We thank the reviewers for their insightful comments. We provide a

advertisement
We thank the reviewers for their insightful comments. We provide a point-by-point
response. We believe that their comments have helped us improve the quality of this
manuscript.
Title:Increasing obesity and comorbidity in patients undergoing primary total hip
arthroplasty in the U.S.: A 13-year study of time trends Version:2Date:2 September
2014
Reviewer:Michael A. Mont
Reviewer's report:
I think that this is an interesting report that can be suitable for publication in BMC
Musculoskeletal Disorders. I only have a few comments, but I would vote for
acceptance.
I think the conclusions in the Abstract should reflect somewhat that this is still data that
was based on the Mayo Clinic Registry and the authors should make note of the fact that
these results occur throughout the country and at other centers, then their conclusions
are solid. They do mention on Page 12 the similarity of some of their findings to the NIS
sampling and perhaps this could be mentioned in the Abstract. Otherwise, I think this is
an excellent report that can be quite useful.
Response: We have made this change suggested by the reviewer in the abstract. “Our
cohort characteristics are similar to previously described characteristics of national U.S.
cohort, suggesting that these trends may be national rather than local trends. This is
important information for policy makers to take into account for resource allocation.”
Level of interest:An exceptional article
Quality of written English:Acceptable
Statistical review:No, the manuscript does not need to be seen by a statistician.
Declaration of competing interests:
I declare that I have no competing interests
Reviewer's report
Title:Increasing obesity and comorbidity in patients undergoing primary total hip
arthroplasty in the U.S.: A 13-year study of time trends Version:2Date:21 November
2014
Reviewer:Jamie E Collins
Reviewer's report:
This is an interesting paper that addresses an important issue in orthopedic care – How
are demographic and clinical variables changing for THA recipients over time? There are
some inconsistencies in the paper, which I’ve outlined below, along with some
questions/clarifications and some minor points to address.
Minor Essential Revisions
1. The methods (paragraph 1) state that “Age was categorized as previously (#60, >6070, >70-80 and >80 years)” but in the results age is stratified into <50, 50-<65, 65-<80,
>=80.
Response: We regret this typographical error and have corrected is as suggested by the
reviewer.
“Age was categorized as <50, 50-<65, 65-<80, ≥80, to allow comparison to other studies
using Medicare data that only include patients 65 years and older.”
2. The introduction (paragraph 3) states that “One of the most remarkable changes in
the epidemiology of primary THA in the last two decades is the expanding indication of
THA to both younger and older patients.” In this context, is change in the mean age over
time (Results, Time Trends in socio-demographic and clinical characteristics and
comorbidity) meaningful?
Response: We agree and have now noted this in the discussion.
“Our study confirms several previous findings including a reduction in mean age and
increase in Deyo-Charlson index for patients undergoing primary THA.”
3. The abstract states that “Compared to 1993-95, significantly more patients (by >2times for most) in 2002-05 had: BMI#40, 2.3% vs. 6.3%;” but it does not appear that this
was ever explicitly tested. Table 2 reports on the overall association between BMI group
and study period (p<0.001) and Table 4 reports on the percent with BMI>=40 within
each age subgroup group across the study periods, but I cannot find a formal test of
percent with BMI>=40 over time period across all age groups.
Response: We used chi-square test to compare these, we have added to methods and
results.
“We calculated summary statistics as proportion or mean (standard deviation).
We used chi-square tests to compare the proportions of patients with each category of
patient characteristic (BMI, underlying diagnosis, Deyo-Charlson index, depression and
anxiety) over time, and also comparisons of categorical variables in 1993-95 vs. 200508. We used the analysis of variance to compare means (age, BMI, Deyo-Charlson
index) in various time periods.
We also examined the time trends in the proportion of patients with extreme
obesity (BMI ≥40; obesity class III) and high comorbidity (Deyo-Charlson index ≥3). We
examined the time trends in BMI ≥40, RA/inflammatory arthritis as underlying diagnosis,
Deyo-Charlson index of ≥3, depression and anxiety, within each age group, using
multivariable-adjusted logistic regression analyses. Time-period was considered in four
category ordinal variable (1993-1995, reference category), and the BMI <40, other
diagnoses, Deyo-Charlson index of ≤2 and absence of depression or anxiety serving as
the reference category for each respective outcome. We performed statistical analyses
using Statistical Package for Social Sciences (SPSS, version 21, Chicago, IL). A pvalue <0.05 was considered significant.”
Results:
“Compared to 1993-95, significantly more patients (by >2-times for most) in 2002-05
had: BMI≥40, 2.3% vs. 6.3%; depression, 4.1% vs. 9.8%; and anxiety, 3.4% vs. 5.7%;
and significantly fewer had an underlying diagnosis of rheumatoid/inflammatory arthritis,
4.2% vs. 1.5% (p≤0.01 for all comparisons using chi-square tests).”
4. Introduction, paragraph 2. It looks like this sentence is missing an “a”: A recent study
in a TKA cohort found that rates of extreme obesity...
Response: We have corrected this.
“A recent study in TKA cohort found that rates of extreme obesity, Deyo-Charlson index
score >=3, depression and anxiety increased 2-3 fold from 1993-2005 (26).”
5. Methods, Study cohort selection. Tense of “ensure”: “Since 1993, trained, dedicated
Total Joint Registry staff has captured and entered these data into electronic database
and ensured the completeness of data.” This sentence is awkward and the authors may
wish to re-phrase.
Response: We regret this typographical error and have corrected is as suggested by the
reviewer.
“…and ensured the completeness of data.”
6. Discussion, paragraph 2: “To our knowledge, no systematic changes occurred in
referral patterns in the U.S. during the study period that would explain the time-trends in
these characteristics.” Are the authors referring to the increasing prevalence of obesity
among THA patients? Could this simply be explained by the growing obesity epidemic in
the United States?
Response: We were referring to referral patterns related to THA from primary care to
orthopedic surgery and have clarified is as suggested by the reviewer.
“To our knowledge, no systematic changes occurred in referral patterns from primary
care physicians to orthopedic surgery or policy for reimbursement for primary THA
(except slight uniform reduction in compensation across the health care system) in the
U.S. during the study period that would explain the time-trends in these characteristics.”
Major Compulsory Revisions
1. Please clarify the statistical methods.
a. The abstract (methods) states that multivariable logistic regression was used, but this
is not mentioned in the methods. The methods section (Statistical Analysis paragraph)
state that ANOVA was used. Logistic regression results are presented in the results
section and in Table 5. Do these include time period as an ordinal (test for trend) or
nominal variable?
Response: We have clarified this now. ANOVA was used for comparison of crude
means and chi-squared for categorical variables. Logistic regression treated time-period
as an nominal variable.
“We calculated summary statistics as proportion or mean (standard deviation).
We used chi-square tests to compare the proportions of patients with each category of
patient characteristic (BMI, underlying diagnosis, Deyo-Charlson index, depression and
anxiety) over time, and also comparisons of categorical variables in 1993-95 vs. 200508. We used the analysis of variance to compare means (age, BMI, Deyo-Charlson
index) in various time periods.
We also examined the time trends in the proportion of patients with extreme
obesity (BMI ≥40; obesity class III) and high comorbidity (Deyo-Charlson index ≥3). We
examined the time trends in BMI ≥40, RA/inflammatory arthritis as underlying diagnosis,
Deyo-Charlson index of ≥3, depression and anxiety, within each age group, using
multivariable-adjusted logistic regression analyses. Time-period was considered in four
category nominal variable (1993-1995, reference category), and the BMI <40, other
diagnoses, Deyo-Charlson index of ≤2 and absence of depression or anxiety serving as
the reference category for each respective outcome. The overall p-value for the variable
was considered as an indicator of whether there was a period effect; visual inspection of
the odds ratios was interpreted for time-trend. We performed statistical analyses using
Statistical Package for Social Sciences (SPSS, version 21, Chicago, IL). A p-value
<0.05 was considered significant.”
b. Tables 2, 3, 4 – are these p-values from ANOVA, with time period as a categorical
independent variable? Or did you test for a linear trend for time? The methods
(Statistical Analysis) state that time trends were examined but does not clarify which
method was used or which p-values are presented.
Response: For tables 2 and 3, p-values are from ANOVA for continuous and chi-sqaured
tests for categorical variables. We have added this to methods and the table legends.
We clarify the source of p-value for each analysis, please see the detailed response
above to #1a.
2. The discussion (paragraph 10) references sensitivity analyses (“robustness of our
findings across several sensitivity models.”). These do not appear to be presented in the
paper.
Response: We have clarified this now, this refers to the multiple multivariable-adjusted
logistic regression models.
“…robustness of our findings across several sensitivity models (age- vs. age- and sexvs. age, sex- and comorbidity-adjusted models).”
3. The authors should add to the limitations that anxiety and depression were assessed
based on ICD-9 diagnostic codes and are therefore likely under-reported. This is
especially important in evaluating time trends – are clinicians more likely to diagnose
anxiety and depression in later study years compared to earlier years (e.g., 2005 vs.
1993)? Meaning, are we actually seeing an increase in anxiety and depression over
time, or just better reporting?
Response: We appreciate this comment and have added this to the discussion.
“We used ICD-9 diagnostic codes to define anxiety and depression, and therefore these
are likely under-reported. In addition, the increasing incidence of anxiety and depression
over time might indicate either a better capture/reporting of these or a true increase in its
incidence in this patient population. We are unable to distinguish between these two
possibilities.”
Level of interest:An article of importance in its field
Quality of written English:Needs some language corrections before being
published
Statistical review:Yes, and I have assessed the statistics in my report. Declaration of
competing interests:
I declare that I have no competing interests
Reviewer's report
Title:Increasing obesity and comorbidity in patients undergoing primary total hip
arthroplasty in the U.S.: A 13-year study of time trends Version:2Date:28 November
2014
Reviewer:Ali Mobasheri
Reviewer's report:
The aim of this study was to examine trends and timelines in key demographic and
clinical characteristics of patients undergoing primary total hip arthroplasty (THA) in the
United States. The authors used data from the Mayo Clinic Total Joint Registry from
1993-2005 to examine the time-trends in demographics (age, body mass index (BMI)),
medical (Deyo-Charlson index) and psychological comorbidity (anxiety, depression) and
underlying diagnosis of patients undergoing primary THA. This is the first time that such
has study has been carried out using this valuable database. The data analysis
performed suggests that obesity, medical/psychological comorbidities and underlying
diagnosis changed rapidly in primary THA patients over 13-years. The authors propose
that studies of THA outcomes should take these rapidly changing patient characteristics
into account. This is an article of outstanding merit and interest in its field. The data
presented should be of great interest to healthcare providers and funders of arthritis
research, especially considering influencing factors and comorbidities.
Response: We thank the reviewer.
Level of interest:An article of outstanding merit and interest in its field Quality of
written English:Acceptable
Statistical review:Yes, but I do not feel adequately qualified to assess the statistics.
Declaration of competing interests:
I do not have any competing interests.
Download