NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Gellad WF, Maggard Gibbons M, Miake-Lye IM, et al. A Comparison of Joint Replacement Disparities in VA and Non-VA Settings: A Systematic Review [Internet]. Washington (DC): Department of Veterans Affairs (US); 2011 Sep.

Cover of A Comparison of Joint Replacement Disparities in VA and Non-VA Settings: A Systematic Review

A Comparison of Joint Replacement Disparities in VA and Non-VA Settings: A Systematic Review [Internet].

Show details


General Comments
Issues such as waiting time, access to orthopedics consultations, etc. have not been clearly documented, and it is unknown if they may relate to some of the differences observed. If these barriers indeed exist, they should be addressed through quality improvement measures.There is a new section in the limitations that addresses this.
Abbreviations not used consistentlyAfter double-checking for inconsistencies, some remain due to the literature itself.
Missing: Hausmann, et al, Arthritis Care & Rheum 2011, p635-642Now included, found in update search
In general the ‘summary of findings’ sections read better than the ‘data’ sections. The data sections need revision. Revisions to the data sections should be directed at reading fluency to better convey the intended message. Substantial editorial attention to writing, paragraph structure, style, grammar, and typographical errors is suggested.Edits have been made to address this issue.
Authors may want to do a brief re-search, using the same database and keywords, for the period from January through June 2011 as a final update prior to publication.Update is now included
I have some reservations about the level of evidence available to make any decisions of consequence. Although there is no documentation of the levels, it appears that most cited studies are at minimum a level III or IV. It would be helpful to document by the standard definitions and the numerical system to be sure everyone is on the same page. In addition, grades of recommendation for the consensus should use a standard nomenclature such as A,B,C,I, again for clarity. When levels of evidence are so low or the mass of supporting evidence is so low and recommendation can only be I, the only conclusion would be we need focused research of the highest quality and nothing more. In this current context, the supposed disparity may have no other basis than personal preference, cultural beliefs and population bias which may not be alterable.The ESP program uses the GRADE system, which does not use the level I, II, III approach. The limitations of the evidence are reflected in the overall “quality of evidence,” and most of these are rated as low or very low.
Executive Summary
The summary was a little vague with respect to results. Since many individuals may only access the summary, a more precise summary of results would be appropriate, such as including how many studies were available for each key question, and whether they included or not VA populations. In general, the findings are reported as ‘few studies….’ or ‘most….’. Including number of studies and participants would be useful. This is all included in the main report, but would be useful in the summary.We have revised the executive summary to include more detail
2nd paragraphs notes there are “disparities” in TJR use in non-VA settings. It would be useful here to mention types of disparities that are being alluded to (e.g., race, gender, ethnicity…)This paragraph has been updated to be more specific.
page 2, Key Question #1, 2nd paragraph notes that future research is unlikely to change confidence on the estimate of the effect. It should be qualified here (and elsewhere in the synthesis where this is also mentioned) that future research is still important for evaluating whether there are any temporal trends in disparities (e.g., do these change over time in response to any policies, interventions, etc).Very good point, we have updated the relevant sections accordingly.
On page 2 in the executive summary----it would be nice in the summary to add the number of studies contributing to the literature for each KQThese numbers have been added at the beginning of the results section.
In the Background on page 1, page 5, and elsewhere the report refers to “disparities.” This description is too general because the report only addresses gender and racial disparities. Clarification of this usage to use language such as “gender and racial disparities” in place of “disparities” is suggested.We revised the draft to clarify.
The authors are limited by research papers which primarily address only two racial / ethnic groups (White, and African-American) and don't clearly address educational, socioeconomic and regional effects.This is a limitation of the primary literature. Almost all the disparity literature deals with gender and race. Even within gender disparities, VA data are very scarce. These additional potential sources of disparities are now noted in the limitations.
Search strategy includes ‘peer-reviewed’ articles. How is this assessed? Do the authors mean original publications? If reviews were included, how was it determined if they were peer-reviewedWe revised this to indicate that anything indexed on PubMed was potentially eligible
In the Methods (page 6) under the heading ‘search strategy’ more detailed description of the search terms should be provided. At a minimum indicate the surgical procedures THR and TKR.We have added some of our specific terms, and the entire search strategy is in Appendix A.
Figure 3 Literature Flow seems to have a discrepancy in the number of articles categorized in the bottom row of boxes. There are 22+35+1 = 58 studies categorized in the bottom row. The row above indicates there were 69 articles assessed. So it seems there are 11 articles (69 minus 58 = 11) that are not categorized.Additional explanations have been added to clarify the overlapping nature of the categories, which accounts for the numerical discrepancy.
Figure 3 - It may also be useful to provide a breakdown of which or how many articles addressed racial disparities and how many addressed gender disparities. The current breakdown seems to indicate only 1 article addressed gender disparitiesThis is correct, there was only one gender article.
I can't follow Figure 3 and the numbers. It says 69 articles were assessed but the numbers below don't add up. Please clarify/fix. Also, I think it would help the reader to explain the literature groups below the figure.Additional explanations have been added to clarify the overlapping nature of the categories, and the groups are now referred to by key question, rather than generation, for clarity.
Study Design
Authors state that study design was not used as inclusion/exclusion criterion. However, Figure 3 includes inappropriate study design as a rejection criterion.We have reworded the figure to be more specific.
The first set of results for Key Question #2 is related to a comparison of VA and non-VA county hospitals. It would be helpful to the reader if there is a clear statement about how these data relate to the key questions (e.g. differences according to a system-level factor?).Updates have been made to address this.
Page 12 – 2nd paragraph under VA Data, 3rd sentence – it is not completely clear what the comparison is for the OR, and there is a grammar issue in the sentence.This sentence has been updated.
Page 12, three studies are alluded to – is the 2nd paragraph in this section about the third study?Yes, this has been noted in the text.
Page 12, 2nd paragraph – which ICD-9 codes were used?ICD-9 codes have been added where necessary
Page 13, 3rd paragraph under Non-VA data: some discussion of the magnitude of differences in TKR rates would be helpful.The Kane review (published in arthritis care and research), does not list actual rates of use, just that rates in one group are larger or smaller than others. We think that the rates presented in the following studies below can present a sense of magnitude of differences in rates.
Page 17, Summary of Findings – doesn't fully summarize the results (or lack thereof) regarding patient, provider, and system level factors.Changes have been made.
In some places it is clear which cohort is being referred to, in other places it is less clear (e.g., “another VA cohort” on p18, 2nd paragraph). Throughout, it would be helpful to have a consistent way of referring to each study in the table.Updates have been made to better identify the cohort (e.g. the Cleveland cohort)
It is not clear that the general information under Non-VA Data that starts on p21 is needed. It seems a bit out of place here.This section has been updated for better flow.
P24, Summary of findings – it would be helpful to compare / contrast this with VA data, mentioning any different findings or just areas in which there are may be more data for non-VA vs. VA.Updates have been made to address this.
For KQ1 there is one ‘summary of findings’ section at the end that includes both VA and non-VA data. For KQ2, the structure differs and was confusing at first – that is, within the KQ2 sections there are three ‘summary’ sections for each of VA + non-VA, VA, and non-VA.This structure was used due to the volume of literature in the sections.
There is only one study examining gender disparities in the VA. This finding of limited research related to gender is not highlighted in the summary of findings.The summary has been updated to reflect this.
Note: on page 12, first sentence under VA data---I think you want a “the” before VA. Also, in the 2nd paragraph, note there is a comma rather than a period in the pt estimate of 0.3 %Changes have been made.
page 16---para 30--- fix tense of first sentence. Note also that the last sentence of this paragraph does not explain what the 2 fold higher odds are of????This sentence has been removed.
Page 17. Please clarify last sentence of 2nd paragraphEdits have been made to address this issue.
Page 18. 2nd paragraph, 2nd to last sentence-----I think you mean TJR rather than OAOA has been verified.
Page 19-last sentence----take out “thus”Fixed
Page 23----2nd paragraph----review the middle sentence that states “social support between various racial groups after undergoing a hip fracture…‥This sentence has been reworded for grammar and clarity.
It is hard to get too excited about KQ 3 since there seemed to be little good evidence about disparities in the VA. I might be clearer about the limitations of the VA data on disparities as you discuss an intervention to improve them in KQ3.Noted
Note also that in the first sentence the word “joint” probably doesn't belong there or you need to add TJRSentence has been verified.
Page 26 paragraph 4----since you are talking about the disparities, even if the data are not robust, I think you might as well say what you found in terms of the disparities. Also, I thought some of the differences were decreased with adjustment for confounders? This is a good place to reiterate that.This is discussed in the “clinical need” paragraph. The differences may decrease after adjustment, but they don't go away entirely.
Page 27----KQ3. If you are going to talk about the one published study it makes sense to me to summarize what it showedThis has been added.
Recommendations for Future Research
It would be appropriate to have more specific recommendations at the end of the review, arising from the evidence, or lack thereof. For instance: 1) areas with conflicting findings; 2) areas needed to be studied in Veterans, for which little information is available (e.g. women are mentioned, how about Hispanics); and 3) potential interventions that should be evaluated on the basis of the findings – patient-based or QI.We have revised the future research section
This is a very comprehensive and detailed review of the literature, and it would be very helpful if more specific recommendations could be drafted in summary of the review. There may also be recommendations that could be made with respect to implementation, but given the current state of the research, it seems that more evidence base is needed regarding interventions to address disparities, before these are put into clinical practice. I think it will help readers / stakeholders to get more out of the evidence synthesis if a more detailed “take home” message is provided with respect to what is still needed.We have revised the future research section
The report does not identify anything to implement. The call for more research seems appropriateWe have revised the future research section
I am not sure I agree with the recommendations for further research. it seems to me that if the evidence base is limited for first and second generation disparity studies that these should be conducted prior to suggesting more third generation research. it is not completely clear to me that there are disparities at the VA. I would like to see a Discussion section (it can be short) in this paper with some discussion of the problems with this evidence base. In particular, I am struck by how often point estimates of disparity were either reduced or eliminated by adjusting for confounders. I think this deserves more synthesis and discussion.We have revised the future research section
Appendix F
Appendix F. Number of articles is 57? The number does not match up with the numbers in Figure 3Numbers have been updated.


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (945K)

Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...