The question to discuss is one about how the data was analyzed. There is not a "right" answer to this discussion - and we are really just looking for you to comment on what may (or may not) be issues with what was done.
The methods say that they used non-parametric approaches - however the authors did not clearly describe what they were.
Table 1 from the paper shows that there were 1 Asian participant in the Obese group, 2 Total Uninsured participants in the analyses (0 Uninsured participants in the Obese group), and 5 Total Military insured participants in the analyses (2/3 in the Non-Obese and Obese groups, respectively).
If we then examine Table 3 and 4 - and focus on the information provided for those 3 variables (Asian, Total Military, Uninsured) please discuss how we can interpret the confidence intervals shown (i.e., in Table 4 we see Military 4.42 (0.45-43.06)).
Also, comment on whether you might suggest any alternative analyses that could have been performed using these data.
Do Obese Patients Present With More Advanced Breast Cancer? 2021, Vol. 87(1) 56 –60 1Department of Surgery, Yale University School of Medicine, New Haven, CT, USA Corresponding Author: Anees B. Chagpar, MD, MSc, MPH, MA, MBA, Department of Surgery, Yale University School of Medicine, New Haven, CT 06510, USA. Email: anees. chagpar@ yale. edu The American Surgeon © The Author(s) 2020 Article reuse guidelines: sagepub. com/ journals- permissions DOI: 10. 1177/ 0003 1348 20949510 journals. sagepub. com/ home/ asu Do Obese Patients Present With More Advanced Breast Cancer? Melinda Wang, BA1, Julian Huang, AB1, and Anees B. Chagpar, MD, MSc, MPH, MA, MBA1 Abstract Background: Obesity is a known risk factor for breast cancer development; however, it is unclear whether obesity is associated with more aggressive disease. We sought to determine the relationship between obesity and tumor characteristics in breast cancer patients. Methods: Medical records of invasive breast cancer patients undergoing mastectomy at our institution between January 2010 and April 2018 were reviewed. Patients who had received neoadjuvant therapy were excluded. Patients were separated into obese (body mass index [BMI] ≥ 30 kg/m2) and nonobese (BMI < 30="" kg/m2)="" categories="" and="" com-="" pared="" using="" nonparametric="" statistical="" analyses.="" results:="" of="" the="" 415="" patients="" in="" this="" cohort,="" 124="" (29.9%)="" were="" obese.="" obese="" patients="" were="" more="" likely="" to="" present="" with="" larger="" tumors="" (tumor="" size="" ≥="" 2="" cm:="" 53.2%="" vs="" 41.0%,="" p=".024)" and="" more="" node-="" positive="" disease="" (46.8%="" vs="" 28.9%,="" p=".001)." controlling="" for="" confounders,="" obesity="" was="" independently="" associated="" with="" node-="" positive="" disease="" (odds="" ratio="" [or]="2.00;" 95%="" ci="" 1.24-3.21,="" p=".004)," but="" not="" with="" tumor="" size="" ≥="" 2="" cm="" (or="1.38;" 95%="" ci="" 0.87-2.21,="" p=".174)." conclusion:="" obesity="" is="" associated="" with="" node-="" positive="" disease="" at="" presentation="" independent="" of="" other="" factors.="" keywords="" obesity,="" breast="" article="" introduction="" according="" to="" the="" centers="" for="" disease="" control="" and="" prevention,="" obesity="" has="" been="" increasing="" in="" the="" united="" states;="" currently,="" over="" 40%="" of="" adults="" in="" this="" country="" are="" obese.1="" importantly,="" obesity="" is="" known="" to="" be="" associated="" with="" a="" num-="" ber="" of="" ailments,="" including="" breast="" cancer.2="" while="" there="" tends="" to="" be="" a="" positive="" correlation="" between="" breast="" cancer="" incidence="" and="" obesity,="" the="" effect="" of="" obesity="" on="" tumor="" characteristics="" has="" not="" been="" consistent.="" some="" studies="" have="" found="" that="" obese="" women="" present="" with="" larger,="" node-="" positive="" cancers,="" that="" are="" more="" often="" hormone-="" receptor="" positive="" compared="" with="" nonobese="" women3-6;="" other="" studies,="" however,="" have="" shown="" no="" association="" with="" these="" factors.7="" given="" this="" variability="" in="" the="" literature,="" we="" sought="" to="" determine="" the="" effect="" of="" obesity="" on="" tumor="" charac-="" teristics="" in="" breast="" cancer="" patients="" at="" our="" institution.="" methods="" medical="" records="" of="" invasive="" breast="" cancer="" patients="" under-="" going="" mastectomy="" at="" our="" institution="" between="" january="" 2010="" and="" december="" 2017="" were="" reviewed.="" neoadjuvant="" therapy="" may="" be="" more="" common="" in="" obese="" patients="" and="" clearly="" affects="" pathologic="" tumor="" and="" lymph="" node="" charac-="" teristics;="" however,="" it="" is="" less="" clear="" as="" to="" whether="" the="" impact="" of="" neoadjuvant="" therapy="" on="" these="" pathologic="" features="" is="" dif-="" ferent="" in="" obese="" and="" nonobese="" patients.="" as="" this="" would="" potentially="" distort="" the="" interpretation="" of="" the="" impact="" of="" obe-="" sity="" on="" tumor="" and="" lymph="" node="" characteristics,="" patients="" who="" underwent="" neoadjuvant="" therapy="" were="" excluded="" from="" this="" study.="" clinicopathologic="" data="" were="" extracted="" and="" patients="" were="" separated="" into="" obese="" (body="" mass="" index="" [bmi]="" ≥30="" kg/m2)="" and="" nonobese="" categories="" (bmi=""><30 kg/="" m2).="" nonparametric="" analyses="" were="" performed="" using="" spss="" version="" 24="" (ibm="" statistics;="" armonk,="" ny,="" usa).="" continuous="" data="" were="" presented="" using="" the="" median="" and="" range.="" for="" clinicopathologic="" variables="" that="" were="" found="" to="" http://crossmark.crossref.org/dialog/?doi="10.1177%2F0003134820949510&domain=pdf&date_stamp=2020-09-13" wang="" et="" al="" 57the="" american="" surgeon="" 00(0)2="" be="" significantly="" correlated="" with="" obesity="" on="" bivariate="" anal-="" ysis,="" multivariate="" analyses="" were="" then="" performed="" to="" evalu-="" ate="" whether="" the="" effect="" of="" obesity="" on="" these="" variables="" was="" independent="" of="" potential="" confounders.="" spearman’s="" cor-="" relation="" was="" performed="" for="" continuous="" variables.="" this="" study="" was="" approved="" by="" the="" human="" investigations="" committee="" at="" yale="" university.="" results="" of="" the="" 415="" patients="" in="" this="" cohort,="" 124="" (19.9%)="" were="" obese.="" the="" median="" bmi="" of="" nonobese="" versus="" obese="" patients="" was="" significantly="" different="" between="" the="" 2="" groups="" (24.30="" kg/m2="" [range="" 16.91="" kg/m2-29.86="" kg/m2)="" vs="" 34.01="" kg/m2="" [range="" 30.04="" kg/m2-65.11="" kg/m2],="" respectively,="" p="">< .001).="" sociodemographic="" factors="" for="" each="" group="" are="" shown="" in="" table="" 1.="" on="" bivariate="" analysis,="" obese="" patients="" were="" more="" likely="" to="" present="" with="" the="" node-="" positive="" disease="" compared="" with="" nonobese="" patients="" (p=".001;" table="" 2).="" controlling="" for="" potential="" confounders,="" including="" race,="" insurance="" status,="" tumor="" size,="" and="" triple-="" negative="" phenotype,="" obesity="" remained="" independently="" associated="" with="" node-="" positive="" disease="" (p=".004;" table="" 3).="" there="" was="" a="" positive="" correla-="" tion="" between="" bmi="" and="" the="" number="" of="" positive="" lymph="" nodes="" (spearman’s="" rho="0.217," p="">< .001="" figure="" 1).="" in="" terms="" of="" the="" actual="" number="" of="" positive="" lymph="" nodes="" among="" node-="" positive="" patients,="" however,="" there="" was="" no="" dif-="" ference="" between="" obese="" and="" nonobese="" patients="" (median="" number="" of="" positive="" nodes="" 2.00="" [range="" 1-31]="" vs="" 1.00="" [range="" 1-25],="" respectively,="" p=".136)." obese="" patients="" were="" also="" more="" likely="" to="" present="" with="" tumors="" ≥="" 2="" cm="" compared="" with="" nonobese="" patients="" (p="">< .001,="" table="" 2).="" on="" multivariate="" analysis,="" however,="" obesity="" no="" lon-="" ger="" was="" associated="" with="" tumor="" size="" (p=".174," table="" 4).="" compared="" with="" nonobese="" patients,="" obese="" patients="" also="" trended="" toward="" presenting="" with="" more="" triple-="" negative="" can-="" cers="" (p=".066," table="" 2).="" however,="" on="" multivariate="" analy-="" sis,="" we="" found="" no="" independent="" predictor="" of="" triple-="" negative="" disease="" (table="" 5).="" conclusion="" obesity,="" in="" our="" study,="" is="" associated="" with="" node-="" positive="" dis-="" ease.="" we="" found,="" on="" bivariate="" analysis,="" that="" obese="" patients="" were="" more="" likely="" to="" present="" with="" the="" node-="" positive="" disease,="" and="" this="" association="" persisted="" on="" multivariate="" analysis="" con-="" trolling="" for="" potential="" confounding="" variables.="" this="" finding="" is="" consistent="" with="" other="" studies.4,5,8="" for="" example,="" chung="" et="" al="" found="" that="" among="" 8742="" patients="" treated="" surgically="" for="" stage="" i="" to="" iii="" breast="" cancer,="" rates="" of="" node="" positivity="" increased="" with="" obesity="" (34.7%="" vs="" 39.8%="" vs="" 44.1%="" vs="" 42.5%="" for="" under-="" weight,="" normal="" weight,="" overweight,="" and="" obese="" breast="" can-="" cer="" patients,="" respectively,="" p="">< .001).8="" haakinson="" et="" al="" table="" 1.="" demographic="" features="" of="" obese="" versus="" nonobese="" invasive="" breast="" cancer="" patients.="" factor="" nonobese="" (n="291)" obese="" (n="124)" p="" value="" age,="" years,="" median="" (range)="" 53.00="" (24-93)="" 54.00="" (28-90)="" .519="" race,="" n="" (%)="" .001="" white="" 243="" (83.5)="" 97="" (78.2)="" black="" 14="" (4.8)="" 20="" (16.1)="" asian="" 15="" (5.2)="" 1="" (0.8)="" other="" 19="" (6.5)="" 6="" (4.8)="" hispanic,="" n="" (%)="" 23="" (7.9)="" 8="" (6.5)="" .687="" insurance="" type,="" n="" (%)="" .001="" uninsured="" 2="" (0.7)="" 0="" (0.0)="" medicaid="" 22="" (7.6)="" 26="" (21.0)="" medicare="" 47="" (16.2)="" 18="" (14.5)="" military="" 2="" (0.7)="" 3="" (2.4)="" private="" 218="" (74.9)="" 77="" (62.1)="" ever="" smoker,="" n="" (%)="" 107="" (36.8)="" 56="" (45.2)="" .124="" hypertension,="" n="" (%)="" 48="" (16.6)="" 42="" (10.2)=""><.001 diabetes="" mellitus,="" n="" (%)="" 12="" (4.1)="" 19="" (15.3)=""><.001 table 2. tumor characteristics of breast cancer in obese versus nonobese patients.a factor nonobese (n = 291) obese (n = 124) p value positive node, n (%) 84 (28.9) 58 (46.8) .001 invasive tumor size ≥ 2 cm, n (%) 119 (41.0) 66 (53.2) .024 triple negative, n (%) 17 (5.9) 14 (11.3) .066 invasive tumor grade, n (%)b .959 grade 1 54 (18.7) 24 (5.8) grade 2 173 (59.9) 74 (18.0) grade 3 62 (21.5) 25 (20.3) ajcc stage (7th edition), n (%) .004 stage i 159 (54.6) 45 (36.3) stage ii 96 (23.1) 53 (42.7) stage iii 35 (12.0) 26 (21.0) stage iv 1 (0.3) 0 (0.0) multifocality/multicentricity 90 (31.1) 44 (35.8) .361 presence of dcis, n (%) 220 (75.6) 90 (72.6) .539 er+, n (%) 252 (86.6) 106 (85.5) .757 pr+, n (%) 221 (76.2) 97 (78.2) .704 her2+, n (%) 37 (13.0) 8 (6.6) .06 brca+, n (%) .394 positive 13 (4.5) 6 (4.8) negative 187 (64.3) 71 (57.3) not tested 91 (31.3) 47 (37.9) abbreviations: ajcc, american joint committee on cancer; dcis, ductal carcinoma in situ; er, estrogen receptor; pr, progesterone receptor. abased on final pathology. bpatients had unknown invasive tumor grade. 58 the american surgeon 87(1)wang et al 3 similarly found that among 1352 patients with invasive breast cancer, obese patients were more likely to present with node- positive disease than nonobese patients (31% vs 25%, respectively, p = .026).5 while they found the rates of neoadjuvant chemotherapy were the same in both obese and nonobese patients (3%), the potential interaction between neoadjuvant therapy and obesity on lymph node status could make the interpretation of these data more dif- ficult. our study, however, excluded patients undergoing neoadjuvant chemotherapy, and still found a correlation between obesity and node- positive status. while we found that the total number of lymph nodes positively correlated with increasing bmi in our entire cohort, when looking only at node- positive patients, there did not seem to be a relationship between obesity and the number of positive nodes. this suggests that the effect of obesity is primarily on lymph node status (positive vs negative), rather than on the number of positive nodes. loi et al, in a study of 1360 women, similarly found a correlation between obesity and a higher number of posi- tive nodes across their entire cohort (44% vs 56% node negative, 31% vs 27% for 1-3 positive nodes and 21% vs 14% for 4 or more positive nodes in obese vs nonobese patients, respectively, p = .02). it is difficult to discern from their data, however, whether this significance was driven by the difference between node- positive and node- negative groups, or whether there truly table="" 2.="" tumor="" characteristics="" of="" breast="" cancer="" in="" obese="" versus="" nonobese="" patients.a="" factor="" nonobese="" (n="291)" obese="" (n="124)" p="" value="" positive="" node,="" n="" (%)="" 84="" (28.9)="" 58="" (46.8)="" .001="" invasive="" tumor="" size="" ≥="" 2="" cm,="" n="" (%)="" 119="" (41.0)="" 66="" (53.2)="" .024="" triple="" negative,="" n="" (%)="" 17="" (5.9)="" 14="" (11.3)="" .066="" invasive="" tumor="" grade,="" n="" (%)b="" .959="" grade="" 1="" 54="" (18.7)="" 24="" (5.8)="" grade="" 2="" 173="" (59.9)="" 74="" (18.0)="" grade="" 3="" 62="" (21.5)="" 25="" (20.3)="" ajcc="" stage="" (7th="" edition),="" n="" (%)="" .004="" stage="" i="" 159="" (54.6)="" 45="" (36.3)="" stage="" ii="" 96="" (23.1)="" 53="" (42.7)="" stage="" iii="" 35="" (12.0)="" 26="" (21.0)="" stage="" iv="" 1="" (0.3)="" 0="" (0.0)="" multifocality/multicentricity="" 90="" (31.1)="" 44="" (35.8)="" .361="" presence="" of="" dcis,="" n="" (%)="" 220="" (75.6)="" 90="" (72.6)="" .539="" er+,="" n="" (%)="" 252="" (86.6)="" 106="" (85.5)="" .757="" pr+,="" n="" (%)="" 221="" (76.2)="" 97="" (78.2)="" .704="" her2+,="" n="" (%)="" 37="" (13.0)="" 8="" (6.6)="" .06="" brca+,="" n="" (%)="" .394="" positive="" 13="" (4.5)="" 6="" (4.8)="" negative="" 187="" (64.3)="" 71="" (57.3)="" not="" tested="" 91="" (31.3)="" 47="" (37.9)="" abbreviations:="" ajcc,="" american="" joint="" committee="" on="" cancer;="" dcis,="" ductal="" carcinoma="" in="" situ;="" er,="" estrogen="" receptor;="" pr,="" progesterone="" receptor.="" abased="" on="" final="" pathology.="" bpatients="" had="" unknown="" invasive="" tumor="" grade.="" 58="" the="" american="" surgeon="" 87(1)wang="" et="" al="" 3="" similarly="" found="" that="" among="" 1352="" patients="" with="" invasive="" breast="" cancer,="" obese="" patients="" were="" more="" likely="" to="" present="" with="" node-="" positive="" disease="" than="" nonobese="" patients="" (31%="" vs="" 25%,="" respectively,="" p=".026).5" while="" they="" found="" the="" rates="" of="" neoadjuvant="" chemotherapy="" were="" the="" same="" in="" both="" obese="" and="" nonobese="" patients="" (3%),="" the="" potential="" interaction="" between="" neoadjuvant="" therapy="" and="" obesity="" on="" lymph="" node="" status="" could="" make="" the="" interpretation="" of="" these="" data="" more="" dif-="" ficult.="" our="" study,="" however,="" excluded="" patients="" undergoing="" neoadjuvant="" chemotherapy,="" and="" still="" found="" a="" correlation="" between="" obesity="" and="" node-="" positive="" status.="" while="" we="" found="" that="" the="" total="" number="" of="" lymph="" nodes="" positively="" correlated="" with="" increasing="" bmi="" in="" our="" entire="" cohort,="" when="" looking="" only="" at="" node-="" positive="" patients,="" there="" did="" not="" seem="" to="" be="" a="" relationship="" between="" obesity="" and="" the="" number="" of="" positive="" nodes.="" this="" suggests="" that="" the="" effect="" of="" obesity="" is="" primarily="" on="" lymph="" node="" status="" (positive="" vs="" negative),="" rather="" than="" on="" the="" number="" of="" positive="" nodes.="" loi="" et="" al,="" in="" a="" study="" of="" 1360="" women,="" similarly="" found="" a="" correlation="" between="" obesity="" and="" a="" higher="" number="" of="" posi-="" tive="" nodes="" across="" their="" entire="" cohort="" (44%="" vs="" 56%="" node="" negative,="" 31%="" vs="" 27%="" for="" 1-3="" positive="" nodes="" and="" 21%="" vs="" 14%="" for="" 4="" or="" more="" positive="" nodes="" in="" obese="" vs="" nonobese="" patients,="" respectively,="" p=".02)." it="" is="" difficult="" to="" discern="" from="" their="" data,="" however,="" whether="" this="" significance="" was="" driven="" by="" the="" difference="" between="" node-="" positive="" and="" node-="" negative="" groups,="" or="" whether="" there="">