≥ symbol is replaced by just the = symbol in the RTF document produced. The ≥ symbol is part of a string in a character variable. Any suggestions? Thanks, KC

kc · ‎05-25-2016

Is it possible to display an "inset" box outside the plot area, to the right of the graph?

kc · ‎04-07-2016

I am trying to change the order of the legend entries to match the visual order of the bar themselves using proc sgplot. I did go through one of the posts on the forums here https://communities.sas.com/t5/SAS-GRAPH-and-ODS-Graphics/SGPLOT-legend-order/m-p/193197#M7129. But SAS 9.4 doesn't seem to recognize the "sortorder=descending" option mentioned in the keylegend statement given in the example of how to match up the legend entries with the visual order (ran the code as is from the first post). Any help is appreciated!

kc · ‎10-15-2015

Worked perfectly, thanks!

kc · ‎10-13-2015

I did run this code below with no luck.

kc · ‎10-13-2015

How do I keep the scale of y-axis from 0 to 10 but only display tick values at 2 4 6 8. Essentially, I want to have 5 intervals (from 0 to 10) instead of 3, so my scatter plot is placed in the middle of the graph instead of being more spread out. My dataset has only 4 observations at (2 4 6 8). Please find an attachment with display of plots.

Online Status	Offline
Date Last Visited	‎09-26-2024 03:14 PM

SAS Support Communities

Re: Trend test for multinomial variable

Re: Trend test for multinomial variable

Re: Trend test for multinomial variable

Re: Trend test for multinomial variable

Re: Trend test for multinomial variable

Trend test for multinomial variable

Re: Test Proportional Hazards Assumption in Recurrent Event Model

Re: Test Proportional Hazards Assumption in Recurrent Event Model

Test Proportional Hazards Assumption in Recurrent Event Model

at-risk table in proc phreg - recurrent event analysis

Re: proc nlmixed - test if the interaction term is significant at a sp...

Re: Simulating Survival Data - issue with overestimation

Re: Proc IML - evaluating multiple conditions

Beta - Binomial Bayesian Model

Re: proc sgplot - axis scale and tick marks

Re: Piecewise Linear Regression

Piecewise Linear Regression

proc freq - agree option for matched pairs

Re: proc mixed 3-way overall interaction p-value

proc mixed 3-way overall interaction p-value

Re: ODS RTF ≥ symbol in a plot

Re: ODS RTF ≥ symbol in a plot

Re: ODS RTF ≥ symbol in a plot

Re: ODS RTF ≥ symbol in a plot

ODS RTF ≥ symbol in a plot

SGPLOT Inset - Outside the plot area

proc sgplot legend entries

Re: proc sgplot - axis scale and tick marks

Re: proc sgplot - axis scale and tick marks

proc sgplot - axis scale and tick marks

Follow Us

What is...