Assessing Internet-Based Information used to aid patient decision-making for pelvic exenteration for locally advanced and recurrent rectal cancer

Background: To establish what online decision aids are currently available for patients contemplating pelvic exenteration (PE) for locally advanced and recurrent rectal cancer (LARC and LRRC). Methodology: A grey literature review was carried out using the Google Search TM engine undertaken using a predened search strategy (PROSPERO database CRD42019122933). Written health information was assessed using the DISCERN criteria and International Patient Decision Aids Standards (IPDAS) with readability content assessed using the Flesch-Kincaid score. Results: Google search yielded 27,782,200 results for the predened search criteria. 131 sources were screened resulting in the analysis of 6 sources. No sources were identied as a decision aid according to the IPDAS criteria. All sources were deemed acceptable quality of written health information, scoring a global score of 3. for the DISCERN written assessment. The median Flesch-Kincaid ease was 50.85 (32.5-80.8) equating to a reading age of 15-18 years. Conclusions: This study has found that there is a paucity of online information for patients contemplating PE. Sources that are available are aimed at a high health literate patient. Given the considerable morbidity associated with this surgery there is a desperate need for high quality relevant information in this area. We would suggest the development of a PDA to improve decision making and ultimately improve patient experience.


Background
In the UK, 14,000 new cases of rectal cancer are diagnosed every year with an estimated 704,000 new cases worldwide in 2018 (1), of which 5-10% are locally advanced (LARC) at presentation (2,3). The incidence of locally recurrent rectal cancer (LRRC) after treatment is 4-8% (4-6). Without treatment the prognosis for both LARC and LRRC is poor, with median survival estimated at less than 1 year and only 5% of patients surviving 5 years (7,8). Pelvic exenteration (PE) represents the only potentially curative option in this patient group. who would otherwise only be suitable for palliative intervention. Complete oncological resection (R0) confers the greatest survival bene t (9,10) with survival rates of 40-50% at 5 years (11)(12)(13)(14)(15).
Whilst PE is the only potentially curative treatment option, it is associated with a signi cant risk of mortality of up to 25% (median 2.2%) and morbidity of up to 100% (median 57%) (8). Whilst the surgical and oncological outcomes have been well reported, there is a paucity of research on survivorship and the psychological impact of exenterative surgery (8) despite the growing number of PE survivors dealing with the effects of radical surgery (16-18).
Shared Decision Making (SDM) between patients and clinicians represents the epitome of patient-centred care. Analysing and understanding evidence-based medicine whilst ensuring that the patient's intentions, values and preferences are considered during the decision-making process are fundamental to SDM. A good decision is one that is informed, consistent with personal values, acted upon and with which patients are satis ed (19). For patients to be able to perform part of the integral decision they need clear, evidence based patient information (EBPI) that is easy to understand, informative and discusses the treatment options that include supportive care (20). The Royal College of Surgeons of England (RCSEng) encourage surgeons to allow su cient time for patients to read further material on their condition and available treatment, including accessing online information (21). It has been demonstrated that around two-thirds of patients use the internet to seek health information (22,23). Previous work has shown the majority of patients choose GoogleTM as an initial starting point when looking for online health information (24). The internet provides patients with a platform to search for disease-related information, supportive communication, practical tips on living with cancer, to nd out more about their disease and other general health issues, and to search for information regarding diagnosis and treatments (25)(26)(27). The internet is an extensive and complex source of information however, all pages do not go through a peer review process and can therefore vary in quality and consistency.
Whilst the principles of SDM are well documented, guidance about the integration into routine clinical practice is limited. One method is to use patient decision aids (PDA). Fundamentally, PDAs provide choice, describe treatment options and assist patients in exploring their preferences to enable decisions. They therefore support a process of deliberation and aid understanding of decisions that should be in uenced by what matters most to patients as individuals to develop informed preferences (28).
Given the complexity of the decision to undergo a PE, a PDA appears to be a suitable method to introduce SDM for this cohort of patients. The aim of this study was to review the presence and quality of current PDAs including internet resources available to LARC/LRRC patients contemplating PE.

Methodology
This review was carried out in accordance with Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. This literature search was registered on the PROSPERO database CRD42019122933 (29).
The Decision Aid Library Inventory (DALI) database (30) was searched for any available decision aids for pelvic exenteration. A search of the grey literature was also performed using Google Search™ engine. Separate searches were undertaken between the 8th January 2019 and 16th January 2019 using the following terms: 1) information about pelvic exenteration; 2) patient decision aid for pelvic exenteration; 3) treatment for advanced rectal cancer; 4) treatment for recurrent rectal cancer 5) information lea et pelvic exenteration; 6) consent for pelvic exenteration; 7) decision making in pelvic exenteration.
The search results were limited to the rst two pages, based on the assumption that internet users rarely go beyond the rst page of search results (24). The articles were screened according to the eligibility criteria and all duplicates removed.
The eligibility criteria included sources that discussed surgical management of PE in LARC and LRRC that were aimed towards patients and written in English. Academic sources targeted for healthcare professionals, advertisements sponsored by Google and articles not written in the English language were excluded. Websites were viewed, screened for relevance and studies that did not meet the inclusion criteria were excluded. Sources assessed as potentially relevant were analysed further.
Sources were screened for appropriateness and checked for eligibility according to the de ned inclusion criteria by two researchers (AW) and (AC). Any con ict between the two researchers were resolved by a third party, (DH). Data were extracted using a prede ned extraction spreadsheet on Excel® version 16.16.16. Data were extracted on the To determine whether the identi ed patient information sources quali ed as a PDA they were assessed according to the criteria published by International Patient Decision Aid (IPDAS) Collaboration (31). This tool assesses the quality of PDAs in terms of their development process and shared decision-making design components (32). The IPDAS assessment of PDAs is reported in three categories: Qualifying (6 items), Certifying (10 items) and Quality (28 items).
Only Qualifying and Certifying domains are mandatory to de ne a decision aid therefore sources were judged solely on these domains. The Qualifying criteria are required in order for an intervention to be considered a decision aid whilst the Certifying criteria ensure the decision aid has no harmful bias. Items 7-10 in the Certifying criteria were not applicable to this study due to its relevance for screening tests and therefore were excluded from analysis.
The quality of the written health information was assessed using the validated DISCERN tool (33). DISCERN has been designed to help users of consumer health information determine the quality of written information about treatment choices (34,35). The tool is formed from 16 questions; the rst 8 questions assess the reliability of the publication and the latter 8, assess the quality of information on treatment choice. Each question rated the information on a 5-point scale ranging from No to Yes. The rating scale has been designed to help decide whether the quality criterion in question is present or has been 'ful lled' by the publication. The global score indicates the assessor's overall conclusion of the quality of the source in providing written health information and can only be scored a 1, 3, or 5.
Readability was assessed using the Flesch-Kincaid reading ease score using an online tool (36, 37). Readability is scored 0-100 and corresponds inversely with school years. The challenge in using the Flesch-Kincaid Reading Ease is that test results are not immediately meaningful and a conversion table is needed to make sense of the score (37). These di culties were recti ed by implementing the Flesch-Kincaid Grade Level which calculates a score that is proportional to school years i.e. grade 1, UK school year 2 age 6-7, grade 2, UK school year 3 age 7-8. The results of the two tests correlate approximately inversely: a text with a comparatively high score on the Reading Ease test should have a lower score on the Grade-Level test.

Results
The DALI database identi ed six PDA's for colorectal cancer, all of which were ineligible for pelvic exenteration (30). A Google search for PDA's yielded 27,782,200 potential sources, of which 131 sources ful lled the eligibility criteria.
Ninety-nine sources were screened, of which 19 were excluded as advertisements. Eighty sources were subsequently assessed for eligibility. Seventy-four sources were excluded for the following reasons: 22 sources did not mention PE within the information; 24 sources were not aimed at LRRC/LARC; 25 sources were not written for patients and three sources had inaccessible websites at the time of analysis. Six sources were therefore eligible for inclusion (Fig. 1).
The six eligible sources were designated webpages, varying between HTML and PDF format. The sources were: 1) Canadian Cancer Charity; 2) Oxford University Hospitals NHS Foundation Trust; 3) Texas Oncology; 4) MD Anderson Cancer Centre; 5) Comprehensive Cancer Centre; 6) London Northwest Healthcare Trust (38-43). All sources were produced from centres in the Western world including; three from the USA, two from the UK and one from Canada.
Two of the sources were PDF format and four were HTML format. Patient targeted information was uploaded by four hospital/specialty associations, one specialist cancer centre and one cancer charity (Table 1). Description of surgical options and oncological treatments were reported in the majority of sources (n = 5). Five sources did not discuss the comparison of surgery vs no surgery nor the overall prognosis. Quality of life, bene ts and risks of surgery, description of the preoperative and recovery periods, length of recovery, nor pain were discussed in 4 of the 6 sources. Half of the sources mentioned surgical complications within their information (Table 2).

DISCERN
The quality of the written information provided by all information sources was acceptable, with all sources scoring a global score of 3 on the DISCERN tool (Table 4). Despite the acceptable quality of the written information, there was signi cant variation in reliability (score range 18-34) and quality of information (score range [11][12][13][14][15][16][17][18][19][20][21][22] provided. The majority of sources did not have clear aims or report when the publication was produced. Four of the 6 sources met the criteria for relevance and provided details of additional information. Half of the sources did not meet the criteria for: source of information used to compile the publication; reference the areas of uncertainty the bene ts and risks of each treatment described; or explain how treatment choices would affect quality of life. Five of 6 sources were scored 4-5 for their balanced and unbiased representation of information. All sources were scored 1-3 for description of how each treatment works. Two sources did not meet the criteria for clarity on whether there may be more than one possible treatment choice available. 16. Based on the answers to all of the above questions, rate the overall quality of the publication as a source of information about treatment choices.

Readability
The average reading age of the UK population is 11 years (44) yet patient information is recommended to be written at a level of a 14 year old (45). The median Flesch-Kincaid reading ease was 50.85 (32.5-80.8). This equates to 10-12th grade reading, age 15-18 years i.e. fairly di cult to read. The median Flesch-Kincaid grade level score was 7.65 (3-9.7), age 13-14 reading age (Table 5).

Discussion
This is the rst review of the literature to focus exclusively on the information available to patients with LRRC or LARC who require PE. There are no PDA's and limited online resources available to assist patients in this particular decision-making process. Most of the available information is of poor quality and does not aid patient decisionmaking, as re ected in low DISCERN and IPDAS scores.
From a patient's perspective, the decision whether to undergo PE surgery is complex. Patients require support to understand the proposed care, treatment and non-operative options available. Supplied information must include the risks, bene ts and consequences of treatment options in order for them to make an informed decision and participate in SDM. The principal concept of SDM is a joint decision made between a clinician and an informed patient without the use of the internet, however, patients continue to access online health information for education.
None of the sources in this study described how life would change after surgery in detail. The consequences of exenterative surgery, length of recovery, QOL and post-operative pain were only discussed in 2 available resources and complications in 3 resources. SDM cannot be achieved with such limited evidence based-literature available.
With 2 resources not comparing all available treatment options and not discussing risks and bene ts associated with each option, a balanced decision based on the information away from the consultation is impossible. Without a PDA or reliable evidence-based literature, patients have no other alternative than to use the internet which currently has poor content.
Less than half of the 12 IPDAS criteria were discussed in the available internet sources. London Northwest Healthcare Trust currently has the most informative piece of literature, however, only met 42% of the IPDASi criteria. This resource lacks su cient balanced detail to qualify as a decision aid and insu cient information to exclude any harmful bias. The overall quality of all sources was adequate according to the DISCERN tool. The lack of high-quality patient information is not limited to LARC/LRRC. Poor quality online health information about other conditions including perianal Crohn's stulae (46), ulcerative colitis (47), breast surgery (48) and full thickness rectal prolapse surgery (49) have been reported. This is alarming when so many patients turn to online resources and trust these sources of information.
The readability assessment of online health information for PE surgery is not patient-friendly with a median Flesch-Kincaid reading ease of 50.85. Information should be aimed towards the lowest health literate patient in order to support patients to understand their treatment options and what is known of the bene ts, harms, consequences and burdens of those options. The text intended for readership by the general public should aim for a schooling age 13-14 years (36) and some of the literature analysed in this study required the reader to have attended University to understand the text. Further efforts are required to invest in health literate decision support materials and ensure that healthcare professionals tailor conversations to take account potential low health literacy.
Limitations of the study include non-inclusion of printed PDAs that individual health care providers have produced but which aren't available on the internet. Further limitations were that searches were limited to the rst two pages and a more extensive search may have resulted in more literature. TO and CCC literature were analysed separately for this study as they were identi ed as different websites with different descriptors, URL, upload source and purpose. However, the content and layout of both were very similar which may introduce some bias to the results.
The importance of PDAs in clinical practice has been acknowledged by NICE, who have developed 39 decision aids which cover conditions that may be surgically treated, including breast cancer, hip and knee osteoarthritis and gynaecological procedures (50). Whilst there are a few surgically based NICE PDAs available, the majority have been developed for conditions usually managed medically for example atrial brillation, coronary heart disease, stroke, heartburn, dementia etc. Despite the increasing acceptance and use of PDA's, as things stand, currently nothing exists for patients speci cally diagnosed with LARC or LRRC. It is evident from the literature there is a clear need for accurate and informative PDA highlighting the risks, bene ts and range of treatments available to these patients.

Conclusion
Current online health information relating to PE surgery for LARC or LRRC is of limited use in aiding patient decisionmaking, as re ected in low DISCERN and IPDAS scores. The current information available is scant and of variable but predominantly poor quality. Given the peri-operative morbidity and mortality risks of PE surgery and lifelong QOL implications, there is a need to develop and disseminate high quality information to facilitate SDM. Further studies need to determine whether local literature is being used at different centres and given the complexity of the decision making, ideally a universally acceptable PDA should be developed.

List Of Abbreviations
CaCC The datasets used and/or analysed during the current study are aviable form the corresponding author on reasonable request.

Ethics approval and Consent to Participate
Ethical approval was obtained for this study by NHS Grampian, North of Scotland Research Ethics Service. IRAS 257890.

Consent for publication
All authors consent to this publication.