Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
Tim Evans <Tim.Evans@wmciu.nhs.uk> |

To |
"'statalist@hsphsun2.harvard.edu'" <statalist@hsphsun2.harvard.edu> |

Subject |
RE: st: How to handle large (or small) odds ratios |

Date |
Fri, 18 May 2012 12:53:17 +0100 |

Hi David, Thanks for your response - my table formatting was not quite like that when it was sent! Having reviewed the data behind the analysis, Ive decided the prudent approach is to remove the staging analysis for those particular adjusted O/Rs as they were based on small numbers. I've also had a look at: 'How should one interpret 1.00E+00 as the LCI and UCI for Complex mixed and stromal? Is the standard error of that OR so small that the LCI and UCI both round to 1.00? Many other CIs are fairly wide.' Again, having looked at the number of cases, they are based upon either a single case, or a very small fraction of the total and appear to be behaving in the exact manner as the reference group. Given the apparent lack of robustness here, I've removed these also. Thanks for helping me to see the obvious. The results I have now are much more sensible. Best wishes Tim -----Original Message----- From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of David Hoaglin Sent: 18 May 2012 12:09 To: statalist@hsphsun2.harvard.edu Subject: Re: st: How to handle large (or small) odds ratios Hi, Tim. When your message arrived, the columns in the table didn't line up well, and some entries for LCI and UCI were missing (e.g., Stage 2, Stage 4, and Unknown stage --- I assume that Stage 1 is the reference category). The OR for Stage 3 seems to be 1.00E+00. Does that have an explanation when the ORs for Stage 2 and Stage 4 are 4.75E+05 and 1.66E+06, respectively? All the ORs are adjusted ORs; but if the OR for Stage 3 is actually 1.00E+00, it needs investigation. In the particular area of oncology, does other research indicate that having a stage other than Stage 1 at diagnosis increases the risk of dying so greatly? Only the ORs for the Stages are very LARGE. In the body of the table, you could note that they are large, and give the details in a footnote. Do articles in the target journal have any examples of how the journal prefers to handle large ORs? Putting them in a footnote should not be a problem when you submit your manuscript; the editor can tell you what the journal prefers. How should one interpret 1.00E+00 as the LCI and UCI for Complex mixed and stromal? Is the standard error of that OR so small that the LCI and UCI both round to 1.00? Many other CIs are fairly wide. David Hoaglin On Fri, May 18, 2012 at 5:06 AM, Tim Evans <Tim.Evans@wmciu.nhs.uk> wrote: > Hi all, > > Not specifically a Stata question, however, I have run a series of logistic regressions (in Stata) comparing the effects of delay in treatment and a persons likelihood of dying within a given time period. I have exported the analysis in readiness to produce tables, but what I wanted to know is whether there is a convention for dealing with very large odds ratios when producing publication tables, in order to produce neat legible tables. Many of these large odds ratios are very significant. Any advice gratefully received. > > Best wishes > > Tim > > Here is an example of my table: > > > Characteristic OR LCI UCI > Delay category > Not Delayed * 1.00 > Delayed 1.12 0.47 2.67 > Morphological Type > Adenomas and Adenocarcinomas 1.00 > Squamous Cell 1.00 1.00 1.00 > Cystic, Mucinous and Serous 5.01 0.87 29.05 > Complex mixed and stromal 1.00 1.00E+00 1.00E+00 > Other 1.69 0.57 5.06 > Stage at diagnosis > Stage 1 1.00 > Stage 2 4.75E+05 0.00E+00 > Stage 3 1.00E+00 1.00E+00 1.00E+00 > Stage 4 1.66E+06 0.00E+00 > Unknown stage 3.65E+05 0.00E+00 > Age at diagnosis (per year) *1.08 1.03 1.13 > Deprivation quintile > Most deprived 1.00 > More deprived 1.08 0.29 4.01 > Average deprivation 1.32 0.41 4.24 > Less deprived 1.18 0.35 3.94 > Least deprived 1.41 0.39 5.05 > Sex > Male *1.00 > Female 0.37 0.15 0.92 > Admission method > Elective 1.00 > Emergency 1.31 0.45 3.78 > Charlson combordity > 0 **1.00 > 1 0.79 0.27 2.31 > 2 0.86 0.17 4.26 > 3 13.77 4.87 38.92 > Diagnosis year > 2006 1.00 > 2007 1.03 0.26 4.07 > 2008 1.24 0.33 4.59 > 2009 1.17 0.31 4.39 > 2010 1.04 0.30 3.65 * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ _DISCLAIMER: This email and any attachments hereto contains proprietary information, some or all of which may be confidential or legally privileged. It is for the exclusive use of the intended recipient(s) only. If an addressing or transmission error has misdirected this e-mail and you are not the intended recipient(s), please notify the author by replying to this e-mail. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this e-mail or any attachments, as this may be unlawful. * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

**Follow-Ups**:**Re: st: How to handle large (or small) odds ratios***From:*David Hoaglin <dchoaglin@gmail.com>

**References**:**st: Hosmer-Lemeshow and other Pseudo Rsquares***From:*Joseph Padgett <josephpadgett@gmail.com>

**st: How to handle large (or small) odds ratios***From:*Tim Evans <Tim.Evans@wmciu.nhs.uk>

- Prev by Date:
**Re: st: Looping and saving regression outputs** - Next by Date:
**Re: st: How to handle large (or small) odds ratios** - Previous by thread:
**Re: st: How to handle large (or small) odds ratios** - Next by thread:
**Re: st: How to handle large (or small) odds ratios** - Index(es):