1. pain prevalence stratified by population type (workers/general/elderly), for each pain type (8); "Pain Tables by Type global review 12 aug TJ.docx"
  2. gender odds ratio meta analysis; need number of subjects, crude (unadjusted) odds ratio, proportion of males, overall prevalence of pain, and consistent pain type (e.g., "All"); "gender global review 12 aug TJ.docx"
  3. depression/anxiety/PTSD odds ratio meta analysis; need same as above for gender odds ratio meta analysis; "Depression, anxiety, ptsd~1.docx‎"
  4. other (rural/education/smoking/income/mental illness) odds ratio meta analysis; need same as above for gender odds ratio meta analysis; "other associations global review 12 aug TJ.docx"
  5. odds ratio for disability in pain versus no pain; need all items listed above for each study; this outcome is difficult because disability is not well defined by each study; "disability table global review 12 aug TJ.docx"
  6. age odds ratio quantitative meta analysis is difficult because studies do not use the same age categorizations; "Age table global review 12 aug TJ.docx"
  7. do not use "global review spreadsheet ...xlsx"

Notes from 8/26/2014

  1. review LBP papers in mood (depression) tables, try to identify unadjusted odds ratio

Notes from 8/28/2014 (Xue)

  1. Download LBP studies (six) which may have association study with mood (depression). There is one study by Tavafian.2008 which is a focus group study, this one was not included in the excel file but available in the word document. Not sure whether we should consider this study or not.

Author/Year Nationality Mood+ LBP+ Mood- LBP+ Mood+ LBP- Mood- LBP- Crude Odds Ratio Time (min) Notes
Sharma et al. (2003) India 1738 856 190 810 8.65 12 "psychosocial issues" or "psychological issues"; paper is very short
Altinel et al. (2008) Turkey 72 966 30 967 2.40 8 "depression"; adjusted OR: 1.846
Bejia et al. (2005) Tunisia 75 101 56 118 1.56 15 "Distributed psychological profile" Adjusted OR 1.93
EL-Sayed et al. (2010) Ethiopia 93 57 232 518 3.64 10 "depression",adjusted OR 3.44: paper also talk about anxiety and PTSD
Tavafian (2008) Iran X X X X   8 Focus group study, counts for cells not identified.only information available is there are totally 24 subjects in the study
Vuuren (2005) South Africa 8 25 11 66 1.92 10 Different definition for LBP, (used FRI definition), "Fear avoidance belief" to measure the mood disorder, adjusted OR, 2.35


  • Send a updated report to Tracy, including 1) the meta proportion analysis on LBP for all subjects, general subjects, elder subjects, workers respectively. 2) the meta OR analysis for the association between LBP and depression using the information from above 6 articles. meta.9.05lbp.pdf


  1. pain prevalence meta-analysis is easiest, once data are cleaned; Xue will coordinate with Tracy and student in this regard
  2. the gender association meta-analysis is the most clean, but need to get all of the required data; Xue will give an example of the required data
  3. if the depression/anxiety/PTSD outcomes are well defined, then these meta-analyses can be done as well; still need required information
  4. disability and age are both problematic; recommend examining these last, or omitting

TODO 10/08/2014

1. In figure 2.2 and 2.3 we need to remove demyttenaere as that is a study of 18 countries not just LMIC

2. In the HA tables and FMS tables, Jen had separated it out, but this was not reflected in the meta-analysis. Here are the categories we needed to separate: ha.prevalence=4 categories: A. headache (this prevalence will be higher) B. chronic daily headache C. chronic migraine D. chronic TTH (you could combine B-D in a "chronic, all" category if necessary, but if just as easy, please do all 3 separately as well)

3. For HA, de Sigueira looked at both gen pop prevalence and elderly gen pop prevalence, I separated these into 2 columns, one for each, it doesn't look like the elderly figure from that paper got included in the elderly portion of the metaanalysis.

4. Also for fms.prevalence, MUS,FMS, and WSP was combined in one category in jen's table called FMS while there was another for msk.prevalence that includes Joint, arthritis, MSK. The current metanalysis combines all this under "FMS" but FMS/WSP prevalence going to be much lower than MSK/joint, that is why we need those separated. In fact, I have separated them out into 4 columns: FMS, WSP, MSK, and joint. I think this clears up the picture and shows what the literature does: FMS is less than WSP, and joint pain prob is equal or slightly higher than MSK pain.

5. For FMS/MSK, the Abegunde and Kayode studies look like they are reporting from the same population, the prevalence is the same, but not sure if we should include it twice or if it matters. However, Jen had reported them differently. I separated them both out for accuracy, but again, may only need to include one of them.

6. The Gomez Olive study (line 5 of FMS/MSK table) had 14.7 transposed instead of 41.7. This has been corrected in the spreadsheet.

7. Line 23 FMS, assumed 50/50 M/F as not reported to calculate prevalence

8. I deleted the "indigestion" study from the AP tables, after reviewing the paper again, I don't think that is looking at the same thing.

9. I sorted all the tables by population and took out any that were "clinic patients" or populations other than gen pop. elderly gen pop, and workers (ie females). Those were skewing our "all" results, I think, and clinic patients do not reflect what we are trying to analyze as overall general prevalence. If you could rerun "all" without using anything but the 3 pops (general, general elderly, and workers) I think it will be better., and I'll account for the reasoning for this in my discussion.

10 REgarding #5 below, the Kayode study actually doesn't exist, that's an error. Anything that says Kayode should be changed to Abegunde. It turns out they were the same study after all.

11 Oladeji in the LBP group is listed in error as a general population, it is an ELDERLY general population. This should be removed from the LBP gen pop met analysis and added to the gen elderly one. Also, Kayode in the elderly gen pop meta analysis should be changed to the name abegunde as described below.

-- MattShotwell - 22 Aug 2014
Topic attachments
ISorted ascending Attachment Action Size Date Who Comment
Sharma.2003.India.LBP.htmlhtml Sharma.2003.India.LBP.html manage 68.2 K 28 Aug 2014 - 11:53 XueHan  
Altinel.2008.Turkey.LBP.pdfpdf Altinel.2008.Turkey.LBP.pdf manage 286.2 K 28 Aug 2014 - 11:48 XueHan  
Bejia.2005.Tunisia.LBP.pdfpdf Bejia.2005.Tunisia.LBP.pdf manage 134.3 K 28 Aug 2014 - 11:48 XueHan  
ELsayed.2011.Ethiopia.NP.pdfpdf ELsayed.2011.Ethiopia.NP.pdf manage 292.3 K 28 Aug 2014 - 11:48 XueHan  
Tavafian.2008.Iran.LBP.pdfpdf Tavafian.2008.Iran.LBP.pdf manage 111.7 K 28 Aug 2014 - 11:48 XueHan  
Van_Vuuren.2005.SouthAfrica.LBP.pdfpdf Van_Vuuren.2005.SouthAfrica.LBP.pdf manage 228.2 K 28 Aug 2014 - 11:48 XueHan  
meta.9.05lbp.pdfpdf meta.9.05lbp.pdf manage 233.9 K 09 Sep 2014 - 15:55 XueHan  
Topic revision: r10 - 08 Oct 2014, XueHan

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback