Shared decision making in breast cancer treatment guidelines: Development of a quality assessment tool and a systematic review

Abstract Background It is not clear whether clinical practice guidelines (CPGs) and consensus statements (CSs) are adequately promoting shared decision making (SDM). Objective To evaluate the recommendations about SDM in CPGs and CSs concerning breast cancer (BC) treatment. Search strategy Following protocol registration (Prospero no.: CRD42018106643), CPGs and CSs on BC treatment were identified, without language restrictions, through systematic search of bibliographic databases (MEDLINE, EMBASE, Web of Science, Scopus, CDSR) and online sources (12 guideline databases and 51 professional society websites) from January 2010 to December 2019. Inclusion criteria CPGs and CSs on BC treatment were selected whether published in a journal or in an online document. Data extraction and synthesis A 31‐item SDM quality assessment tool was developed and used to extract data in duplicate. Main results There were 167 relevant CPGs (139) and CSs (28); SDM was reported in only 40% of the studies. SDM was reported more often in recent publications after 2015 (42/101 (41.6 %) vs 46/66 (69.7 %), P = .0003) but less often in medical journal publications (44/101 (43.5 %) vs 17/66 (25.7 %), P = .009). In CPGs and CSs with SDM, only 8/66 (12%) met one‐fifth (6 of 31) of the quality items; only 14/66 (8%) provided clear and precise SDM recommendations. Discussion and conclusions SDM descriptions and recommendations in CPGs and CSs concerning BC treatment need improvement. SDM was more frequently reported in CPGs and CSs in recent years, but surprisingly it was less often covered in medical journals, a feature that needs attention.


| INTRODUC TI ON
Breast cancer (BC) is the most common cancer in women, with 2.1 million new cases each year (25% of all female cancers), and it also causes the greatest number (about 670000 in 2018, 15%) of cancer-related deaths among women 1,2 . Mortality and morbidity from BC have decreased in recent years thanks to early diagnosis and the combination of new treatments in a growing array of different strategies 3,4 . The best BC treatment must be personalized 4,5 , and choosing the ideal approach requires a high degree of specialization, scientific-technical updating, multidisciplinary coordination and patient participation [6][7][8][9] .
This participation in shared decision making (SDM) is considered a keystone in the achievement of sustainable high-quality cancer care, and it becomes especially important when separate treatment options with overall similar potential can yield very different results depending on patients' preferences 9,10 . In developed countries, SDM is a legal obligation [11][12][13] , and it has been shown to increase the satisfaction of the patient 9 , improve cost-effectiveness 9 and reduce malpractice lawsuit 14 . It is claimed to be a keystone to guarantee good quality cancer care 9 , and it is highly recommended by medical associations [15][16][17] .
The implementation of SDM has persistent barriers [18][19][20][21][22] , and it is still poor 23,24 . Many authors have proposed strategies for promotion and practical application of SDM 10,21,[25][26][27][28] . A threestep model introducing choice, describing options and exploring preferences has been suggested 10 . Another proposal involves encouraging patients to make their own care goals that clinicians translate into treatment plans 21,25 . Option Grids and other decision aids are thought to make the SDM process easier 26,27 .
Measuring SDM as a quality indicator and reimbursing professionals that actually use SDM have been floated as another idea involving incentivization 28 .
This important subject should be adequately covered in clinical practice guidelines (CPGs) and consensus statements (CSs), especially in those that are published in a medical journal. The aim of this systematic review was to evaluate the characteristics of CPGs and CSs with SDM compared to those without, to develop an SDM quality assessment tool and to collate the specific information and recommendations about SDM concerning BC treatment in women.

| ME THODS
This systematic review was carried out following protocol registration (Prospero No: CRD42018106643) and using a prospective protocol developed based on recommended methods for literature searches and assessment of guidelines. During the course of the work, no SDM assessment tool was identified in the literature, so we developed such a tool for data extraction in our work. It was reported according to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) 29,30 (see Appendix 1).

| Data sources and searches
A systematic search combining MeSH terms "shared decision making", "clinical practice guidelines", "guidelines", "consensus", "breast cancer", "breast cancer treatment" and including word variants was conducted using MEDLINE covering the period January 2010 to December 2019, without language restrictions. We further searched online databases (EMBASE, Web of Science, Scopus, CDSR, etc.), 12 guideline-specific databases and 51 websites of relevant professional societies (see Appendix ). For completeness, we searched on the World Wide Web and the bibliographies of known relevant publications to identify additional studies of relevance to the review.

| Study selection and data extraction
We included CPGs and CSs about BC management, produced by governmental agencies or national and international professional organizations and societies. We excluded CPGs and CSs about screening and diagnosis, obsolete guidelines replaced by updates from the same organization, and CPG and CSs for education and information purpose only.
Two reviewers (MMC and IMMN) independently considered the potential eligibility of each of the titles and abstracts from the citations and requested full-text versions. Working independently, reviewers assessed the full text to confirm eligibility. Disagreements were resolved by consensus or arbitration by a third reviewer (MMD).
Duplicate articles were identified and removed. Where multiple versions of a CPG or CS were retrieved, the most recent version was reviewed. Data were extracted from selected CPGs and CSs in duplicate, independently. The intraclass correlation coefficient (ICC) was used to assess consistency between reviewers in data extraction, and the reliability level was excellent >0.90 31 . Authoritative guidance 32 on systematic review methods recommends inter-reviewer reliability assessment that is designed to compare measurements obtained by two or more reviewers extracting data from the same papers.

| Guideline quality assessment and data extraction
We conducted a search to identify a quality assessment tool for SDM. No relevant tools were identified, so we constructed one using consensus to create a checklist from a long list of items identified in the literature searches. The quality of CPGs and CSs for SDM to manage patients with BC was independently evaluated by two different reviewers (MMC and IMMN) using a piloted data extraction form. Disagreements between the two authors (MMC and IMMN) over the risk of bias for particular studies were solved by group discussion involving an arbitrator (MMD) who took the final decision.

| Data synthesis
Two authors (MMC and IMMN) synthesized the data extracted to summarize key information within using a piloted data extraction form concerning characteristics of CPGs and CSs with the SDM information and recommendations contained within them. Rate data were compared using chi-square test to examine whether CPGs and CSs with SDM were different to those without SDM.

| Study selection
Of the 4116 potential citations identified, a total of 167 documents (139 CPGs 33-171 and 28 CSs  were identified for final evaluation ( Figure 1). ICC for reviewer agreement was 0.97.

| Development of a quality assessment tool
Individual quality items were scattered across a number of tools for guidelines assessment 200,201 . A long list of items was compiled and presented to a group of four BC and SDM specialists in a consensus meeting. This process including several revisions and iterations which led to a 31-item checklist grouped into thirteen domains (see Appendix ). Of these, 68% (n = 21) were identified from the AGREE 201 and 48% (n = 15) from the RIGHT 200 tools.
Only 13% (n = 4) of these items did not appear in any of these two tools. However, the expert consensus advised their inclusion after examining other literature in the bibliography of interest about SDM 9,21,24,25,27 . The consensus meeting following approval of the 31-item checklist recommended that each item be examined for compliance. The greater the percentage of items complied with, the greater the quality for SDM in the CPG or CS assessed. The consensus meeting did not recommend the construction of a formal score or a cut point for defining quality.

| Study characteristics
The distribution by countries of CPGs and CSs that speak about SDM was irregular ( Figure 1). Europe stood out with a total of 25 were the most recently updated (highlighted in Table 2). Overall, the last update of the CPGs and CSs with SDM was more recent than that

| SDM in CPGs and CSs concerning BC
The analysis of the compliance of the items valued is presented in

| Main findings
We developed a standardized quality assessment tool for assessing the coverage of SDM in recommendation documents. Our review and analysis showed that SDM description, clarification and recommendations CPGs and CSs concerning BC treatment were poor, leaving a large scope for improvement in this area. SDM more frequently reported in CPGs and CSs in recent years but surprising SDM was less often covered in medical journals (Figure 3). Because the items mainly came from two wide-used indexes 200,201 , demonstrably our tool should be considered to have face validity.

| Strengths and weaknesses
Therefore, we are confident that our finding of poverty of SDM information in practice recommendations is trustworthy and merits further consideration.
Inter-examiner reliability should be calculated in systematic reviews as the data extracted should be the same by different reviewers 203 . Intra-examiner reliability is a pre-condition for inter-observer reliability, and so was not calculated or reported 31 . In our paper, the inter-examiner reliability score was found to be excellent (ICC = 0.97).

| Implications
To our knowledge, information and recommendations about SDM in BC CPGs and CSs have not been systematically analysed previously.
Neither did we find a tool to evaluate SDM reporting quality. This is surprising because SDM is a legal obligation 11-13 and a key component for high-quality patient-centred cancer care [6][7][8][9][10] .
Breast cancer is the paradigm of the situation where a two-way exchange not only of information but also of treatment preferences is needed to find the best option for a particular patient, as different strategies may show a priori similar advantages and disadvantages but possible outcomes are deeply related to the patient's values and personal situation 10,203 .
Formal recommendations should promote SDM application in clinical routine practice, but this has proved difficult and slow [18][19][20][21]23,24 . It would require changing attitudes, acquiring new skills, developing specific tools and ensuring an environment where communication and sharing perspectives are valued 10,21,[25][26][27] . Effective implementation strategies could be underpinned by SDM detailed in CPGs and CSs as these documents should be expected to provide this specific content [11][12][13] . Our work has identified a gap that offers an important contribution in directing further research and debate, including assessment of risk of bias in guidelines. It highlights the need for more objective-specific tools for SDM assessment, evaluation of their psychometric properties and promotion in CPGs and

F I G U R E 3 Comparison between the
year of publication of the guide according to whether or not SDM appearance CSs for diverse malignancies. Future studies should be required in that direction.

| CON CLUS IONS
This systematic review found that BC treatment CPGs and CSs insufficiently addressed SDM. Implementation of this practice is important for high-quality patient-centred cancer care, but lack of knowledge is a known barrier. SDM descriptions and recommendations in CPGs and CSs concerning BC treatment need improvement. SDM was more frequently reported in CPGs and CSs in recent years, but surprisingly it was less often covered in medical journals, a feature that needs attention. In the future, SDM should be suitably explained and encouraged and specific tools should be applied to assess its dealing and promotion in specific cancer treatment CPGs and CSs. Medical journals should play a strong role in promoting SDM in CPGs and CSs they publish in the future.

ACK N OWLED G EM ENTS
We gratefully thank the 'Programa Beatriz Galindo. Modalidad Senior. Ministerio de Ciencia, Innovación y Universidades' for making possible to link the distinguish researcher Khalid S. Khan to the University of Granada.

CO N FLI C T S O F I NTE R E S T
The study was conducted in Granada, Spain. There are no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
Each author certifies that he/she has made a direct and substantial contribution to the conception and design of the study, development of the search strategy, the establishment of the inclusion and exclusion criteria, data extraction, analysis and interpretation.
MMC was involved in the design of the study, literature search, data collection and analysis, quality appraisal and writing. IMMN was involved in the literature search and data collection. MMD was involved in the design of this study, analysis of data and writing. LM was involved in writing. KSK was involved in the design of this study, conducted the quality appraisal, in the writing, and provided critical revision of the paper. ABC was involved in the design of this study and provided critical revision of the paper. All authors read and provided the final approval of the version to be published.

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.