Barriers and enablers to patient recruitment for randomised controlled trials on treatment of chronic wounds: A systematic review

Abstract

Randomised controlled trials represent the gold standard in intervention efficacy evaluation. However, suboptimal recruitment affects completion and the power of a therapeutic trial in detecting treatment differences. We conducted a systematic review to examine the barriers and enablers to patient recruitment for randomised controlled trials on chronic wound treatment. Review registration was under PROSPERO 2017:CRD42017062438. We conducted a systematic search of Ovid MEDLINE, EBSCOhost CINAHL, Ovid Cochrane Library, Ovid EMBASE, and Ovid PsycINFO databases in June 2017 for chronic wound treatment randomised controlled trials. Twenty‐seven randomised controlled trials or qualitative studies met the inclusion criteria. Among the 24 randomised controlled trials, 21 were assessed as low quality in relation to recruitment, and 3 were assessed as high quality. All 27 studies reported barriers to recruitment in chronic wound randomised controlled trials. The reported barriers to recruitment were: study‐related, patient‐related, clinician‐related, health system‐related, and/or operational‐related. No study reported recruitment enablers. To enhance randomised controlled trial recruitment, we propose the need for improved integration of research and clinical practice. To alleviate the problems arising from inadequate reporting of randomised controlled trials, the Consolidated Standards of Reporting Trials Statement could include an additional item on recruitment barriers. This approach will allow for increased awareness of the potential barriers to recruitment for Randomised controlled trials (RCTs) in both wound management and other health care research.

1. INTRODUCTION

1.1. Rationale

Chronic non‐healing wounds are a growing and costly problem to the health system in industrialised countries.1 Randomised controlled trials (RCTs) represent the gold standard in evaluating the efficacy of therapeutic interventions for successful treatment. However, among the key challenges in chronic wound research is the timely completion of clinical trials that often experience slow or insufficient enrolment. Recruitment to chronic wound RCTs is particularly challenging because of the diversity of aetiologies and patient comorbidities, the plethora of treatment options, and varied outcome measures.2 These factors may impact the ability to adequately conduct a high‐quality RCT, which may lead to inaccurate conclusions that do not contribute to knowledge.2

Recruitment to RCTs requires a balance between achieving full enrolment and ensuring an appropriate study sample. A review of 114 multicentre clinical trials reported that less than a third of included studies achieved the target sample size. In addition, more than half reported that a time extension was needed to complete the recruitment target.3, 4 Clinical trial investigators face a number of challenges, and the failure to address the challenges means that novel and innovative interventions that have the potential to improve patient outcomes and decrease health costs may take longer to be implemented and may come at a higher cost. Trial results that may not be applicable beyond the study was highlighted by the streptokinase RCT, where an estimated 10 000 unnecessary deaths occurred from acute myocardial infarction following delays in recruitment.5

Barriers to recruitment have been identified as study‐related, participant‐related, researcher‐related, and organisational/health system‐related.5, 6 Study‐related barriers include narrow inclusion and exclusion criteria and overestimation of potential recruits. Comorbid conditions, language proficiency, cognitive ability, and patient's lack of interest are some of the participant‐related factors.7 Researcher‐related factors include differences in perception between principal investigators and research nurses.7 Organisational factors include a busy clinical practice, lack of clinician time, and inadequate incentives for clinician to engage in recruitment.5, 6, 7

Reported enablers, or facilitators, of trial recruitment include: a motivated and experienced study team; good coordination and communication between the research team, clinicians, and recruitment sites; “real‐world” inclusion criteria; a dedicated research study nurse at each site; coordinated trial management; positive trial publicity; and study trial information that is clear and simple.6, 7 Reported strategies to facilitate recruitment include trial extension, recruitment criteria revision, and adding extra recruitment sites.

1.2. Objectives

The aim of this study was to identify and examine the barriers to and enablers of patient recruitment in chronic wound RCTs.

2. METHODS

2.1. Protocol and registration

This systematic review was conducted in accordance with the preferred reporting items for systematic reviews and meta‐analyses (PRISMA).8 The protocol registered with PROSPERO 2017:CRD42017062438.

2.2. Eligibility criteria

We included RCTs and quasi‐randomised trials that evaluated interventions for adult patients (as defined by study authors) undergoing treatment for a skin ulcer, leg ulcer, varicose ulcer, and diabetic foot ulcer. We also included studies with a qualitative research design that examined the issue of recruitment to RCTs on chronic wound treatment. We excluded intervention studies for the treatment of pressure ulcers, surgical wounds, or burns as these were considered to have a different aetiology. Studies were also excluded if they were conducted in an inpatient hospital setting for the trial duration.

2.3. Information sources and search strategy

We searched 5 electronic databases: Ovid MEDLINE (January 1, 1997–June 12, 2017), EBSCOhost CINAHL (1982–June 12, 2017), Ovid Cochrane Library (inception—June 12, 2017), Ovid EMBASE (1974–June 12, 2017), and Ovid PsycINFO (1806–June 12, 2017).

Medical Subject Heading terms and keywords associated with the concepts of “wound healing,” “patient selection,” and “randomised controlled trial” were used to develop the search strategy (Table 1).

Table 1.

Search strategy

1 Wound Healing/ 2 (wound adj4 healing).tw. 3 1 or 2 4 Re‐Epithelialization/ 5 (reepithelialization or re‐epithelialization).tw. 6 4 or 5 7 Ulcer/ 8 ulcer*.tw. 9 7 or 8 10 Leg Ulcer/ 11 (leg adj2 ulcer*).tw. 12 10 or 11 13 venous ulcer/ 14 (venous adj3 ulcer*).tw. 15 13 or 14 16 Varicose Ulcer/ 17 (varicose adj3 ulcer*).tw. 18 16 or 17 19 ((stasis adj2 ulcer*) or (lower adj3 extremit* adj3 ulcer) or (crural adj2 ulcer*)).tw. 20 foot ulcer/ 21 (foot adj2 ulcer*).tw. 22 20 or 21 23 diabetic foot/ 24 (diabetic adj2 foot).tw. 25 23 or 24 26 skin ulcer/ 27 (skin adj2 ulcer*).tw. 28 26 or 27 29 (ulcer adj2 healing).tw. 30 (chronic adj2 wound*).tw. 31 (complex adj1 wound*).tw. 32 (wound* adj2 care).tw. 33 3 or 6 or 9 or 12 or 15 or 18 or 19 or 22 or 25 or 28 or 29 or 30 or 31 or 32 34 Patient Selection/ 35 (patient adj2 selection).tw. 36 34 or 35 37 (patient* adj4 (recruit* or allocat* or enlist* or enrol* or involv* or invit* or accru* or screen*)).tw. 38 (participa* adj4 (recruit* or allocat* or enlist* or enrol* or involv* or invit* or accru* or screen*)).tw. 39 37 or 38 40 36 or 39 41 Randomized Controlled Trial/ 42 (rct* or randomized or randomised).tw. 43 41 or 42 44 Clinical Trial/ 45 clinical trial*.tw. 46 44 or 45 47 43 or 46 48 33 and 40 and 47 49 limit 48 to yr="1997 ‐Current"

Database: Ovid MEDLINE(R) 1946 to present with daily update.

2.4. Study selection

We collated the results of all databases using a reference management software (EndNote X8, Clarivate Analytics, 2016), and duplicates were removed using a standard function. We exported the remaining records to an online systematic review platform (Covidence, Veritas Health Innovation Ltd, 2017). Two reviewers (L.B. and R.M.) independently screened the titles and abstracts. The full texts of studies that appeared to fit the eligibility criteria were reviewed independently by 2 reviewers (L.B. and J.K.L.) for potential inclusion. We conducted a bibliographic review of included studies to identify additional studies that may meet the inclusion criteria. We resolved discordance between reviewers by discussion, and when necessary, a third reviewer (R.M.) helped reach consensus.

2.5. Data collection process and data items

Two reviewers (L.B. and J.K.L.) reviewed the full text of included studies and extracted the following data into Microsoft Excel software (Version 16, Microsoft Corporation, 2016): study characteristics, participant characteristics, study design, and intervention details. Because recruitment was central to this review, recruitment strategies, target sample size, number of patients screened and enrolled, reasons for refusal, reasons for achieving/not achieving the target sample size, approaches used to improve recruitment, and recommendations for future studies were extracted. To obtain details not found in the full text, attempts were made to retrieve information from the trial registration and published reports. Corresponding authors were contacted by email when a factor was mentioned to have affected recruitment but was lacking in details for further analysis. Any area of discrepancy was resolved via consensus.

2.6. Assessment of study quality in relation to recruitment

Two reviewers (L.B. and J.K.L.) independently assessed the recruitment reporting quality of each study using criteria adapted by Forster and colleagues.9 The study was assessed based on each of the 5 following quality metrics: (1) recruitment location, (2) who conducted recruitment, (3) time spent planning/preparing recruitment, (4) time spent conducting recruitment, and (5) if target population was specified. For each metric, a value of 0 was assigned if the metric was absent or inadequately described and a value of 1 if the metric was present or explicitly described. A 5‐point score was created where studies with a score of 4 or 5 were deemed as high quality, while those with a score of <4 were considered low quality.

2.7. Synthesis of results

To describe recruitment barriers and enablers in the context of the studies conducted, we examined the study characteristics, participant characteristics, intervention details, and outcomes. A thematic analysis of the study text describing the barriers and enablers to recruitment was also conducted.

2.8. Protocol deviations

As the main intent of conducting this review was to examine factors related to recruitment, not to examine intervention efficacy, the review questions, quality assessment, and strategy for data synthesis were refined in the protocol. First, some of the research questions were determined to be less relevant to participant recruitment and how recruitment can be maximised. These questions were replaced with “What information is required to be reported about participant recruitment for RCTs?” Second, the quality assessment tool was amended from an assessment of the risk of bias to recruitment reporting quality assessment. Third, the data synthesis was refined to reflect that a narrative synthesis was conducted as the proposed statistical analyses were deemed beyond the scope of this review.

3. RESULTS

3.1. Study selection

The initial search yielded 6614 records. Of these, 382 were selected for full‐text review, which led to further exclusion of 355 studies because of ineligibility (Figure 1). Ultimately, 27 studies were included in the review.10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36

Figure 1.

PRISMA flowchart of search results

3.2. Study characteristics

Included studies were published between 2003 and 2017. Characteristics of included studies are presented in Table 2. Studies were conducted in the United Kingdom (n = 12, 44.4%)10, 11, 14, 15, 19, 25, 26, 27, 30, 31, 33, 34; Australia (n = 4, 14.8%)17, 23, 28, 29; the United States (n = 2, 7.4%)12, 36; and 1 each from Canada (3.7%),35 Germany (3.7%),18 India (3.7%),20 Italy (3.7%),32 Mexico (3.7%),22 and New Zealand (3.7%).21 The remaining 3 studies (11.1%) were conducted in more than 1 country: the United States and Canada13 or the United Kingdom24 and the European Union and Australia.16 Intervention and control groups are defined in Table 2; the participants' characteristics, recruitment rate, and quality metric score are outlined in Table 3. Of those which explicitly stated the target sample size (n = 25), only 3 studies (12.0%) achieved the desired number of participants,11, 12, 27 while 4 studies (16.0%) achieved less than 20% of target recruitment.10, 18, 19, 30 The recruitment rate, defined as a percentage of the number of participants enrolled over the number of patients screened, ranged from 0.5%19 to 95.2%.21 Outcomes related to recruitment strategies are presented in Table 4.

Table 2.

Study characteristics

Author (y)	Country(ies), number of sites	Study design	Wound type	Study period (wk)	Intervention group (IG)	Comparator group (CG)
Davies et al. (2004)10	UK, 17 sites	RCT methodology	Venous leg ulcer	52 or longer if ulcer recurrence occurred, an additional 52 to 156	Compression bandaging plus venous surgery	Compression bandaging alone
Davies et al. (2015)11	UK, 1 site	RCT	Venous leg ulcer	12 or earlier when the ulcer had healed	4‐layer compression bandaging plus larvae	4‐layer compression bandaging alone
Driver et al. (2006)12	USA, 14 sites	RCT	Diabetic foot ulcer	24	Platelet‐rich plasma gel	Saline gel
Driver et al. (2017)13	USA & Canada, 22 sites	RCT	Diabetic foot ulcer	12	Moist wound therapy plus transdermal continuous oxygen therapy	Moist wound therapy plus sham device
Dumville et al. (2009)14	UK, 22 sites	3‐armed, open‐label RCT	Venous or mixed venous and arterial leg ulcer	Up to 52	Loose larvae or bagged larvae	Hydrogel
Eccles et al. (2005)15	UK, –	RCT	Leg ulcer	12	UlcerCare leg wrap	Sham non‐magnetic leg wrap
Edmonds (2009)16	EU and AUS, multiple sites	Open‐label RCT	Diabetic foot ulcer	24	Standard therapy plus Apligraf	Standard therapy
Edwards et al. (2005)17	AUS, multiple sites	Pilot RCT	Venous leg ulcer	12	Community‐based clinic “leg Club”	Home‐based care
Herber et al. (2009)18	Germany, multiple sites	Qualitative on an open‐label RCT	Leg ulcer	52 or earlier when the wound had healed	Physician‐nurse specialist collaboration, tandem practice	Physician alone
Idris et al. (2005)19	UK, 1 site	Pilot RCT	Diabetic foot ulcer	20	One‐to‐one education with a diabetes specialist nurse and dietician to achieve tight glycaemic control	Standard therapy
Jain et al. (2011)20	India, 2 sites	RCT	Lower extremity wounds	13	Injected and over‐sprayed with a total of 5 cm3 of autologous bone marrow‐derived cells	Injected with 5 cm3 of autologous peripheral blood
Jull et al. (2009)21	New Zealand, 1 site	Pilot, open‐label RCT	Venous leg ulcer	12	Compression plus 12‐week progressive resistance exercise programme	Compression
Lammoglia‐Ordiales et al. (2012)22	Mexico, 1 site	RCT	Venous leg ulcer	8	Hydrogel with Mimosa tenuiflora cortex extract (MTC‐2G)	Hydrogel
Leach (2003)23	AUS, 1 site	RCT methodology	Venous leg ulcer	12	Oral administration of horsechestnut seed extract	Standard therapy
Lipsky et al. (2012)24	USA, 8 sites; UK, 1 site	Pilot, open‐label RCT	Moderately infected diabetic foot ulcers	Up to 6	Gentamicin‐collagen sponge plus systemic antibiotic therapy	Systemic antibiotic therapy
Michaels et al. (2009)25	UK, multiple sites	RCT	Venous leg ulcer	52	Silver‐donating dressing	Standard dressing
Moffatt et al. (2003)26	UK, 5 sites	Open‐label RCT	Venous leg ulcer	24	4‐layer high‐compression elastic bandage systems	2‐layer high‐compression elastic bandage systems
Mudge et al. (2014)27	UK, multiple sites	Open‐label RCT	Venous (VLU) or mixed arterial/venous (MLU) leg ulcers	12 or until the ulcer had healed, if sooner	Larval therapy dressing	Standard debridement technique using a hydrogel
O'Brien et al. (2013)28	AUS, 1 site	Pilot, open‐label RCT	Venous leg ulcer	12	12‐week home‐based progressive resistance exercise programme	Standard therapy. At baseline, instructed not to change their physical activity habits during the 12‐week period
O'Brien et al. (2017)29	AUS, 3 sites	Open‐label RCT	Venous leg ulcer	12	12‐week exercise intervention with telephone coaching. At baseline, received an author‐developed exercise booklet, a “taking care of your legs” brochure, a pedometer and “keep it up” worksheets to diarise exercises and daily steps.	Usual care and telephone calls. At baseline, received a “taking care of your legs” brochure, a pedometer and “keep it up” worksheets to diarise daily steps.
O'Hare et al. (2010)30	UK, 1 site	RCT	Venous leg ulcer	24	4‐layer compression bandaging plus foam sclerotherapy	4‐layer compression bandaging
Ormerod et al. (2015)31	UK, 39 sites	RCT	Pyoderma gangrenosum	Up to 26	Oral ciclosporin 4 mg/kg/d in 2 divided doses	Oral prednisolone 0.75 mg/kg/d in a single dose
Polignano et al. (2004)32	Italy, 4 sites	RCT	Venous leg ulcer	24 or until the ulcer had healed, if sooner	4‐layer compression bandaging	A non‐compliant, plaster‐type bandage comprising Viscopaste and Tensoplast
Reynolds et al. (2004)33	UK, 3 sites	RCT	Exuding wounds	4.1 (29 d)	Drawtex dressing	Standard dressing
Robson et al. (2009)34	UK, 1 site	Open‐label RCT	Leg ulcer or other chronic wound	24	Honey	Standard therapy
Smith et al. (2010)35	Canada, multiple sites	Pilot RCT	Leg ulcer	13	4‐layer compression bandaging	Short‐stretch compression bandaging
Weed et al. (2004)36	USA, 1 site	Pilot RCT	Leg ulcer	28	Platelet lysate product mixed with collagen	Platelet‐poor plasma mixed with collagen

Country: UK = United Kingdom; USA = United States of America; EU = European Union; AUS = Australia; Study design: RCT = Randomised controlled trial.

General: (–) indicates not specified/Not reported.

Table 3.

Participant characteristics, recruitment rate, and quality metric score

Author (y)	Mean age (SD) or age group (n)	Gender (% male)	Target N	Time spent recruiting (mo)	Screened (n)	Enrolled (n)	Recruitment rate (%)a	Quality metric score
Davies et al. (2004)10	74.0b	57%b	1000	18	759	75	9.9	N/A
Davies et al. (2015)11	IG: 75.6 (10.4)CG: 78.1 (10.2)	IG: 35%CG: 30%	40	31	601	40	6.7	4
Driver et al. (2006)12	IG: 56.4 (10.2)CG: 57.5 (9.1)	IG: 80%CG: 84%	72	—	129	72	55.8	2
Driver et al. (2017)13	IG: 58.8 (9.4)CG: 58.6 (12.31)	IG: 70%CG: 77%	160	37	183	130	71.0	3
Dumville et al. (2009)14	IG1: 74.1 (12.9)IG2: 73.5 (12.2)CG: 74.3 (12.8)	IG1: 38%IG2: 34%CG: 51	370 or 270 (90% versus 80% power)	34	1712	267	15.6	3
Eccles et al. (2005)15	IG: 79 (8.0)CG: 81 (8.0)	IG: 75%CG: 67%	100	—	—	28	—	1
Edmonds (2009)16	IG: 56.4 (11.6)CG: 60.6 (9.8)	IG: 89%CG: 85%	240	23	106	82	77.4	2
Edwards et al. (2005)17	<60 (3), 61–70 (7), 71–80 (11), >80 (12)b	52%b	—	—	—	33	—	3
Herber et al. (2009)18, 49	IG: 69 (40, 89)cCG: 75 (62, 86)c	IG: 64%CG: 50%	300d	12d	—	29	—	N/A
Idris et al. (2005)19	—	—	50	6	200	1	0.5	3
Jain et al. (2011)20	IG: 54.3CG: 58.6	IG: 68%CG: 61%	60	24	—	48	—	3
Jull et al. (2009)21	IG: 54.6 (19.9)CG: 53.5 (19.9)	IG: 24%CG: 42%	50	6	42	40	95.2	3
Lammoglia‐Ordiales et al. (2012)22	IG: 58.2 CG: 62.6	IG: 41%CG: 32%	50	24	—	41	—	3
Leach (2003)23, 50	IG: 76.2 (9.5)dCG: 78.9 (9.1)d	IG: 70%dCG: 44%d	120d	7	155d	42	—	N/A
Lipsky et al. (2012)24	IG: 57.9 (11.5)CG: 54.7 (2.8)	IG: 61%CG: 83%	75	>12	56	36	64.3	3
Michaels et al. (2009)25	IG: 68.8 (16.7)CG: 72.4 (13.7)	IG: 50%CG: 42%	284	32	304	213	70.1	3
Moffatt et al. (2003)26	IG: 70.2 (14.4);CG: 71.8 (11.3)	IG: 42%CG: 44%	120	—	—	112	—	2
Mudge et al. (2014)27	IG: <60 (8), 61–70 (8), 71–80 (14), >80 (16)CG: <60 (9), 61–70 (4), 71–80 (14), >80 (15)	IG: 41%CG: 45%	88	—	—	88	—	2
O'Brien et al. (2013)28	IG: 66.0 (6.0)CG: 63.6 (20.0)	IG: 50%CG: 60%	40	6	16	13	81.3	3
O'Brien et al. (2017)29	IG: 71.3 (15.8)CG: 71.7 (13.4)	52%b	110	21	340	63	18.5	3
O'Hare et al. (2010)30, 51	69 (33, 90)b , c		200d	23	326	40	12.3	3
Ormerod et al. (2015)31	IG: 57.2 (16.9)CG: 51.3 (15.2)	IG: 29%CG: 42%	140	41	499	121	24.2	3
Polignano et al. (2004)32	IG: 68.4 (13.9)CG: 68.6 (9.6)	IG: 41%CG: 34%	100	24	—	68	—	3
Reynolds et al. (2004)33	IG: 75.5 (12.5)CG: 76.3 (11.1)	IG: 33%CG: 34%	200	—	778	142	18.3	1
Robson et al. (2009)34	IG: 66.4 (16.1)CG: 68.2 (15.5)	IG: 37%CG: 32%	200	27	114	105	92.1	3
Smith et al. (2010)35	<75 (4), >75 (8)b	33%b	—	3	49	12	24.5	4
Weed et al. (2004)36	IG: 67.6 (11.9)CG: 57.8 (18.2)	IG: 53%CG: 67%	80	—	—	26	—	2

Groups: IG = Intervention group; CG = Comparator group.

General: (–) indicates not specified/not reported; N/A = non‐applicable.

Quality metric score: studies that scored 4–5 were considered high‐quality studies, while studies that scored 1–3 were considered low quality.

^aRecruitment rate = Enrolled/screened × 100.

^bData within groups not reported.

^cMedian years (Minimum, Maximum).

^dInformation was obtained from a related, secondary source.

Table 4.

Recruitment strategies

Author (y)	Recruitment setting	Recruitment party	Strategies to improve recruitment post‐commencement	Suggestions made post‐trial
Davies et al. (2004)10	Hospital outpatient clinics	Local coordinator (professional role was not defined)	(1) Identified possible stumbling blocks with consultants, local coordinators, and community nurses; (2) amended selection criteria; (3) site visits to address site‐specific issues; and (4) instead of paying for data, a financial incentive of £256 per month was given to a local coordinator at each Centre to actively search for suitable patients	—
Davies et al. (2015)11	Specialist community‐based leg ulcer service	Leg ulcer nurse specialists	—	(1) amend selection criteria
Driver et al. (2006)12	Wound care physicians' and podiatrists' offices, outpatient wound care centres, a university‐based college of podiatric medicine clinic, Veteran's Administration wound care clinics, and an Army hospital limb preservation programme	—	—	—
Driver et al. (2017)13	Outpatient clinic	Site investigator or coordinator (professional role was not defined)	—	(1) amend selection criteria
Dumville et al. (2009)14	Community nurse‐led services, hospital wards, and hospital outpatient leg ulcer clinics	The research nurses, tissue viability teams, district nurses, and hospital outpatient staff for recruiting participants (professional role was not defined)	(1) Extended time and (2) extended funding	—
Eccles et al. (2005)15	Clinics	—	—	(1) amend selection criteria
Edmonds (2009)16	Hospitals	—	—	—
Edwards et al. (2005)17	Community‐based nursing services	Community nurses, medical practitioners, or self‐referred	—	—
Herber et al. (2009)18	General practitioners' clinics and community‐based specialist clinics	Medical practitioners based in the community, including general practitioners and specialists	(1) Initiated additional recruitment sites by involving dermatologists and phlebologist	—
Idris et al. (2005)19	Hospital‐based multidisciplinary foot clinic	Viv Savage, Bernie Kirk, Sarah Stevenson, Linda Altman, and Maureen Smith (professional role was not defined)	—	—
Jain et al. (2011)20	Surgical outpatient department	Researchers who were also surgeons	—	—
Jull et al. (2009)21	Community nurse leg ulcer service	—	(1) Extended recruitment period	—
Lammoglia‐Ordiales et al. (2012)22	Hospital interdisciplinary wound and ostomy care Centre	—	—	—
Leach (2003)23	Community nursing service	Community nurses	(1) Amended selection criteria, (2) extended recruitment period, (3) provided reminder notices to staff, (4) the researcher approached each nurse individually to increase awareness and to addressed trial‐related concerns, (5) researcher identified eligible participants for staff to approach, and (6) direct mailing of information sheets to all patients	—
Lipsky et al. (2012)24	Podiatrists' offices and hospitals	Investigators and study staff (professional role was not defined)	(1) Widened the patient age range and (2) amended protocol by decreasing the number of patients requiring post‐dose pharmacokinetic sampling and reducing the duration of post‐dose sampling	—
Michaels et al. (2009)25	Specialist primary care leg ulcer clinics	Clinical staff (professional role was not defined)	(1) Extended trial duration and (2) revised target sample size from 284 to 212 because attrition rate was lower than expected	—
Moffatt et al. (2003)26	Community leg ulcer clinics	—	—	—
Mudge et al. (2014)27	Hospital wards, outpatient clinics, community leg ulcer clinics, and community nurse caseloads	—	(1) Extended time and (2) initiated additional recruitment sites	—
O'Brien et al. (2013)28	Hospital outpatient clinic	Principal researcher (professional role was not defined)	—	—
O'Brien et al. (2017)29	Outpatient wound services and a community nursing service	—	—	—
O'Hare et al. (2010)30	One‐stop nurse‐led leg ulcer clinic	—	—	—
Ormerod et al. (2015)31	Hospitals	Recruiting dermatologist
Polignano et al. (2004)32	Hospital wards and outpatient clinic	—	—	—
Reynolds et al. (2004)33	Hospitals	—	—	—
Robson et al. (2009)34	Hospital wards and outpatient clinic	—	—	—
Smith et al. (2010)35	Not‐for‐profit nursing agency (nurse leg ulcer clinics and home care visits) and community‐based company of nurses	Nurses	(1) inconsistent screening was found, so all nurses were instructed to screen all patients with a leg ulcer to capture an accurate profile of persons ineligibility and (2) amended selection criteria Post‐trial: (1) initiate additional recruitment sites, (2) amend selection criteria	(1) Initiate additional recruitment sites, (2) amend selection criteria
Weed et al. (2004)36	Outpatient clinic	—	—	—

General: (–) indicates Not reported.

3.3. Study quality in relation to recruitment

Recruitment reporting quality of 3 included studies was not assessed as these were methodology papers10, 23 or a qualitative study.18 As summarised in Table 3, among the 24 studies where a quality assessment was conducted, 3 (12.5%) were assessed as “high” quality,11, 20, 35 and 21 (87.5%) were classified as “low” quality in relation to recruitment description.12, 13, 14, 15, 16, 17, 19, 21, 22, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 36 All studies explicitly outlined the eligibility criteria of the targeted population, and a majority of the studies specified the recruitment setting (n = 20, 83.3%).11, 12, 13, 14, 17, 19, 20, 21, 22, 24, 25, 26, 27, 28, 29, 30, 31, 32, 34, 35, 36 Four studies (16.7%) specified the role recruiters played in usual patient care,11, 17, 31, 35 and no study (0.0%) reported the time spent planning/preparing the recruitment. Although no study was excluded on the basis of the quality of reporting, assessments are included in this review to inform and improve future reporting of barriers to recruitment.

3.4. Synthesis of results

The information retrieved from each study in relation to factors that affected recruitment was limited. Contact was made with the corresponding authors of 4 studies,16, 20, 22, 26 but only 1 author responded to elaborate on the cause of poor recruitment.37 Although this review aimed to review both enablers and barriers to recruitment, studies with a good recruitment outcome did not report enabling factors. Hence, only the barriers to recruitment were thematically synthesised and categorised into 5 categories: study factors, health system factors, patient factors, operational factors, and clinician factors. Slightly more than half of the studies (15 of 27, 55.6%) attributed poor recruitment outcome to more than 1 factor (Table 5).

Table 5.

Barriers to recruitment: synthesis of results

Barriers to recruitment	Reference
	10	11	12	13	14	15	16	17	18	19	20 a	21	22	23	24	25	26	27	28	29	30	31	32	33	34	35	36	n
Study design factors																												16
Lack of eligible patients		•	•	•	•			•		•				•	•			•			•			•		•
Overestimated the number of eligible participants	•				•	•			•									•			•				•
Health system factors																												11
Change in practice	•	•			•																•		•		•
Ethics‐ or organisation‐enforced requirements						•	•				•			•				•
Operational factors*													•				•											11
Lack of funding for research	•						•		•					•								•
Time constraints to meet project deadline											•								•	•		•
Multiple research trials recruiting at the same site									•	•									•
Patient factors																												9
Study burden could be overwhelming			•							•		•							•								•
Treatment preference																								•	•		•
Logistically difficult to participate								•			•
Clinician factors																												5
Limited time for non‐clinical activities			•						•							•
Staff gatekeeping										•				•
Lack of interest/incentive									•
Factor groups	3	2	3	1	2	2	2	2	3	4	4	1	1	4	1	1	1	2	2	1	2	1	1	1	3	1	1	3

General: • indicates particular theme identified in the study.

^aInformation was obtained from personal communication with the corresponding author; *not specified.

Study‐related factors were the most commonly reported barriers to recruitment (n = 16, 59.3%). Among these 16 studies, 12 identified that having defined a set of “too restrictive” selection criteria was a major issue, which led to a lack of eligible patients.11, 12, 13, 14, 17, 19, 23, 24, 27, 30, 33, 35 To improve recruitment, 3 studies amended the participant selection criteria.10, 23, 35 Seven studies reported that fewer patients were eligible,10, 14, 15, 18, 27, 30, 34 even when on‐site clinicians were consulted during study conception. Authors concluded that patients seen during the recruitment period were different from those previously seen. Studies also partly attributed the difference in patient cohorts to a change in best practice “following the reported success of compression therapy”.10 This was mentioned as a causal factor in 3 other studies11, 14, 30 and was categorised under “health system factors”.

Eleven studies (40.7%) attributed the cause of poor recruitment to factors related to health system, a change in practice,10, 11, 14, 30, 32, 34 and requirements imposed by the ethics committee or health organisation.15, 16, 20, 23, 27 How a change in practice can impact a trial was well illustrated in the trial conducted by Robson and colleagues,34 evaluating the effect of an antibacterial honey on wound healing. Their trial was prematurely terminated because honey products were added to the list of reimbursable items, and patients or nursing staff could obtain the products on prescription without having to participate in the trial.34 On the other hand, the process of gaining approval to access patients could be “a lengthy, convoluted process”.15

Operational factors that prevented authors from achieving their target sample size were reported in 11 studies (40.7%),10, 16, 18, 19, 20, 22, 23, 26, 28, 29, 31 including a lack of funding (n = 5), time constraints (n = 4), and having to compete with other trials concurrently recruiting patients (n = 3). First, the small amount of funding resources meant that minimal incentives could be provided to the recruitment site, which would be unattractive as “commercially funded trials often offer reimbursement to the host institution at a 10‐20‐fold greater level” than what researchers could offer.10 Second, studies with a strict deadline, such as those that were completed as part of a PhD project29 or had a defined timeframe,37 could not extend the recruitment period. Of all the included studies, 5 reported to have extended the recruitment period, but only 1 study attained the target sample size,27 and the rest did not.14, 21, 23, 25 Third, concurrent trials impacted the number of eligible patients that could be approached.19, 28 Idris and colleagues found that, of those who were seen at the recruitment site, “29% were debarred by involvement in other studies”.19

Patient‐related factors, such as study burden, treatment preference, and logistic difficulty, were reported in 9 studies (33.3%).12, 17, 19, 20, 21, 28, 33, 34, 36 The study burden outlined by the authors was either related to intervention, such as an intensive intervention involving vigilant self‐monitoring19 and a 12‐week home‐based exercise regimen,28 or to outcome measurement requiring extensive testing12 or requiring extra visits.36 In 3 studies, it was reported that, even if eligible, patients had a preference for a product that they regarded as effective for their wound and were, therefore, unwilling to participate in an RCT knowing that they could be receiving their disfavoured product.33, 34, 36 A minority of patients were precluded from participating in any trial as they lived too far away.17, 20

Five studies (18.5%) had difficulty obtaining agreement from the clinicians to assist with recruitment.12, 18, 19, 23, 25 Considering the workload of clinicians, it was difficult for clinicians to invest their time and resources to take part in research activities.12, 18, 25 In addition, authors thought that clinicians were only approaching patients who they deemed suitable.19, 23 In a study conducted in Germany, physicians did not allow their practices to be included as a recruitment site because “there were no financial rewards offered” and “a general lack of interest in research”.18 Davies et al and Leach et al attempted to address clinician‐related barriers by having site visits to speak to clinicians, but none of these studies achieved the target sample size.

4. DISCUSSION

4.1. Summary of evidence

We were able to assess 24 studies (88.9%) that reported recruitment description; 21 (87.5%) of these studies were assessed as low quality and 3 studies (12.5%) as high quality. Some studies reported 1 or more barriers to recruitment to chronic wound RCTs related to: the trial design (16 studies, 59.3%) and operation (5 studies, 18.5%), health system (11 studies, 40.7%), patient cohort (9 studies, 33.3%), and clinicians (5 studies, 18.5%).

These findings are supported by previous research on RCTs that have examined the factors that impact recruitment in any health‐related field. Trial design factors were the most commonly reported barrier to recruitment, particularly the narrow inclusion criteria that resulted in a lack of eligible patients.38 Selection criteria present a significant challenge to the conduct of RCTs as a balance between low risk of bias and relevance of results to relevant patient population.39, 40

Patient‐related recruitment barriers include study burden and logistics, particularly for disadvantaged populations such as the elderly,41, 42 and patients’ treatment preferences.41 Lack of clinical equipoise presents an obstacle to completion and may also contribute to the termination of trials because of inadequate patient enrolment.43 The overestimation of the pool of potential patients is a well‐recognised phenomenon in the recruitment to randomised trials.44 General practitioners, nurse practitioners, and wound care consultants play an important role in the recruitment of a patient. Clinician recruiters have been reported to influence the impact on recruitment.40 For example, doctors may express their discomfort about individual groups or patients’ participation in the clinical trial even if the patient is eligible if the trial intervention clashes with their clinical judgement or preference for a particular intervention, resulting in a perceived or actual loss of professional autonomy. In addition, clinical research may be perceived as an inferior intervention to their perceived clinical judgement and preferred treatment plan.41

While no study reported which factors enabled positive recruitment outcomes, 9 studies reported strategies to improve recruitment based on their experiences of implementing a clinical trial.10, 14, 18, 21, 23, 24, 25, 27, 35 Strategies included recruitment period extension,14, 21, 23, 25, 27 amendment of selection criteria,10, 23, 24, 35 initiation of additional recruitment sites,18, 27 revision of target sample size,24, 25 and conducting more frequent site visits.10 Implementing strategies during the trial to improve patient recruitment may be necessary; however, the intended approach should be optimised prior to trial commencement. In other disciplines, recruitment‐retention protocols have been suggested.38 The recruitment retention plan includes evidence‐based strategies with social marketing principles to address key barriers to patient recruitment and include a range of suggestions. These include the need for a triage algorithm, broad eligibility criteria, a dedicated recruitment nurse, simplified consent documents and information booklets, planned visits that included standardised practicing through role plays; well‐defined consenting procedures encompassing proxy scenarios; integration with clinical services; plans for staff relief; focused outreach; public/provider education; plans to address gatekeeping; and continuous monitoring and quality assurance.38 This approach has been implemented in a palliative care RCT in oncology and reported an increased recruitment rate.45 This recruitment approach may be useful in future chronic wound RCTs. In recent years, there have been recommendations for clinical trials to be embedded into routine health care.46 In this context, both patients and health care professionals would expect to be involved in clinical research. This approach would overcome some of the clinician‐ and patient‐related barriers identified in this and other reviews.41

4.2. Strengths and limitations

We applied a broad literature search in this review to capture relevant literature, and the systematic approach minimised the likelihood that we missed any relevant trials. Two review authors independently selected studies, extracted data, and assessed the quality of the description of reported barriers to recruitment. Given this systematic approach, the risk of missing studies meeting the inclusion criteria was minimal. We invite readers to notify us of any studies, published or unpublished, that meet our criteria. In addition, the chronic wound RCTs excluded from this review (over 250 studies) may have experienced barriers and/or enablers, although these were not reported, and the included studies may have failed to report the full extent of barriers and enablers experienced during the trial.

4.3. Implications

Literature on barriers to recruitment is not well indexed because the number of studies is quite small, whereas in trials, barriers to recruitment are not often reported, and it is unclear whether it is measured. The Consolidated Standards of Reporting Trials (CONSORT) Statement does not require authors to report barriers or enablers to patient recruitment. Reporting of barriers and enablers to recruitment in RCTs could improve the reporting standard on these important aspects.47 We suggest that barriers to recruitment be included in the limitations section of the CONSORT Statement. In addition, trials could consider including in their final report the enablers if they recruited the target sample within the funded period.

4.4. Conclusions

The recruitment process is multifaceted with factors that intertwine with trial design and operation, clinicians, patients, and health system converging to influence participation.48 Recruitment is 1 of the most arduous tasks in a clinical trial process; it warrants considerable planning as part of the study design. Without rigorous clinical evidence, new treatments cannot be introduced into practice. Equally, without support from health care institutions and clinicians, patient recruitment is challenging.23 Increased clinical staff involvement in wound care research may facilitate the implementation of evidence‐based practice in chronic wound management. The CONSORT Statement could include an additional item on barriers to recruitment. This approach will allow for increased awareness of the potential barriers to recruitment for RCTs in both wound management and other health care research.

Bugeja L, Low JK, McGinnes RA, Team V, Sinha S, Weller C. Barriers and enablers to patient recruitment for randomised controlled trials on treatment of chronic wounds: A systematic review. Int Wound J. 2018;15:880–892. 10.1111/iwj.12940