Patch augmentation surgery for rotator cuff repair: the PARCS mixed-methods feasibility study

Protocol and registration

Evidence synthesis was carried out in accordance with the recommendations of the Cochrane Handbook for Systematic Reviews55 and the Centre for Reviews and Dissemination (CRD) guidance for undertaking reviews in health care, and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). 55–57 The review protocol and search strategy has been previously been registered and published in full. 58

Search strategy

A previous Cochrane systematic review had carried out a comprehensive search prior to April 2006. 16 Based on this search we searched the following databases between April 2006 and February 2017 (and updated our search in August 2018): EMBASE, MEDLINE, the Cochrane Library, incorporating Central Register of Controlled Trials (CENTRAL), Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects (DARE), the HTA database and the NHS Economic Evaluation Database (NHS EED). In addition, the reference lists of all identified articles and reviews were checked for relevant articles. 17,27,39,44,59 The search strategy was initially developed for EMBASE (see Appendix 1) and has previously been published. 58 Our strategy was subsequently modified for use in MEDLINE and the Cochrane Library databases.

Inclusion and exclusion criteria

Study selection

Two authors (MB and NSN) independently screened all of the titles and abstracts identified from the search strategy. Full reports for all relevant studies identified were then reviewed and assessed against the eligibility criteria. A third independent reviewer (GG) was available to resolve any disagreements regarding study inclusion. Reasons for exclusion are detailed in the PRISMA flow diagram (Figure 4).

FIGURE 4.

The PRISMA flow chart of study selection. PRP, platelet-rich plasma; SR, systematic review.

Data extraction

Two authors (MB and NSN) extracted the following data from all eligible studies: general study information (authors, publication year and study location), study population (sample size, age, sex and tear size), study characteristics (study design, inclusion/exclusion criteria, duration of clinical and radiological follow-up, surgical technique and patch characteristics), all primary and secondary outcomes for each study and adverse events or complications. Each reviewer independently checked the results of the data extraction process.

Risk-of-bias assessment

The risk of bias was independently assessed by two authors (MB and NSN) and discrepancies were discussed with a third reviewer (GG), allowing resolution based on unanimous decision. RCTs were assessed using the risk-of-bias tool (2011 update) provided by the Cochrane Collaboration. 55 Each domain was rated as having a ‘low’, ‘high’ or ‘unclear’ risk of bias before the study was assessed as a whole. Observational comparative studies were assessed using the Risk Of Bias In Non-randomized Studies – of Interventions (ROBINS–I) tool. 62 The risk of bias for each domain was judged as low, moderate, serious, critical or no information, followed by an overall judgement of bias based around the judgements from each individual domain. Single-arm studies were not formally assessed for risk of bias.

Data analysis

Identified studies were grouped (RCTs, observational comparative and non-comparative) and a narrative summary of results was reported in accordance with the standards set out in the PRISMA checklist. 20 Data from all available studies were utilised in the quantification of complications. All studies that compared the outcomes of RCR with graft augmentation with standard RCR were considered for meta-analysis. A meta-analysis was conducted only for outcomes consistently reported across studies and reported using Review Manager version 5.3 (RevMan, the Cochrane Collaboration, The Nordic Cochrane Centre, Copenhagen, Denmark). Regardless of the observed statistical heterogeneity, we conducted an analysis for each patch type (xenograft, allograft, autograft or synthetic) when each type was represented by at least two comparative studies. Given the known controversy surrounding xenograft isolated from small intestinal submucosa (SIS), the analysis for xenografts was further divided into SIS-derived and non-SIS. There were insufficient study numbers to permit further subdivision based on graft configuration (on-lay or bridging). Complications (including patch-related adverse events) were grouped together given how they were reported across the included studies. They were not formally meta-analysed and only crudely summarised as overall numbers for augmentation and non-augmentation groups across all variations in patches and non-patches and reported events.

Statistical analysis

For dichotomous parameters included in the meta-analysis, the risk ratio (RR) with 95% confidence interval (CI) was calculated for each graft type. For continuous variables, such as shoulder-specific functional outcome scores, the effect was reported as the mean difference with 95% CI. Owing to the significant heterogeneity in the specific functional shoulder scores utilised between studies, a meta-analysis was conducted using the most frequently used score across all studies at final follow-up. In each patch type, if no single functional outcome score was consistently used, scores were combined and a standard mean difference was reported with 95% CI. Studies in which no standard deviation was calculable, or in which only subcomponents of functional outcome scores were reported, were reported only descriptively. Heterogeneity was characterised by use of the I²-statistic and a random-effect analysis used to incorporate heterogeneity among studies.

Patient involvement

Patient representatives were full members of the PARCS Study Steering Committee and provided critical feedback on the study protocol. 1

Study selection

The search strategy identified 939 articles, of which 56 were duplicates (see Figure 4). A total of 883 abstracts were reviewed in detail, with 44 appearing to meet inclusion criteria. After full-text review, 27 articles were excluded based on the following criteria: included an abstract only (n = 3), treatment was a platelet-rich plasma-derived matrix lacking the structural properties of patch augmentation (n = 4), RCR did not involve any form of augmentation (n = 9) and the article was a duplicate (n = 4). A further 28 articles were identified from existing systematic reviews, which generated a total of 52 studies for inclusion. No economic evaluations of patch-augmented rotator cuff surgery were identified.

Study characteristics

Surgical characteristics

Shoulder pain and function

Re-tear (including radiological assessments)

The integrity of the surgical repair was assessed by all RCTs and seven observational comparative studies, with a re-tear rate ranging from 10% to 73% following patch implantation and from 18% to 65% following a standard RCR (see Appendix 2, Table 21). Magnetic resonance imaging (MRI) was the commonest imaging modality (62%) utilised to diagnose recurrent tears, with a magnetic resonance arthrogram utilised in a further 23% of studies. The majority of studies undertook postoperative imaging after 1–2 years; however, there was considerable heterogeneity existing in the radiological classification of re-tears, and four studies did not provide any details. Definitions of re-tears could be broadly categorised into two themes: the presence of any tear or the presence of tears greater than the residual intraoperative defects. Five studies also attempted to subcategorise recurrent tears into ‘partial’ or ‘complete’. For example, the study by Iannotti et al. 35 described a third ‘partially healed’ group, which was defined as a smaller rotator cuff lesion than that observed during preoperative imaging.

Although the RCT investigating human allograft (GRAFTJACKET) demonstrated a significantly lower failure rate in the augmentation arm, neither of the RCTs investigating the xenograft patch Restore found any reduction in re-tear rate. 26,35,65 In conflict with these findings, a multipatch comparative study found no difference in failure rate between three different patches: xenograft (Restore), human allograft (GRAFTJACKET) or synthetic (Artelon). 30 Among the observational comparative studies, significantly lower rates of re-tears were reported after augmentation with synthetic (Repol Angimesh), autograft (fascia lata) or allograft patches (Arthroflex and Allocover), whereas no improvement in re-tears was observed following augmentation with a long head of biceps tendon autograft or for the Restore patch. 33,39–43,67,71

Shoulder pain

Health-related quality of life

Patch-related adverse events

A total of 43 studies provided data on complications, of which 21 studies reported the occurrence of complications in a total of 73 patients. The more commonly reported events across the studies included superficial and deep infections, and inflammatory response. Other reported complications were shoulder bursitis, biceps rupture, fibrosis, unexplained fever, shoulder manipulation, wound erythema, shoulder stiffness, persistent pain, skin reaction, biceps deformity, ossification, cardiac event and possible inflammatory response. One study,69 which used a quadriceps autograft-based patch, reported knee-related problems and nerve injury.

Other secondary outcomes

No data on other secondary outcomes of interest were reported.

Meta-analysis

Risk of bias

Assessment of bias was conducted for all RCTs and comparative studies (Tables 2 and 3). Only the study by Bryant et al. 65 was at a low risk of bias, with the remaining RCTs assessed as having an unclear risk. These findings are based on a lack of study methodology detail, in particular surrounding blinding of patients and outcome assessors. All observational comparative studies had a serious risk of bias, which centred around the potential for confounding, bias in patient selection and outcome measurement.

Strengths and limitations of the study

Strengths of this review include a priori published protocol,58 a comprehensive search strategy, inclusion of non-English language articles, duplicate assessment of eligibility, a risk-of-bias assessment and data extraction. Nonetheless, there remain several limitations to the current review, which are mainly a reflection of the quality of the published primary research available. Only four RCTs have been published, of which two relate to a product (Restore) that has now been withdrawn from market because of safety concerns. 26 In addition, substantial heterogeneity between studies was observed, with the majority of studies also judged to have a high risk of bias, which seriously limited our ability to draw firm recommendations. An exhaustive exploration of the heterogeneity has not been undertaken and indeed such an analysis was not declared a priori in our protocol paper. 58 However, separating studies by patch type did influence the degree of heterogeneity and we would, therefore, recommend that patch type should be considered in the design of future reviews.

In comparison with previous systematic reviews, we have included one additional RCT63 and three observational comparative studies representing 278 patients not otherwise identified. 27,44,59,63,66,67,71 Results from our meta-analysis are, in part, consistent with a previous analysis that found an overall reduction in re-tear rate and improved ASES scores following patch augmentation. 44 The substantial number of additional studies included in this current review provide greater precision and, although a subgroup analysis was not originally specified in our protocol, they have allowed us to hypothesise that patch type may have an effect on patient outcomes. The occurrence of adverse events with only certain patch types adds some credibility to this notion. Previous reviews of augmented RCR have, on the basis of a presumed effect on patient outcome, excluded studies based on the size of rotator cuff tear or surgical technique (on-lay or bridging). 27 It is possible that each technique reflects different patient cohorts; for example, the use of bridging scaffolds may represent larger, more chronic or even recurrent rotator cuff tears. However, we were unable to detect any overall difference in patient-reported outcomes, re-tear rate or pain scores between studies reporting on-lay or bridging techniques. It should be noted that differences in terminology makes comparison of these surgical techniques challenging; the terms ‘irreparable’, ‘bridging’, ‘interposition’ or ‘reconstruction’ were used interchangeably in a number of studies or to refer to the same approach. To help facilitate the future interrogation of the relationship between surgical technique and outcomes we would suggest that only the terms ‘on-lay’ (defined as repair augmentation) or ‘bridging’ (defined as repair reconstruction/augmentation) be utilised in accordance with previously published definitions. 27

There are a growing number of patches available for the augmentation of RCR. Despite the safety-related withdrawal of certain patches, as well as wider concerns surrounding medical device and mesh implantation, rigorous clinical evaluation of patch augmentation is lacking. 41,110 We were particularly concerned by the absence of publicly available research for several patches currently in clinical use [e.g. dCELL^® (Tissue Regenix Group plc, Leeds, UK) and Leeds–Kuff™ (Neoligaments, Leeds, UK)]. Although some studies have indicated promise for specific patches, firm recommendations in terms of patch type or surgical technique cannot be made at present. There remains a need for well-designed comparative studies (preferably multicentre RCTs) that are capable of robustly evaluating the effectiveness and safety of multiple patch types. Furthermore, routine reporting of patch registry data could address the current lack of robust safety data for cuff-augmented RCR. 111

Stage 2: British Elbow and Shoulder Society membership survey

The BESS members are predominantly clinically active shoulder surgeons. The BESS office used the e-mail list to invite participants to complete the survey, avoiding the unnecessary sharing of personal data. Information about the PARCS study and a hyperlink to the survey was provided. There was no minimum number of responses required as the study was opportunistic in terms of sample size and was not driven by statistical testing.

The e-mail invitation was sent out in April 2017 and available to complete until the end of August 2017. Surgeon members of the BESS attending the annual meeting in June 2017 were also offered an opportunity to complete the survey at an exhibition stand. To assess respondent demographics, participants were asked about their grade (i.e. consultant, trainee or other) and place of work (i.e. district general hospital, university teaching hospital, private practice or other). To determine their familiarity with and experience of augmented RCR, participants were asked about their preferred surgical technique for RCR (predominantly open, predominantly arthroscopic or substantial amount of both open and arthroscopic repairs), whether or not they had previously used patch-augmented rotator cuff surgery (i.e. no, yes within 6 months or yes but not within 6 months) and the total number of augmented cuff repairs that they had undertaken. Further questions sought to determine the types of patch commonly in use and to investigate the factors influencing patch selection. Two separate free-text questions were posed: ‘Which patches have you used?’ and ‘Why did you use these specific patches?’. A final free-text box was provided to allow further comments about the choice of patch to be recorded.

To gather opinion on patient selection for augmented RCR, respondents were asked to consider discrete patient subgroups. Four different tear sizes (i.e. small, medium, large and massive) were combined with different four different ages (50, 60, 70 and 80 years) to produce 16 combinations. Participants were asked if they considered each patient combination appropriate for patch augmentation. An answer of ‘yes’, ‘no’ or ‘unsure’ could be provided for each scenario. A free-text box was also provided to capture further comments relating to patient suitability.

Respondents were then asked to consider participation in a future clinical trial in augmented RCR. Responders were asked if they would be interested in participating in a RCT of patch-augmented surgery (i.e. yes, no or maybe). Members were asked what factors could be addressed to encourage participation in a RCT.

The survey was piloted with four members of the shoulder and elbow surgical team who perform patch repair. The survey took approximately 10 minutes to complete.

Stage 4: surgeon triallist survey

The surgeon triallist survey was directed at surgeons who had taken part in previous large multicentre NHS shoulder trials and, therefore, the subset of surgeons who were most likely to participate in a RCT of patch-augmented rotator cuff surgery. Eligible participants were identified through a network of surgeon triallists who have participated in previous NHS-based shoulder surgical trials [i.e. Can Shoulder Arthroscopy Work (CSAW)?, UK Frozen Shoulder Trial (UKFroST) and UKUFF]. 18,112,113 They were invited to complete the survey by the PARCS project management group through a personalised e-mail or face-to-face invitation. The survey was open between June and August 2018. Non-responders received up to two e-mail reminders asking them to complete the survey.

The intention was to invite at least 30 research-active orthopaedic shoulder surgeons. This was considered large enough to meet the aim of this component of the project and ensured that a range of surgeons and centres were included. This would also be close to the number of surgeons needed to participate in a future trial of patch-augmented rotator cuff surgery.

The surgeon triallist survey was developed based on previous stages, including the systematic review, the above BESS membership survey and the stakeholder focus groups. Uncertainties related to the patient population, intervention, control and outcome (PICO) elements, and trial practicalities were incorporated into the survey design.

The survey started by asking if the triallists currently used a patch to augment a RCR on any of their patients. If they answered yes, the surgeons were then invited to respond to questions on how they typically used a patch (bridge, on-lay or other) and whether or not the state of the subscapularis affected their decision to use a patch. The next section included questions about a trial of patch-augmented rotator cuff surgery and which patients they would be willing to randomise. Other questions included what the comparator in a trial would be and more detail about patient characteristics, including tear size, age ranges, presence of atrophy, glenohumeral OA and cuff arthropathy. The survey also asked questions about logistical trial procedures, including the timing of randomisation, surgical repair technique, types of patch they would be willing to use, postoperative rehabilitation and length of patient follow-up.

The survey was reviewed and tested by members of the study team, including surgeons and trainees. The survey took approximately 15 minutes to complete. During the survey, participants were asked to register their interest in taking part in further stages of the PARCS study.

Stage 2: British Elbow and Shoulder Society membership survey

A total of 550 medically qualified members of the BESS (of whom 481 were consultant orthopaedic surgeons) were invited to participate, with 105 (19%) responding. The respondents were mostly consultant surgeons (97%), with the majority working at district general hospitals (48%) (Table 4). Most participants (95%) worked within the NHS but with some reporting additional work within the private sector (32%).

Most respondents undertook arthroscopic RCRs (66%), with only a minority solely undertaking open repairs (14%) (Table 5). When asked whether or not they had ever used a patch to augment rotator cuff surgery, over half (58%) had done so. The majority of patch users had undertaken an augmented repair within the last 6 months (70%). The utilisation of patches among surgeons performing open repairs was slightly lower (40%) than for those reporting an arthroscopic (56%) or mixed open and arthroscopic practice (76%).

A varied surgical experience was reported among those who had performed patch-augmented RCR. Most surgeons reported low use, with a median of five rotator cuff augmentation procedures performed; however, the maximum reported was 200 procedures.

Responses relating to the patch types used can be seen in Table 6. When asked about the patch types and products utilised during RCR, 13 different products were reported. Decellularised dermis accounted for 85% of the different patches used, and non-degradable synthetic meshes made up the remaining 15%. Human decellularised products were more frequently used, with only 13% of decellularised patches being porcine derived (the rest being human, except for one for which it was not clear what the source material was). All synthetic scaffolds that were reported were non-degradable and produced from a variety of polymers (i.e. polyester, polypropylene and polyurethane). Overall, GRAFTJACKET was the most commonly reported device (55%) with the Leeds-Kuff Patch (10%), Arthroflex (8%) and dCELL (8%) the next most common.

Reported factors influencing patch selection are given in Table 7. The device’s perceived efficacy was an important theme, with clinical evidence (24%) and personal and peer experience (8% and 4%, respectively) cited as important determinants. Product characteristics formed another dominant theme, with patch usability, strength and material influencing selection. A product’s cost and availability within the local hospital was also important. In addition, the specific characteristics of a rotator cuff tear may also determine patch choice.

Participants responded with their opinions on patient selection for a RCT of patch-augmented RCR (Figure 8). Responses for patients aged 50 or 60 years tended to be similar regarding patient suitability for receiving a patch. However, the effect of age on patient suitability was clearly influenced by tear size. Among older patients (aged 70 or 80 years) with small and medium-sized tears, a greater (although still relatively small) proportion of upper limb surgeons either would consider augmentation or were unsure (12–26%).

FIGURE 8.

The BESS membership survey responses relating to suitability on patient selection. Reproduced with permission from Baldwin et al. 114 © 2020 Baldwin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited (https://creativecommons.org/licenses/by/4.0/). This includes minor additions and formatting changes to the original figure.

Reproduced with permission form Baldwin et al. 114 © 2020 Baldwin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited (https://creativecommons.org/licenses/by/4.0/). This includes minor additions and formatting changes to the original figure.

Conversely, the reverse trend was observed in large and massive tears, with a greater proportion of surgeons considering augmentation appropriate in the 50 or 60 years age groups (39–59%).

Overall, tear size seemed to be more important than age in assessing patient suitability. Just over half of respondents (range 19–59%) would use augmentation in large and massive tears, compared with ≤ 10% (range 3–11%) for small and medium-sized rotator cuff tears. However, it is worth noting that considerable uncertainty remains. Around one-fifth (range 19–27%) of respondents were unsure as to the role of augmentations in medium, large and massive tears.

Additional free-text comments were provided by 48 participants, with six dominant themes emerging (Table 8). Tear characteristics remained an important consideration during patch augmentation. As well as tear size, the degree of fatty atrophy, the intraoperative ability to mobilise the tendon and the tension of the repair were often mentioned as important factors during tear classification. Surgical assessment of tendon quality was reported as important by 17% of respondents. A lack of glenohumeral OA (13%) and the failure of a standard repair (27%) were provided as other important qualifiers during augmented repair consideration. For older patients (aged ≥ 70 years) with large tears, 10% of respondents justified avoiding augmented repair because of the perceived success of reverse shoulder arthroplasty.

When asked if they would actively participate in a RCT of patch augmentation, half of the respondents confirmed an interest, with a further 22% undecided and the remainder not interested. Twelve respondents provided additional comments that explored barriers to participation. Further trial details were mentioned in almost half (46%) of comments and specific limits were placed on the inclusion/exclusion criteria by 23%, for example ‘no compulsion to use patch in small/medium tears’. A further 31% of comments listed concerns over the type of intervention or comparison that would be utilised. For example, ‘compared to balloon interposition’ or ‘comparing reverse [total shoulder arthroplasty] with patch repair’ were suggested.

Stage 4: surgeon triallist survey

Of the surgeons invited, 24 out of 31 (77%) completed the survey. All of the surgeons who responded had acted as the principal investigator for at least one of the UKUFF,2 CSAW112 and UKFroST113 trials.

Patch use was not as common in this population of surgeons as implied in the BESS membership survey (Table 9). Only 11 of the 24 surgeons (46%) regularly used patches. Nine (38%) do not currently use patches but would be willing to for the purpose of a trial and four do not use patches at all and would not consider using them in a trial setting. Responses from the four surgeons unwilling to use a patch in a trial have been excluded from the analysis. How the surgeons used a patch varied: six out of the 11 surgeons used patches as a bridge to fill a persistent defect after a standard repair and five used it as on-lay to reinforce a standard repair.

Among the 20 surgeons, there was a division of opinion related to the question on whether or not the state of the subscapularis muscle would affect their decision to use a patch, with 11 (55%) stating yes and 9 (45%) stating no. Comments related to these answers indicated that some surgeons would use a patch only if the subscapularis was intact or repairable, whereas others would make the decision based on other patient characteristics. Furthermore, some stated that they would never consider the state of the subscapularis when making a decision to use a patch.

Surgeons were asked to consider which patients they would be prepared to randomise in a number of given scenarios. The first scenario proposed a two-arm study, comparing RCR plus a patch with a cuff repair with no patch. The patient characteristics for consideration within this design included patients with medium, large or massive tears, patients having revision surgery, patients aged between 50 and 60 years, 60 and 70 years, 70 and 80 years and those aged ≥ 80 years. The second scenario proposed a three-arm study, comparing a standard RCR plus a specific patch (patch A) with a standard RCR plus a different patch (Patch B) versus a standard RCR with no patch. The same patient characteristic options were given as for the two-arm trial (Table 10). Surgeons had the option of indicating if they would not randomise any patients into either study.

Large and massive tears were the most popular tear size group in both trial scenarios, with > 50% of surgeons opting for these patient groups to be included in a study. Revision surgery would also be considered within a trial patient population. Patients aged between 50 and 70 years make up the most common age range, with patients aged > 80 years being the less favoured to be included in a trial.

Further questions were asked about specific patient characteristics that surgeons felt should be excluded from a trial of patch-augmented RCR surgery. These included what degree of muscle atrophy would need to be present to consider exclusion. The grading of atrophy was based on the Goutallier Classification. 115 Surgeons were also asked to consider the degree of glenohumeral OA that would determine patient exclusion. This grading was based on the Kellgren–Lawrence Classification. 116,117 The presence of cuff arthropathy was also considered in this section of the survey.

Of the responses, 50% indicated the exclusion of patients with grade 3 and 4 atrophy (> 50% of fatty muscle atrophy). However, additional comments provided in this question implied that patients with a degree of atrophy could be included if other characteristics indicated patch use. Three of the responses indicated that atrophy was not an important consideration for patches (Table 11). Over half of the surgeons felt that patients with glenohumeral OA greater than grade 2 should be excluded from a trial of patch use. The presence of cuff arthropathy provided a definitive response as 86% of surgeons indicated that these patients should be excluded from a trial assessing patches in rotator cuff surgery. Additional comments suggested that surgeons would opt for a reverse total shoulder replacement procedure on patients with cuff arthropathy.

Following on from patient characteristics that surgeons felt should be excluded from the trial, participants were asked questions related to the practical elements involved in designing a trial (Table 12). Surgeons were varied in their responses about the timing of randomisation: prior to the day of surgery was favoured by 35% and a further 40% opted for during the operation, once the shoulder pathology had been assessed. The remainder mostly favoured either in the anaesthetic room (5%) or at the start of the operation (15%). One surgeon (5%) suggested that randomisation should be completed at the time that patients are listed for surgery. There were practical reasons listed in the additional comments for wanting randomisation to be completed prior to the day of surgery. Most of the surgeons (85%) felt that an agreed standardised repair technique was optimal for a trial of this nature. However, they also indicated that uptake of this may be dependent on what the technique involves. Use of a specific brand of patch was supported by 75% of surgeons. There was 95% support to standardise postoperative rehabilitation in a trial. There was a fairly even split between opinions on the length of follow-up, with 50% of surgeons suggesting that 12 months would be sufficient and another 40% believing that 24 months would be more appropriate.

A few respondents stated that they were currently involved in or planning a study on patch augmentation.

Strengths and limitations of the surveys

As with all survey-based data collection, there is potential for a response bias. In addition, the survey options potentially limit the generalisability of the findings of this survey. The achieved response rate to the BESS membership survey was low, but not dissimilar to that achieved in similar surveys of clinical professional groups. It is also consistent with responses to other surveys sent to the BESS surgical membership in the experience of the PARCS study team. BESS members, in particular those who are more likely to respond to this survey, are not necessarily representative of the wider upper-limb surgical community and may include more research-oriented surgeons. In this case, respondents may be more familiar with and supportive of patch use in RCR. Although steps were taken to ensure that there was anonymity, it is possible that respondents may have answered questions in a way that did not exactly reflect their personal beliefs. Few respondents were clearly not in favour of patch use. Moving forward, a wider range of upper-limb surgeons could be consulted to increase the robustness of studies such as this. The proportion who participated are probably more interested and perhaps more positive about the use of patches than the BESS surgical community as a whole.

The triallist survey sample was as expected (approximately 20 surgeons) and surgeons were invited personally, based on their previous experience in surgical trials. Their responses should not be taken as indicative of the surgical community as they were selected intentionally to be the subgroup of surgeons who would, in principle, be more likely to participate in a RCT. Furthermore, there may have been other research-active surgeons willing to participate but could not as they were unknown to the PARCS study team. This is a key limitation of the method. The survey was designed to be deliberately short and concise to keep surgeons engaged while also providing relevant data. Patient characteristics that indicate failure in standard RCR surgery could have been explored further in the surgeon triallist survey. Results suggest that surgeons would not be willing to conduct patch-augmented surgery on patients with muscle atrophy nor with glenohumeral OA present, yet it is unknown whether or not these patients would benefit from an augmented repair.

Conclusion

Despite the aforementioned limitations, there were a number of important findings from both of the surveys. It was clear that there are a substantial number of patches in clinical use across the NHS and that this includes patches of different types. Furthermore, a strong theme from the respondents was the lack of evidence to inform current practice, and the need for multicentre clinical trials. There was appetite to partake in a prospective RCT and emphasis that patches commonly used within the NHS (e.g. GRAFTJACKET) needed to be included. The general opinion was that more research is required to inform the use of patch augmentation, with a focus on assessing patient safety and efficacy. From the triallist survey, there was some agreement in the patient population and outcome timings, but what the final trial design incorporates remains uncertain. How a patch is used and at what point the decision is made to use one were also highlighted as questions requiring further exploration. The uncertainty around these aspects was taken forward into the consensus process.

Recruitment of patients and carers

Consultant orthopaedic surgeons (and PARCS investigators) based at the Nuffield Orthopaedic Centre in Oxford and the James Cook University Hospital in South Tees approached potential participants through their units. Patients had to meet the following inclusion criteria to be approached: willing and able to give informed consent for participation in the study; aged ≥ 18 years; have the ability to understand and communicate (read, speak, and write) in English at a level that permits effective interaction; and have active or previous shoulder problems involving the rotator cuff. Patients were invited regardless of whether or not they had undergone surgery, and their carers were also invited if available. At each site, a balance of men and women, and of patients and carers, was sought. A convenience sampling approach for recruitment was used from the Oxford-based patient clinics. At South Tees, potential participants were identified from (1) a list of patients who had previously registered interest to be involved in future patient and public involvement (PPI) activity as part of the pre-grant submission PPI evaluation conducted for the PARCS study, (2) a list of patients who received treatment for rotator cuff in the past 3 years within the unit and (3) a small number identified by prospective convenience sampling of patients attending the investigator surgeons’ outpatient clinic. Once identified, all participants were approached directly at their outpatient/physiotherapy appointments or via telephone by the PPI co-applicant at this site.

Those who responded positively and met the inclusion criteria were included in the PARCS study focus groups. Identification of potential participants continued until the target number of participants had been recruited (between four and eight participants per focus group).

Recruitment of other stakeholder groups

Representatives of regulatory bodies and industry were identified via local networks, approached and invited to participate directly. We also used snowballing as information was passed on to potential participants via professional acquaintances. 132

Focus groups

Ahead of the focus group session, potential participants were provided with a study information sheet specifically tailored to their stakeholder group, describing the aim of the focus group and what to expect. Each focus group lasted around 2 hours, with a break for refreshments. The cost of travel was reimbursed, and participants were offered a shopping voucher to the value of £25 as a token of appreciation. Each participant was asked to personally sign and date a consent form, of which they were given a copy.

During each focus group session, the aims of the PARCS study, and focus groups specifically, were briefly introduced. Participants were asked to consider a number of key issues, scenarios or vignettes. These included key items of information about the possible trial design options, such as the different kinds of patches available and their acceptability, the choice of comparative study arms, the most appropriate outcome measures, and the methods of data collection. The way in which this information was delivered was adapted according to participant group. For example, a more technical approach was used for the industry stakeholders. Participants were provided with the following definition of a patch: ‘an implantable human, synthetic or animal material which is used with the aim of improving tissue healing and patient outcome’.

We also asked focus group participants to provide some basic background information about themselves (i.e. sex, age, relevant experience and previous treatments). This information was anonymous. Participants were provided with a plain opaque envelope in which to place the completed ‘background information form’. They were asked to place this envelope into a box as they left.

The focus groups were facilitated by an appropriately trained member of the PARCS study team (NM). Discussions were audio-recorded and one or two observers (LK, JCC or CC) took notes to aid in the transcription of audio files and analysis. The resulting audio files and accompanying field notes were transcribed verbatim. Transcription was performed by an external transcription company. Any identifying information appearing in focus group transcripts was removed as soon as possible following transcription to minimise the risk of participant identification. The audio files and transcripts were encrypted for secure storage.

Analysis

The focus group transcripts were analysed by two members of the PARCS team (CC and JCC) using thematic analysis. 133 The following steps were followed:

1. Analysts familiarised themselves with the transcript.

2. One analyst (CC) initially coded the transcripts by hand, both deductively (guided by themes included in the focus group topic guides) and inductively (allowing unanticipated themes and subthemes to emerge). The first three transcripts were also independently coded by a second analyst (JCC). The codes were subsequently compared, discussed and agreed with the first analyst. The emphasis of the analysis was on the acceptability of the proposed trial and on factors that might influence such acceptability.

3. Analysts reviewed the coding and agreed a working thematic framework. This was applied to subsequent transcripts using QSR NVivo 10 software (QSR International, Warrington, UK) and evolved as analysis progressed. In accordance with the study design objectives, the themes were grouped into PICO elements, that is their relationship to patient population, intervention, control and/or outcome (including timing of measurements). The framework also included practical considerations to take forward.

The final thematic framework and its content were used to inform the questions posed in the triallist survey, which followed the focus groups (see Chapter 3).

Focus group participants

A total of 24 people took part in the focus groups (Table 13). Thirteen participants were patients who had previously been treated in the NHS for their shoulder problem. Ten of the patients had undergone shoulder surgery. Five patients reported ongoing shoulder problems (of whom three had had surgery), whereas two reported that their problems had resolved (of whom one had had surgery). Two carers (partners) of patient representatives also participated.

Views on patient population for a trial assessing patches in rotator cuff repair surgery

Patients and regulatory body representatives were eager to discuss which characteristics patients would need to be present for the patch to be a treatment option:

To me it’s got to go on the size of the tear and how much discomfort they’re getting because you can’t treat everyone the same.

P504 – patient, South Tees

Would age be a big factor in this?

P505 – patient, South Tees

So, the surgery, the inclusion criteria, would you say that they’ve done everything possible, like physio?

P401 – regulatory representative, Ethics

Some patients felt that all patients having RCR surgery should be offered a patch:

. . . so you’re going to get . . . you’re getting a high percentage [of success] at the moment, 60% [success rate after standard RCR] as against 40% [re-tear rate after standard RCR], so you’ve already got that percentage of success so my argument is if you’ve got that surely anybody that’s prepared to take the trial should just go for the patch . . .

P101 – patient, Thames Valley

It seems silly, like you said, why aren’t they just using [patches] now, if they’re using [patches] elsewhere in the body . . .

P506 – patient, South Tees

Other participants felt that the patient population needed to be more specific. Industry representatives were strong in their opinion that, for a patch’s effectiveness to be assessed, confounding factors and certain patient characteristics should be excluded:

. . . in the end it’s a clinical decision and also the patient, patient age, bone quality, tissue quality, smoker or not, again . . . keep them out of the way. Steroid abusers need them out the way as well.

P203 – industry representative, company 2

To me it’s got to go on the size of the tear and how much discomfort they’re getting because you can’t treat everyone the same [each patient is different].

P504 – patient, South Tees

Patients’ reactions to being hypothetically offered a patch were generally positive, as they focused on the idea that patches augment repairs and, therefore, ‘can only be better’ (P101 – patient, Thames Valley). Patients had a risk-seeking approach to manage their pain:

Can I go on your list first to have it done, please? (Addressing Investigator 3)

P504 – patient, South Tees

I mean it’s not as if you’re having a heart transplant or anything like that which you would discuss, or cancer, it’s not is it, it’s something which is just . . . you’ve got pain there and you know there’s a relief, you can have an operation and the extra patches, yeah, let’s go for it.

P501 – patient, South Tees

So something different [like patches] might just assist it.

P103 – patient, Thames Valley

However, there were also some cautious questions from the patient groups about what happens to the patch after the operation:

Does that stay in then, that patch, or not?

P506 – patient, South Tees

Are the patches likely to break down in compound over a period of time?

P103 – carer, Thames Valley

There were no apparent negative views towards using a patch in rotator cuff surgery. However, some patients expressed a preference for human-derived patches. One patient was concerned about the safety of animal-derived patches:

I would probably steer to one that comes from human beings . . . probably not cows and pig due to CJD [Creutzfeldt–Jakob disease] . . . foot and mouth because we don’t really still understand those two diseases . . .

P102 – patient, Thames Valley

Views on potential interventions and controls in a trial assessing patches in rotator cuff repair surgery

Focus group participants had varying perspectives on what a trial intervention and control group should look like. The issue of randomisation was discussed, as were patch elements and types that may be used within a trial. These topics were then related to willingness, of both industry and patients, to participate in a trial.

Industry stakeholders had variable opinions on what a trial of patch use should look like. One recognised the need to provide an evidence base before progressing with patch development further:

I think [a patch A vs. patch B trial] is something a bit down the line isn’t it, I think really we need to know that this [patch] works in the first place before you start enhancing it.

P202 – industry representative, company 2

Other industry representatives felt that current surgical practices had evolved further than this and that current trends needed to be represented in any study design involving the use of patches in rotator cuff surgery:

If clinical practice is moving onto augmentation already for these sorts of cuff tears, the question in the industry is who’s got the best patch . . .

P203 – industry representative, company 2

I think [surgeons] would want to know which avenue to take as opposed to cuff versus patch as a broad thing, I think we’d want a little bit more than that, I think we’re at that position now.

P202 – industry representative, company 2

There was also discussion among the patients about whether or not there was a need for a RCT at all, given the known failure rate of standard repair and widespread use of the patches in other parts of the body:

. . . why you’re doing this research is basically because of the patches, so put the standard repair out of the window as far as I’m concerned and go right, we’re doing nothing but patches here, that’s all it’s about.

P501 – patient, South Tees

If the normal process [standard repair] was only available and there was nothing else then I’d probably go down that [trial participation] route but because of listening to the percentages [40% failure rate for standard repairs] I still don’t see the logic of having that [repair with no patch] as against the patch. Because you’ve moved on . . .

P101 – patient, Thames Valley

If you know the success rate of a patch on a certain [different] part of the body is 99% or whatever or 90% then you go, ‘yeah, we can safely use that one as a guide’.

P501 – patient, South Tees

In line with their initial risk-seeking reaction to being hypothetically offered a patch, some patients said that they would seek to participate in a trial, regardless of what treatment they would potentially receive:

I’d go into trial to get rid of the pain, regardless of what they were going to do, I just have to have something done.

P107 – patient, Thames Valley

Excerpt from South Tees patient focus group:

If [it] was thought I’d have 50 : 50 chance of getting the patch or whatever even to have the normal surgery, I’d go for it.

P506 – patient, South Tees

Yeah, so would I.

P504 – patient, South Tees

Because it’s something or nothing.

P506 – patient, South Tees

Yeah, yeah, you can’t lose.

P504 – patient, South Tees

However, consistent with their positive reaction to the idea of being offered a patch, patients who had experienced numerous rounds of conservative treatment tended to dislike the concept of randomisation and the consequential risk that they may not receive a patch:

. . . [if] people decide to go for the trial they’re not going to know which one they’ve had [patch or no patch]. I wouldn’t even go down that route . . . If I had the patch I’m quite prepared to go down that route, but I’m not prepared to go down the route of I don’t know what’s happening.

P101 – patient, Thames Valley

The CRN discussions highlighted the importance of considering existing patch use at participating trusts when determining the content of the trial arms:

So if you standardise and say you have to use, say, these five [patches] and that’s not what they’re using then there could be a sourcing issue.

P302 – regulatory representative, CRN

Whatever they use as part of standard care if they’re just using that and they’re just taking another one [patch] off the shelf it’ll go through [local trust approvals] a lot easier.

P303 – regulatory representative, CRN

The willingness of those in the industry to offer their products to a potential trial of patch augmentation was complex. Their involvement would depend on all the PICO elements that were proposed and what the definition of a patch was within the trial. The type of repair products used during the cuff repair was flagged up as an important element by one member of the industry focus group, in this case the anchors and tape (to anchor down the tendon) used in the operation:

There are just too many variables out of our control and potentially out of your control as in the ideal study, playing devil’s advocate, so the ideal study would be [surgeons] all have to use [company Z] anchors with tape ‘cause it’s best [general laughter] with their [company Z] artificial patch or [company Z] dermal graft and that’s the study, that’s easy.

P203 – industry representative, company 2

Manufacturing processes, rehabilitation guidelines and the purpose of a patch would also need to be considered for industry representatives to participate and willingly put forward their product to be assessed:

The way [patches are] kind of weaved and that kind of thing all affects its strength and how it works, how it then heals in the patient that kind of thing . . . so we probably wouldn’t want ours bunched in with [other patches in one treatment arm] . . .

P202 – industry representative, company 2

The industry group highlighted the importance of declaring conflicts of interest and of ensuring that there was transparency of involvement in any trial.

The regulatory body representatives also expressed some concerns about the components of a patch to be used within a trial and how these could exclude some communities within an otherwise eligible patient population.

Excerpt from regulatory representatives’ discussion:

That’s going to raise some practical difficulties though, because if you’re at a centre and they are using patches that have got animal products in, and then they have a patient consent to take part in a study, but then they don’t want one of those particular patches, will you, the study, then provide [non-animal derived] patches to that centre? . . . Or you exclude that patient.

P401 – regulatory representative, ethics

[Excluding those patients] can be difficult. For instance, we’ve had studies where they use a standardised fatty diet, which contains bacon, and they’ve actually excluded anyone who won’t eat bacon. . . . Our [ethics] committee did actually have an issue with that, because, obviously, that excludes all Jewish people and Muslim people. So, then we were saying, ‘Well, hang on, are you denying access to clinical trials for taking part in research to particular ethnic or religious groups?’ It’s just something that, if that’s the direction you’re going to go down, you just have to be prepared to have, to make, that argument.

P402 – regulatory representative, ethics

Views on the outcomes to explore in a trial assessing patches in rotator cuff repair surgery

Stakeholder opinions on outcomes of importance appeared to stem from their own experiences and knowledge about treatment and products. Patients highlighted that pain was the most important issue to address with treatment, regardless of what that treatment involved:

That’s the bottom line . . . that pain is the bottom line and that’s it.

P501 – patient, South Tees

If I didn’t have the pain that I’ve got but I had the restricted movement I could probably live with that, it’s living with the pain.

P506 – patient, South Tees

Certainly pain relief. I think that’s the main issue with all these things because how much pain you can tolerate . . .

P101 – patient, Thames Valley

Other outcomes that were important to patients included sleep and being able to perform everyday activities.

. . . my pain was different actually because it manifested itself at night, I couldn’t sleep.

P104 – patient, Thames Valley

No, I was the same, I didn’t sleep for about 6 months.

P107 – patient, Thames Valley

During the day, for some reason, it didn’t really irk me but at night it really was a problem.

P104 – patient, Thames Valley

Mine was . . . about 6 months I didn’t sleep. Because the pain was so bad.

P107 – patient, Thames Valley

You might be doing things around the house like women do and I know men do as well, and you think ‘oh, I’ll just go and do that because it’s all right’ and you do it and you think, ‘ooh, I shouldn’t have done that’, that sort of thing. It catches you off the cuff . . .

P102 – patient, Thames Valley

Patients, along with industry representatives, discussed the importance of preventing further surgery. With the current re-tear data in mind (40% re-tear rate following standard rotator cuff surgery), there was the feeling that treatment needed to be definitive and not used as a back-up if the initial repair failed:2

Safety seat belts, maybe it’s overkill but the patient wants one procedure not two. [In discussion about why a patch would be used.]

P203 – industry representative, company 2

I think [patches] should be used from the start.

P506 – patient, South Tees

If it goes again how many times can you actually have that repaired the ordinary way before you end up with so much scar tissue that it’s not going to happen at all? As against if you have the patch you’ve got a higher rate of success and you may not ever have to have it done again . . .

P101 – patient, Thames Valley

Outcome measurement was also discussed in some focus groups. Potential methods ranged from radiological imaging (‘you’re not going to get a second look’ – P203 industry representative) to patient-reported questionnaires, safety measurements and reporting of failure (i.e. further surgery). Regulatory stakeholders highlighted the importance of product safety monitoring and minimising the burden of follow-up for patients:

Keep the study as simple as possible, so that the burden on the patient is as little as it can be. So, things like biopsies, if that’s what you need to be able to answer your outcomes, then do the biopsies and explain it, but if you don’t need to do the biopsies to answer your question, then I would say consider them very carefully. Also, I don’t know if we addressed how painful they are. If you can do something that’s quick and painless, then that’s not really a burden to the patient. If the biopsies are going to be as painful as, like, a lumbar puncture, then make sure you’ve got a really, really good reason for doing it.

P402 – regulatory representative, ethics

So, the key thing I’ve got is about safety and monitoring, because whatever patch you do, or whatever CE [Conformité Européene] mark, how do they currently monitor whether it’s doing the job it’s meant to be? Obviously, there’s subjective [indicators], where you fill out questionnaires . . . but showing that the patch hasn’t moved or slipped, do you do imaging?

P401 – regulatory representative, ethics

Practical considerations when designing a trial assessing patches in rotator cuff repair surgery

Practical elements of trial design and trial set-up were discussed within the PICO themes above. Participants also suggested other considerations in trial design and set-up, including what information about patches and treatment should be presented to patients when informing them about the trial.

The patient and the regulatory focus groups raised questions about the type of information patients may seek about patches, how the patch is physically inserted into the shoulder and what other research is being carried out:

Have you actually tested the patches on anybody else yet or . . . ?

P101 – patient, Thames Valley

Are the patches likely to break down in compound over a period of time?

P104 – patient, Thames Valley

Is your research part of a worldwide programme or is it just part of [named academic institution]?

P102 – patient, Thames Valley

I think you could say that there’s different patches, and then tell them to speak to their doctor about which one they’re likely to get at their centre.

P402 – regulatory representative, ethics

Patients also raised some questions about the safety that would need to be considered when developing patient information. This included the permanency and ease of removal of a patch:

Does that stay in then, that patch, or not?

P506 – patient, South Tees

If you have the patch done and it really gives you hell afterwards is it easy enough to get out again?

P105 – patient, Thames Valley

Strengths and limitations

This qualitative study enabled us to elicit and understand the views of a broad range of stakeholders, including patients with experience of shoulder problems in northern and southern England. This would not have been possible through the use of crude surveys. The focus group participants included a balanced mix of men and women, and the patients had varied experiences of shoulder surgery (no surgery, surgery without a patch and surgery with a patch). The participation of industry representatives was a strength of this study, providing an insight into product development and use that is not often available to people outside the clinical setting.

However, because of the time-limited nature of the focus groups, we were unable to explore some complex issues in more depth. For example, the circumstances under which patients would or would not be willing to take part in a RCT of patch-augmented surgery were touched on but were not exhaustively investigated. The study team acknowledge that this could be an issue and will review it further during the design of recruitment and training processes in a definitive trial. Participation in the focus groups depended on patients’ ability and willingness to travel to NHS sites outside routine appointments and on a weekday, potentially excluding those with daytime commitments or lacking transportation. It is possible that focus groups held at a community venue, at a weekend or in a different part of the UK may have resulted in a wider range of views being captured. The semistructured discussions with regulatory body representatives (CRN and ethics committee members) would have ideally included more participants. However, there were challenges in identifying appropriate people who were working in regulatory bodies, such as the Medicines and Healthcare products Regulatory Authority. The study team were also unable to organise a discussion with people who work in procurement in the NHS.

Conclusion

Overall, there was general support for patch use in RCR surgery from all stakeholders who participated in the focus groups. There was acknowledgement of the risk involved in receiving an implant such as the patch, and advocacy for product safety monitoring during the trial. However, there was some discrepancy among stakeholders about what the patient population, intervention and control arms of the trial should be. Patients may be unwilling to participate in a trial with a ‘no patch’ control arm if access to a patch is available within routine care. Improvement in pain and function and preventing further treatment were important outcomes identified in the focus groups. How and when these would be measured in a trial needs further exploration. The uncertainties and further questions that emerged from the focus groups were used to inform the surgeon triallist survey (see Chapter 3) that followed. These included (1) who should have a patch, and how are patches used by surgeons? and (2) what could a patch be compared with within a trial?

Stage 5: Delphi study

A Delphi study to develop a consensus on the best way to design a clinical trial of patch-augmented rotator cuff surgery was conducted. The Delphi method is a structured process of obtaining information from a group of experts using a series of related questionnaires, each one refined using respondents’ feedback from a previous version. 143 It is a well-known and increasingly common method used in the clinical setting to establish a consensus. 130,143,144 A multiple-stage online Delphi survey consisting of at least two but no more than three rounds was conducted. The surveys were developed and conducted online.

Participants involved in stages 2–4 of the PARCS study were invited to take part in stage 5, according to stakeholder group.

Given the nature of the study, there has been no formal sample size calculation, but around 50–80 responses were originally anticipated. 1 There are generally no accepted guidelines for the optimal sample size needed to achieve consensus in Delphi studies. 145 This sample size was based on previous experience of conducting this type of survey and anticipated attrition rates at each round. Substantial loss from the initial to the final round is not unusual. 144,146

All participants invited into the Delphi survey had been involved in earlier stages of the project. The Delphi study proposed different options for elements of RCT design. The aim of this was to help reach a degree of agreement to take forward to the final consensus meeting. Findings from stages 1–4 were used to determine the trial design elements to be included in the first round of the Delphi study.

Initially it was intended for all stakeholders to take part in both rounds. Based on feedback received from the patient representatives on the project team, the initial round was sent to the non-patient stakeholders only (i.e. surgeons, physiotherapists, industry and researchers). This was because of concern that the initial round would be too ambiguous and would be unfair to the individual to assess. Patients were involved in the second round only (a third round was not anticipated to be necessary once the initial round had been outlined). Two versions of the second-stage survey were used, one for patient and public stakeholders and one for professional stakeholders (e.g. surgeons and researchers). The patient and public version had a subset of the full set of questions that were most pertinent to this stakeholder group and were presented using more accessible language and avoiding, as far as possible, technical terminology.

An e-mail was sent to each participant containing a personalised link that enabled access for convenient survey completion. For a subset of potential participants’ paper copies were sent as per the participants’ preference, or a generic link was sent to a specific group of stakeholders.

During completion of the first round, survey participants were asked their stakeholder group and their place of work (professional stakeholders only). Data were extracted to Microsoft Excel and summarised for data analysis.

Participants were presented in the survey with proposed trial design elements (e.g. choice of two- or three-arm trial design, eligibility criteria for participation and information on the timing of the outcome data collection) and asked to score agreement with each using a 1–5 scale, where 1 represents complete disagreement and 5 represents complete agreement. A ‘do not know/not relevant’ option was provided as applicable. Participants were given the opportunity to comment further in free-text boxes. The aim was to allow the justification of any answers and the communication of new or proposed adjustments to the design elements. No new elements were suggested by participants in round 1 of the Delphi survey, although a number of the existing ones were revised.

For each section on the second round Delphi study, a summary of the findings from the previous round was offered. Participants then scored their agreement on each design element. The final set of proposals, areas of provisional consensus and remaining disagreement and uncertainty were then taken forward to the consensus meeting in stage 6 and used as the basis for discussion.

Where necessary, in both rounds of the Delphi survey non-responders received a maximum of two reminder messages. The final reminders contained a specific deadline for survey closure. 147 Each survey took approximately 15 minutes to complete. Scores (range 1–5) from each round were calculated as a percentage of the total responses. Consensus for a proposed design element was defined as > 70% of responses scoring the element 4 or 5, and no more than 15% of responses scoring the element one. Median and ranges were also produced for the scores. We explored similarities and differences across stakeholder groups. Textual responses were summarised narratively.

Stage 6: consensus meeting

Findings from stages 1–5 were fed into, and informed the structure of, a 2-day face-to-face consensus meeting. This meeting sought to agree on an acceptable and feasible trial design for a definitive RCT to assess the clinical effectiveness and cost-effectiveness of a patch to augment surgical RCR. A range of stakeholders were involved, including surgeons, triallists and patient and public representatives, who took part in stages 2–5 of the study. An independent academic surgeon was invited to chairperson the meeting. Participants were selected for invite based on their perspectives and experience to ensure that there was a variety or representation. For example, surgeons who do currently use patches to augment RCR were invited along with those who would be potentially willing to do so for a trial. To ensure that a robust decision was made, approximately 30 stakeholders were invited to participate. 1

Ahead of the consensus meeting, participants were sent a proposal of a trial scenario for consideration, based on the results from the Delphi study. Patient and public representatives were reimbursed for expenses and compensated for their time. 148 The meeting was structured to ensure that key areas of uncertainty and disagreement were reviewed and discussed. Consensus on key elements of the trial design was sought, namely patient eligibility, intervention and control definitions, surgeon requirements, outcomes and target difference. Previous stages informed draft guidance, options and recommendations for a RCT assessing patch-augmented rotator cuff surgery. A post-meeting report was drafted and circulated to participants for their review and comments. The report included details of the key design decisions made at the meeting with points for clarification.

Stage 5: Delphi study

Twenty-nine out of the 43 (67%) individuals who were invited to take part in the Delphi study took part: 18 surgeons, two industry representatives, four physiotherapists, two medical researchers and three patient representatives. Table 14 gives a breakdown of the groups and the location of work for the participants. Round 1 had 23 responses and round 2 had 24, including three patient participants. Findings from the two rounds are presented in turn below.

Stage 6: consensus meeting

The consensus meeting was held on 29 and 30 January 2019 in Oxford. A total of 22 individuals attended the meeting: representatives from industry (2), patients (2), surgeons (8), researchers (4), project team members (5), and an independent chairperson. Prior to the meeting, the agenda and a summary of a trial proposal using the round 2 Delphi survey were circulated. The proposal had only minor adjustments from the round 2 Delphi survey, with the explicit age restriction removed and imaging clarified to take place only at 12 months. In addition, a number of issues related to each component raised in the Delphi study were included in the proposal, highlighted for discussion in the meeting. Towards the end of the first day it became clear that the proposal trial design needed revision to address the varied use of patches and a revised proposal was presented at the beginning of the second day. The overall decision of the meeting was that a RCT of RCR with a patch was both acceptable and feasible. However, it was agreed that a single study or randomised comparison would not be able address all the key questions. The outline of two RCTs was agreed in principle at the meeting, as was the desire for a study to capture safety data for patch use irrespective of involvement in a trial (e.g. registry study). Following the meeting a summary of the proposal was circulated and agreed.

Study proposal

An outline of the agreed study proposal is given in Tables 17 and 18. Two areas for a RCT assessing the use of patches in rotator cuff surgery were identified. The first one assesses the use of a patch for patients with a completed RCR and the second relates patients with a partial RCR. Figure 10 illustrates the patient flow, along with patients who potentially could contribute data to a registry to inform on the safety of specific patches. The two randomised comparisons could potentially be within one more comprehensive trial or conducted separately.

FIGURE 10.

Proposed study patient eligibility pathway.

Patient eligibility

Patient eligibility provided the most disagreement in the study. A number of aspects were agreed on, such as the exclusion of patients with other shoulder problems and with clinically significant OA. There was a variety of specific exclusion criteria proposed by individual respondents, with a range of views on age, muscle atrophy, tear size and having a previous RCR. The relationship between patient eligibility and patch technique was noted by a number of participants. The final proposal reflects an inclusive approach whenever possible. The broad criteria were accepted, although the details were not fully resolved.

Types of patches

A variety of patches are available for clinical use, reflecting different materials, processes and designs. The systematic review (see Chapter 2) identified 28 different patches, defined as an implantable human, synthetic or animal material that is used with the aim of improving tissue healing and/or patient outcome via some form of mechanical support. Of these, 22 could be classed as a product, and six were a tissue graft from either the patient or (in one case) a cadaver. There was comparative evidence for only 12 different patches. The survey of the surgical BESS membership identified 13 different patches currently in use in the UK. Most are produced from decellularised human dermis, with the rest being made form porcine or synthetic materials. It is concerning that these two groups did not overlap fully, with at least two patches (dCELL and Leeds–Kuff) in current clinical use that have no clinical evidence identified. No single type of patch could be considered either to be dominant in terms of use or to have compelling evidence in its favour. There was some evidence for the support of allograft and synthetic patches to improve the re-tear rate and for synthetic patches to reduce pain. Mostly the evidence is non-existent, or too weak to draw even tentative conclusions.

With regard to running a large definitive trial, it was clear that the use of a specific patch would be unwarranted but a decision about which types of patch to allow within the study would be important. Any secondary evaluation of evidence, such as safety of the included patches, would be advantageous. Uncovering the need for preliminary safety evaluation of patches was unexpected in this study. The requirement for a non-animal patch in the trial design, particularly an alternative to a porcine based patch, was noted.

Surgical use of a patch

The surgical use of patches in rotator cuff surgery falls broadly into two groups. In an on-lay technique the patch is sutured on top of the tendon-to-bone repair, whereas in a bridging technique the patch is sutured into the exposed area following a partial repair. Although to some degree a partial repair can be anticipated in advance, this is not always the case, as the quality of the tissue is not entirely clear until the repair has been attempted.

The distinction between the two techniques was clear in some of the studies in the systematic review. This became increasingly apparent as surgeon stakeholders were engaged through stages 2–6. To reconcile the two techniques, two RCTs for patch use were proposed: one for on-lay use and one for bridging. Different patch types and designs may be thought to be better suited to each approach.

Beyond this, there was overall support for flexibility for surgeons to conduct patch-augmented RCR according to their personal practice. A standardised postoperative regime was supported by the surgeon triallists in stage 4.

Conducting a randomised controlled trial

Specific information related to how a RCT might be conducted was gained from the surgeon triallist survey, the Delphi study and the consensus meeting. The Delphi study and the consensus meeting showed most support for a two-arm trial of RCR with and without patch use. Although differences between patch types were noted by a number of participants as being of interest, the difference in how patches are used appeared to be a higher priority for evaluation. This is reflected in the research recommendations of the study.

Randomisation during the operation was most supported, once the RCR had been attempted. This has the benefit of confirming the presence and nature of a tear and knowing whether or not the repair could be completed.

Participants in the surgeon triallist survey, Delphi study and consensus meeting supported use of both a patient-reported pain and function measure and imaging as trial outcomes. There was support for a 24-month follow-up. The timing of assessments within this period indicate support for an ‘early’ assessment around 4 months, followed by further assessments at 12 and 24 months. It would not be necessary to have the same outcomes at every time point.

Economic evaluation considerations

A future definitive RCT of patch use should consider embedding an economic evaluation of the patches under investigation to assess their cost to the NHS as well as their benefits to patient HRQoL.

Little evidence was available to shape a RCT economic-based evaluation. Considerations about the types of data (e.g. inpatient/outpatient visits, medication), frequency and intensity of patient data collection (e.g. 1 and 4 months post surgery) as well as the means of data collection (e.g. paper/electronic forms) have to be made when designing a prospective RCT. Resource utilisation not related to the surgery, such as use of non-NHS care to deal with daily activities, as well as loss of income caused by surgery, are also recommended to be included in the patient data collection process. Although SF-36 was the only HRQoL measure to be captured in the current literature, a future economic evaluation should also administer the EQ-5D measure, which is recommended by NICE.

Further research priorities

As outlined in Chapter 5, there were four clear research questions that require evaluation. Proposed research studies to address these are as follows:

1. a RCT to assess the on-lay use of a patch in rotator cuff surgery once the surgical repair of the rotator cuff has been completed

2. a RCT to assess the bridging use of a patch where the surgical repair of the rotator cuff is partial

3. a registry or observational study to assess the safety of all patches available for clinical use

4. an economic evaluation of the use of patches in RCR preferably embedded within a RCT or as a standalone evaluation.

An outline of the proposed studies is given in Tables 17 and 18 and in Figure 10.