Breast Cancer Research and Treatment (2020) 181:499–518 https://doi.org/10.1007/s10549-020-05657-8 REVIEW Healthcare providers’ adherence to breast cancer guidelines in Europe: a systematic literature review Ena Niño de Guzmán1 · Yang Song1 · Pablo Alonso‑Coello1,2 · Carlos Canelo‑Aybar1,2 · Luciana Neamtiu3 · Elena Parmelli3 · Javier Pérez‑Bracchiglione4 · Montserrat Rabassa1 · David Rigau1 · Zuleika Saz Parkinson3 · Iván Solà1 · Adrián Vásquez‑Mejía5 · Ignacio Ricci‑Cabello2,6,7 Received: 25 February 2020 / Accepted: 27 April 2020 / Published online: 6 May 2020 © The Author(s) 2020 Abstract Purpose Clinical guidelines’ (CGs) adherence supports high-quality care. However, healthcare providers do not always comply with CGs recommendations. This systematic literature review aims to assess the extent of healthcare providers’ adherence to breast cancer CGs in Europe and to identify the factors that impact on healthcare providers’ adherence. Methods We searched for systematic reviews and quantitative or qualitative primary studies in MEDLINE and Embase up to May 2019. The eligibility assessment, data extraction, and risk of bias assessment were conducted by one author and cross-checked by a second author. We conducted a narrative synthesis attending to the modality of the healthcare process, methods to measure adherence, the scope of the CGs, and population characteristics. Results Out of 8137 references, we included 41 primary studies conducted in eight European countries. Most followed a retrospective cohort design (19/41; 46%) and were at low or moderate risk of bias. Adherence for overall breast cancer care process (from diagnosis to follow-up) ranged from 54 to 69%; for overall treatment process [including surgery, chemo- therapy (CT), endocrine therapy (ET), and radiotherapy (RT)] the median adherence was 57.5% (interquartile range (IQR) 38.8–67.3%), while for systemic therapy (CT and ET) it was 76% (IQR 68–77%). The median adherence for the processes assessed individually was higher, ranging from 74% (IQR 10–80%), for the follow-up, to 90% (IQR 87–92.5%) for ET. Inter- nal factors that potentially impact on healthcare providers’ adherence were their perceptions, preferences, lack of knowledge, or intentional decisions. Conclusions A substantial proportion of breast cancer patients are not receiving CGs-recommended care. Healthcare pro- viders’ adherence to breast cancer CGs in Europe has room for improvement in almost all care processes. CGs development and implementation processes should address the main factors that influence healthcare providers’ adherence, especially patient-related ones. Registration: PROSPERO (CRD42018092884). Keywords Breast neoplasms · Guidelines as topic · Evidence-based medicine · Guideline adherence · Systematic review Abbreviations CGs Clinical guidelines RT Radiotherapy Ena Niño de Guzmán and Yang Song have contributed equally to CT Chemotherapy work and should be regarded as co-first authors. ET Endocrine therapy or hormonal therapy Electronic supplementary material The online version of this BCS Breast-conserving surgery article (https​://doi.org/10.1007/s1054​9-020-05657​-8) contains MT Mastectomy supplementary material, which is available to authorised users. ALND Axillary lymph node dissection TNBC Triple-negative breast cancer * Ena Niño de Guzmán enino@santpau.cat SLNB Sentinel lymph node biopsy HER2 Human epidermal growth receptor 2 * Luciana Neamtiu luciana.neamtiu@ec.europa.eu Extended author information available on the last page of the article 13 Vol.:(0123456789) 500 Breast Cancer Research and Treatment (2020) 181:499–518 Background Information sources and search strategy Clinical guidelines (CGs) are defined as “systematically We designed the search strategy and conducted the elec- developed statements to assist healthcare providers and tronic searches, for both systematic reviews and primary patients’ decisions about appropriate health care for spe- studies, in MEDLINE (accessed through PubMed) and cific clinical circumstances” [1]. These recommendations Embase (accessed through Ovid), from inception to May are intended to optimise patient care, reduce inappropriate 2019. The strategy searched for terms related to adherence, practice variation, enhance the transition of research into clinical guidelines, and breast cancer. Search strategies practice, and improve healthcare quality and safety [2]. are outlined in Additional file 2. Reference sections from Despite the availability of CGs with different presentation retrieved systematic reviews were used to identify additional formats, a constant production or updating process, their primary studies. uptake, adherence, or compliance by healthcare providers is variable [3], and sometimes reported as suboptimal [4–6]. Eligibility criteria and selection of studies For example, it has been estimated that only 50% of patients in the United States receive CGs-compliant healthcare [4]. We sought for primary studies conducted in European coun- Shared decision-making may be influenced by the com- tries. The addressed population was healthcare providers plexity of cancer care (e.g. tumour related features), patient’s of breast cancer care (including primary care, oncology, characteristics, and limitations in the evidence base [5]. Bar- radiotherapy, or others). Selected studies should measure riers to healthcare providers’ adherence to CGs could be the adherence of healthcare providers’ indications to breast personal barriers, as the healthcare provider’s knowledge cancer CGs, with any method (quantitative or qualitative) (lack of awareness or familiarity with CGs) and the pro- or source of data (self-reported, medical records’ assess- vider’s attitude towards change in practice, and external bar- ment, interviews) or explore the barriers to adherence from riers (the type of guideline, patient, or environment) [6]. A healthcare providers’ perspective. better understanding of barriers to CGs adherence might The extension of assessment could be only one recom- help healthcare providers comply with CGs recommenda- mendation or a complete guideline. We sought for studies tions, thereby improving the quality and cost-effectiveness reflecting usual care conditions. We excluded studies assess- of health care [7]. ing patients’ adherence to treatment or indications, clinical In 2018, over 400,000 incident breast cancer cases were trials, or studies to implement programmes to improve CGs estimated in Europe [8]. Because of the high burden of the adherence. One author (IR) screened titles and abstracts disease and the enormous health-economic impact [9], to select potentially relevant references to be evaluated on several breast cancer CGs in Europe have been developed, full text. Then, two authors (AVM, ENDG) independently reflecting decades of intensive research [2]. However, their assessed whether these studies met the eligibility criteria. methodological quality, evaluated with the AGREE II instru- Discrepancies were resolved through discussion. We man- ment [10], reported low scores for the “rigour of develop- aged references with Endnote version X7 software (Thom- ment” and “applicability” domains. The latter includes son Reuters, New York, USA). guideline implementation and resource implications [11]. Despite adherence to breast cancer CGs in Europe being Data extraction and risk of bias assessment associated with better survival outcomes [12, 13], healthcare providers’ adherence in usual care has not been systemati- We used tabular formats to extract main study characteristics cally explored yet. The objective of this systematic literature (e.g. country, publication year, objective, year of study data, review is twofold: (i) to evaluate the extent of healthcare name of the guideline, guideline scope, adherence definition, providers’ adherence to breast cancer CGs in Europe, and number of patients, and patients’ characteristics). Primary (ii) to identify the barriers to CGs adherence from their outcomes were the proportion of patients receiving breast perspective. cancer adherent care according to CGs recommendations, by treatment modality, and the factors that impact on provider’s adherence. Methods We applied risk of bias assessment tools based on the study design: the AXIS tool for cross-sectional studies [16], We conducted a systematic review [14] and used the the Quality Assessment Tool for Before-After (Pre-Post) PRISMA guidance for its reporting [15]. PRISMA check- Studies with No Control Group for non-controlled before- list is provided in the Additional file 1. We registered the after studies [17], the Newcastle–Ottawa Quality Assess- research protocol in PROSPERO (CRD42018092884). ment Scale for prospective and retrospective cohort studies 13 Breast Cancer Research and Treatment (2020) 181:499–518 501 [18], and the Critical Appraisal Skills Programme (CASP) We grouped findings based on similarities of treatment checklist for the qualitative study [19]. We followed each modality. To summarise the proportion of adherence, we tools’ recommendations for the risk of bias assessment. The calculated descriptive statistics (the median, Interquartile AXIS tool [16] assigns 0 to 10 scores and has three cat- range (IQR) (including the median), and range). Data anal- egories: low (1–4), moderate (5–7), and high (8–10). The ysis included estimating if possible the overall adherence Quality Assessment Tool for Before-After (Pre-Post) Studies when reported disaggregated (e.g. as subgroups); selecting With No Control Group [17] includes 12 criteria, and three the least restrictive definition when several definitions were final categories: good, fair, or poor. The Newcastle–Ottawa reported; and selecting the most recent publication or the Quality Assessment Scale [18] uses scores from 0 (the high- one with the longest period for studies with more than one est risk of bias) to 9 (the lowest risk of bias). And, the Criti- publication with a similar objective. To explore barriers to cal Appraisal Skills Programme (CASP) [19] includes ten adherence, we summarised the main findings from quantita- domains, with three possible answers: “yes”, “can’t tell”, tive studies reporting factors that were significantly associ- and “no”. ated with non-adherence, and narratively integrated these Six authors (AVM, IR, ENDG, JP; MR, YS) were findings with the qualitative findings. We classified them involved in the data extraction and risk of bias assess- as internal or external (related to healthcare providers or ment. One author extracted data, and then two authors (YS, not). We explored sources of heterogeneity by the source of ENDG) contrasted the accuracy of data against full texts, data (i.e. self-reported or medical records based), treatment cross-checked the risk of bias assessment, and completed modality, and subtype of tumour. We used Microsoft Excel missing information. Disagreements were solved by discus- for data analysis. For data reporting, we present tabular sum- sion or with the help of a third author (IR). maries and graphical representations. Data synthesis and analysis Results We considered inappropriate pooling quantitative findings due to the high variability in adherence definitions and Study selection healthcare processes assessed. Instead, we summarised results in a narrative synthesis, for which we defined and Our search yielded 10,558 references. After removing dupli- categorised breast cancer care as follows: cates, 8137 references were screened based on titles and abstracts. Of these, 102 references were selected for full-text • The overall breast cancer care: the whole healthcare pro- appraisal. We finally included 57 references representing 41 cess from diagnosis to follow-up primary studies. Three studies were reported in more than • The overall treatment process: primary and adjuvant one publication: BRENDA I [12, 13, 20–28], BRENDA II therapy (if applicable) including surgery, CT, ET, and [29–31], and OncoDoc2 [32–34]. (Fig. 1). Excluded studies RT. Follow-up procedures not included. alongside their exclusion rationale are available in Addi- • Systemic therapy: comprising both CT and ET tional file 3. • Pre-treatment procedures/diagnosis: staging and HER2 status assessment before the treatment process Study characteristics • Surgical procedures: breast-conserving surgery (BCS), mastectomy (MA), sentinel lymph node biopsy (SLNB), Included studies were published between 1997 [35] and axillary lymphonodectomy (ALND), or other surgery 2019 [36, 37]. These were conducted in eight European recommendations, either assessed separately or as a countries, with two countries leading in frequency: The group. Netherlands (n = 13, 31.7%), and Italy (n = 8, 19.5%). Most • Therapies (assessed independently): CT and associated followed a retrospective cohort design (n = 19; 46.3%), the therapies (e.g. antiemetics); ET; Targeted anti-HER2 remaining included cross-sectional studies (n = 13; 31.7%); therapy; or RT non-controlled before-after studies (n = 5; 12.2%); prospec- • Supportive measures during therapies: defined as any tive cohorts (n = 4; 4.9%); a case study (n = 1; 2.4%), and a therapy aimed to prevent side effects for specific treat- qualitative study (n = 1; 2.4%). ments Two cross-sectional studies were based on a national • Follow-up/monitoring: consultations or procedures after audit [38, 39]. Still, they reported findings for two different receiving treatment with monitoring purposes subpopulations. Similarly, eleven retrospective studies used • Primary prevention strategies: interventions to promote data from the National Cancer Registry of the Netherlands early diagnosis of breast cancer in asymptomatic patients. but analysed different regions or periods [40–50]. 13 502 Breast Cancer Research and Treatment (2020) 181:499–518 Fig. 1  Flow chart of the screening process to identify relevant studies Most studies (n = 36, 87.8%) assessed adherence based on criterion, reference to the currency of guideline, or to the medical records (from hospital databases, specific registers, relevance of patients’ profile (Additional file 5). or national cancer registries), and their sample sizes ranged from 131 [51] to 104,201 records [44]. The remaining five Risk of bias studies (12.2%) were based on self-reported data, through surveys or interviews to healthcare providers [52–56], and We assessed the risk of bias for 40 studies; the case study their samples sizes ranged from 10 [56] to 202 participants [32–34] was not considered as suitable for this assessment. [53]. Adherence to CGs recommendations was the primary Most were at low (25/40, 62.5%) or moderate risk of bias outcome in 38 studies (92.6%), while three focused on fac- (14/40, 35.0%). Only one study (1/40, 6%) was considered tors that influence adherence [32, 47, 56]. Sixteen studies to be at high risk of bias, mainly due to concerns regard- (39.0%) measured adherence to CGs for more than one pro- ing selection bias in data analyses [64] (Table 1, Additional cess of care (Table 1, Additional file 4). file 6). Thirty-one studies (75.6%) reported how researchers measured the adherence to CGs. Most of these studies Adherence to breast cancer CGs recommendations (22/31, 71%) assessed adherence contrasting healthcare pro- for healthcare processes viders’ indications for breast cancer patients against a selec- tion of CGs recommendations. Six studies (6/31, 19.4%) Main findings reported by treatment modality, and clinical addressed adherence as a process indicator integrated in the guidelines are available in Table 2 and Fig. 2. quality assurance programme of their institution [57–62]. In three studies (3/31, 9.7%), a “tumour board”, or a multidisci- Overall breast cancer care plinary team of physicians, was involved in the clinical path- way of treatment decision [29, 31, 32, 63]. The definitions of Adherence to CGs for the overall breast cancer care adherence were variable across studies. They included one was measured only in three studies with a range from or more of the following components: non-adherence defi- 54 and 69% [35, 57, 58] and included patients receiving nition, classification for non-adherent treatment, threshold treatment from 1995 to 2012. These studies varied in 13 Breast Cancer Research and Treatment (2020) 181:499–518 503 Table 1  Characteristics of selected studies (n = 41) Characteristics n % References Country The Netherlands 13 31.7 [40–50, 60, 65] Italy 8 19.5 [38, 39, 52, 55, 57, 58, 61, 66] Germany1 6 14.6 [12, 13, 20–31, 36, 62, 64, 67, 68] France1 6 14.6 [33–35, 51, 63, 69, 70] UK 4 9.8 [53, 54, 56, 71] Denmark 2 4.9 [72, 73] Croatia 1 2.4 [37] Belgium 1 2.4 [74] Study design Retrospective ­cohort1 19 46.3 [13, 25–27, 36, 40, 41, 44–51, 57, 62, 66, 71, 73] Cross-sectional 13 31.7 [37–39, 42, 52–55, 58, 63–65, 70] Non-controlled before-after 5 12.2 [35, 59–61, 69] Prospective ­cohort1 2 4.9 [29–31, 74] Case ­study1 1 2.4 [32–34] Other (qualitative) 1 2.4 [56] Publication ­yeara 1999–2008 12 21.1 [35, 38, 39, 41, 45, 46, 52, 53, 60, 69, 71] 2009–2011 13 22.8 [13, 21, 24, 25, 32, 40, 50, 51, 58, 63–66] 2012–2014 11 19.3 [12, 22, 26, 33, 34, 42, 48, 55, 61, 68, 70] 2015–2017 16 28.1 [20, 23, 27, 28, 30, 31, 43, 44, 47, 49, 56, 59, 62, 67, 72, 74] 2018–2019 5 8.8 [29, 36, 37, 54, 73] Guideline ­scopeb Treatment1 31 75.6 [12, 13, 20–30, 32, 35, 36, 40, 41, 43–45, 47, 48, 50, 51, 57–64, 66, 68, 69, 71–74] Follow-up 4 9.8 [35, 42, 55, 57, 58, 69] Preventive measures 4 9.8 [37, 49, 53, 56, 70, 74] Diagnosis 4 9.8 [35, 58, 61, 65] Risk of ­biasc Low 25 62.5 [13, 30, 35, 36, 40–44, 46, 48, 49, 55–58, 61, 65, 66, 70–72, 74, 75] Moderate 14 35.0 [37–39, 45, 47, 50–54, 59, 60, 63, 69] High 1 2.5 [64] 1 Reported in more than one separate publications a Percentages calculated over the total number of publications (n = 57) b The total exceeds 100% since some guidelines comprised more than one scope c Percentages over 40 studies, a case study was not evaluated what process they considered as part of overall care: one Overall treatment process included RT, CT, ET, initial examination, and follow-up indications and found that only half of the clinicians were Six studies addressed the overall treatment process (surgery, adherent to CGs (54%) [35]; the second study evaluated CT, ET, and RT). These studies represented patients receiving nine quality indicators for diagnosis, surgery, therapy, treatment in the period from 1991 to 2009 [28, 32, 41, 48, 59, and follow-up, and found 64% of adherence to CGs [58]; 63]. The median adherence was 57.5% (IQR 38.8–67.3%), and and the third measured seven process indicators of breast ranged from 29 [63] to 91% [32]. A subgroup analysis of the cancer care including follow-up and found 69% of adher- BRENDA I study [22] found that only 15% of patients with ence with the 80% of cut-off, and 38% when it increased bilateral breast cancer (BBC) received a compliant treatment, to 90% [57]. requiring 100% of compliance to define adherence. 13 Table 2  Adherence levels by treatment modality 504 Study1 Guideline Overall care* or ST Pre-treat- Surgical CT ET RT Anti HER2 Supportive Follow -up overall treatment ment/diag- procedures measures 13 nosis Andreano ESMO 2015 69%* 76% 81% 97% – – RT after BCS – – 84% 2017 [57] 84% Balasubrama- North Trent 82% – – 90% – 74% – – – nian 2003 breast can- [71] cer group guidelines for treat- ment and referral (1998) Barni 2011 Italian 64%*–71% 93% BCS 74%; 100% 90% 92% – 0% [58] Association ALND of Medical 68%; SLNB Oncology 57% (AIOM) Boskovic Bone health – – – – – – – 32–75% – 2017 [74] guidelines BRENDA I German 67% vs 47% – 86.3% vs 88% vs 68% 91% vs 87% 95% vs 83% – – – Ebner 2015 National 88.3% [67] Step 3 (S3) (younger vs older) BRENDA I German 61% – – 91% 74% 75% 83% – – – Ebner 2015a National [20] Step 3 BRENDA I German – – BCS: 87%; Recom- Recom- Recom- – – – Hancke 2010 National MT: 52% mended: mended: mended: [21] Step 3 56%; not 88%; not 86%; not recom- recom- recom- mended: mended: mended: 95% 92% 88% BRENDA I German 33% vs 57% – – BCS: 86% 74% – 84% – – – Schwentner National overall: 54% 2012a [12] Step 3 BRENDA I German BBC 15% – – – – – – – – – Schwentner National 2012b [22] Step 3 BRENDA I German – TNBC: 44%; – TNBC 86%; TNBC 53%; – TNBC: 90%; – – – Schwentner National non-TNBC: non-TNBC non-TNBC non-TNBC 2013 [68] Step 3 70% 87% 87% 92% Breast Cancer Research and Treatment (2020) 181:499–518 Table 2  (continued) Study1 Guideline Overall care* or ST Pre-treat- Surgical CT ET RT Anti HER2 Supportive Follow -up overall treatment ment/diag- procedures measures nosis BRENDA I German – 59% – 87.6% 78% 91% 91% – – – Van Ewijk National 2015 [23] Step 3 BRENDA I German – 54% – 87% 80% 81% 87% – – – Varga 2010 National [24] Step 3 BRENDA I German 52% – – BCS: 85% 71% 85% 84% – – – Wöckel 2010 National MT: 85% [25] Step 3 2004 BRENDA I German – – – – – – 87% – – – Wöckel 2014 National [26] Step 3 Breast Cancer Research and Treatment (2020) 181:499–518 BRENDA I German 48% – – – – – – – – – Wökel 2010a National [13] Step 3 BRENDA I German – Charlson 0: – – – – – – – – Wollschlager National 68% 2017 [27] Step 3 Charlson 1–2: 66% Charlson ≥ 3: 58% BRENDA I German 53% – – – – – – – – Wolters 2015 National [28] Step 3 BRENDA II German – – – – – 92% – – – – Stuber 2017 National [31] Step 3 BRENDA German – – – – High risk: – – – – – II Leinert National 90%; 2019 [29] Step 3 low risk: 100% BRENDA II German – – – – 85% – – – – Schwentner National 2016 [30] Step 3 Bucchi 2009 1995, 1998, 53% – – – – – – – – [66] 2001 Gallen Interna- tional Conferences guidelines 13 505 Table 2  (continued) 506 Study1 Guideline Overall care* or ST Pre-treat- Surgical CT ET RT Anti HER2 Supportive Follow -up overall treatment ment/diag- procedures measures 13 nosis de Munck NABON – – – – – – 94% – – 2011 [40] (National Breast Cancer Organiza- tion of The Nether- lands) de Roos 2005 CCN guide- 69% – – – – – – – – – [41] lines DURTO 2003 St Gallen – – – – – 96% – [38] Interna- tional Consensus Panel Heins 2017 Dutch – – – – 90–99% 90–95% – – – – [43] National Cancer Treatment guidelines for breast cancer Holm- Danish Breast – – – SLNB 76% – – – – – – Rasmussen Cancer 2017 [72] Group guideline Jacke 2015 S1-guidelines 15% vs 34% – – ALND 81% 75% vs 93% 70% vs 84% BCS + RT: – – – [59] (1996– vs 81% 29% vs 53% 1997) vs S3-guide- lines (2003– 2004) Jensen 2018 Danish Breast – – 99% 95% 97% 96% 80–97% – – – [73] Cancer Group guidelines Breast Cancer Research and Treatment (2020) 181:499–518 Table 2  (continued) Study1 Guideline Overall care* or ST Pre-treat- Surgical CT ET RT Anti HER2 Supportive Follow -up overall treatment ment/diag- procedures measures nosis Kuijer 2017 Dutch AST – – – – 78% 88% – – – – [44] guidelines: 2005–2008, 2008–2012 and 2012–2014 Lebeau 2011 French 12–29% – – 65% 73% 88% 48% – – – [63] National CPGs Liebrich 2011 2006 AGO- – – – – – – – 77% – – [64] S3 German Cancer Society Breast Cancer Research and Treatment (2020) 181:499–518 Mille 2000 Centre – – – – – – – – – First–year 16% [69] Regional vs 74% Leon Bernard (CRLB) guideline (1993 vs 1995) Mylvaganam British best – – – 17–50% – – – – 19–100% – 2018 [54] practice guidelines for implant- based breast reconstruc- tion 2015 Natoli 2014 National and – – – – – – – – – 10% [55] Interna- tional Oncology societies OncoDoc2 Cancer Est 91% – – – – – – – – – Bouaud CPGs 2011 [32] Seroussi 2012, 2013 [33, 34] 13 507 Table 2  (continued) 508 Study1 Guideline Overall care* or ST Pre-treat- Surgical CT ET RT Anti HER2 Supportive Follow -up overall treatment ment/diag- procedures measures 13 nosis Ottevanger CCCE – – – BCS 39% vs 91% vs 88% – 75% vs 93% – – – 2004 [60] National 35%; MT: guideline 56% vs 54% (1988 vs 1996) Poncet 2009 French – – – – – – – National – – [51] National 31%– guidelines Regional and regional 83% CGs Ray-Coquard Guideline in 19% vs 54%* – 75% vs 86% 96% vs 92% 71% vs 85% 72% vs 94% 83% vs 93% – – 31% vs 80% 1997 [35] 1993 (1993 vs 1995) Ray-Coquard French and – – – – 95% – – – – – 2012 [70] interna- tional guidelines to manage CT-induced anaemia Roila 2003 St Gallen – – – – 60% 81% – – – – [39] Interna- tional Consensus Panel Sacerdote Piedmont – – – BCS 86% vs 66% vs 64% 94% vs 94% 88% vs. 88% – – – 2013 [61] guidelines 93% (2002 vs 2004) Schaapveld Regional – – – – 89% – – – – – 2004 [46] treatment CCCN Guidelines Schaapveld CCCN Guide- – – – BCS: 95%; – – – – – – 2005 [45] lines MRM: 90.6% Schrodi 2015 German – – – BCS 72%; – – – – – – [62] National SLNB: Step 3 0.2% vs (1999 vs 51% 2010) Breast Cancer Research and Treatment (2020) 181:499–518 Table 2  (continued) Study1 Guideline Overall care* or ST Pre-treat- Surgical CT ET RT Anti HER2 Supportive Follow -up overall treatment ment/diag- procedures measures nosis van de Water Dutch AST < 65 years: – – – – – – – – – 2012 [48] guidelines: 62% ≥ 75 years: 2005–2008, 56% 2008–2012 and 2012–2014 Van Ryckeg- European – – – – – – – – 75% – hem 2019 Organisa- [37] tion for Research and Treat- ment of Cancer Breast Cancer Research and Treatment (2020) 181:499–518 guidelines (EORTC 2010) Vercauteren Dutch – – Adding US: – – – – – – – 2010 [65] national 94%; not CGs for additional: breast can- 83% cer screen- ing and diagnosis, American College of Radiology CGs Visser 2016 HERA – – – – – – – – 63% – [49] (Herceptin adjuvant study) Weggelaar CCN guide- – 77% – – 90% 76% 91% – – – 2011 [50] lines Wimmer German – – – – – – 97% – – 2019 [36] National Step 3 (2017) 1 Sorted in alphabetical order *Including follow-up BBC bilateral breast cancer, BCS breast-conserving surgery, CT chemotherapy, ET endocrine therapy, HER2 human epidermal growth receptor, MRM modified radical mastectomy, MT mastec- tomy, SLNB sentinel lymph node biopsy, TNBC triple-negative breast cancer, US ultrasonography. Preventive measures procedures described in the text 13 509 510 Breast Cancer Research and Treatment (2020) 181:499–518 Fig. 2  Median adherence proportions for overall breast cancer care represent the “minimum” and “maximum” values. Outliers are shown and individual therapies. The square inner line represents the median, as circles. CT chemotherapy, ET endocrine therapy, RT radiotherapy while the upper and lower borders, the interquartile ranges. The bars Systemic therapy Surgical procedures Five studies addressed systemic therapy (CT and ET Three studies assessed compliance for more than one sur- indications). These studies included patients receiving gical procedure. These studies included patients receiving treatment in the period from 1992 to 2012 [27, 50, 57, treatment in the period from 1992 to 2008 [20, 35, 63].The 66, 71]. The median adherence for systemic therapy was median adherence for surgical procedures was 86.3% (IQR 76% (IQR 68–77%), and ranged from 53 [66] to 82% [71]. 75.7–89.2%), and ranged from 65 [63] to 92% [35]. Moreover, eleven studies measured adherence for indi- vidual surgical procedures which included (1) breast-con- Adherence to breast cancer CGs—procedures serving surgery (BCS) [45, 58, 60–62], the median adher- or therapies (assessed separately) ence was 74% (IQR 75.7–93%), and ranged from 35 [60] to 95% [45]; (2) mastectomy (MA) [25, 45, 60], the adherence Pre‑treatment procedures ranged from 54 [60] to 91% [45]; (3) SLNB [58, 62, 72], from 51 [62] to 76% [72]; (4) ALND [58, 59, 73], from 68 Five studies addressed the procedures before starting [58] to 81% [59]; and (5) other indicators included organi- treatment. [35, 57, 58, 65, 73]. These procedures were sational indicators [54] or “breast surgery without needing initial examination [35], indicating mammography before a second surgery” [57], which reported 17% and 97% of surgery [57, 58]; using ultrasonography after mammogra- adherent treatment, respectively. phy when applicable [65]; and assessing HER2 receptors status before surgery [73]. The median adherence for pre- Chemotherapy treatment procedures was 86% (IQR 82–96%), and ranged from 81%, for indicating mammography [57], to 99%, for Fifteen studies addressed CT, including patients receiving HER2 status assessment [73]. treatment in the period between 1992 and 2016 [20, 29, 35, 13 Breast Cancer Research and Treatment (2020) 181:499–518 511 39, 43, 44, 46, 50, 58–61, 63, 71, 73]. The median adher- One study reported that “consultations and mammograms ence was 88% (IQR 80–90%), and ranged from 60 [39] to for follow-up purposes were excessive”, and this was the 100% [58]. Additionally, one study [70] assessed compli- reason for non-compliance [42]. ance with CGs recommendations for the treatment of CT- induced anaemia and found that 95% of patients received Primary prevention strategies erythropoiesis-stimulating agents (ESAs) according to CGs. One study [45] addressed primary prevention strategies in Endocrine therapy primary care, related to the management of women with concerns about familial breast cancer, and found that 80% of Twelve studies addressed recommendations for ET, includ- the general practitioners were compliant with CGs. ing patients receiving treatment in the period between 1992 and 2016 [20, 31, 35, 39, 43, 44, 50, 58, 59, 61, 63, 73]. The Barriers for healthcare providers’ adherence median adherence was 90% (IQR 87–92.5%), and ranged to breast cancer CGs from 76 [50] to 96% [73]. Sixteen studies described and analysed potential barriers for Targeted anti‑HER2 therapy adherence to CGs, and factors associated with non-adherent indications. These barriers were categorised as internal and Three studies assessed recommendations for the use of anti- external factors [6]; the latter included patient-related fac- HER2 therapy, including patients receiving treatment in the tors and structural/organisational barriers (Table 3, Fig. 3). period between 2003 and 2009 [40, 51, 64]. The adherence Internal factors, represented by healthcare provider- values were 31% [51], 77% [64], and 94% [40]. related factors for non-adherence to CGs, include their per- ceptions, preferences, knowledge, and attitudes regarding Radiotherapy CGs recommendations. A qualitative study [56] explored the barriers for the implementation of CGs for preventive treat- Twelve studies addressed recommendations for RT, includ- ment for women at increased risk of breast cancer. Accord- ing patients receiving treatment in the period between 1992 ing to providers’ perceptions, the reasons for non-adherence and 2016 [26, 35, 36, 50, 57–61, 63, 71, 75]. The median were the perceived lack of benefit of the interventions, being adherence was 89.5% (IQR 81.5–93%), and ranged from 48 poorly informed and finding difficulties interpreting recom- [34] to 97% [36, 73]. mendations [56]. Another retrospective study reported that non-adherent indications in surgery reflected providers’ pref- Supportive measures during therapies erences for using specific techniques [45]. In other cases, these non-adherent indications would be Five studies [37, 38, 49, 54, 74] addressed procedures intentional and conscious healthcare providers’ decisions, as aimed to prevent adverse events during treatment, including described in a case study [34] conducted within an [“opti- patients receiving treatment in the period between 1996 and mal”] setting, where a CGs-based clinical decision support 2017. The median adherence was 75% (IQR 63–75%, and system (OncoDoc2) was routinely applied. They found that ranged from 19 [46] to 96% [37]. These recommendations all non-compliant decisions concerned mostly a group of included (1) using antiemetics to avoid acute or delayed patients and decisions (i.e. elderly patients in pre-surgery emesis induced by chemotherapy [37]; (2) applying meas- decisions, patients with micro-invasive tumour in re-excision ures to prevent surgical infections [46]; (3) using primary decisions, and patients with positive hormone receptors and prophylaxis with granulocyte colony-stimulating factors in HER2+ in adjuvant decisions). These discordances were patients receiving chemotherapy [27]; (4) monitoring left found mainly in areas where scientific evidence is lacking, ventricular ejection fraction (LVEF) during anti-HER2 and the non-adherence behaviour, in this case, was actually therapy [76], and (5) indicating vitamin D and calcium in intentional and conscious [34] (Table 3, Fig. 3). patients receiving adjuvant non-steroidal aromatase inhibi- External factors include the patient-related factors and tors [75]. structural factors. The patients’ age was the patient-related factor that appeared consistently associated with non- Follow‑up adherent treatment [20, 21, 30, 31, 33, 34, 40, 45, 48, 50, 63, 64, 68, 72]. In comparison with younger patients, older Six studies addressed follow-up indications, including women were less likely to receive CGs concordant surgery, patients receiving treatment in the period between 1995 and CT and RT, but were more likely to receive guideline- 2013 [35, 42, 55, 57, 58, 69]. The median adherence was concordant ET. The intrinsic tumour characteristics were 74% (IQR 10–80%), and ranged from 0 [58] to 84% [57]. also associated with non-adherent behaviours, such as the 13 512 Breast Cancer Research and Treatment (2020) 181:499–518 Table 3  Barriers that impact on healthcare providers’ adherence to breast cancer CGs Factor Main findings Internal factor Practitioners’ perceptions and preferences ∙ Clinicians were poorly informed about preventive therapy or perceived lack of benefit of preventive therapy and experienced difficulties interpreting guidelines [56] ∙ The considerable variation in BCS rates is more consistent with variations in surgeon preferences than the patient’s choice [45] ∙ Intentional and conscious healthcare providers’ decisions [34] External factor—patient-related factors Patient’s age ∙ Treatment adherence was significantly lower for surgery, RT, and systemic therapy in women aged 80 years and older, and all modalities were applied much less frequently, except for endocrine therapy which was more frequently applied in the oldest [50] ∙ Non-compliance with clinical decisions for treatment was associated with older patient age [63] ∙ Deviations from the initial therapy decision were more frequent in older patients (≥ 75 years) than in younger ones [30] ∙ Non-compliance has concerned elderly patients age > 80 years in pre-surgery decisions [33, 34] ∙ Adherence to treatment guidelines was affected by age at diagnosis [48] ∙ The effect of non-adherence was stronger in the oldest [20] ∙ Non-adherence on RT and CT increased with age [21] ∙ One reason to withhold trastuzumab was older age [40, 64] ∙ Adherence was markedly lower for elderly patients; either ALND or RT was omitted [45] ∙ Guideline adherence was significantly lower in TNBC, most pronounced in the > 65 years subgroup [68] ∙ Tumour board decision against ET was associated with the younger age of patients [31] ∙ The use of SLNB was significantly higher in younger patients (< 40 years) [72] Comorbidities ∙ Non-guideline-adherent treatment was associated with comorbidities [20], Charlson index [27] “particular cases” [32] ∙ Most common reasons to withhold trastuzumab were cardiovascular disease [40] ∙ Variation was in part but not entirely attributable to comorbidities [64] ∙ Tumour board recommendation against CT was significantly more frequent in patients with cognitive impairment [29] Tumour stage/tumour characteristics ∙ Non-guideline-adherent treatment was associated with higher tumour stages [20] ∙ Non-compliance on RT was associated with lymph node involvement or peritumoral vascular inva- sion. Within the overall treatment sequence, it was associated with positive lymph nodes, and grade III versus grade I [63] ∙ Non-compliance was associated with a micro-invasive tumour in re-excision decisions, and with HR + and Her2+ in adjuvant decisions [33, 34] ∙ Compliance with SLNB with BCS was significantly higher in patients with tumour size ≥ 50 mm, Van Nuys classification group III, palpable lesion, and upper lateral quadrant of the breast location [72] Gene expression profile ∙ The use of gene expression profiles (GEP) was independently associated with an increased risk of receiving CT in clinical low-risk patients and with a lower risk of CT administration in high-risk patients [47]. Adherence to the GEP result was higher in high-risk patients with a discordant GEP result compared to low-risk patients with a discordant GEP result [47] Quality of life (QoL) ∙ Tumour board decision against ET was associated with reduced QoL [31]. If the QoL was good, higher age was not related to deviation [30] Previous treatment ∙ Women who received RT had excessive follow-up consultations compared to did not receive it [42] ∙ Non-adherence was associated with the absence of prior axillary surgery in adjuvant decisions [33] ∙ In high-risk febrile neutropenia category, adherence to primary prophylaxis with granulocyte colony- stimulating factors (PPG) was more common among patients receiving dose-dense therapy than those receiving the classical chemotherapy [37] Socioeconomic status ∙ Low socioeconomic status (SES) patients were more likely to be undertreated with CT than high SES patients, however no association with ET. For ethnicity, no association with CT or ET was observed [44] External factor—structural and Organisation factors Geographic region and academic status ∙ Non-compliant decisions were mainly “choices of multidisciplinary staff meetings” [34] ∙ More adherence in research centres, in Northern Italy [38], and in one region of care versus another [63] ∙ The use of SLNB in patients who underwent BCS was significantly higher in low-volume depart- ments. While for women who underwent a mastectomy, SLNB was higher in high-volume depart- ments [72] BCS breast-conserving surgery, RT radiotherapy, CT chemotherapy, ET endocrine therapy, SLNB sentinel lymph node biopsy, ALND axillary lin- fatic node dissection, TNBC triple-negative breast cancer, GEP gene expression profile, QoL quality of life, SES socioeconomic status 13 Breast Cancer Research and Treatment (2020) 181:499–518 513 Fig. 3  Barriers for healthcare providers’ adherence to breast cancer CGs. Based on main categories proposed by Cabana et al. [6] gene expression profile [47] having triple-negative breast Discussion cancer [12, 68] or being diagnosed with higher tumour stages [20, 63]. Non-adherence was also associated with Main findings comorbidities [20, 27, 29, 32, 40, 64] [e.g. in patients with cardiovascular diseases, it was more frequent to withhold We synthesised the literature exploring the healthcare pro- trastuzumab [40], or in patients with cognitive impair- viders’ adherence to breast cancer CGs published in the last ment it was more frequent the delay starting chemotherapy 20 years across eight European countries. Most followed a [29]]. Other patient-related factors included the quality of retrospective cohort design and were considered to have a life (QoL) and low socioeconomic status (SES) [31]. Poor low or moderate risk of bias. Adherence for overall breast QoL was associated with non-adherence to ET [31]. If the cancer care process (from diagnosis to follow-up) ranged QoL was good, older age was not related to deviation [30]. from 54 to 69%, for the overall treatment process (including Low SES patients were more likely to be undertreated with surgery, CT, ET, and RT) the median adherence was 57.5% CT than patients with higher SES; however, an association (IQR 38.8–67.3%), while for systemic therapy it was 76% with ET was not found [29, 44]. Previous treatment also (IQR 68–77%). The median adherence values for the pro- influenced non-adherence, e.g. RT was associated with an cesses assessed individually were higher, ranging from 74% increased follow-up consultation [42], the absence of prior (IQR 10–80%) for the follow-up to 90% (IQR 87–92.5%) for axillary surgery, with non-adherence to adjuvant decisions ET. Factors that potentially impact on healthcare provid- [33], and the type of CT received [37] with non-indicating ers’ adherence were internal: their perceptions, preferences, therapies to prevent side effects. lack of knowledge, or intentional decisions; and external: the The structural factors represent the environmental patient-related and structural factors. The most consistent or organisational characteristics of the healthcare sys- factor for non-adherence was the age of patients. tem were associated with non-adherent behaviour. One study reported that non-compliant decisions were mainly Our results in the context of previous research “choices of multidisciplinary staff meetings” [34]. Other factors associated with adherence to CGs recommenda- Previous systematic reviews addressing adherence of health- tions were the academic activity in the organisation [38], care providers to CGs recommendations but out of the scope geographical location [63], and the volume of departments of breast cancer care [76–79], found similar results to our [72] (Table 3, Fig. 3). review, these studies reported a wide range of adherence 13 514 Breast Cancer Research and Treatment (2020) 181:499–518 values, depending on the specific criteria evaluated or the values. We identified only one qualitative study exploring treatment modality. From 8.2 to 65.3%, for a set of CGs [76]; factors for non-adherence to breast cancer CGs, possibly from 5 to 95% for different recommendations in the manage- explained by the type of databases we selected. Another ment of acute coronary syndromes [77]; and from 0.3 to important limitation is that guideline non-adherence data did 100% in antibiotic prophylaxis [78]. One study found 50% not differentiate deliberate guideline deviations from unjusti- of adherence for cancer-associated thrombosis [79], lower fied practice variation. This was not possible to explore as than the median estimated for the overall treatment process most of the included studies did not specify the reasons for in breast cancer care (58%). non-adherence. Other factors, like changes between versions Previous systematic reviews have also explored the main or discrepancies between several national and international barriers to healthcare providers’ adherence in other health- guidelines, could have potentially influenced guideline devi- care areas [77, 78, 81, 82]. The factors they reported are ations. However, this analysis was not feasible. consistent with our findings: the internal factors related to health professionals’ limited skills or competences to use a therapy, and their poor knowledge, a low level of awareness, Implications for practice and research familiarity or confidence with the commonly used therapy (prior to CGs-recommended therapy), or a perceived lack of This systematic review summarises for first-time health- benefit of the "new" therapy [80–82]. They also pointed out care providers’ adherence to breast cancer guidelines in as internal factors those related to intentional non-adherence Europe. Even though observed median proportions seem [77]. Those highlighted as external factors include patient- to be acceptable for most specific treatments, there is still related factors, such as comorbidities and patient-level bar- room for improvement for healthcare providers’ adherence, riers [78, 80], as well as structural factors, such as organi- especially for supportive measures during therapy as well sational characteristics like being a teaching hospital [78]. as during follow-up. Advances in breast cancer screening Other additional factors identified in previous studies and treatment have reduced breast cancer mortality across included poor organisational- or institutional-level sup- the age spectrum in the past decade [73, 74]. However, we port, inadequate peer support among health professionals, identified that the most consistent external factor associated the complex nature of some therapies or guidelines [80], with non-adherence to CGs was the older age of patients. the expectation that compliance is mandatory [4], or the Breast CGs might not adequately address this subpopulation, patients’ preferences or demands [4, 81]. or they may represent a population where the evidence to develop breast cancer guidelines is scarce. Strengths and limitations We found high variability in the methods used to measure, define, and report adherence. Future studies should provide Our study has several strengths. The broad eligibility cri- the rationale to define adherence with enough transparency teria we applied helped to capture the complete picture of and should always consider the strength of a recommenda- the healthcare providers’ adherence to breast cancer CGs tion in their selection. In usual care, healthcare providers in Europe. We included studies with any design reporting may not be aware of the standards required for CGs and the measure of adherence even for single CGs recommen- may be tempted to select low-quality recommendations [74]. dations. The detailed data extraction process helped us to Hence, to facilitate the use of evidence supporting healthcare provide a synthesis of adherence proportions by treatment recommendations, guideline developers should use rigor- modality, although definitions or selected recommendations ously developed presentation formats (e.g. decisions aids, were variable. We could identify different reporting elements clinical decision trees). In agreement with other authors [75, across definitions and methods to measure adherence. For 76], we believe it is necessary to provide a selection of rel- factors associated with healthcare providers’ adherence, we evant, reliable, and reproducible definitions for unwarranted adapted a conventional classification, which highlights fac- clinical variation in healthcare [76], in this case for breast tors to be considered by breast cancer CGs developers and cancer care. Use of robust methods to measure adherence, other relevant stakeholders. avoiding the selection or performance bias, with appropriate Our study also has some limitations. Included studies blinding of assessment, will help to evaluate this process in were conducted in eight European countries, with more than a more reliable and reproducible way. The development of half of them coming from two countries (The Netherlands more qualitative research to capture breast cancer health- and Italy). Therefore, we should be cautious when interpret- care providers’ perspective should be fostered. Furthermore, ing results and cannot extrapolate our findings to other Euro- health care providers should register the reasons for non- pean countries. Since we included studies of patient cohorts adherence to facilitate that real-world data inform guideline from the 1990s, when the implementation of guidelines updating. was starting, we might be underestimating CGs adherence 13 Breast Cancer Research and Treatment (2020) 181:499–518 515 Conclusions References A substantial proportion of breast cancer patients appear not 1. Institute of Medicine (US), Committee on Clinical Practice Guide- lines, Field MJ, Lohr KN (1992) Guidelines for clinical practice: to be receiving CGs-recommended care. Aligning healthcare from development to use. National Academies Press (US), Wash- provider’s decisions with breast cancer CGs recommenda- ington (DC) tions in European countries should be improved for almost 2. Clinical Practice Guidelines We Can Trust (2011) National Acad- all processes of care, especially for preventive therapies emy Press (US), Washington DC. https​://www.ncbi.nlm.nih.gov/ books​/NBK20​9546/. Accessed 31 July 2019 and follow-up. Knowing the reasons for non-compliance is 3. Gurses AP, Marsteller JA, Ozok AA, Xiao Y, Owens S, Pro- essential to understand these deviations. The development novost PJ (2010) Using an interdisciplinary approach to identify and implementation of CGs for breast cancer patients should factors that affect clinicians’ compliance with evidence-based address relevant patient-related factors to enhance the appli- guidelines. Crit Care Med 38(8 Suppl):S282–S291 4. McGlynn EA, Asch SM, Adams J, Keesey J, Hicks J, DeCristo- cability of CGs in clinical care. faro A et al (2003) The quality of health care delivered to adults in the United States. N Engl J Med 348(26):2635–2645 5. Levit L, Balogh E, Nass S, Ganz PA (2013) Committee on Author contributions ENDG, YS, PAC, CC, LN, JP, MR, DR, IS, AV, improving the quality of cancer care: addressing the challenges and IR conducted the systematic review. IR, DR, IS, CC, PAC, and of an aging population BoHCS, institute of medicine. Deliver- LN contributed to the definition of the research protocol. IS conducted ing high-quality cancer care: charting a new course for a system the search strategy. IR, AV, ENDG, JP, YS, and MR conducted study in crisis. National Academies Press (US), Washington (DC) selection, data extraction, and quality appraisal of included studies. 6. Cabana MD, Rand CS, Powe NR, Wu AW, Wilson MH, ENDG and YS conducted the statistical analysis. IR, ENDG, AV, JP, Abboud PA et al (1999) Why don’t physicians follow clinical YS, MR, IS, DR, CC, PAC, and LN contributed to the interpretation practice guidelines? A framework for improvement. JAMA and reporting of the results. YS and ENDG drafted the first version 282(15):1458–1465 of the article. All authors reviewed critically reviewed and provided 7. van Fenema EM (2017) Assessment of guideline adherence and comments on subsequent versions of the article. All authors read and quality of care with routine outcome monitoring data. Tijdschr approved the final manuscript prior to submission. Psychiatr 59(3):159–165 8. Ferlay J, Colombet M, Soerjomataram I, Dyba T, Randi G, Funding The systematic review was carried out by the Iberoamerican Bettio M et al (2018) Cancer incidence and mortality patterns Cochrane Center under Framework Contract 443094 for procurement in Europe: estimates for 40 countries and 25 major cancers in of services between European Commission Joint Research Centre and 2018. Eur J Cancer 103:356–387 Asociación Colaboración Cochrane Iberoamericana. YS is funded by 9. Luengo-Fernandez R, Leal J, Gray A, Sullivan R (2013) Eco- China Scholarship Council (No. 201707040103). MR is funded by Sara nomic burden of cancer across the European union: a popula- Borrell contract (CD16/00157). AVM received a training Grant D43 tion-based cost analysis. Lancet Oncol 14(12):1165–1174 TW007393 Fogarty International Centre of the US National Institutes 10. Brouwers MC, Kho ME, Browman GP, Burgers JS, Cluzeau F, of Health for the Emerging Diseases and Climate Change Research Feder G et al (2010) AGREE II: advancing guideline develop- Unit of the School and Public Health Administration at Universidad ment, reporting and evaluation in health care. J Clin Epidemiol Peruana Cayetano Heredia. 63(12):1308–1311 11. Gandhi S, Verma S, Ethier JL, Simmons C, Burnett H, Alibhai SM (2015) A systematic review and quality appraisal of inter- Data availability The datasets used or analysed during the current study national guidelines for early breast cancer systemic therapy: are are available from the corresponding author on reasonable request. recommendations sensitive to different global resources? Breast 24(4):309–317 Compliance with ethical standards 12. Schwentner L, Wolters R, Koretz K, Wischnewsky MB, Kreien- berg R, Rottscholl R et al (2012) Triple-negative breast cancer: Conflict of interest Zuleika Saz Parkinson, Luciana Neamtiu, and the impact of guideline-adherent adjuvant treatment on sur- Elena Parmelli are current employees of the Joint Research Centre, vival–a retrospective multi-centre cohort study. Breast Cancer European Commission; Ignacio Ricci-Cabello, Ena Niño de Guzmán, Res Treat 132(3):1073–1080 Javier Pérez-Bracchiglione, Yang Song, Montserrat Rabassa, Iván 13. Wockel A, Kurzeder C, Geyer V, Novasphenny I, Wolters R, Solà, David Rigau, Carlos Canelo-Aybar, and Pablo Alonso-Coello are Wischnewsky M et al (2010) Effects of guideline adherence employees of the Iberoamerican Cochrane Center. in primary breast cancer–a 5-year multi-center cohort study of 3976 patients. Breast 19(2):120–127 14. Ganann R, Ciliska D, Thomas H (2010) Expediting systematic Open Access This article is licensed under a Creative Commons Attri- reviews: methods and implications of rapid reviews. Implement bution 4.0 International License, which permits use, sharing, adapta- Sci 5:56 tion, distribution and reproduction in any medium or format, as long 15. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P (2009) as you give appropriate credit to the original author(s) and the source, Preferred reporting items for systematic reviews and meta-anal- provide a link to the Creative Commons licence, and indicate if changes yses: the PRISMA statement. BMJ 339:b2535 were made. The images or other third party material in this article are 16. Downes MJ, Brennan ML, Williams HC, Dean RS (2016) included in the article’s Creative Commons licence, unless indicated Development of a critical appraisal tool to assess the quality of otherwise in a credit line to the material. If material is not included in cross-sectional studies (AXIS). BMJ Open 6(12):e011458 the article’s Creative Commons licence and your intended use is not 17. National Institute of Health (2014) Study quality assessment permitted by statutory regulation or exceeds the permitted use, you will tools USA. https​://www-nhlbi​-nih-gov.are.uab.cat/healt​h-topic​ need to obtain permission directly from the copyright holder. To view a s/study​-quali​ty-asses​sment​-tools​. Accessed 31 July 2019 copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/. 13 516 Breast Cancer Research and Treatment (2020) 181:499–518 18. Wells GSB, O’Connell D, Peterson J, Welch V, Losos M, Tug- 32. Bouaud J, Seroussi B (2011) Revisiting the EBM decision model well P (2013) The Newcastle-Ottawa Scale (NOS) for assessing to formalize non-compliance with computerized CPGs: results in the quality of nonrandomized studies in meta-analyses. https​ the management of breast cancer with OncoDoc2. AMIA Annu ://www.ohri.ca/progr​a ms/clini​c al_epide​m iolo​g y/oxfor​d .asp. Symp Proc 2011:125–134 Accessed 31 July 2019 33. Seroussi B, Laouenan C, Gligorov J, Uzan S, Mentre F, Bouaud J 19. Critical Appraisal Skills Programme CASP Qualitative Check- (2013) Which breast cancer decisions remain non-compliant with list (2018). https​: //casp-uk.net/wp-conte​n t/uploa​d s/2018/03/ guidelines despite the use of computerised decision support? Br J CASP-Quali​tativ​e-Check​list-2018_filla​ble_form.pdf. Accessed Cancer 109(5):1147–1156 31 July 2019 34. Seroussi B, Soulet A, Messai N, Laouenan C, Mentre F, Bou- 20. Ebner F, Hancke K, Blettner M, Schwentner L, Wockel aud J (2012) Patient clinical profiles associated with physician A, Kreienberg R et al (2015) Aggressive Intrinsic Sub- non-compliance despite the use of a guideline-based decision types in Breast Cancer: a predictor of guideline adherence support system: a case study with OncoDoc2 using data mining in older patients with breast cancer? Clin Breast Cancer techniques. AMIA Annu Symp Proc 2012:828–837 15(4):e189–e195 35. Ray-Coquard I, Philip T, Lehmann M, Fervers B, Farsi F, Chauvin 21. Hancke K, Denkinger MD, Konig J, Kurzeder C, Wockel A, Herr F (1997) Impact of a clinical guidelines program for breast and D et al (2010) Standard treatment of female patients with breast colon cancer in a French cancer center. JAMA 278(19):1591–1595 cancer decreases substantially for women aged 70 years and older: 36. Wimmer T, Ortmann O, Gerken M, Klinkhammer-Schalke M, a German clinical cohort study. Ann Oncol 21(4):748–753 Koelbl O, Inwald EC (2019) Adherence to guidelines and benefit 22. Schwentner L, Wolters R, Wischnewsky M, Kreienberg R, of adjuvant radiotherapy in patients with invasive breast cancer: Wockel A (2012) Survival of patients with bilateral versus uni- results from a large population-based cohort study of a cancer lateral breast cancer and impact of guideline adherent adjuvant registry. Arch Gynecol Obstet 299:1131–1140 treatment: a multi-centre cohort study of 5292 patients. Breast 37. Van Ryckeghem F, Haverbeke C, Wynendaele W, Jerusalem 21(2):171–177 G, Somers L, Van den Broeck A et al (2019) Real-world use 23. Van Ewijk R, Wockel A, Gundelach T, Hancke K, Janni W, of granulocyte colony-stimulating factor in ambulatory breast Singer S et al (2015) Is guideline-adherent adjuvant treatment cancer patients: a cross-sectional study. Support Care Cancer an equal alternative for patients aged %3e65 who cannot par- 27(3):1099–1108 ticipate in adjuvant clinical breast cancer trials? A retrospective 38. DURTO (2013) Antiemetic prescription in Italian breast cancer multi-center cohort study of 4,142 patients. Arch Gynecol Obstet patients submitted to adjuvant chemotherapy. Support Care Can- 291(3):631–640 cer 11(12):785–789 24. Varga D, Wischnewsky M, Atassi Z, Wolters R, Geyer V, Strunz K 39. Roila F, Ballatori E, Patoia L, Palazzo S, Veronesi A, Frassol- et al (2010) Does guideline-adherent therapy improve the outcome dati A et al (2003) Adjuvant systemic therapies in women with for early-onset breast cancer patients? Oncology 78(3–4):189–195 breast cancer: an audit of clinical practice in Italy. Ann Oncol 25. Wockel A, Varga D, Atassi Z, Kurzeder C, Wolters R, Wisch- 14(6):843–848 newsky M et al (2010) Impact of guideline conformity on breast 40. de Munck L, Schaapveld M, Siesling S, Wesseling J, Voogd AC, cancer therapy: results of a 13-year retrospective cohort study. Tjan-Heijnen VC et al (2011) Implementation of trastuzumab in Onkologie 33(1–2):21–28 conjunction with adjuvant chemotherapy in the treatment of non- 26. Wockel A, Wolters R, Wiegel T, Novopashenny I, Janni W, metastatic breast cancer in the Netherlands. Breast Cancer Res Kreienberg R et al (2014) The impact of adjuvant radiotherapy on Treat 129(1):229–233 the survival of primary breast cancer patients: a retrospective mul- 41. de Roos MA, de Bock GH, Baas PC, de Munck L, Wiggers T, ticenter cohort study of 8935 subjects. Ann Oncol 25(3):628–632 de Vries J (2005) Compliance with guidelines is related to better 27. Wollschlager D, Meng X, Wockel A, Janni W, Kreienberg R, local recurrence-free survival in ductal carcinoma in situ. Br J Blettner M et al (2017) Comorbidity-dependent adherence to Cancer 93(10):1122–1127 guidelines and survival in breast cancer-Is there a role for guide- 42. Grandjean I, Kwast AB, de Vries H, Klaase J, Schoevers WJ, line adherence in comorbid breast cancer patients? A retrospective Siesling S (2012) Evaluation of the adherence to follow-up care cohort study with 2137 patients. Breast J 24:120–127 guidelines for women with breast cancer. Eur J Oncol Nurs 28. Wolters R, Wischhusen J, Stuber T, Weiss CR, Krockberger M, 16(3):281–285 Bartmann C et al (2015) Guidelines are advantageous, though 43. Heins MJ, de Jong JD, Spronk I, Ho VKY, Brink M, Korevaar not essential for improved survival among breast cancer patients. JC (2017) Adherence to cancer treatment guidelines: influence of Breast Cancer Res Treat 152(2):357–366 general and cancer-specific guideline characteristics. Eur J Pub 29. Leinert E, Schwentner L, Blettner M, Wockel A, Felberbaum R, Health 27(4):616–620 Flock F et al (2019) Association between cognitive impairment 44. Kuijer A, Verloop J, Visser O, Sonke G, Jager A, van Gils CH et al and guideline adherence for application of chemotherapy in older (2017) The influence of socioeconomic status and ethnicity on patients with breast cancer: results from the prospective multi- adjuvant systemic treatment guideline adherence for early-stage center BRENDA II study. Breast J 25:386–392 breast cancer in the Netherlands. Ann Oncol 28(8):1970–1978 30. Schwentner L, Van Ewijk R, Kuhn T, Flock F, Felberbaum R, 45. Schaapveld M, de Vries EG, Otter R, de Vries J, Dolsma WV, Blettner M et al (2016) Exploring patient- and physician-related Willemse PH (2005) Guideline adherence for early breast cancer factors preventing breast cancer patients from guideline-adherent before and after introduction of the sentinel node biopsy. Br J adjuvant chemotherapy-results from the prospective multi-center Cancer 93(5):520–528 study BRENDA II. Support Care Cancer 24(6):2759–2766 46. Schaapveld M, de Vries EG, van der Graaf WT, Otter R, Wil- 31. Stuber T, van Ewijk R, Diessner J, Kuhn T, Flock F, Felberbaum lemse PH (2004) Quality of adjuvant CMF chemotherapy for R et al (2017) Which patient- and physician-related factors are node-positive primary breast cancer: a population-based study. J associated with guideline adherent initiation of adjuvant endo- Cancer Res Clin Oncol 130(10):581–590 crine therapy? Results of the prospective multi-centre cohort study 47. Schreuder K, Kuijer A, Rutgers EJT, Smorenburg CH, van Dalen BRENDA II. Breast Cancer 24(2):281–287 T, Siesling S (2017) Impact of gene-expression profiling in 13 Breast Cancer Research and Treatment (2020) 181:499–518 517 patients with early breast cancer when applied outside the guide- 64. Liebrich C, Unger G, Dlugosch B, Hofmann S, Petry KU (2011) line directed indication area. Eur J Cancer 84:270–277 Adopting guidelines into clinical practice: implementation of 48. van de Water W, Bastiaannet E, Dekkers OM, de Craen AJ, West- trastuzumab in the adjuvant treatment of breast cancer in lower endorp RG, Voogd AC et al (2012) Adherence to treatment guide- Saxony, Germany, in 2007. Breast Care (Basel) 6(1):43–50 lines and survival in patients with early-stage breast cancer by age 65. Vercauteren LD, Kessels AG, van der Weijden T, Severens JL, van at diagnosis. Br J Surg 99(6):813–820 Engelshoven JM, Flobbe K (2010) Association between guideline 49. Visser A, van de Ven EM, Ruczynski LI, Blaisse RJ, van Halteren adherence and clinical outcome for patients referred for diagnostic HK, Aben K et al (2016) Cardiac monitoring during adjuvant tras- breast imaging. Qual Saf Health Care 19(6):503–508 tuzumab therapy: guideline adherence in clinical practice. Acta 66. Bucchi L, Foca F, Ravaioli A, Vattiato R, Balducci C, Fabbri C Oncol 55(4):423–429 et al (2009) Receipt of adjuvant systemic therapy among patients 50. Weggelaar I, Aben KK, Warle MC, Strobbe LJ, van Spronsen with high-risk breast cancer detected by mammography screening. DJ (2011) Declined guideline adherence in older breast cancer Breast Cancer Res Treat 113(3):559–566 patients: a population-based study in the Netherlands. Breast J 67. Ebner F, van Ewijk R, Wockel A, Hancke K, Schwentner L, Fink 17(3):239–245 V et al (2015) Tumor biology in older breast cancer patients–what 51. Poncet B, Colin C, Bachelot T, Jaisson-Hot I, Derain L, Magaud L is the impact on survival stratified for guideline adherence? A et al (2009) Treatment of metastatic breast cancer: a large obser- retrospective multi-centre cohort study of 5378 patients. Breast vational study on adherence to French prescribing guidelines and 24(3):256–262 financial cost of the anti-HER2 antibody trastuzumab. Am J Clin 68. Schwentner L, Wockel A, Konig J, Janni W, Ebner F, Blettner M Oncol 32(4):369–374 et al (2013) Adherence to treatment guidelines and survival in 52. Aristei C, Amichetti M, Ciocca M, Nardone L, Bertoni F, Vidali triple-negative breast cancer: a retrospective multi-center cohort C (2008) Radiotherapy in Italy after conservative treatment of study with 9,156 patients. BMC Cancer 13:487 early breast cancer. A survey by the Italian Society of Radiation 69. Mille D, Roy T, Carrere MO, Ray I, Ferdjaoui N, Spath HM et al Oncology (AIRO). Tumori 94(3):333–341 (2000) Economic impact of harmonizing medical practices: com- 53. de Bock GH, Vliet Vlieland TP, Hakkeling M, Kievit J, Springer pliance with clinical practice guidelines in the follow-up of breast MP (1999) GPs’ management of women seeking help for familial cancer in a French Comprehensive Cancer Center. J Clin Oncol breast cancer. Fam Pract 16(5):463–467 18(8):1718–1724 54. Mylvaganam S, Conroy EJ, Williamson PR, Barnes NLP, Cutress 70. Ray-Coquard I, Morere JF, Scotte F, Cals L, Antoine EC (2012) RI, Gardiner MD et al (2018) Adherence to best practice con- Management of anemia in advanced breast and lung cancer sensus guidelines for implant-based breast reconstruction: results patients in daily practice: results of a French survey. Adv Ther from the iBRA national practice questionnaire survey. Eur J Surg 29(2):124–133 Oncol 44(5):708–716 71. Balasubramanian SP, Murrow S, Holt S, Manifold IH, Reed MW 55. Natoli C, Brocco D, Sperduti I, Nuzzo A, Tinari N, De Tursi M (2003) Audit of compliance to adjuvant chemotherapy and radio- et al (2014) Breast cancer "tailored follow-up" in Italian oncology therapy guidelines in breast cancer in a cancer network. Breast units: a web-based survey. PLoS ONE 9(4):e94063 12(2):136–141 56. Smith SG, Side L, Meisel SF, Horne R, Cuzick J, Wardle J (2016) 72. Holm-Rasmussen EV, Jensen MB, Balslev E, Kroman N, Tved- Clinician-reported barriers to implementing breast cancer chemo- skov TF (2017) The use of sentinel lymph node biopsy in the prevention in the UK: a qualitative investigation. Public Health treatment of breast ductal carcinoma in situ: a Danish population- Genom 19(4):239–249 based study. Eur J Cancer 87:1–9 57. Andreano A, Rebora P, Valsecchi MG, Russo AG (2017) 73. Jensen MB, Laenkholm AV, Offersen BV, Christiansen P, Kroman Adherence to guidelines and breast cancer patients survival: a N, Mouridsen HT et al (2018) The clinical database and imple- population-based cohort study analyzed with a causal inference mentation of treatment guidelines by the Danish Breast Cancer approach. Breast Cancer Res Treat 164(1):119–131 Cooperative Group in 2007–2016. Acta Oncol 57(1):13–18 58. Barni S, Venturini M, Molino A, Donadio M, Rizzoli S, Maiello 74. Boskovic L, Gasparic M, Petkovic M, Gugic D, Lovasic IB, Soldic E et al (2011) Importance of adherence to guidelines in breast Z et al (2017) Bone health and adherence to vitamin D and cal- cancer clinical practice. The Italian experience (AIOM). Tumori cium therapy in early breast cancer patients on endocrine therapy 97(5):559–563 with aromatase inhibitors. Breast 31:16–19 59. Jacke CO, Albert US, Kalder M (2015) The adherence paradox: 75. Jensen A, Mikkelsen GJ, Vestergaard M, Lynge E, Vejborg I guideline deviations contribute to the increased 5-year survival of (2005) Compliance with European guidelines for diagnostic breast cancer patients. BMC Cancer 15:734 mammography in a decentralized health-care setting. Acta Radiol 60. Ottevanger PB, De Mulder PH, Grol RP, van Lier H, Beex LV 46(2):140–147 (2004) Adherence to the guidelines of the CCCE in the treat- 76. Arts DL, Voncken AG, Medlock S, Abu-Hanna A, van Weert ment of node-positive breast cancer patients. Eur J Cancer HC (2016) Reasons for intentional guideline non-adherence: a 40(2):198–204 systematic review. Int J Med Inform 89:55–62 61. Sacerdote C, Bordon R, Pitarella S, Mano MP, Baldi I, Casella 77. Engel J, Damen NL, van der Wulp I, de Bruijne MC, Wagner C D et al (2013) Compliance with clinical practice guidelines for (2017) Adherence to cardiac practice guidelines in the manage- breast cancer treatment: a population-based study of quality-of- ment of non-ST-elevation acute coronary syndromes: a systematic care indicators in Italy. BMC Health Serv Res 13:28 literature review. Curr Cardiol Rev 13(1):3–27 62. Schrodi S, Niedostatek A, Werner C, Tillack A, Schubert-Fritschle 78. Gouvea M, Novaes Cde O, Pereira DM, Iglesias AC (2015) G, Engel J (2015) Is primary surgery of breast cancer patients Adherence to guidelines for surgical antibiotic prophylaxis: a consistent with German guidelines? Twelve-year trend of popula- review. Braz J Infect Dis 19(5):517–524 tion-based clinical cancer registry data. Eur J Cancer Care (Engl) 79. Mahe I, Chidiac J, Helfer H, Noble S (2016) Factors influencing 24(2):242–252 adherence to clinical guidelines in the management of cancer- 63. Lebeau M, Mathoulin-Pelissier S, Bellera C, Tunon-de-Lara C, associated thrombosis. J Thromb Haemost 14(11):2107–2113 Daban A, Lipinski F et al (2011) Breast cancer care compared 80. Baatiema L, Otim ME, Mnatzaganian G, de-Graft Aikins A, with clinical Guidelines: an observational study in France. BMC Coombes J, Somerset S (2017) Health professionals’ views on the Public Health 11:45 13 518 Breast Cancer Research and Treatment (2020) 181:499–518 barriers and enablers to evidence-based practice for acute stroke for the management of low back pain: a systematic review and care: a systematic review. Implement Sci 12(1):74 metasynthesis of qualitative studies. Clin J Pain 32(9):800–816 81. Egerton T, Diamond LE, Buchbinder R, Bennell KL, Slade SC (2017) A systematic review and evidence synthesis of qualita- Publisher’s Note Springer Nature remains neutral with regard to tive studies to identify primary care clinicians’ barriers and jurisdictional claims in published maps and institutional affiliations. enablers to the management of osteoarthritis. Osteoarthr Cartil 25(5):625–638 82. Slade SC, Kent P, Patel S, Bucknall T, Buchbinder R (2016) Bar- riers to primary care clinician adherence to clinical guidelines Affiliations Ena Niño de Guzmán1 · Yang Song1 · Pablo Alonso‑Coello1,2 · Carlos Canelo‑Aybar1,2 · Luciana Neamtiu3 · Elena Parmelli3 · Javier Pérez‑Bracchiglione4 · Montserrat Rabassa1 · David Rigau1 · Zuleika Saz Parkinson3 · Iván Solà1 · Adrián Vásquez‑Mejía5 · Ignacio Ricci‑Cabello2,6,7 Yang Song Ignacio Ricci‑Cabello yangsong@cochrane.es nacho.ricci.cabello@gmail.com Pablo Alonso‑Coello 1 Iberoamerican Cochrane Centre ‑ Department of Clinical PAlonso@santpau.cat Epidemiology and Public Health, Biomedical Research Carlos Canelo‑Aybar Institute Sant Pau (IIB Sant Pau), Sant Antonio María Claret carlos.canelo.ay@gmail.com 167, 08025 Barcelona, Spain 2 Elena Parmelli CIBER de Epidemiología y Salud Pública (CIBERESP), Elena.PARMELLI@ec.europa.eu Madrid, Spain 3 Javier Pérez‑Bracchiglione European Commission, Joint Research Centre (JRC), Via javier.perez.b@gmail.com E. Fermi 2749, 21027 Ispra, VA, Italy 4 Montserrat Rabassa Interdisciplinary Centre for Health Studies (CIESAL), MRabassa@santpau.cat Universidad de Valparaíso, Valparaíso, Chile 5 David Rigau Facultad de Medicina Humana, Universidad Nacional Mayor drigau@santpau.cat de San Marcos, Lima, Peru 6 Zuleika Saz Parkinson Balearic Islands Health Research Institute (IdISBa), Palma, Zuleika.SAZ‑PARKINSON@ec.europa.eu Spain 7 Iván Solà Primary Care Research Unit of Mallorca, Balearic Islands isola@santpau.cat Health Service, Palma, Spain Adrián Vásquez‑Mejía adrianvas.1995@gmail.com 13