Parts of this chapter are reproduced with permission from Cruickshank et al.89 and Innes et al. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.
Introduction
Recommendations from the recent NIHR guideline on Gallstone Disease66 have clearly demonstrated that insufficient high-quality information has been produced for patients on the effect of cholecystectomy on patient outcomes. The guideline recommends that ‘research is needed to establish the long-term patient benefits and harms, so that appropriate information can be provided to patients to aid decision-making and long-term management of their condition’.66 However, due to the way they have been designed and conducted many completed trials are not as informative on patient care, as they could be due to lack of standardisation across studies, outcome definition, collection and reporting. This heterogeneity of outcomes across studies also hampers useful synthesis of primary studies in meta-analyses and ultimately negatively impacts on decision-making by all stakeholders. In addition to the heterogeneity of outcomes currently reported and the problems this causes, measuring the wrong outcomes (i.e. those that are not valued by clinicians or, more importantly, patients) could also be a real risk for many studies if stakeholders are not consulted during the trial design process. One way that these problems with heterogeneity of trials and the consequent relevance of their results to stakeholders can be addressed is through the development and use of COSs.90,91 At study inception there were no registered or published COS for symptomatic uncomplicated gallstone disease.
Core outcome set aims to define a minimum set of outcomes that should be considered ‘core’ for the evaluation and reporting of specific interventions or conditions (i.e. the set of outcomes that should always be considered and ideally measured in effectiveness trials).90,91 There is a growing body of literature to provide support for development of COS.90 Specifically, they are developed using consensus methods involving stakeholder groups, such as health professionals and patients, so as to ensure that the outcomes being defined are both clinically and personally relevant for the individuals involved.90,91 Generation of a COS is not expected to be mutually exclusive to the measurement of other outcomes. However, a core set will foster greater consistency in outcome reporting between studies and lead to more meaningful data being available to contribute to future meta-analysis.90,91 Moreover, COS can minimise the threat of outcome reporting bias by ensuring consistency between what is measured and what is reported.90,91 Ultimately, they should improve the overall efficiency of trials through focused outcome collection and reporting, and the quality of the evidence they produce, enabling better informed healthcare decisions to be made.
The general methodology describing the development of a COS (a systematic review of the literature to identify outcomes reported to date; interviews with stakeholders to explore additional outcomes of importance; and a consensus-based approach to determine which outcomes should be considered core) was adopted in this research to develop a COS for symptomatic uncomplicated gallstone disease. Details of this project were registered in advance and included in the Core Outcome Measures in Effectiveness Trials (COMET) Initiative database.92
Aims and objectives
Aim
The aim of this study was to develop a COS for uncomplicated symptomatic gallstone disease effectiveness trials, to recommend which outcomes should be measured and reported as a minimum and reflecting the interests of all relevant stakeholders.
Objectives
The study-specific objectives are:
- to identify a list of outcomes from those: reported in studies of the treatment of uncomplicated symptomatic gallstones from a systematic review of the literature; or reported in qualitative studies that have explored the lived experience of those with a diagnosis of uncomplicated symptomatic gallstones with specific reference to outcomes of importance
- to explore additional outcomes relevant to the treatment of uncomplicated symptomatic gallstones using semistructured interviews with stakeholders (which may include patients, clinicians, specialist nurses, RNs)
- to prioritise outcomes from both a clinician and patient perspective
- to compare patient and clinician prioritised outcomes
- to integrate patient and clinician outcomes into a combined COS.
Methods
Study overview
The development of the COS followed best practice and involved two sequential stages.93 The methods for each stage are outlined below and presented in Figure 16. The scope was restricted to interventions (surgical and non-surgical) that treat symptomatic uncomplicated gallstone disease in adults.
FIGURE 16 Core outcome set development overview. a, 1 Delphi participant did not specify a clinical role. (Figure not reproduced here; see the original article.)
Stage 1. Identification of outcomes relevant for symptomatic uncomplicated gallstone disease: generation of the outcome list.
The identification of relevant outcomes was informed by two sources: existing evidence and new primary research. The specifics of these are detailed below.
Identification of outcomes from existing literature
To generate the initial long list of outcomes, we conducted three linked reviews of existing literature: (1) a systematic review to identify outcomes reported in trials of interventions for symptomatic uncomplicated gallstone disease; (2) a content analysis of individual items within disease-specific patient-reported outcome measures (PROMs); and (3) outcomes reported by patients with a lived experience of symptomatic uncomplicated gallstone disease in exploratory studies. This multipronged approach presents an efficient and robust approach to providing the first step towards generation of a COS.
Types of studies
The search strategies developed for the health technology assessment conducted previously by our team were updated.35 In particular, we aimed to identify additional randomised trials, or studies reporting the development of PROMs, in symptomatic uncomplicated gallstone disease (e.g. trials of early vs. delayed laparoscopic cholecystectomy for uncomplicated biliary colic). We included RCTs from reference lists of relevant Cochrane reviews (early vs. delayed laparoscopic cholecystectomy for biliary colic; robot assistant vs. human or another robot in laparoscopic cholecystectomy; miniport vs. standard ports for laparoscopic cholecystectomy; fewer-than-4 vs. 4 ports for laparoscopic cholecystectomy; early vs. delayed laparoscopic cholecystectomy for acute cholecystitis).94–98 Reference lists of all reviews were searched for relevant studies. Exploratory studies (using interviews, focus groups and other methods) that have explored any aspect of living with symptomatic uncomplicated gallstone disease were also included.
Search methods for identification of studies
Extensive electronic searches were undertaken to identify trials for a project on the clinical effectiveness of cholecystectomy and these are reported in full elsewhere.35 The databases searched included MEDLINE (1946 to week 37 2012), MEDLINE-in-process (10 September 2012), EMBASE (1974 to 10 September 2012), Science Citation Index (1970 to 12 September 2012), BIOSIS (1956 to 12 September 2012) and the Cochrane Central Register of Controlled Trials (Issue 9–12, 2012). Studies identified from these searches were used to elicit reported outcomes. The MEDLINE and EMBASE searches were updated in May 2016 (September 2012–May 2016) to identify more recent relevant trials.
A specific search of MEDLINE and EMBASE was undertaken to identify studies that report PROM outcome data for cholecystitis, biliary colic and gallstones, with records retrieved by the main search for trials excluded to avoid duplication of the results. This search was undertaken in May 2016 (1980 to May 2016).
In addition, a search for relevant qualitative studies was undertaken in August 2016 in the Ovid versions of MEDLINE (from 1966 to 2016). The search strategies combined search terms for cholecystitis, cholecystectomy, biliary colic and gallstones, with terms for qualitative research.
The search strategies for MEDLINE and EMBASE are reported in Appendix 5.
Inclusion criteria for eligible studies were as follows:
- Participants: Adults aged over 18 years with symptomatic uncomplicated gallstone disease.
- Intervention: Any intervention (surgical or non-surgical management, i.e. expectant management or dietary advice or medical therapy) used to manage symptomatic uncomplicated gallstone disease in adults.
- Outcomes: All reported outcomes well eligible for inclusion.
Excluded studies included those focusing on asymptomatic gallstone disease or on acute severe cholecystitis, cholangitis or pancreatitis, which were not considered suitable for inclusion. In addition, studies including ‘complex’ gallstone cases that is, empyema, ascending cholangitis and gallstone ileus, were excluded. Reports published in non-English languages for which a translation could not be organised were also excluded.
Study selection and data extraction
For all three reviews of existing literature conducted, citations identified through the search were independently assessed by one reviewer with another assessing a 10% random sample. All full-text papers considered potentially eligible were screened by one reviewer and checked by a second reviewer.
For the review of interventions for symptomatic uncomplicated gallstone disease, one reviewer extracted details of all outcomes reported (verbatim) and any reported definition of outcomes provided by the authors (e.g. operating time may have been defined and reported by some studies as ‘interval between initial skin incision and in closure’ others ‘duration of surgery’, etc.). Disease-specific outcomes identified as being assessed with validated tools (whether self-reported or not) were reviewed and the specific domains contained within the tool were used to present the outcomes rather than reporting the tool itself.
Data were recorded in Microsoft Excel for Microsoft 365. A 10% sample was checked by a second reviewer. Other relevant data (i.e. study and participant characteristics) were extracted by one reviewer and checked by a second reviewer. At all stages, disagreement between reviewers was resolved by discussion.
From the list of outcomes reported in trials of interventions for symptomatic uncomplicated gallstone disease described above, disease-specific PROMS were identified and supplemented with the studies identified in the search. Data were extracted by one reviewer who recorded the name of the PROM(s), the reported PRO scales and individual verbatim items.
For the studies reporting participants’ experiences, data on study characteristics such as author, publication date, country, focus of investigation, data collection methods, number of participants and details of sample size were extracted. Additionally, two authors independently extracted data from two main sources reporting study findings: (1) direct quotes from participants and (2) authors interpretations of participants’ quotes. These data were recorded verbatim and analysed to identify ‘descriptive’ thematic codes.
Data analysis
All data are summarised and presented in tabular form. Data from the review of interventions for symptomatic uncomplicated gallstone disease are presented verbatim.
The individual verbatim items from each PROM were analysed using an inductive content analysis approach and informed by previous PROM coding work.99 All PROM items were systematically categorised into conceptual health domains according to the aspect which they aim to capture. Health domains were generated inductively from the identified individual items. Domain mapping was conducted independently by two authors, with any conflicts resolved through discussion.
For the qualitative studies, the constant comparison method was used to compare findings across studies and an inductive thematic synthesis was undertaken to generate a list of themes and subthemes (focused on outcomes) from the data, to map across the presurgery and postsurgery timeline.100,101 Throughout this process, the description and wording of the themes were continually revised, and notes made as to how themes and/or subthemes related and how some could be merged. These findings were discussed further with the research team to finalise the themes across the studies, and these were considered, where appropriate, as domains relevant for inclusion in the development of the COS.
Identification of outcomes of importance to stakeholders from new primary research
In addition to outcomes reported in existing literature, we conducted primary qualitative research to further inform the identification of outcomes of relevance to patients. Data were sought from interviews with patients with a diagnosis of symptomatic uncomplicated gallstone disease; a focus group with patients who had undergone cholecystectomy; and audio consultations from participants taking part in the C-GALL trial.
Participant identification and recruitment
Potential participants for the interviews were identified and recruited from those participating in the C-GALL trial. Participants were provided with a study PIL either in the clinic or posted to the participant if a decision about C-GALL trial entry was made later. The PIL contained a detachable reply slip to complete and return to the researcher (in a reply-paid envelope) indicating whether they would like to discuss participating in the interview study exploring various aspects of trial participation and included outcomes of importance. Patients being approached to participate in the C-GALL trial were asked (for trial purposes) if they would consent to their consultation being audio-recorded; more information in Chapter 6. If consent was obtained, these audio recordings were then analysed by the research team for the identification of outcomes.
A second group of participants took part in a focus group discussion, who had experienced gallstone disease, were identified through the Scottish Health Research register (SHARE www.registerforshare.org/) and sent an invitation letter asking them to contact the research team if they were interested in participating. Following initial contact, a researcher phoned the interested participants and ensured they understood what taking part in this study entailed and arranged a suitable time for the focus group.
Data collection
One author conducted the interviews over the telephone between April and August 2017. The focus group was conducted by two members of the trial team in July 2017. Trial consultations from the four sites open to recruitment (at time of COS development) were sampled to inform outcome identification. Informed (written and recorded) consent was obtained from all participants (for interviews, focus groups and consultations) prior to data collection, and confidentiality of the participants was assured.
Participants were encouraged to consider what aspects of their disease or treatment impacted them most, both in terms of physical and psychological functioning and what improvements they would wish to see in terms of their outcomes. The interviews, focus group and audio consultations were audio-recorded and transcribed verbatim (targeted transcription for the consultations) using a professional transcription service.
Data analysis
All transcripts were imported into NVivo (V.10, 2013: QSR International, Warrington, UK) and analysed using conventional content analysis (i.e. coding categories are derived directly from the text data and are used to interpret meaning from the content).99 Various themes and subthemes were generated by one researcher based on the contents of the transcripts to identify the outcomes stated by the participants, and these were then further discussed with another member of the team to finalise the list of outcomes identified across the primary qualitative data. The analysis was oriented to address the aim of identifying the range of outcomes that might be considered important to patients and the reasons used to justify assessment of them as important.
Categorisation of identified outcomes into outcome domains
The list of potential outcomes generated from both the systematic evidence search and the primary qualitative research formed the basis of a ‘long’ list of outcomes used to refine the items into a final ‘short’ list for inclusion in the consensus stage of the COS development. Outcomes were first grouped and reduced according to original source, that is, the initial long list from the systematic evidence review was reduced for direct duplication by two members of the research team. A similar process was conducted to deduplicate the outcomes identified from the PROM coding, qualitative evidence synthesis and primary qualitative research. These outcome lists were then merged and reduced (through iterative group discussions) based on areas of overlap to produce a final shortlist of individual outcomes of relevance in this context.
These individual outcome items were further grouped into broader concept-level headings to categorise outcome domains. These concept-level headings were informed by other outcome categorisation work in the area of COS and supplemented through discussion with the project management group.102–104
Stage 2. Agreement of outcomes relevant for symptomatic uncomplicated gallstone disease: defining the core set.
As yet, there is no agreed methodology on the best approach to achieve consensus when developing a COS for use in effectiveness trials.90 A range of consensus-generating approaches have been used in existing COS, ranging from survey-based approaches (e.g. Delphi) to consensus meetings (Nominal Group Technique) and variations in between.90 This study chose to use a mixture of consensus meeting with patients and a Delphi consensus survey approach with healthcare professionals. These methods are detailed below.
Consensus meeting with patients
Patients who were members of the C-GALL PPI Group (established to provide patient input into the C-GALL trial) were invited to contribute to the consensus meeting. In advance of the meeting, these patients were e-mailed information, including a brief description of COS and the purpose of the activity. They were also provided with the list of identified outcomes from stage 1 of this study and asked to identify their top three outcomes to discuss at the meeting. The group discussion was facilitated by the Health Services Research Unit PPI co-ordinator and the C-GALL trial PPI grant holder. Verbal consent was sought at the start of the meeting, with all discussion audio-recorded. Everyone shared their top three identified outcomes with a short explanation of why they felt it was important. All outcomes identified as important were discussed further, allowing each person to put forward their feelings about their chosen outcomes. After discussion, the individuals then rated the outcome as ‘high’, ‘medium’ or ‘lower’ importance. It was stressed that naturally, these were all important, but the purpose was to focus on the single most important outcome, which should be collected every time. The outcomes were then ranked based on the average scoring and the consensus definition applied to determine inclusion of outcomes in the core set. After all initially identified outcomes were discussed, the group were given time to check whether there were any other outcomes they felt would be important to include.
Online Delphi survey with healthcare professionals
A Delphi survey was used to seek agreement on the relative importance of outcomes identified from all previous stages. Each outcome generated from Stage 1 was listed together with a plain language definition on the online Delphi Manager platform.105 Although there is no formal guidance on sample size for the panel in Delphi surveys, a minimum of 10–18 has been suggested, and as such, we aimed to have a final sample larger than 18.106,107 Healthcare professionals were invited to participate through e-mail distribution lists of professional societies and through social media, The Association of Surgeons of Great Britain, Association of Upper Gastrointestinal Surgery of Great Britain and Ireland and Association of Laparoscopic Surgeons of Great Britain and Ireland (ALSGBI) were specifically requested to send the e-mail invite to their memberships. General Practitioners at Scottish sites involved in the C-GALL trial were also sent invites through the Scottish Primary Care Research Network.
The e-mail invite contained a brief information sheet and a link to the Delphi website, providing further information and allowing interested participants to register. Potential participants were reminded of the importance of completing all rounds of the process. Those who agreed to participate were asked to register their details online, which generated a unique identifier against which their data was stored and reminder e-mails for non-response were generated. Participants were not able to identify other participants or individual responses from other participants. Consent was implicit by completion and return of the questionnaire.
Three rounds of the survey (R1, R2 and R3) were completed. In R1 the outcomes were listed alphabetically by domain (to minimise potential weighting due to ordering effects), and participants were asked to consider what outcomes should be measured and reported as a minimum to develop a COS for uncomplicated symptomatic gallstone disease effectiveness trials. They then scored outcome importance using a 9-point Likert scale, where 1–3 was not important and 7–9 was essential, as per recommended methodology for COS.93 Participants were also invited to submit any additional outcomes during R1, which were reviewed by the study team and considered for inclusion in R2. All participants were then asked to complete R2. Participants were provided with their R1 score and an anonymised distribution of the group’s scores. Participants were asked to consider this information when scoring the outcome again in R2. A final round of rating (R3) was completed providing feedback as per R2. However, R3 also provided responders with patient scores (from the patient’s consensus meeting) and asked them to consider the importance of the outcome considering their own scores, the scores of the other responders and the scores of the patients. No outcomes were removed between the rounds.
Analysis of round 3
For the final analysis, for each outcome the number of participants who scored the item and the distribution of scores were summarised, alongside the number of respondents who scored the items across all three rounds. For each outcome, the proportion of respondents scoring 1–3, 4–6 and 7–9 on the Likert scale was calculated for each outcome. Each outcome was classified as: ‘consensus in’ (i.e. consensus that the outcome should be included in a core set), ‘consensus out’ (i.e. consensus that the outcome should not be included in a core set) or ‘no consensus’ (i.e. items that are equivocal and require further research for clarification).
The original consensus definition was based on existing COS studies that required 70% or more of the group to agree an outcome as important (or not) and < 15% score in the opposite direction. However, while blinded to outcome identity, the consensus definition was amended given the large number of outcomes meeting the original consensus definition and considered ‘consensus in’ (i.e. 21). As such the stringency for consensus on outcomes being considered ‘consensus in’ and included in the core set was increased to more than 95% of respondents scoring outcomes as 7–9 (i.e. essential) and ‘consensus out’ considered as 50% or less scoring 7–9. Following the Delphi, the outcomes that had reached consensus from the patient meeting were compared and combined with the outcomes meeting consensus for being included in the core set from the Delphi survey (Table 31).
| Consensus classification | Description | Original definition | Amended definition |
|---|---|---|---|
| Consensus in | Consensus that outcome should be included in the COS | ≥ 70% scoring 7–9 AND < 15% scoring 1–3 | ≥ 95% scoring 7–9 |
| Consensus out | Consensus that outcome should not be included in the COS | ≥ 70% scoring 1–3 AND < 15% scoring 7–9 | ≤ 50% scoring 7–9 and no new compelling reasons in the comment boxes |
| No consensus | Uncertainty about importance of outcome | ≤ 50% scoring 7–9 and no new compelling reasons in the comment boxes |
Results
Generation of outcome list
Sample characteristics
The search of existing literature identified 137 studies (129 from trials and PROMs and 8 qualitative studies) eligible for inclusion (Figure 17). A total of 137 publications, from 129 RCTs with 4 reporting gallstone disease-specific PROMs and 8 qualitative studies were included. The majority of the included trials involved one type of surgery versus another type of surgery.108–113 There were only a few non-surgical interventions.
FIGURE 17 PRISMA systematic review diagram for evidence synthesis. (Figure not reproduced here; see the original article.)
The PROMS varied in the aspect of focus of the measure with only two PROMs focused on gallstone disease,114–116 one on gastrointestinal diseases,117 and the other on quality of life after abdominal surgery.118 All of the PROMs reported to measure multidimensional constructs, for example, QoL. The PROMs varied in the number of constructs they aimed to assess (ranging from 4 to 8) and the number of items they asked participants to report on (ranging from 5 to 41, median = 27).
Of the eight included studies using qualitative methods, seven used interviews [face to face (n = 5) or telephone (n = 2)] and one used focus groups for data collection. All of the treatments being explored in the studies were surgical, but the types of surgery varied.117,119,120 Table 32 provides a summary of the characteristics of participants included in the quantitative and qualitative literature review.
| Characteristic | Quantitative review | Qualitative review |
|---|---|---|
| Number of study participants | (n = 119 studies) | (n = 8 studies) |
| Median | 75 | 19.5 |
| Range | 14–618 | 6–100 |
| Total | 10,757 | 256 |
| No of males/females | (n = 96 studies) | (n = 7 studies) |
| Median | 21/53.5 | 4/16 |
| Range | 0–255/15–415 | 2–15/4–37 |
| Total | 2,632/6,166 | 43/113 |
| Age (years) | (n = 14 studies) | – |
| Median (years) | 46.1 | – |
| Range (years) | 40–53 | 19–81 |
| Country | (n = 128 studies) | (n = 8 studies) |
| UK | 7a | 1 |
| EEA (excluding UK) | 44a | 3 |
| USA | 12a | 1 |
| Other | 66 | 3 |
| No. of centres | (n = 129 studies) | (n = 8 studies) |
| Single centre | 113 | 8 |
| Multicentre | 16 | |
| Total no of outcomes reported | 330 | 17 |
The primary qualitative research included 6 individual interviews, 1 focus group (n = 5 participants) and analysis of 20 consultations (5 each from 4 different hospital sites) which provided data from a total sample of 31 patients (Table 33). They included 26 women and 5 men, 26 of whom had been approached to take part in the C-GALL trial (12 trial consenters and 14 trial non-consenters), and 5 patients who had not been approached about the trial but who had all had a cholecystectomy.
| Research type and number of participants (n) | Gender | Approached to take part in the C-GALL trial, Yes/No and number (n) | Trial consenters (allocated to receive surgery or observation/conservative management)a or decliners |
|---|---|---|---|
| Interviews at recruitment (n = 6) | Female = 6 | Yes (n = 6) | Allocated surgery = 4Allocated medical management = 1Trial non-consenter = 1 |
| Audio consultations (n = 20) | Female = 16Male = 4 | Yes (n = 20) | Allocated surgery = 4Allocated medical management = 3Trial non-consenter = 13 |
| Focus group participants (n = 5) | Female = 4Male = 1 | No | Not applicable |
Outcomes identified from existing literature and new primary research
A total of 386 individual recorded outcomes were identified across the combined evidence from existing literature and new primary research: 330 outcomes (which were reported 1147 times) from trials evaluating interventions, 22 outcomes from studies reporting PROMs, 17 outcomes from existing qualitative studies and 17 outcomes from primary qualitative research.
The review of existing qualitative literature identified five additional outcomes for inclusion in the long list, not identified by the trials or studies reporting PROMS, namely: dizziness, fainting, trust, weight and prevention of additional disease. The primary qualitative research identified a further three additional outcomes (breathing problems, cough and mortality) that were not in the previous patient-focused evidence (studies reporting PROMs and qualitative literature), with two of these (breathing problems, cough) making unique contributions to the overall outcome list.
The 390 individual reported outcomes across the four data sources were reduced into a ‘short’ list of outcomes which could be measured in comparative effectiveness trials (i.e. phase III pragmatic effectiveness trials) of interventions to treat uncomplicated symptomatic gallstone disease (see Appendix 6). Several outcomes were dropped from the long list as they were deemed not eligible as clinical endpoint outcomes for use in trials of this type (e.g. system and process outcomes such as duration of surgery which might be important in earlier phase trials). Therefore, the final list covered 27 broad outcome domains that contained 54 distinct outcomes.
Defining the core set
Consensus meeting with patients
Five patients with uncomplicated gallstone disease, three of whom had previously had surgery (two of whom had resolution of symptoms, one who did not) and two patients who had not had surgery, participated in the consensus meeting. Following sharing of the prioritised outcomes, the group agreed that three outcomes should be considered as having ‘high’ importance for inclusion (i.e. in a COS), namely: QoL, overall pain and overall health state. When discussing the overall health state, there were also discussions about anxiety. The group felt that any measurement of overall health state would be more important than anxiety alone, but that it was important to take into consideration when measuring overall health state, the anxiety aspects of the disease and overall mental strain when dealing with the symptoms and length of time waiting for diagnosis and treatment.
Online Delphi survey with healthcare professionals
Round 1 of the Delphi survey was completed by 82 participants (66 surgeons and 15 GPs and 1 participant who did not specify a clinical role) and R2 by 54 (65% of those from R1). The R3 participant sample comprised 38 surgeons, 1 GP and 1 respondent who did not specify a clinical role (a total of 70% of those from R2) (Table 34). The majority were male (82.5%), aged 45–64 (55%) and based in England (85%) (see Table 34). The Delphi survey was open to responses across the three rounds from June 2019 to January 2021.
| N = 40 | |
|---|---|
| Subgroup | |
| Surgeon | 38 (95.0) |
| GP | 1 (2.5) |
| Unknown | 1 (2.5) |
| Gender | |
| Male | 33 (82.5) |
| Female | 7 (17.5) |
| Age (years) | |
| 18–44 | 17 (42.5) |
| 45–64 | 22 (55.0) |
| 65–84 | 1 (2.5) |
| Place of residence | |
| England | 34 (85.0) |
| Scotland | 4 (10.0) |
| Wales | 2 (5.0) |
| Years of treating uncomplicated gallstones | |
| 1–3 years | 33 (82.5) |
| Missing | 7 (17.5) |
Following R1 scoring, 41 additional outcomes were suggested for consideration. Following discussion with the COS study team, nine were agreed as new and taken forward for scoring in R2 and R3. These additional nine included: re-admission, reoperation, the need for ERCP, percutaneous drain, use of analgesia, need for an outpatient appointment, steatorrhea, pancreatitis and appetite. The other suggested outcomes were excluded due to being out of scope or duplicates of existing outcomes. Scores from 40 participants in R3 were included in the final analysis, an attrition rate of 37% from R1, and 23% from R2. Attrition analysis indicated no significant difference between responders and non-responders between the Delphi rounds.
Core outcome set
At the end of R3 of the Delphi, 10 outcomes achieved consensus for inclusion in the COS. Two of these 10 outcomes had also been identified at the consensus meeting with patients. An additional outcome (overall pain) was also identified as consensus during the patient meeting but did not achieve consensus during the healthcare professional Delphi. The final COS includes eleven outcomes grouped across four domains (Table 35). These were QoL, overall health state, overall satisfaction, overall pain, common bile duct injury, biliary leak, haemorrhage, need for endoscopic retrograde cholangiopancreatography, intra-abdominal collections, admission/re-admission for problems and reoperation.
| Outcome (n = 11) | Domain |
|---|---|
| QoL a | Generic health |
| Overall health statea | |
| Overall satisfactionbOverall painc | |
| Common bile duct injuryd | Intraoperative AE |
| Biliary leakd | |
| HaemorrhagedNeed for ERCPd | |
| Intra-abdominal collectionsd | Intra- and postoperative AE |
| Admission/re-admission for problemsb | Cost-effectiveness |
| Reoperation |
Discussion
COSs for use in effectiveness trials comparing surgical approaches (to a range of comparators) have seen a significant increase in development over recent years. For example, the COMET registry contains 183 entries including the word ‘surgery’ in its database and includes a broad range of specialties from cancer to trauma and cosmetic surgery.121 The value of the COS developed in the project reported here is for trials of interventions to treat uncomplicated gallstone disease, of which there are many trials evaluating various interventions.35 While the COS reported in this paper was designed for use in effectiveness trials, like many other COS (both within surgery and other specialties), it could have value for research and clinical practice.122,123 However, further work to determine international relevance, identification of measures to collect the outcomes, and further validation with additional stakeholders is also of importance.
Core outcomes set focus on the ‘what’ of outcome measurement and reporting rather than the ‘how’ which would include defining time points at which the measurement should be made. Therefore, future research should determine how to measure the outcomes identified in this COS. Some of this work has been conducted. A recent systematic review assessing the methodological quality of PROMs in patients undergoing laparoscopic cholecystectomy identified six validated measures.124 That review was not able to recommend a specific PROM for use in laparoscopic cholecystectomy due to the limited number of studies and poor quality of the measures identified.124 A more recent review also identified considerable variation in the measurement and reporting of patient-reported outcomes after laparoscopic cholecystectomy. This more recent review identified the need for a COS that would incorporate patient-reported outcomes and the consideration of longer-term outcomes.125
Strengths and limitations
Given the challenges with recruitment and sustained participation of the Delphi, it will be important to engage with a wide range of healthcare professionals and trialists to ensure this COS for uncomplicated symptomatic gallstone disease is consulted on as widely as possible prior to implementation. Further work may also be required among additional clinical stakeholder groups (e.g. general practitioners, service commissioners) to ensure the core outcomes represent outcomes they would also consider core when evaluating a range of interventions for the treatment of uncomplicated symptomatic gallstones. In addition, determining the transferability of this COS to other settings (i.e. low- and middle-income countries) would also be important to ensure wide applicability and implementation.
A key strength of the work is the extensive outcome mapping exercise on which the COS builds. Even though a small number of patients contributed to the consensus meeting, the outcomes that had been previously identified as of importance to participants were informed by an extensive synthesis of patients reported relevant outcomes through systematic literature review and primary research.89
Conclusions
The study has developed the first COS for use in effectiveness trials evaluating interventions for uncomplicated symptomatic gallstone disease. It was prospectively registered on the COMET Initiative database and development, and reporting has been informed by existing standards for COSs.90,126 The final COS includes 11 outcomes deemed critically important by both patients and healthcare professionals. Further work is required to identify appropriate ways to measure the core outcomes and to verify its findings in a wider stakeholder group.