Miret et al. BMC Health Services Research (2019) 19:940 https://doi.org/10.1186/s12913-019-4789-3 RESEARCH ARTICLE Open Access Factors associated with readmissions in women participating in screening programs and treated for breast cancer: a retrospective cohort study Carme Miret1,2,3, Laia Domingo2,4* , Javier Louro2,4, Teresa Barata5, Marisa Baré3,4,6, Joana Ferrer7, Maria Carmen Carmona-García8,9,10, Xavier Castells2,3,4 and Maria Sala2,4 Abstract Background: We aimed to identify the risk factors associated with early, late and long-term readmissions in women diagnosed with breast cancer participating in screening programs. Methods: We performed a multicenter cohort study of 1055 women aged 50–69 years participating in Spanish screening programs, diagnosed with breast cancer between 2000 and 2009, and followed up to 2014. Readmission was defined as a hospital admission related to the disease and/or treatment complications, and was classified as early (< 30 days), late (30 days-1 year), or long-term readmission (> 1 year). We used logistic regression to estimate the adjusted odds ratios (aOR), and 95% confidence intervals (95% CI) to explore the factors associated with early, late and long-term readmissions, adjusting by women’s and tumor characteristics, detection mode, treatments received, and surgical and medical complications. Results: Among the women included, early readmission occurred in 76 (7.2%), late readmission in 87 (8.2%), long- term readmission in 71 (6.7%), and no readmission in 821 (77.8%). Surgical complications were associated with an increased risk of early readmissions (aOR = 3.62; 95%CI: 1.27–10.29), and medical complications with late readmissions (aOR = 8.72; 95%CI: 2.83–26.86) and long-term readmissions (aOR = 4.79; 95%CI: 1.41–16.31). Conclusion: Our results suggest that the presence of surgical or medical complications increases readmission risk, taking into account the detection mode and treatments received. Identifying early complications related to an increased risk of readmission could be useful to adapt the management of patients and reduce further readmissions. Trial Registration: ClinicalTrials.govIdentifier: NCT03165006. Registration date: May 22, 2017 (Retrospectively registered). Keywords: Breast cancer, Screening, Readmission, Complication * Correspondence: ldomingo@parcdesalutmar.cat 2 Department of Epidemiology and Evaluation, IMIM (Hospital del Mar Medical Research Institute), Passeig Marítim, 25-29, 08003 Barcelona, Spain 4 Research Network on Health Services in Chronic Diseases (REDISSEC), Av. de Monforte de Lemos, 5, 28029 Madrid, Spain Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Miret et al. BMC Health Services Research (2019) 19:940 Page 2 of 10 Background Materials and methods Breast cancer is the most frequently diagnosed malig- Setting and study population nant tumor and the main cause of cancer death in We conducted a retrospective cohort study in a sample women worldwide [1, 2]. However, over the last 20 years, of 1086 women with breast cancer who underwent there has been a reduction in breast cancer mortality breast cancer screening in 4 screening programs in Spain rates in Western countries, which has mainly been at- (Parc de Salut Mar, Girona, Sabadell and Canary tributed to the widespread use of breast cancer screening Islands), who were diagnosed with breast cancer be- practices, advances in adjuvant treatments, and improve- tween 2000 and 2009, and who were followed-up until ments in healthcare quality, especially through the intro- June 2014. This cohort encompasses information on tu- duction of functional units and multidisciplinary teams mors diagnosed by routine screening mammograms and in hospitals [3–5]. This mortality reduction implies an during the interval between 2 mammograms (CAMISS increasing number of cancer survivors who face the pos- Cohort; ClinicalTrials.gov NCT03165006). sibility of experiencing a breast cancer recurrence, an- Following the recommendations of the European other cancer diagnosis, or adverse effects of treatment, Guidelines for quality assurance in breast cancer screen- which translates into increased use of health services [6]. ing and diagnosis [22, 23], the Spanish breast cancer Women participating in breast cancer screening pro- screening program offers free biennial mammograms to grams are more likely to be diagnosed at earlier stages women aged 50–69 years. Two-view mammography is [7, 8], showing better survival [7, 9, 10] and requiring performed and a double reading is made using the BI- less aggressive treatment modalities [11, 12] than women RADS (Breast Imaging Reporting and Data System) clas- diagnosed symptomatically, and have higher percentages sification or equivalent [24]. Once a tumor has been of breast-conserving surgery [11]. Treatment approaches confirmed histologically, women enter the hospital cir- for patients with early-stage and locally advanced breast cuit for treatment and follow-up and are no longer in- cancer require a multidisciplinary treatment approach vited to participate in screening. that may include surgery, radiation therapy, and systemic Data were collected through a protocol approved by therapy. Whereas the impact of screening and early de- the clinical research ethics committee of Parc de Salut tection have been widely evaluated in terms of recur- Mar (Barcelona), and the rest of participant institutions. rences and mortality [5, 13], less attention has been paid No informed consent was required. to their impact on readmissions and treatment-related complications. Readmission Most studies on readmissions in women with breast For the purpose of our study, readmission was defined cancer have been conducted in the US context and as a hospital admission related to the disease and/or have focused on short-term readmissions after surgery complications of treatment that occurred at the same [14–18], without considering the broader therapeutic hospital where the patient was diagnosed and treated for approach employed in women with breast cancer. breast cancer. Only a couple of studies have evaluated readmissions Readmission time (in days) was calculated from the due to the complications of adjuvant treatment re- date of the surgical treatment and the date of the first ceived by patients with breast cancer [19, 20]. In readmission. For those women not undergoing surgery addition, to our knowledge, all previous studies have (n = 6), the date of the first treatment (which was focused on symptomatic women, and there are no chemotherapy in all patients) was used to calculate the studies evaluating readmissions in women participat- readmission date. We classified readmissions as: early ing in screening programs. Given the increasing num- readmissions (those occurring in the first 30 days after ber of women participating in breast cancer screening first treatment), late readmissions (between 30 days and programs, and therefore with cancers diagnosed at 1 year after first treatment), and long-term readmissions earlier stages [21], information on treatment compli- (> 1 year after first treatment). cations, readmissions and health services’ use in this We also collected the reason for readmission associ- subset of women may be useful to provide more ac- ated with the disease, categorized as: tumor re-excision, curate information to women and to better forecast scar dehiscence, drainage of an abscess, disease progres- the use of health services. sion, complications due to adjuvant treatment, and The aim of this study was to explore risk factors asso- others (including scar dislocation and abscess drainage). ciated with early and late readmissions in women diag- We excluded women with no available information nosed with breast cancer participating in screening (n = 2) on readmissions. We also excluded women re- programs, taking into account women’s and tumor char- admitted for plastic reconstruction (n = 29), since recon- acteristics, detection mode, the treatments received, and structions were performed immediately in some treatment-related complications. hospitals and were deferred in others. Finally, we Miret et al. BMC Health Services Research (2019) 19:940 Page 3 of 10 included 234 women with readmissions and 821 without epidermal growth receptor 2 (Her2): luminal A (ER + / readmissions (Fig. 1). Her2- or PR + / Her2-); luminal B (ER + / Her2 + or PR + / Her2 +); Her2 (ER- / PR- / Her2 +); and triple-negative Study variables (ER-, PR-, Her2-) [26]. Ki67 expression was not taken into Women’s characteristics were obtained from the screen- account to define phenotypes. ing program databases, hospital-based tumor registries, Treatment-related information was obtained from the population-based cancer registries, hospital administra- clinical records review. The surgical approach was catego- tive databases, and clinical records review. Age at diag- rized as: 1) conservative surgery without lymphadenec- nosis was calculated from the date of birth and the date tomy; 2) conservative surgery with lymphadenectomy; 3) of cancer diagnosis. The presence of comorbidities at radical surgery with or without lymphadenectomy; and 4) diagnosis was obtained from the clinical records review. no surgery and/or adjuvant treatment. Adjuvant treatment To measure the burden of disease, the Charlson Comor- combinations were categorized as: 1) chemotherapy, bidity Index (CCI) was used [25], collapsed into 2 radiotherapy, and hormonal therapy; 2) radiotherapy and categories: CCI = 0; CCI ≥1. hormone therapy; 3) radiotherapy and chemotherapy; 3) Information on the breast cancer detection mode was chemotherapy, radiotherapy, and hormone therapy; and 4) obtained from the databases of the screening programs. other treatments. Anti-Her 2 therapy could not be in- We differentiated between breast cancers detected by cluded in our analysis. Information on neoadjuvant and routine screening mammograms (i.e., screen-detected adjuvant treatments was also collected. cancers) and cancers detected between 2 screening mammograms (i.e., interval cancers). Follow-up information Tumor-related information was obtained from the Data on complications associated with treatment, recur- population-based cancer registries, the hospital-based can- rences, and vital status at the end of the follow-up were cer registry, and the clinical records review. Information obtained through the clinical records review. We catego- was collected on invasiveness (ductal carcinoma in situ, in- rized treatment-related complications as follows: general vasive cancer); focality (unifocal and multifocal and/or mul- complications (including pain and anxiety), surgical ticentric); tumor-node-metastasis (TNM) stage (stage 0- complications (including lymphedema, adhesions, skin IV); tumor grade (well differentiated, moderately differenti- infection, and soft tissue necrosis), and medical compli- ated, and poorly differentiated/undifferentiated); and tumor cations (including treatment toxicity, endometrial alter- phenotype. Tumors were classified into 4 phenotypes ac- ations, hypothyroidism, mycosis, vascular insufficiency, cording to the expression of hormone receptors [estrogen asthenia, palpitations, mastitis, and depression). Recur- receptor (ER) and progesterone receptor (RP)] and rences included local, regional, and distant metastasis. Fig. 1 Flowchart of the study Miret et al. BMC Health Services Research (2019) 19:940 Page 4 of 10 Vital status at the end of follow-up was classified as alive surgery without lymphadenectomy than other study or dead. groups (p < 0.001), and higher percentages of radical sur- gery (p < 0.001). Women with long-term readmissions Statistical analysis were more likely to have received neoadjuvant treat- Univariate and bivariate descriptive analyzes were per- ments (p < 0.001), whereas no statistically significant formed comparing women’s and tumoral characteristics, differences were found according to the adjuvant treat- complications, and the treatments received between ments received. However, chemotherapy, radiotherapy women with and without readmission. We differentiated and hormone therapy were more common in women between early, late and long-term readmissions through- with early and late readmissions (39.5 and 32.2%; re- out the study. Statistical significance was estimated using spectively), while radiotherapy and hormone therapy the chi-squared test. For those variables showing statisti- were more common in women not requiring readmis- cally significant differences, two-sided equality tests for sion (35.4%). column proportions were calculated to assess which The percentages of patient complications associated categories were statistically different. with treatment, recurrences and vital status according to To evaluate factors associated with early, late and long- type of readmission are shown in Table 3. Women with term readmissions, we fitted 3 logistic regression models readmissions showed higher percentages of general and adjusted by age, Charlson Index, detection mode, TNM medical complications than those without readmissions. stage, focality, tumor grade, tumor phenotype, surgical The most common medical complications associated and adjuvant treatment, general complications, surgical with late and long-term readmission were treatment tox- complications, medical complications, and screening pro- icity (N = 20 and N = 11, respectively) and fatigue (N = gram. Finally, additional analyses were performed, exclud- 11 and N = 16, respectively). In the case of early re- ing complications from the final model. admission was fatigue (N = 10). Data management and statistical analyses were per- Women with early readmissions showed a higher pro- formed using the IBM SPSS Statistics program, version portions of surgical complications than other study 25. Statistical significance was set at p < 0.05. groups, although this difference was not statistically sig- nificant (p = 0.076). In addition, recurrence and death Results were significantly more frequent in patients with early, We included 1055 women diagnosed with breast cancer, late, and long-term readmissions than in those without of whom 76 had an early readmission (7.2%), 87 had a readmissions. late readmission (8.2%), and 71 had long-term readmis- The reasons for readmission associated with disease sions (6.7%) (Table 1). across readmission subgroups are shown in Table 4. Women’s and tumor characteristics were compared Over 80% of early readmissions were due to tumor re- between women with early, late, and long-term readmis- excision and mastectomy/lymphadenectomy, whereas sions and those of women with no readmissions (Table 1). most long-term readmissions were related to disease Age tended to be younger (< 60 years) in women with progress (78.9%). early readmissions than in those with no readmissions Multivariate analyses to assess risk factors for early, (64.5% vs 53.1%; p = 0.013). The percentages of multifocal late and long-term readmissions, and adjustment by and moderately differentiated tumors was higher in women’s and tumor characteristics, treatment received, women with early readmissions than in those with no complications, and screening program, are shown in readmissions (p = 0.05). The percentage of luminal A tu- Table 5. Conservative surgery without lymphadenectomy mors was higher in women with early and late readmis- increased the risk of early readmission (aOR = 2.91; sions than in those women with no readmissions (58.0%, 95%CI = 1.13–7.52) compared with conservative surgery 64.8 vs 51.8%, respectively), while the percentage of Her2 with lymphadenectomy. However, there was no associ- tumors was higher in women with long-term readmissions ation in women who underwent radical surgery. Surgical than in those with no readmissions (23.9 vs 8.1, respect- complications were associated with an increased risk of ively; p = 0.017). Although not statistically significant, early readmission (aOR = 3.62; 95%CI = 1.27–10.29), but women with readmissions tended to present with cancer no association was found with women’s and tumor at advanced TNM stages, and specifically with higher per- characteristics. centages of lymph node involvement (data not shown). Medical complications were associated with a signifi- No significant differences were found according to comor- cantly increased risk of late readmission (Table 5) bidities at diagnosis or detection mode. (aOR = 8.72; 95%CI = 2.83–26.86), as well as with long- The treatments received according to type of readmis- term readmission (aOR = 4.79; 95%CI = 1.41–16.31). sion are summarized in Table 2. Women with long-term Additional analysis excluding complications showed readmissions showed lower percentages of conservative non-significantly different results to those presented in Miret et al. BMC Health Services Research (2019) 19:940 Page 5 of 10 Table 1 Women’s and tumoral characteristics according to readmission Readmitted, n (%) Not readmitted, p-value1 (n = 234) n (%) (n = 821) Early readmission Late readmission Long-term readmission Total 76 (7.2) 87 (8.2) 71 (6.7) 821 (77.8) Age (years) 50–54 24 (31.6) 22 (25.3) 32 (45.1)a 224 (27.3)a 0.013 55–59 25 (32.9) 27 (31.0) 16 (22.5) 212 (25.8) 60–64 21 (27.6) 19 (21.8) 17 (23.9) 225 (27.4) 65–69 6 (7.9) 19 (21.8) 6 (8.5) 160 (19.5) Missing 0 0 0 0 Charlson index 0 61 (80.3) 61 (70.1) 47 (66.2) 603 (73.4) 0.245 1 15 (19.7) 26 (29.9) 24 (33.8) 218 (26.6) Missing 0 0 0 0 Detection mode Screen-detected 50 (65.8) 62 (71.3) 42 (59.2) 568 (69.2) 0.306 Interval Cancer 26 (34.2) 25 (28.7) 29 (40.8) 253 (30.8) Missing 0 0 0 0 Tumor-node-metastasis stage In situ 6 (8.1) 8 (9.6) 5 (7.4) 84 (10.4) 0.069 a a,b b Stage I 36 (48.6) 38 (45.8) 18 (26.5) 368 (45.5) Stage II 25 (33.8) 26 (31.3) 28 (41.2) 249 (30.8) Stage III/IV 7 (9.5) 11 (13.3) 17 (25.0)a 108 (13.3)a Missing 2 4 3 9 Focality Unifocal 43 (70.5)a 54 (75.0) 49 (81.7) 594 (85.3)a 0.005 a a Multifocal and/or multicentric 18 (29.5) 18 (25.0) 11 (18.3) 102 (14.7) Missing 15 15 11 125 Tumor grade Well differentiated 13 (21.0) 18 (27.3)a 4 (7.0)a,b 193 (29.2)b 0.001 Moderately differentiated 33 (53.2) 34 (51.5) 27 (47.4) 248 (37.6) Poorly differentiated/ undifferentiated 16 (25.8) 14 (21.2)a 26 (45.6)a 219 (33.2) Not applicable 7 4 1 49 Missing 7 17 13 112 Tumor phenotype Luminal A 29 (58.0) 35 (64.8) 19 (41.3) 324 (51.8) 0.017 Luminal B 13 (26.0) 9 (16.7) 10 (21.7) 181 (28.9) HER2 3 (6.0) 3 (5.6) 11 (23.9)a 51 (8.1)a Triple-negative 5 (10.0) 7 (13.0) 6 (13.0) 70 (11.2) Missing 26 37 25 195 1 Chi-square test a,b Values in the same row and subtable sharing the same subscript were significantly different (p < .05) in the two-sided equality test for column proportions our main analysis. In particular, there were no differ- Discussion ences in the association between the surgical approach We aimed to identify the risk factors associated with and the risk of early, late and long-term readmissions early, late and long-term readmissions in women diag- (Additional file 1: Table A). nosed with breast cancer participating in screening Miret et al. BMC Health Services Research (2019) 19:940 Page 6 of 10 Table 2 Treatments provided according to readmission Readmitted, n (%) Not p-value1 readmitted, Early Late Long-term n (%) readmission readmission readmission (n = 821) (n = 76) (n = 87) (n = 71) Neoadjuvant therapy No 72 (94.7)a 80 (92.0)b 52 (73.2)a,b,c 731 (89.1)c < 0.001 a b a,b,c c Yes 4 (5.3) 7 (8.0) 19 (26.8) 89 (10.9) Missing 0 0 0 1 Surgical treatment Conservative2 surgery without lymphadenectomy 30 (40.5)a 30 (34.9)b 10 (14.5) a,b,c 267 (32.8)c 0.001 2 Conservative surgery with lymphadenectomy 27 (36.5) 40 (46.5) 31 (44.9) 386 (47.4) Radical3 surgery with or without lymphadenectomy 17 (23.0) 13 (15.1)a 25 (36.2) a,b 149 (18.3)b No surgery and/or adjuvant treatment 0 (0.0) 3 (3.5) 3 (4.3) 12 (1.5) Missing 2 1 2 7 Adjuvant treatment post-surgery Chemotherapy, radiotherapy and hormone therapy with/without anti- 30 (39.5) 28 (32.2) 20 (29.0) 229 (28.1) 0.159 Her2 therapy Radiotherapy and hormone therapy with/without anti-Her2 therapy 14 (18.4) 27 (31.0) 22 (31.9) 288 (35.4) Radiotherapy and chemotherapy with/without anti-Her2 therapy 9 (11.8) 5 (5.7) 9 (13.0) 71 (8.7) Other treatments 23 (30.3) 27 (31.0) 18 (26.1) 226 (27.8) Missing 0 0 2 7 1 Chi-square test 2 Included: quadrantectomy, tumorectomy and segmentectomy 3 Included: simple mastectomy, radical mastectomy, and modified radical mastectomy a,b,c Values in the same row and subtable sharing the same subscript were significantly different (p < .05) in the two-sided equality test for column proportions Table 3 Patient complications associated with treatment, recurrence and vital status according to readmission Readmitted, n (%) Not p-value1 (n = 234) readmitted, n (%) Early readmission Late readmission Long-term readmission (n = 821) (n = 76) (n = 87) (n = 71) Complications associated with treatment2 Any complication3 34 (44.7)a 44 (50.6)b 38 (53.5)c 182 (22.2)a,b,c < 0.001 4 a b c General complications 20 (26.3) 21 (24.1) 22 (31.0) 105 (12.8)a,b,c < 0.001 5 Surgical complications 18 (23.7) 14 (16.1) 13 (18.3) 110 (13.4) 0.076 Medical complications6 22 (28.9)a 29 (33.3)b 26 (36.6)c 80 (9.7)a,b,c < 0.001 Recurrence No recurrence 66 (86.8)a 71 (81.6)b 16 (22.5)c 743 (90.5)a,b,c < 0.001 a b c a,b,c Recurrence 10 (13.2) 16 (18.4) 55 (77.5) 78 (9.5) Vital status Alive 66 (86.8)a 75 (86.2)b 30 (42.3)c 728 (88.7)a,b,c < 0.001 a b c Dead 10 (13.2) 12 (13.8) 41 (57.7) 93 (11.3)a,b,c 1 Chi-square test 2 The categories are not exclusive as the same patient could have more than one complication 3 Number of women with at least one complication 4 Included: pain and anxiety 5 Included: lymphedema, adhesions, skin infection and soft tissue necrosis 6 Included: treatment toxicity, endometrial alterations, hypothyroidism, mycosis, vascular insufficiency, asthenia, palpitations, mastitis, and depression a,b,c Values of the same row and subtable that share the same subscript are significantly different in p < .05 in the two-sided equality test for column proportions Miret et al. BMC Health Services Research (2019) 19:940 Page 7 of 10 Table 4 Reason of readmission associated with the disease, according to time until readmission Early readmission, Late readmission, Long-term readmission, p-value1 n (%) n (%) n (%) (N = 76) (N = 87) (N = 71) Tumor re-excision 27 (35.5)a 26 (29.9)b 4 (5.6)a,b < 0.001 a b a,b Mastectomy and/or lymphadenectomy 36 (47.4) 34 (39.1) 3 (4.2) Due to medical treatment 5 (6.6) 15 (17.2) 4 (5.6) Due to disease progression 3 (3.9)a 8 (9.2)b 56 (78.9)a,b Others (including scar dislocation, drain abscess, bleeding) 5 (6.6) 4 (4.6) 4 (5.6) Plastic surgery reconstructions were excluded because they differ between hospitals, some reconstruct at the time until the surgical intervention 1 Chi-square test a,b Values in the same row and subtable sharing the same subscript were significantly different (p < .05) in the two-sided equality test for column proportions programs. Our findings suggest that women who experi- while those with medical complications had an increased enced surgical or medical complications after breast can- likelihood of late and long-term readmissions. cer treatment had an increased risk of readmission, In our study, descriptive data revealed that early re- taking into account the detection mode and the treat- admission occurred in 7.2% of women. Other studies ments received. Specifically, women with surgical com- calculating the percentages of early readmission after plications had an increased risk of early readmissions, breast cancer surgery have found lower percentages of Table 5 Bivariate and multivariate logistic regression analysis of factors associated with early, late and long-term readmissions Early readmission Late readmission Long-term readmission Odds Ratio (95% CI) Odds Ratio (95% CI) Odds Ratio (95% CI) Unadjusted Adjusted1 Unadjusted Adjusted1 Unadjusted Adjusted1 Detection mode Screen-detected Ref. Ref. Ref. Ref. Ref. Ref. Interval cancer 1.14 (0.70–1.86) 0.73 (0.32–1.66) 0.86 (0.53–1.40) 0.60 (0.23–1.53) 1.54 (0.94–2.53) 0.65 (0.22–1.88) Tumor-node-metastasis stage In situ 0.73 (0.30–1.78) 1.44 (0.14– 0.94 (0.42–2.07) 2.55 (0.24–27.53) 1.25 (0.45–3.46) 5.21 (0.31–88.88) 15.15) Stage I Ref. Ref. Ref. Ref. Ref. Ref. Stage II 0.97 (0.57–1.65) 0.65 (0.26–1.61) 0.96 (0.57–1.61) 1.93 (0.68–5.46) 2.29 (1.25–4.22) 0.93 (0.30–2.92) Stage III/IV 0.61 (0.26–1.39) 0.27 (0.06–1.30) 0.93 (0.46–1.86) 2.71 (0.55–13.42) 3.31 (1.66–6.62) 1.64 (0.33–8.10) Surgical treatment Conservative2 surgery without 1.65 (0.96–2.84) 2.91 (1.13–7.52) 1.09 (0.66–1.78) 1.26 (0.45–3.53) 0.45 (0.22–0.92) 0.46 (0.11–1.87) lymphadenectomy Conservative2 surgery with Ref. Ref. Ref. Ref. Ref. Ref. lymphadenectomy Radical3 surgery with or without 1.54 (0.82–2.89) 2.54 (0.84–7.71) 0.76 (0.40–1.45) 0.46 (0.12–1.76) 2.04 (1.17–3.55) 1.57 (0.46–5.30) lymphadenectomy Surgical complications4 No Ref. Ref. Ref. Ref. Ref. Ref. Yes 1.91 (1.09–3.34) 3.62 (1.27– 1.13 (0.62–2.05) 1.23 (0.36–4.16) 1.33 (0.71–2.49) 0.59 (0.18–1.97) 10.29) Medical complications5 No Ref. Ref. Ref. Ref. Ref. Ref. Yes 2.55 (1.50–4.32) 2.69 (0.84–8.60) 3.28 (2.02–5.32) 8.72 (2.83–26.86) 3.76 (2.24–6.31) 4.79 (1.41–16.31) 1 Model adjusted by age, Charlson Index, detection mode, tumor-node metastasis stage, focality, tumor grade, tumor phenotype, surgical and adjuvant treatment, general complications, surgical and medical complications, and screening program 2 Included: quadrantectomy, tumorectomy, and segmentectomy 3 Included: simple mastectomy, radical mastectomy, and modified radical mastectomy 4 Included: lymphedema, adhesions, skin infection, and soft tissue necrosis 5 Included: treatment toxicity, endometrial alterations, hypothyroidism, mycosis, vascular insufficiency, asthenia, palpitations, mastitis, and depression Miret et al. BMC Health Services Research (2019) 19:940 Page 8 of 10 around 5.5% [15, 17]. The reason for this variability in with, early readmission after curative chemotherapy for the percentage of readmissions may be that these studies breast cancer and reported that the factors associated with included women with broader age ranges and readmis- early readmission included tumor size (patients with T2 sions were analyzed after specific surgical procedures. A were more likely to be admitted), receiving adjuvant ther- study exploring the variation in 30-day readmissions apy versus neoadjuvant therapy, and undergoing fewer after a major surgical procedure, concluded that despite chemotherapy cycles [30]. It may be advisable to provide considerable variation in early readmission across surgi- more intensive follow-up to women with complications cal subspecialties, variation in early readmission was at- due to chemotherapy in order to avoid unplanned read- tributable to non-modifiable patient-level factors [16]. missions and improve the use of health services for this Although there is no single definition of readmission subset of women. [27], most studies on readmissions among women with In contrast with other studies, we found no association breast cancer have analyzed factors related to those occur- between comorbidities and readmissions [15–17]. This ring 30 days [15–17] or 1 year after surgical treatment lack of association may be due to some misinformation, [28]. Our study is consistent with prior series in showing given that data on comorbidities were manually obtained an association between surgical complications (mainly from the clinical records review. Regarding detection wound complications and surgical-site infections) and mode, we observed a non-significant trend for higher early readmissions [14, 16]. An analysis of risk factors re- rates of early and long-term readmissions in women with lated to readmission after immediate breast reconstruction interval cancers. However, this effect was attenuated in surgery found that patients with surgical complications the adjusted models, indicating that both women whose had a 4-fold increased risk of early readmission [15]. An- cancer was detected symptomatically or asymptomatically other study found that most early readmissions after mast- had the same chance of readmission. These results con- ectomy were related to postoperative complications, trast with evidence showing differences in tumor charac- rather than exacerbations of comorbidities [17]. Other teristics according to detection mode [31, 32]. To our studies have shown that additional risk factors for early re- knowledge, this is the first work assessing factors associ- admission were length of stay, payer type, physician vol- ated with readmissions among screened women. However, ume, and active smoking [17, 18]. to better assess the effect of breast cancer screening on In the current adjusted analysis, which included surgi- readmissions, it would be interesting to compare cohorts cal and medical complications, the risk of early readmis- of screened vs unscreened women. sion was higher in women receiving conservative surgery This study has some limitations. First, the number of without lymphadenectomy than in those receiving con- events in some categories of the analysis was relatively servative surgery with lymphadenectomy. In addition, small, hampering the identification of significant associa- most early readmissions were due to tumor re-excision, tions. Nevertheless, the number of cases included in the mastectomy or lymphadenectomy, which were probably study ensured sufficient statistical power to meet the related to previous conservative surgery which lately re- study objectives. Second, the manual and retrospective quired a more aggressive approach. Furthermore, when collection of some variables might have introduced an we excluded complications from the main analysis, the information bias, either due to some misinformation or risk of early readmission was not associated with the to variability in the quality of the information in the clin- surgical approach. Therefore, it seems that early re- ical records in the distinct hospitals. However, the clin- admission risk was not increased by the surgical ap- ical records review was done by trained professionals, proach but rather by the complications themselves. following a common protocol, and the final models were Late and long-term readmissions were more common adjusted by different screening programs to control the in women with medical complications than in the other variability produced by the inclusion of distinct pro- study groups, and in the adjusted analyses medical com- grams. Third, other relevant variables such as obesity, plications became a risk factor for late and long-term re- smoking status, breast reconstruction approach, or re- admission regardless of the treatments received. Most admission not related to breast cancer could not be ex- medical complications consisted of the adverse effects of plored because they were not collected in the CAMISS chemotherapy, such as fever, pain, nausea, vomiting, diar- cohort. However, we did include a number of variables rhea, myelosuppression (neutropenia and leukopenia), and related to tumor characteristics, treatments and compli- anemia [29, 30]. Only one study has evaluated readmis- cations, allowing us to provide an overview of all disease sions after neoadjuvant chemotherapy in patients with approach. Fourth, Ki67 expression was not included in breast cancer and, in the adjusted analyses, found that the definition of phenotypes, since this information was neoadjuvant chemotherapy was not statistically signifi- not available in patients diagnosed at the beginning of cantly associated with early readmissions [19]. Another the study period. Last, because treatment for Her2 tu- study examined the reasons for, and factors associated mors was introduced in 2006, the effect of this treatment Miret et al. BMC Health Services Research (2019) 19:940 Page 9 of 10 could not be examined throughout the analysis. How- Funding ever, women with Her2-positive tumors showed higher Data collection was supported by grants from MINECO-Instituto de Salud Carlos III-FEDER [grant numbers: PS09/01153 (M Sala), PI12/00387 (M Sala), percentages of long-term readmission, which were prob- PI11/01296 (X Castells) and PI15/00098 (X Castells)]. Data analysis was partly ably associated with recurrences in women who could supported by grants the Research Network on Health Services in Chronic not benefit from anti-Her2 treatment. Diseases (REDISSEC) (MINECO-Instituto de Salud Carlos III) [RD12/0001/0015 (M Sala); RD12/0001/0007 (M Baré); RD16/0001/0013 (M Sala)]. The funders had no role in study design, data collection, data analysis, data interpretation, Conclusion or writing of the manuscript. In conclusion, our results suggest that the presence of surgical and medical complications increases the risk of Availability of data and materials The datasets analysed during the current study are not publicly available due early and late readmissions, adjusted by detection mode to privacy regulations, but are available from the corresponding author on and treatments received. To our knowledge, this is the reasonable request. first study that analyzes the factors associated with read- missions among women participating in screening. This Ethics approval and consent to participate The study was approved by the clinical research ethics committee of Parc de information may be useful to improve the management Salut Mar, Barcelona. Consent to participate was not required due to the of the disease, especially among women with complica- retrospective nature of the study. The ethical committee of the Parc de Salut tions due to breast cancer treatment and predict health Mar approved this procedure. services use. Providing more intensive surveillance in women with treatment complications may help reduce Consent for publication Not applicable. further readmissions associated with the disease. Competing interests Supplementary information The authors have no competing interests to declare. Supplementary information accompanies this paper at https://doi.org/10. 1186/s12913-019-4789-3. Author details 1 Preventive Medicine and Public Health Training Unit PSMar-UPF-ASPB, Parc Additional file 1. Table A. Multivariate logistic regression analysis of de Salut Mar, Agència de Salut Pública de Barcelona, Pompeu Fabra factors associated with early, late and long-term readmissions. University, Barcelona, Spain. 2Department of Epidemiology and Evaluation, IMIM (Hospital del Mar Medical Research Institute), Passeig Marítim, 25-29, 08003 Barcelona, Spain. 3Department of Pediatrics, Obstetrics and Abbreviations Gynecology, Preventive Medicine and Public Health, Universitat Autònoma aOR: Adjusted odds ratio; BI-RADS: Breast Imaging Reporting and Data de Barcelona (UAB), 08193 Bellaterra, Barcelona, Spain. 4Research Network on System; CCI: Charlson Comorbidity Index; CI: Confidence intervals; Health Services in Chronic Diseases (REDISSEC), Av. de Monforte de Lemos, 5, ER: Estrogen receptor; Her2: Human epidermal growth factor receptor 2; 28029 Madrid, Spain. 5General Directorate of Health Care Programs, Canary OR: Odds ratio; PR: Progesterone receptor; TNM: Tumor-node-metastasis Islands Health Service, C/ Juan XXIII,13, 35005 Las Palmas de Gran Canaria, Spain. 6Cancer Screening and Clinical Epidemiology, Corporació Sanitària Acknowledgments Parc Taulí, 08208 Sabadell, Spain. 7Department of Radiology, Hospital de The authors acknowledge the dedication and support of the entire CAMISS Santa Caterina, C/ Dr. Castany, s/n, 17190 Salt, Girona, Spain. 8Epidemiology Study Group (alphabetical order): IMIM (Hospital del Mar Medical Research Unit and Girona Cancer Registry, Oncology Coordination Plan, Department of Institute), Barcelona: Xavier Castells, Mercè Comas, Laia Domingo, Francesc Health, Catalan Institute of Oncology, C/ Sol, 15, 17004 Girona, Spain. 9Girona Macià, Marta Roman, Anabel Romero, María Sala. Canary Islands Health Biomedical Research Institute (IDIBGI), C/ Dr Castany s/n, 17190 Salt, Girona, Service: Teresa Barata, Isabel Diez de la Lastra, Mariola de la Vega. Corporació Spain. 10Department of Medical Oncology, Catalan Institute of Oncology, Sanitària Parc Taulí, Sabadell: Marisa Baré, Núria Torà. Hospital Santa Caterina, University Hospital Dr Josep Trueta, Av. França, S/N, 17007 Girona, Spain. Girona: Joana Ferrer. Epidemiology Unit and Girona Cancer Registry: María Carmen Carmona-García. Hospital Galdakao-Usansolo, Vizcaya: Susana García, Received: 11 July 2019 Accepted: 28 November 2019 Maximina Martín, Miren Orive, María Amparo Valverde. Canary Islands Foundation for Health Research: Jeanette Pérez, Amado Rivero, Cristina Valcárcel Hospital Costa del Sol, University of Málaga: María del Carmen References Padilla, Maximino Redondo, Teresa Téllez, Irene Zarcos Hospital Universitario 1. Torre L, Bray F, Siegel RL, Ferlay J, Lortet-tieulent J, Jemal A. Global Cancer Donostia /Biodonostia: Cristina Churruca, Amaia Perales, Javier Recio, Irune statistics, 2012. CA Cancer J Clin. 2015;65(2):87–108. Ruiz, Cristina Sarasqueta, José María Urraca Instituto Oncológico de Guipúzcoa- 2. International Agency for Research on Cancer, World Health Organization. Onkologikoa: María Jesús Michelena Hospital Universitario Basurto: Julio Moreno GLOBOCAN 2018: cancer today. Breast cancer incidence in women, Hospital Universitario Cruces: Gaizka Mallabiabarrena, Patricia Cobos, Borja Otero Spain 2018. Hospital Universitario Txagorritxu: Javier Gorostiaga, Itsaso Troya. 3. Jørgensen KJ, Zahl P-H, Gøtzsche PC. Breast cancer mortality in organised mammography screening in Denmark: comparative study. BMJ. 2010;340(5_ Authors’ contributions Part_1):c1241. CM, LD, MS and XC designed the study, and CM and LD wrote the manuscript. 4. Paci E, EUROSCREENWorking Group. Summary of the evidence of breast JL and CM performed the statistical analyses and MS, LD and XC contributed to cancer service screening outcomes in Europe and first estimate of the the analyses and interpreted the data. TB, MB, JF and MCCG were involved in benefit and harm balance sheet. J Med Screen. 2012;19(SUPPL. 1):5–13. data acquisition and collaborated in drafting the manuscript and revising it 5. Independent UK Panel on Breast Cancer Screening, Marmot MG, Altman critically for important intellectual content. All authors read and approved the DG, Cameron DA, Dewar JA, Thompson SG, et al. The benefits and harms of final manuscript and have agreed both to be personally accountable for the breast cancer screening: an independent review. Br J Cancer. 2013;108: author’s own contributions and to ensure that questions related to the 2205–40. accuracy or integrity of any part of the work, even ones in which the author 6. Puigpinós-Riera R, Castillo Gómez A, Romero Morales A, Aller M, Castells X, was not personally involved, are appropriately investigated, resolved, and the Sala M. Social and clinical determinants of the use of health services in resolution documented in the literature. women with breast cancer (cohort DAMA). Gac Sanit. 2019;33(5):434–41. Miret et al. BMC Health Services Research (2019) 19:940 Page 10 of 10 7. Mook S, Van’t Veer LJ, Rutgers EJ, Ravdin PM, Van De Velde AO, Van 30. Pittman NM, Hopman WM, Mates M. Emergency room visits and hospital Leeuwen FE, et al. Independent prognostic value of screen detection in admission rates after curative chemotherapy for breast Cancer. J Oncol invasive breast cancer. J Natl Cancer Inst. 2011;103(7):585–97. Pract. 2015;11(2):120–5. 8. Choi KS, Yoon M, Song SH, Suh M, Park B, Jung KW, et al. Effect of 31. Domingo L, Salas D, Zubizarreta R, Baré M, Sarriugarte G, Barata T, et al. mammography screening on stage at breast cancer diagnosis: results from Tumor phenotype and breast density in distinct categories of interval the Korea National Cancer Screening Program. Sci Rep. 2018;8(1):1–8. cancer: results of population-based mammography screening in Spain. 9. Sankatsing VDV, van Ravesteyn NT, Heijnsdijk EAM, Looman CWN, van Luijt Breast Cancer Res. 2014;16(1):R3 Available from: http://breast-cancer- PA, Fracheboud J, et al. The effect of population-based mammography research.biomedcentral.com/articles/10.1186/bcr3595. screening in Dutch municipalities on breast cancer mortality: 20 years of 32. Baré M, Torà N, Salas D, Sentís M, Ferrer J, Ibáñez J, et al. Mammographic follow-up. Int J Cancer. 2017;141(4):671–7. and clinical characteristics of different phenotypes of screen-detected and 10. Morrell S, Taylor R, Roder D, Robson B, Gregory M, Craig K. Mammography interval breast cancers in a nationwide screening program. Breast Cancer service screening and breast cancer mortality in New Zealand: a National Res Treat. 2015;154(2):403–15. Cohort Study 1999-2011. Br J Cancer. 2017;116(6):828–39. 11. Tóth D, Varga Z, Tóth J, Árkosy P, Sebő É. Short- and long-term (10-year) results of an organized, population-based breast Cancer screening program: comparative, Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in observational study from Hungary. World J Surg. 2018;42(5):1396–402. published maps and institutional affiliations. 12. Hofvind S, Holen, Aas T, Roman M, Sebuødegård S, Akslen LA. Women treated with breast conserving surgery do better than those with mastectomy independent of detection mode, prognostic and predictive tumor characteristics. Eur J Surg Oncol. 2015;41(10):1417–22. 13. Løberg M, Lousdal ML, Bretthauer M, Kalager M. Benefits and harms of mammography screening. Breast Cancer Res. 2015;17(63):1–12. 14. Lai JK, Martin MA, Meyricke R, O’Neill T, Roberts S. Factors associated with short-term hospital readmission rates for breast Cancer patients in Western Australia: an observational study. J Am Coll Surg. 2007;204(2):193–200. 15. Nelson JA, Fischer JP, Chung C, Wu LC, Serletti JM, Kovach SJ. Risk of Readmission following Immediate breast reconstruction: results from the 2011 American College of Surgeons National Surgical Quality Improvement Program data sets. Plast Reconstr Surg. 2014;134(2):193e–201e. 16. Gani F, Lucas DJ, Kim Y, Schneider EB, Pawlik TM. Understanding variation in 30-day surgical readmission in the era of accountable care. JAMA Surg. 2015;150(11):1042. 17. Chow I, Hanwright PJ, Hansen NM, Leilabadi SN, Kim JYS. Predictors of 30- day readmission after mastectomy: a multi-institutional analysis of 21,271 patients. Breast Dis. 2015;35(4):221–31. 18. Sun SX, Leung AN, Dillon PW, Hollenbeak CS. Length of stay and readmissions in mastectomy patients. Breast J. 2015;21(5):526–32. 19. Landercasper J, Bennie B, Bray MS, Vang CA, Linebarger JH. Does neoadjuvant chemotherapy affect morbidity, mortality, reoperations, or readmissions in patients undergoing lumpectomy or mastectomy for breast cancer? Gland Surg. 2017;6(1):14–26. 20. Gibson S, McConigley R. Unplanned oncology admissions within 14 days of non-surgical discharge: a retrospective study. Support Care Cancer. 2016; 24(1):311–7. 21. International Agency for Research on Cancer. Cancer screening in The European Union (2017). Recommendation on cancer screening 2017. 22. Perry N, Broeders M, de Wolf C, Tornberg S, Holland R, von Karsa L. European guidelines for quality assurance in breast cancer screening and diagnosis. Fourth edition--summary document. Ann Oncol. 2006;19(4):614–22. 23. Amendoeira I, Anttila A, Bellocq J, Bianchi S, Bielska-Lasota M, Boecker W, et al. European guidelines for quality assurance in breast cancer screening and diagnosis (2013). Health & Consum Protec. European Communities/ EUREF, editor. Vol. 19, Health & Consum Protec. 2006. p. 1–432. 24. Ascunce N, Salas D, Zubizarreta R, Almazán R, Ibáñez J, Ederra M, et al. Cancer screening in Spain. Ann Oncol. 2010;21(SUPPL.3):43–51. 25. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83. 26. Goldhirsch A, Wood WC, Coates AS, Gelber RD, Thürlimann B, Senn H-J, et al. Strategies for subtypes--dealing with the diversity of breast cancer: highlights of the St. Gallen international expert consensus on the primary therapy of early breast Cancer 2011. Ann Oncol. 2011;22(8):1736–47. 27. Halfon P, Eggli Y, Prêtre-Rohrbach I, Meylan D, Marazzi A, Burnand B. Validation of the potentially avoidable hospital readmission rate as a routine indicator of the quality of hospital care. Med Care. 2006;44(11):972–81. 28. Hynes DM, Weaver F, Morrow M, Folk F, Winchester DJ, Mallard M, et al. Breast cancer surgery trends and outcomes: results from a National Department of veterans affairs study. J Am Coll Surg. 2004;198(5):707–16. 29. Hassett MJ, O’Malley AJ, Pakes JR, Newhouse JP, Earle CC. Frequency and cost of chemotherapy-related serious adverse effects in a population sample of women with breast cancer. J Natl Cancer Inst. 2006;98(16):1108–17.