Thanks Raoul.
This works and gives almost the same result as the modelbased aproach
- p=0.314 with smrby and p=0.316 with the modelbased.
Roland
2007/11/13, raoul reulen <[email protected]>:
> Hi Roland,
>
> I had the same problem with the command smrby. I changed the ado-file
> and now it works. I think the problem is to do with the weights, they
> need to be integer. I changed 'fweight' into 'iweight' and now it
> works.
>
> You don't need to collapse on the incidence variable, but you do need
> to calculate the expected number before you collapse the data.
>
> The model based option is based on a Wald test and should give you
> similar results.
>
> Hope this helps
>
> Raoul
>
> On 13/11/2007, roland andersson <[email protected]> wrote:
> > I have individual data with multiple records per patient after I have
> > split the follow up time depending on year and age and have merged a
> > variable for the incidence rate per 100.000 personyears. With the
> > command
> >
> > xi:strate i.timeperiod,smr(inc) per (100000)
> >
> > I get the following result
> >
> > +-----------------------------------------------------------+
> > timeperiod D E SMR Lower Upper
> > -----------------------------------------------------------
> > 1 7 0.43 16.102 7.677 33.777
> > 2 5 1.65 3.032 1.262 7.284
> > 3 14 2.11 6.642 3.934 11.215
> > +-----------------------------------------------------------+
> >
> > I want to test for the linear trend over the timeperiods.
> >
> > If I try the smrby on this result I get:
> >
> > . smrby D E, by(timeperiod) trend
> >
> > Observed Expected -- Poisson Exact --
> > timeperiod D E O/E (%) [95% Conf. Interval]
> >
> > 1 7 0.4300 1627.9*** 655 3354
> > 2 5 1.6500 303.0 98 707
> > 3 14 2.1100 663.5*** 363 1113
> > may not use noninteger frequency weights
> > r(401);
> >
> > Why do I get this error message?
> >
> > According to your second choice I use the following model on this result.
> >
> > glm D timeperiod, family(poisson) lnoffset(E) eform
> >
> > Iteration 0: log likelihood = -9.6767233
> > Iteration 1: log likelihood = -9.5071336
> > Iteration 2: log likelihood = -9.5068971
> > Iteration 3: log likelihood = -9.5068971
> >
> > Generalized linear models No. of obs = 3
> > Optimization : ML Residual df = 1
> > Scale parameter = 1
> > Deviance = 7.236772128 (1/df) Deviance = 7.236772
> > Pearson = 6.884656839 (1/df) Pearson = 6.884657
> >
> > Variance function: V(u) = u [Poisson]
> > Link function : g(u) = ln(u) [Log]
> >
> > AIC = 7.671265
> > Log likelihood = -9.506897131 BIC = 6.13816
> >
> >
> > OIM
> > D IRR Std. Err. z P>z [95% Conf. Interval]
> >
> > timeperiod .7561818 .2108008 -1.00 0.316 .4378616 1.305917
> > E (exposure)
> >
> > ie there is no linear trend p=0.316
> >
> > Is this a correct use of the glm model? Or can I use some other method
> > on the original dataset? If I collapse the dataset what happens with
> > the incidensvariable which should not be aggregated but stay the same.
> > Or do I have to collapse the dataset and merge the incidensvariable
> > after the collapse?
> >
> > Roland Andersson
> >
> >
> > 2007/10/18, raoul reulen <[email protected]>:
> > > Hi
> > >
> > > I'm not exactly sure what you want to do, but the title of the message
> > > suggests you want to do a test for linear trend accros several SIRs.
> > > What I normally do is use smrby.ado because it includes a test for
> > > trend and heterogeneity. Alternatively, you might want to model the
> > > SIRs by using a GLM model with a Poisson error structure and then fit
> > > the factor of interest as a consecutive non-negative integer variable.
> > > Make sure you collapse your data before you model it.
> > >
> > > .collapse (sum) _d E pyrs, by(timeperiod)
> > > .xi, noomit:glm _d i.timeperiod if E!=0,fam(pois) lnoffset(E) eform noconstant
> > >
> > > gives you the SIRs
> > >
> > > .xi, noomit:glm _d timeperiod if E!=0,fam(pois) lnoffset(E) eform noconstant
> > >
> > > gives you the p-value for linear trend
> > >
> > > _d is the numner of events, E the number of expected. The p-value for
> > > timeperiod should give you the test for linear trend accross the
> > > groups.
> > >
> > >
> > > Hope it helps
> > >
> > > Raoul Reulen
> > > Cancer Research UK Graduate Training Fellow
> > >
> > >
> > >
> > >
> > >
> > > On 17/10/2007, roland andersson <[email protected]> wrote:
> > > > I have used strate.ado to calculate standardised incidence ratios with
> > > > standardisation for age, sex and timepriod. I want to make inferences
> > > > on the development of the SIRs over three timeperiods. Can you help me
> > > > with instruction how to do this?
> > > > *
> > > > * For searches and help try:
> > > > * http://www.stata.com/support/faqs/res/findit.html
> > > > * http://www.stata.com/support/statalist/faq
> > > > * http://www.ats.ucla.edu/stat/stata/
> > > >
> > >
> > >
> > > --
> > > -------------------------------------------------------
> > > Raoul C. Reulen
> > > Cancer Research UK Training Fellow
> > > *
> > > * For searches and help try:
> > > * http://www.stata.com/support/faqs/res/findit.html
> > > * http://www.stata.com/support/statalist/faq
> > > * http://www.ats.ucla.edu/stat/stata/
> > >
> > *
> > * For searches and help try:
> > * http://www.stata.com/support/faqs/res/findit.html
> > * http://www.stata.com/support/statalist/faq
> > * http://www.ats.ucla.edu/stat/stata/
> >
>
>
> --
> -------------------------------------------------------
> Raoul C. Reulen
> Cancer Research UK Training Fellow
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
