Abstract
Objective
This registry-based systematic review aimed to investigate the use and applicability of adaptive designs in ongoing low back pain (LBP) and osteoarthritis (OA) trials and discuss emerging opportunities of adaptive designs in the field, in light of the expected advantages and possible challenges.
Method
We searched ClinicalTrials.gov for records of ongoing studies in chronic LBP or OA. The applicability of adaptive designs for the included trials was assessed based on the ratio of primary endpoint length to recruitment length, considering a ratio below 0.25 as an indicator of potential feasibility. Meeting this ratio means there may be opportunities to implement adaptations that improve trial efficiency and benefit participants yet to be recruited.
Results
Overall, 579 clinical trials were included (135 trials in LBP and 447 trials in OA, with three overlapping studies among the two conditions). A ratio below 0.25 was observed in 89 (66%) LBP trials and 256 (57%) OA trials, suggesting they could potentially benefit from an adaptive design. However, only two (1.5%) out of 135 LBP trials and eight (1.8%) out of 447 OA trials reported an adaptive design, with the most common types being dose escalation studies (67%), followed by group-sequential designs (22%), and master protocols (11%).
Conclusion
The use of adaptive designs in LBP and OA clinical trials is rare, although many trials would benefit from using this innovative approach to trial conduct. Researchers should be encouraged to consider the benefits and challenges of employing an adaptive design to plan future trials.
Keywords: Adaptive design, Spine, Musculoskeletal, Randomised controlled trial, Clinical trial, Registry-based review
Graphical abstract
1. Introduction
Low back pain (LBP) and osteoarthritis (OA) are amongst the top five most prevalent and burdensome musculoskeletal conditions [1], affecting, respectively, 619 [2] and 595 [3] million people worldwide in 2020. Arguably, both LBP and OA have become research priorities, with over five thousand randomised controlled trials in LBP or OA being published over the last 10 years (PubMed, January 17, 2025). Although well-designed and conducted randomised controlled trials provide the highest level of scientific evidence for determining the effectiveness of health interventions [4], they are inherently time-consuming and resource-intensive, typically requiring large sample sizes with long-term follow-ups, incurring elevated costs, and demanding complex operational infrastructures [5,6]. In contrast, it has been estimated that as much as 85% of the overall medical and health research investment does not translate into meaningful, clinically relevant findings, resulting in research waste [7,8]. In this context, strategies aimed at increasing efficiency, achieving optimal value, and reducing research waste associated with clinical trials are encouraged [9].
Adaptive designs [5,10,11] stand out as a potential strategy to address the shortcomings of conventional clinical trials and mitigate research waste. With applicability ranging from early phase dose-escalation studies to confirmatory trials, adaptive designs enable adjustment to an ongoing study based on accumulating trial data by the interim(s), in accordance with pre-specified decision rules, and without undermining the trial's integrity [5,10,11]. Adaptive designs may also allow investigators to test multiple hypotheses and answer multiple research questions under the same protocol, in a seamless process driven by pre-planned decision rules and adaptations [10,12]. These adaptations potentially lead to reduced sample sizes, increased power, and/or shorter study durations, therefore optimising the use of resources, reducing costs, and accelerating the development timeline of new intervention approaches [5,10,13]. In addition to driving enhanced flexibility and efficiency, some adaptive designs are considered more ethical, as they tend to minimise the exposure of study participants to ineffective or potentially harmful interventions [14].
Although compelling and attractive, adaptive designs are not always advantageous compared to traditional approaches [10,15]. The benefits related to increased efficiency may come at the expense of a considerable increase in operational and logistical complexity, which must not be ignored [10,11,15,16]. As such, the decision-making process for choosing an adaptive design over the traditional approach should carefully consider the study aims and trade off the potential efficiency gains against the increased operational complexity. Several key methodological aspects play an important role when determining whether designing an adaptive trial is worthwhile or not [14,15].
The relationship between the primary endpoint length and the total planned recruitment length, for example, provides valuable information on the feasibility of implementing adaptations that could translate into actual, increased trial efficiency [15,17]. Because adaptive designs rely on the interim analysis of accrued trial data to decide whether to implement the planned adaptations, the interim analysis should include outcome data from an adequate sample size to allow a reliable decision [15]. To enable this, the endpoint used to make a decision at the interim analysis should be sufficiently short in comparison to the total planned recruitment length [15]. When this condition is not met, the trial's recruitment can likely be completed before enough outcome data is accumulated for the interim analysis [10,15]. Alternatively, pausing the trial's recruitment until sufficient primary outcome data have been made available for the interim analysis is not a feasible approach and can even downplay the increased efficiency achieved by the adaptive design. Based on these concepts, adaptive designs are considered most likely to be beneficial when the ratio of primary endpoint length to recruitment length is small [15], with a threshold of 0.25 being proposed as a guide [17,18]. In practical aspects, the time required to observe the primary endpoint should be less than 25% of the total planned time needed to recruit the target sample size (Box 1). Meeting this ratio provides confidence that, following the collection of the primary endpoint, sufficient time will remain to clean and process the data, conduct the interim analysis, and implement the planned adaptations, benefiting future participants without disrupting the recruitment timeline.
Box 1. Practical example of how the ratio of primary endpoint length to recruitment length may influence an adaptive design.
For illustrative purposes, we may consider a hypothetical clinical trial investigating pain intensity at three months as the primary endpoint, with an expected recruitment length of 12 months, and a planned interim analysis halfway through the target sample size. In this scenario, the time required to observe the primary endpoint is 25% of the total planned recruitment period for the target sample size, generating a ratio of 0.25. Assuming a uniform recruitment rate, this trial would require at least nine months to recruit and collect the primary endpoint data from 50% of the participants (i.e., six months to recruit 50% of the target sample size, and three more months to collect the primary outcome data from the last recruited participant). Accounting for the time required to analyse the accrued interim data and, if applicable, modify the trial, the planned adaptations would benefit the individuals recruited in the last few months. Conversely, trials with a relatively long primary endpoint length compared to the total planned recruitment length, especially those with a ratio above 0.25, are less likely to achieve increased efficiency from adaptive designs, as very few or no participants benefit from the planned adaptations. Ultimately, a ratio below 0.25 means that interim analyses are likely to have sufficient primary endpoint information collected to inform meaningful adaptations, and there is likely to be ample opportunity to implement adaptations in a way that improves trial efficiency and benefits participants yet to be recruited.
Alt-text: Box 1
There has been considerable uptake of adaptive designs in oncology [19], cardiovascular diseases [[20], [21], [22], [23]], and infectious diseases [[24], [25], [26], [27]]. Increasingly, they are also used in medical device studies [28] and diagnostic accuracy studies [29]. Nevertheless, adaptive designs appear to remain underutilised in musculoskeletal health [30,31], although they could be particularly useful in advancing clinical research in the field. Therefore, the purpose of this study was to systematically review the current use and applicability of adaptive designs in LBP and OA clinical research, determining the frequency and proportion of ongoing clinical trials employing an adaptive design, and those that could potentially benefit from adaptive designs, as implied by the ratio of primary endpoint length to recruitment length solely. This work also aimed to explore emerging opportunities of adaptive designs in the field, with an emphasis on the expected advantages and possible challenges.
2. Method
2.1. Study design and data source
A registry-based systematic review conducted according to methods proposed previously [17] and reported according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses statement [32].
ClinicalTrials.gov was systematically searched for records containing “chronic LBP” or “osteoarthritis” in the “Condition/Disease” field and “recruiting” in the “Study Status” field. No other restrictions or filters were applied. The searches were conducted on August 27, 2024 for LBP and on September 8, 2024 for OA. Developed and maintained by the U.S. National Institutes of Health through its National Library of Medicine, ClinicalTrials.gov is one of the most comprehensive publicly available online databases and registries of clinical research studies.
The records yielded by the electronic searches were downloaded into a comma-separated values file. A total of 30 data fields were available and selected for download. The downloaded comma-separated values file containing the search results was then converted to an XML file and managed through Microsoft Excel (Microsoft, Washington, United States) for screening, data extraction, calculations, and data analysis.
2.2. Eligibility criteria and study selection
Records were screened by two trained reviewers (AMCS and MW), with support from senior researchers (PF and MH), according to the following eligibility criteria: “interventional” studies, as registered in the “Study Type” field, “recruiting” at the time of the search, as registered in the “Study Status” field, and with participants with LBP or OA, as registered in the “Conditions” field. Given the applicability of adaptive designs for non-randomised [33,34] and single-arm trials [35,36], for example, with the use of concurrent and non-concurrent (historical) external controls [37], or even without a control group [38], these study designs were not excluded. Records that clearly defined themselves as pilot or feasibility studies were excluded, as well as records with insufficient information on the “Study Type”, “Study Status”, or “Study Design” fields. No restrictions were applied on the types of interventions and comparators, or the nature and timing of outcome measures. All included records were attributed a Study ID for the purposes of this study.
2.3. Data extraction and calculations
Data were extracted from the included records by two trained reviewers (AMCS and MW) using a standardised data extraction form and with support from senior researchers (PF and MH). To ensure accuracy, extracted data were independently double-checked for a randomly selected sample of 20% of the included records.
The extracted data included the body area, type of intervention under investigation, primary endpoint length (i.e., the length of time, in days, required to observe the primary endpoint following enrolment), funder type, study sponsor (industry or non-industry), and site configuration (single-centre or multicentre). Because a single study may investigate more than one type of intervention, the sum of counts and proportions by type of intervention may exceed the total number of studies. For studies with more than one primary outcome measure or endpoint registered under the “Primary Outcome Measures” field, a conservative approach was adopted, with the longest endpoint extracted. Additionally, the “Detailed Description” field of each record was extracted from the online registration system and manually transferred to the data extraction form.
The total planned recruitment length, expected recruitment rate, and ratio of primary endpoint length to recruitment length were calculated for all studies based on the registered data. For studies with date fields formatted as month and year, the date was standardised to the first day of the month for calculation purposes. The total planned recruitment length (i.e., the planned length of time, in days, required to recruit the target sample size) was calculated based on the registered “Start Date” and “Primary Completion Date” fields. The expected recruitment rate (i.e., the estimated number of participants recruited per month) was calculated based on the target sample size, as registered in the “Enrolment” field, divided by the calculated total planned recruitment length. Finally, the ratio of primary endpoint length to recruitment length was calculated based on the extracted primary endpoint length divided by the calculated total planned recruitment length.
The study design (i.e., fixed-sample or adaptive) was determined primarily based on the registered “Intervention Model” (i.e., strategy for assigning interventions to participants), as registered in the “Study Design” field. According to the ClinicalTrials.gov Data Element Definitions, the term “sequential” represents an intervention study model where “groups of participants are assigned to receive interventions based on prior milestones being reached in the study, such as in some dose escalation and adaptive design studies”. Studies reporting a “sequential” intervention model were further examined to ascertain whether the term was related to an adaptive design. Additionally, a structured keyword search was conducted in the “Study Title”, “Brief Summary”, “Detailed Description”, and “Study Design” fields to identify potentially relevant mentions of adaptive designs. The list of search terms was established based on a published literature search strategy [12], considering the most common types of adaptive designs [10,11], as well as other innovative trial designs and master protocols, such as basket, umbrella, and platform trials [12]. The search terms included “adaptive”, “platform”, “basket”, “umbrella”, “sequential”, “multi-arm”, “multi-stage”, “MAMS”, “pick-the-winner”, “drop-the-loser”, “re-estimation”, “re-assessment”, “enrichment”, “seamless”, “escalation”, “interim”, “master”, as well as their variations (e.g., singular, plural, hyphenated, non-hyphenated variants). Each instance of the identified keywords was manually examined within its context to determine its relevance to the study design.
2.4. Data analysis and synthesis
The main study characteristics were summarised, and the main findings were presented descriptively. For each condition, data were shown overall and stratified based on the 0.25 cutoff for the ratio of primary endpoint length to recruitment length, allowing comparison of key study characteristics and identification of design implications affecting trial efficiency, as well as potential sources of methodological heterogeneity.
The applicability of adaptive designs for the included studies was determined based on the ratio of primary endpoint length to recruitment length, considering the suggested threshold of 0.25 as an indicator of potential feasibility. Studies with a ratio below 0.25 were reported as most likely to benefit from adaptive designs [15,17]. The distribution of LBP and OA studies according to the ratio of primary endpoint length to recruitment length was shown graphically, and two ongoing studies from each condition were presented graphically to illustrate how the ratio might favour or not the use of an adaptive design.
3. Results
3.1. Study identification
The search for LBP and OA studies returned 191 and 521 records, respectively (Fig. 1). After screening against the eligibility criteria, 579 records were included, 135 in LBP and 447 in OA (Supplementary Material 1). Three overlapping studies (NCT03168802, NCT05986292, and NCT05852808) were identified across LBP and OA datasets and were retained in the reporting of each condition.
Fig. 1.
Study flow diagram showing the steps of identification, screening, and inclusion of LBP and OA clinical trials for this review.
3.2. LBP
Among the 135 LBP studies, the types of intervention involved devices (20.9%), behavioural interventions (17.6%), drugs (12.4%), procedures (10.5%), among others (46.4%). On average, included LBP trials presented a target sample size of 219 participants (range: 20 to 10,000), a primary endpoint length of 128 days (range: 0 days to 2.5 years), and a total planned recruitment length of 2.1 years (range: 53 days to 14.6 years) (Table 1). Assuming a uniform recruitment rate, the expected number of participants recruited per month ranged from 0.3 (i.e., one participant every three to four months) to 115, with a mean of 10 participants per month (Table 1). The median ratio of primary endpoint length to recruitment length was 0.15 (Table 1). Most ongoing LBP trials (89 studies, 65.9%) had a ratio of less than 0.25 and could potentially benefit from adaptive designs (Fig. 2).
Table 1.
Characteristics of the included LBP studies (n = 135).
| Study Characteristic |
Category or Description | Total (n = 135) | r < 0.25 (n = 89) | r ≥ 0.25 (n = 46) |
|---|---|---|---|---|
| Target sample size (counts) | Mean (Min-Max) | 219 (20–10,000) | 265 (25–10,000) | 131 (20–558) |
| Median (IQR) | 83 (50–152) | 86 (50–160) | 82 (47–143) | |
| Primary endpoint length (days) | Mean (Min-Max) | 127.6 (0.0–912.0) | 73.9 (0.0–365.0) | 231.4 (28.0–912.0) |
| Median (IQR) | 84.0 (30.4–182.4) | 56.0 (28.0–91.2) | 182.2 (84.0–364.8) | |
| Total planned recruitment length (days) | Mean (Min-Max) | 767.9 (53.0–5327.0) | 961.7 (148.0–5327.0) | 393.1 (53.0–1704.0) |
| Median (IQR) | 523.0 (274.5–987.0) | 731.0 (364.0–1329.0) | 315.0 (138.3–521.8) | |
| Expected recruitment rate (monthly) | Mean (Min-Max) | 10.0 (0.3–115.1) | 7.9 (0.3–115.1) | 14.1 (1.9–50.9) |
| Median (IQR) | 5.8 (2.9–10.0) | 4.5 (2.5–7.5) | 10.3 (5.5–16.9) | |
| Ratio of primary endpoint length to recruitment length (r) | Mean (Min-Max) | 0.30 (0.00–2.99) | 0.10 (0.00–0.25) | 0.70 (0.25–2.99) |
| Median (IQR) | 0.15 (0.07–0.32) | 0.09 (0.04–0.15) | 0.50 (0.32–0.97) | |
| Study phase [n (%)] | Early Phase I | 1 (0.7%) | 1 (1.1%) | 0 (0%) |
| Phase I | 1 (0.7%) | 1 (1.1%) | 0 (0%) | |
| Phase I-II | 2 (1.5%) | 1 (1.1%) | 1 (2.2%) | |
| Phase II | 4 (3.0%) | 3 (3.4%) | 1 (2.2%) | |
| Phase II-III | 1 (0.7%) | 1 (1.1%) | 0 (0%) | |
| Phase III | 5 (3.7%) | 1 (1.1%) | 4 (8.7%) | |
| Phase IV | 3 (2.2%) | 2 (2.2%) | 1 (2.2%) | |
| Not applicable | 118 (87.4%) | 79 (88.8%) | 39 (84.8%) | |
| Funder Type [n (%)] | U.S. Federal Government Agency | 4 (3.0%) | 4 (4.5%) | 0 (0%) |
| Other Government Agency | 0 (0%) | 0 (0%) | 0 (0%) | |
| Industry | 16 (11.9%) | 9 (10.1%) | 7 (15.2%) | |
| Network | 0 (0%) | 0 (0%) | 0 (0%) | |
| Other | 115 (85.2%) | 76 (85.4%) | 39 (84.8%) | |
| Blanks | 0 (0%) | 0 (0%) | 0 (0%) | |
| Study sponsor [n (%)] | Industry | 16 (11.9%) | 9 (10.1%) | 7 (15.2%) |
| Non-industry | 119 (88.1%) | 80 (89.9%) | 39 (84.8%) | |
| Blanks | 0 (0%) | 0 (0%) | 0 (0%) | |
| Site Configuration [n (%)] | Single centre | 113 (83.7%) | 73 (82.0%) | 40 (87.0%) |
| Multicentre | 22 (16.3%) | 16 (18.0%) | 6 (13.0%) | |
| Blanks | 0 (0%) | 0 (0%) | 0 (0%) |
r: Ratio of primary endpoint length to recruitment length.
Fig. 2.
Distribution of low back pain (LBP) studies according to the ratio (r) of primary endpoint length to recruitment length (n = 135).
Among the 89 LBP trials with a ratio below 0.25, the most common types of intervention were devices (19.2%), behavioural interventions (19.2%), and procedures (12.5%), while for those with a ratio equal to or above 0.25 the most common types of intervention were devices (24.5%), drugs (16.3%), and behavioural interventions (14.3%). LBP trials with a ratio below 0.25 presented, on average, larger sample sizes, shorter primary endpoint length, longer total planned recruitment length, and a slower expected recruitment rate. No relevant trends were observed regarding study phase, funder type, study sponsor, or site configuration. To demonstrate how the ratio of primary endpoint length to recruitment length influences an adaptive design, we plotted these characteristics using two ongoing LBP studies as examples: one likely to benefit from an adaptive design and one unlikely to do so (Supplementary Material 2).
In terms of study design, 14 unique records were identified for further investigation, and only two (1.5%) studies were considered to have an adaptive design (Supplementary Material 1 and 3), NCT05986292, describing a master protocol [12], and NCT06053242, describing a dose escalation study. Only one of them presented a ratio below 0.25 (Supplementary Material 3, Supplementary Table S1).
3.3. Osteoarthritis
Among the 447 OA studies, the majority investigated the knee (64.2%), followed by hip (8.9%), shoulder (7.2%), thumb (4.0%), hand (2.0%), spine (1.8%), wrist (1.6%), and ankle (1.3%). At a smaller proportion, studies also investigated OA in multiple joints (7.8%) or at the elbow, fingers, foot, or the big toe (0.9%). The types of intervention involved devices (24.3%), drugs (23.1%), procedures (19.9%), behavioural interventions (8.3%), biological interventions (6.9%), among others (24.5%). On average, included OA trials presented a target sample size of 291 participants (range: 3 to 20,000), a primary endpoint length of 327 days (range: 0 days to 10 years), and a total planned recruitment length of 3 years (range: 20 days to 13.7 years) (Table 2). Assuming a uniform recruitment rate, the expected number of participants recruited per month ranged from 0.1 (i.e., one participant every 10 months) to 741, with a mean of 10 participants per month (Table 2). The median ratio of primary endpoint length to recruitment length was 0.21 (Table 2). Most ongoing OA trials (256 studies, 57.3%) had a ratio of less than 0.25 and could potentially benefit from adaptive designs (Fig. 3).
Table 2.
Characteristics of the included OA studies (n = 447).
| Study Characteristic |
Category or Description | Total (n = 447) | r < 0.25 (n = 256) | r ≥ 0.25 (n = 191) |
|---|---|---|---|---|
| Target sample size (counts) | Mean (Min-Max) | 291 (3–20,000) | 349 (6–20,000) | 211 (3–9172) |
| Median (IQR) | 90 (51–155) | 90 (50–150) | 90 (52–160) | |
| Primary endpoint length (days) | Mean (Min-Max) | 326.6 (0.0–3650.0) | 110.9 (0.0–730.0) | 615.7 (21.0–3650.0) |
| Median (IQR) | 182.4 (42.0–364.8) | 84.0 (14.0–182.4) | 364.8 (228.0–729.6) | |
| Total planned recruitment length (days) | Mean (Min-Max) | 1090.0 (20.0–4995.0) | 1132.4 (30.0–4079.0) | 1033.2 (20.0–4995.0) |
| Median (IQR) | 898.0 (461.0–1460.0) | 936.0 (593.8–1490.8) | 761.0 (365.0–1333.5) | |
| Expected recruitment rate (monthly) | Mean (Min-Max) | 9.6 (0.1–740.6) | 10.0 (0.2–740.6) | 9.1 (0.1–257.9) |
| Median (IQR) | 3.6 (1.7–7.7) | 3.3 (1.6–6.4) | 4.0 (1.8–9.3) | |
| Ratio of primary endpoint length to recruitment length (r) | Mean (Min-Max) | 0.39 (0.00–10.00) | 0.10 (0.00–0.25) | 0.79 (0.25–10.00) |
| Median (IQR) | 0.21 (0.06–0.48) | 0.08 (0.02–0.16) | 0.51 (0.34–0.77) | |
| Study phase [n (%)] | Early Phase I | 5 (1.1%) | 3 (1.2%) | 2 (1.0%) |
| Phase I | 12 (2.7%) | 6 (2.3%) | 6 (3.1%) | |
| Phase I-II | 14 (3.1%) | 7 (2.7%) | 7 (3.7%) | |
| Phase II | 29 (6.5%) | 17 (6.6%) | 12 (6.3%) | |
| Phase II-III | 6 (1.3%) | 3 (1.2%) | 3 (1.6%) | |
| Phase III | 17 (3.8%) | 8 (3.1%) | 9 (4.7%) | |
| Phase IV | 33 (7.4%) | 27 (10.5%) | 6 (3.1%) | |
| Not applicable | 331 (74.0%) | 185 (72.3%) | 146 (76.4%) | |
| Funder Type [n (%)] | U.S. Federal Government Agency | 5 (1.1%) | 2 (0.8%) | 3 (1.6%) |
| Other Government Agency | 17 (3.8%) | 6 (2.3%) | 11 (5.8%) | |
| Industry | 64 (14.3%) | 28 (10.9%) | 36 (18.8%) | |
| Network | 1 (0.2%) | 1 (0.4%) | 0 (0%) | |
| Other | 359 (80.3%) | 219 (85.5%) | 140 (73.3%) | |
| Blanks | 1 (0.2%) | 0 (0%) | 1 (0.5%) | |
| Study sponsor [n (%)] | Industry | 65 (14.5%) | 28 (10.9%) | 37 (19.4%) |
| Non-industry | 382 (85.5%) | 228 (89.1%) | 154 (80.6%) | |
| Blanks | 0 (0%) | 0 (0%) | 0 (0%) | |
| Site Configuration [n (%)] | Single centre | 351 (78.5%) | 206 (80.5%) | 145 (75.9%) |
| Multicentre | 94 (21.0%) | 50 (19.5%) | 44 (23.0%) | |
| Blanks | 2 (0.4%) | 0 (0%) | 2 (1.0%) |
r: Ratio of primary endpoint length to recruitment length.
Fig. 3.
Distribution of osteoarthritis (OA) studies according to the ratio (r) of primary endpoint length to recruitment length (n = 447).
Among the 256 OA trials with a ratio below 0.25, the most common types of intervention were drugs (28.2%), devices (20.4%), and procedures (18.2%), while for those with a ratio equal to or above 0.25 the most common types of intervention were devices (29.6%), procedures (22.1%), and drugs (16.4%). OA trials with a ratio below 0.25 presented, on average, larger sample sizes, shorter primary endpoint length, slightly longer total planned recruitment length, and slightly faster expected recruitment rate. No relevant trends were observed regarding study phase, funder type, study sponsor, or site configuration. To demonstrate how the ratio of primary endpoint length to recruitment length influences an adaptive design, we plotted these characteristics using two ongoing OA studies as examples: one likely to benefit from an adaptive design and one unlikely to do so (Supplementary Material 2).
In terms of study design, 41 unique records were identified for further investigation, and only eight (1.8%) were considered to have an adaptive design (Supplementary Material 1 and 3). Four studies (NCT06463847, NCT04893174, NCT05211986, and NCT06218784) described a dose escalation design, two studies (NCT06000410 and NCT05761015), a group sequential design, and two studies (NCT05986292 and NCT05929066), a master protocol [12]. Six of them presented a ratio below 0.25 (Supplementary Material 3, Supplementary Table S2).
4. Discussion
To the best of our knowledge, this is the first study providing a landscape analysis of the use and applicability of adaptive designs in the field of musculoskeletal health. Based on 579 LBP and OA ongoing trials registered at ClinicalTrials.gov, the use of adaptive designs in the field is rare (9 studies, 1.6%, with one overlap), despite most studies (59%) presenting methodological characteristics that suggest adaptive designs might be a feasible approach. The most common types of adaptive designs were dose escalation studies (6 studies, 67%), followed by group-sequential designs (2 studies, 22%), and master protocols (1 study, 11%). Additionally, five studies reported a SMART design, where some or all participants are randomised to different treatment options at two or more decision points, depending on their response to previous interventions [39]. Because SMART studies adapt interventions rather than the trial conduct, they were not accounted as adaptive for the purposes of this review. Although adaptive designs (Fig. 4), master protocols (Fig. 5), and SMART designs (Fig. 6) are innovative trial designs that share the goal of improving the efficiency of clinical trials, they differ in their key design features and, consequently, in their potential applications within musculoskeletal health research.
Fig. 4.
Most common types of adaptive designs, with illustrations of their key design features, as well as descriptions of their concept and potential applications. The allocation ratios, sample sizes, number of arms, and number of interim analyses are provided for illustrative purposes to demonstrate the concept of the design; these factors can assume different values depending on the aim of the study and the research question.
Fig. 5.
Most common types of master protocols, with illustrations of their key design features, as well as descriptions of their concept and potential applications. The allocation ratios, sample sizes, number of arms, and number of interim analyses are provided for illustrative purposes to demonstrate the concept of the design; these factors can assume different values depending on the aim of the study and the research question.
Fig. 6.
Sequential, multiple assignment, randomised trial (SMART) design, with an illustration of its key design features, as well as a description of its concept and potential applications. The number and type of arms and the number of decision points are provided for illustrative purposes to demonstrate the concept of the design; these factors can vary depending on the aim of the study and the research question.
Most ongoing studies (342 studies, 59%, with three overlaps) presented a ratio of primary endpoint length to recruitment length below 0.25, suggesting these studies could potentially benefit from adaptive designs. In practical aspects, this would mean more efficiency and flexibility to detect clinically relevant findings, if present. Trials with a ratio below 0.25 tended to present larger sample sizes, shorter primary endpoint lengths, and longer total planned recruitment lengths, trends observed for both LBP and OA studies, and a slower expected recruitment rate in LBP trials compared to a slightly faster expected recruitment rate in OA trials. No relevant trends were observed regarding study phase, funder type, study sponsor, or site configuration for LBP or OA. Methodological considerations of the present study include the conservative approach of using the longest endpoint length registered under “Primary Outcome Measures” to calculate the ratio of primary endpoint length to recruitment length. This approach generated relatively larger ratios, therefore producing a reduced estimate of the frequency and proportion of trials for which adaptive designs could be potentially feasible. Had the shortest endpoint been selected for calculations, these estimates would have been higher.
The underutilisation of adaptive designs in the field of LBP and OA has several potential underlying factors. First, musculoskeletal health research has been markedly and disproportionately underfunded, both relative to the associated disease burden and compared to other disease areas where adaptive designs have become increasingly adopted, such as oncology and cardiovascular diseases [[40], [41], [42]]. Financially overstretched settings, in turn, pose a major barrier to the widespread adoption of novel study designs that require larger and more complex operational infrastructure, advanced technical expertise, and additional resource management and capacity planning [43]. In the current dataset, most trials were single-centre, sponsored by non-industry institutions, and funded by universities, community-based organisations, or individuals. Arguably, these investigator-led trials would have limited resources and restricted access to a lead statistician with expertise in adaptive trial designs to support optimal and efficient planning prior to grant application submissions, and to guide trial conduct and analysis. The type of funder may also influence the choice of adaptive design. For example, sample size re-estimation may be of limited utility in federally funded studies, where resources are allocated over a pre-specified period. Such designs entail acceptance of the potential need to increase the target sample size to answer the research question [43,44]. In contrast, multi-arm multi-stage designs might be useful to efficiently test multi-component interventions [45]. The type of trial sponsorship has been shown to affect the planning of interim analysis, with public-sponsored trials typically utilising interim analyses more often and earlier [44,46]. Therefore, it is important to recognise that the effort required to design and conduct an adaptive trial may not be feasible for some small-to mid-sized, academic-led clinical trials with limited internal data management and statistical support.
Additional layers of contributing factors hampering the uptake of adaptive designs in musculoskeletal health research are related to the inherent study characteristics in the field, which may pose greater challenges to the adoption of adaptive designs compared with other conditions. For example, a considerable proportion of clinical trials in musculoskeletal conditions, particularly of non-pharmacological interventions, are powered for large, overly optimistic effect sizes, resulting in small sample sizes. Proposing interim analyses halfway through a small trial would result in even smaller sample sizes per group, which would not be considered reliable enough to justify a change in practice. Another consideration is the extensive reliance on patient-reported outcome measures, which may be susceptible to bias and measurement error, and exhibit limited responsiveness, repeatability, and validity. These aspects may hinder a consensus among investigators on the choice of outcomes to inform the adaptations. Preferably, when considering self-reported measures at the interim(s), triallists should prioritise those with acceptable responsiveness, reliability, and validity. Finally, the lack of awareness of adaptive designs among stakeholders, including investigators, sponsors, funding bodies, and clinicians, should be recognised. Literature has shown that, even among those who are familiar with the concept of adaptive designs, these are often perceived as complex, challenging to implement, difficult to interpret, and associated with considerable uncertainty [[47], [48], [49]]. To address these barriers and increase the uptake of adaptive designs in musculoskeletal health, funding agencies, research institutions, and regulatory authorities at local, national and international levels should optimise funding schemes for musculoskeletal conditions, incentivise the adoption of innovative study designs by promoting training and education and fostering cross-sectoral partnerships for capacity building, as well as encourage early engagement among stakeholders when co-designing research. In settings constrained by limited funding and workforce shortages, it is recommended to establish clinical trial design centres within research institutes, universities, or at national clinical trial networks, with substantial methodological expertise and the capacity to effectively support research teams in the design, conduct, and evaluation of innovative clinical trials. Additionally, research teams that plan to conduct an internal pilot to assess recruitment feasibility could leverage it to undertake a formal statistical futility analysis that could then optimise resource allocation [46].
It is worth noting that a low ratio of primary endpoint length to recruitment length is neither a sufficient nor a necessary condition to design an adaptive trial. Investigators may consider it worthwhile to adapt a study with a high (i.e., above 0.25) ratio despite the efficiency loss [50]. It is also important to consider that trialists tend to project overoptimistic recruitment estimates at the design stage [51], usually reflected in the trial registration, and that unanticipated disruptions to the recruitment timeline during the conduct stage are not uncommon, thereby generating actual ratios that are lower than those initially projected. In some circumstances, it may be considered advantageous or desirable, albeit not without risk, to adapt the study based on short-term, non-primary endpoints that are highly correlated with a longer-term primary outcome [52]. However, when using a surrogate endpoint at the interim(s), a valid assumption is needed and essential, and further challenges are expected when self-reported outcome measures are involved. In other scenarios, special emphasis should be given to what “total evidence” is required to affect a change in policy or how much evidence is required about secondary endpoints, including harms. Therefore, careful consideration of the trial's overall goals is of utmost importance when determining: 1) whether an adaptive design is appropriate, and 2) the type of adaptations to be implemented. In some circumstances, the challenges and limitations of conducting an adaptive trial may outweigh the alleged efficiency advantages and, therefore, a more straightforward approach would be preferable [15].
This registry-based review has limitations that should be acknowledged. First, the searches were conducted exclusively on the ClinicalTrials.gov registry, which constrains the generalisability of our findings to other contexts. Although ClinicalTrials.gov is one of the largest trial registries, encompassing more than 500,000 studies from over 200 countries and territories, many other trial registries exist globally, such as the ICTRP, ANZCTR, EudraCT, and ISRCTN. Consequently, our findings may not fully capture the breadth of the entire clinical trials landscape. The lack of a systematic search and analysis of published study protocols and statistical analysis plans associated with the included records is another limitation to consider. While the approach was not feasible within the scope of the current study, future research should consider the systematic inclusion of study protocols and statistical analysis plans, where available, for a richer dataset. Another limitation concerns the intrinsic restrictions imposed by most trial registration systems, typically composed of structured and unstructured free text fields, some mandatory, yet others optional. As these templates are designed to capture simplified information, they are limited in their capacity to reflect all protocol details, hampering the use and analysis of registration metadata [53]. Hence, some authors may use adaptive designs without highlighting them in the study registration, as identified in earlier reviews [12,54] and frequently observed when group-sequential stopping boundaries or blinded sample size re-estimation are used. There is a growing interest towards enhancing transparency and strengthening the reporting of adaptive designs, as demonstrated by the publication of the Adaptive Designs Consolidated Standards of Reporting Trials (CONSORT) Extension statement [55]. A similar initiative in the context of trial registers would help address the issue of insufficient reporting and promote appropriate indexing of different study designs, therefore optimising access, dissemination, transparency, and accountability of planned and ongoing adaptive trials. While trial registers should consider expanding the scope of designs covered in structured data elements, such as the “Study Design” field, investigators should clearly identify the study as adaptive by including terms like “adaptive design” or “adaptive trial” in the title, abstract, and design sections, and provide detailed specifications of the pre-planned adaptations, corresponding decision rules, and schedule for interim analyses [10,54]. Finally, a major limitation of the current study design lies in the lack of qualitative data from investigators of the included trials on whether adaptive designs were considered, and if so, the reasons why they were not the preferred approach.
In conclusion, although the ratio of primary endpoint length to recruitment length is an arbitrary heuristic and not the sole factor to be considered when designing an adaptive trial, this feature is a useful and pragmatic methodological indicator of the potential feasibility of an adaptive design over the conventional approach. From a methodological perspective, given that most ongoing trials in LBP and OA have a relatively low ratio, there is evidence prima facie that many trials in this field could potentially benefit from adaptive designs. In practice, the specificities of musculoskeletal clinical research, including the prevailing research funding landscape and the typical study configuration, may limit the proportion of studies for which adaptive designs are feasible and advantageous. While wider awareness and acceptance of adaptive designs in the field of LBP and OA should be facilitated and encouraged, we suggest that researchers consider the trade-offs between trial efficiency and the practical complexity of an adaptive design to plan future trials that are more efficient, ethical, and minimise research waste.
Ethical approval
Ethical approval was not required for this review.
Author contributions
AMCS: Conceptualisation, methodology, software, resources, writing the original draft, visualisation, project administration. MW: Software, resources, review. MH: Conceptualisation, methodology, review. MLF: Review. PRB: Review, supervision. TS: Review. SNL: Review. LB: Review. JMSW: Conceptualisation, methodology, resources, review. HZ: Review. PHF: Conceptualisation, methodology, software, resources, review, supervision, project administration. All authors revised and approved the final version of the manuscript.
AMCS take responsibility for the integrity of the work from inception to the finished manuscript.
Data availability
All data supporting the findings of this study are available within the article and its supplementary materials or from the corresponding author upon reasonable request.
Funding
This research received no specific funding from any public, commercial, or not-for-profit organisations. AMCS holds a Faculty of Medicine and Health Research Centres Stipend Scholarship and an International Tuition Fee Scholarship, both awarded by the University of Sydney. MH holds a National Health and Medical Research Council of Australia Investigator Leadership Grant (#1172928). MLF holds a National Health and Medical Research Council of Australia Investigator Leadership Grant. SNL is funded by Melanoma Institute Australia. JMSW is funded by a National Institute for Health and Care Research Research Professorship (NIHR301614). HZ is funded by Cancer Research UK through a Career Development Fellowship (RCCCDF-May24/100001). PHF holds a Rewarding Research Success Funding Scheme (Grant ID: G216248, PC: RX450), awarded by the Faculty of Medicine and Health, The University of Sydney.
The funders or sponsors had no role in the conceptualisation, design of the study, data collection, management, analysis, interpretation of data, writing and preparation of the report, or the decision to submit the report for publication.
Declaration of competing interests
The authors declare that they have no competing interests.
Acknowledgements
The authors are thankful to the University of Sydney (Australia) for the support provided.
Handling Editor: Professor H Madry
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.ocarto.2026.100777.
Contributor Information
Anelise Moreti Cabral Silveira, Email: anelise.silveira@sydney.edu.au.
Michelle Wang, Email: mwan5189@uni.sydney.edu.au.
Michelle Hall, Email: michelle.hall@sydney.edu.au.
Manuela L. Ferreira, Email: mferreira@georgeinstitute.org.au.
Paula R. Beckenkamp, Email: paula.beckenkamp@sydney.edu.au.
Thomas Snelling, Email: tom.snelling@sydney.edu.au.
Serigne N. Lo, Email: serigne.lo@sydney.edu.au.
Laurent Billot, Email: lbillot@georgeinstitute.org.
James M.S. Wason, Email: james.wason@newcastle.ac.uk.
Haiyan Zheng, Email: hz2075@bath.ac.uk.
Paulo H. Ferreira, Email: paulo.ferreira@sydney.edu.au.
Appendix A. Supplementary data
The following is/are the supplementary data to this article.
References
- 1.Liu S., Wang B., Fan S., Wang Y., Zhan Y., Ye D. Global burden of musculoskeletal disorders and attributable factors in 204 countries and territories: a secondary analysis of the global burden of disease 2019 study. BMJ Open. 2022;12(6) doi: 10.1136/bmjopen-2022-062183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.GBD 2021 Low Back Pain Collaborators Global, regional, and national burden of low back pain, 1990-2020, its attributable risk factors, and projections to 2050: a systematic analysis of the global burden of disease study 2021. Lancet Rheumatol. 2023;5(6):e316–e329. doi: 10.1016/S2665-9913(23)00098-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.GBD 2021 Osteoarthritis Collaborators Global, regional, and national burden of osteoarthritis, 1990-2020 and projections to 2050: a systematic analysis for the global burden of disease study 2021. Lancet Rheumatol. 2023;5(9):e508–e522. doi: 10.1016/S2665-9913(23)00163-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Burns P.B., Rohrich R.J., Chung K.C. The levels of evidence and their role in evidence-based medicine. Plast. Reconstr. Surg. 2011;128(1):305–310. doi: 10.1097/PRS.0b013e318219c171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bhatt D.L., Mehta C. Adaptive designs for clinical trials. N. Engl. J. Med. 2016;375(1):65–74. doi: 10.1056/NEJMra1510061. [DOI] [PubMed] [Google Scholar]
- 6.Farrell B. Efficient management of randomised controlled trials: nature or nurture. BMJ. 1998;317(7167):1236–1239. doi: 10.1136/bmj.317.7167.1236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chalmers I., Bracken M.B., Djulbegovic B., Garattini S., Grant J., Gulmezoglu A.M., et al. How to increase value and reduce waste when research priorities are set. Lancet. 2014;383(9912):156–165. doi: 10.1016/S0140-6736(13)62229-1. [DOI] [PubMed] [Google Scholar]
- 8.Chalmers I., Glasziou P. Avoidable waste in the production and reporting of research evidence. Lancet. 2009;374(9683):86–89. doi: 10.1016/S0140-6736(09)60329-9. [DOI] [PubMed] [Google Scholar]
- 9.Hoffmann J.M., Bauer A., Grossmann R. The carbon footprint of clinical trials: a global survey on the status quo and current regulatory guidance. BMJ Glob. Health. 2023;8(9) doi: 10.1136/bmjgh-2023-012754. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pallmann P., Bedding A.W., Choodari-Oskooei B., Dimairo M., Flight L., Hampson L.V., et al. Adaptive designs in clinical trials: why use them, and how to run and report them. BMC Med. 2018;16(1):1–15. doi: 10.1186/s12916-018-1017-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Thorlund K., Haggstrom J., Park J.J., Mills E.J. Key design considerations for adaptive clinical trials: a primer for clinicians. BMJ. 2018;360(k698) doi: 10.1136/bmj.k698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Park J.J.H., Siden E., Zoratti M.J., Dron L., Harari O., Singer J., et al. Systematic review of basket trials, umbrella trials, and platform trials: a landscape analysis of master protocols. Trials. 2019;20(1):572. doi: 10.1186/s13063-019-3664-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Li Q., Lin J., Lin Y. Adaptive design implementation in confirmatory trials: methods, practical considerations and case studies. Contemp. Clin. Trials. 2020;98 doi: 10.1016/j.cct.2020.106096. [DOI] [PubMed] [Google Scholar]
- 14.Millen G.C., Yap C. Adaptive trial designs: what are multiarm, multistage trials? Arch Dis Child Educ Pract Ed. 2020;105(6):376–378. doi: 10.1136/archdischild-2019-317826. [DOI] [PubMed] [Google Scholar]
- 15.Wason J.M.S., Brocklehurst P., Yap C. When to keep it simple - adaptive designs are not always useful. BMC Med. 2019;17(1):152. doi: 10.1186/s12916-019-1391-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.US Food and Drug Administration . U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER) Center for Biologics Evaluation and Research (CBER). Silver Spring: Food and Drug Administration; 2019. Adaptive designs for clinical trials of drugs and biologics - Guidance for Industry; p. 33. [Google Scholar]
- 17.Mukherjee A., Grayling M.J., Wason J.M.S. Adaptive designs: benefits and cautions for neurosurgery trials. World Neurosurg. 2022;161:316–322. doi: 10.1016/j.wneu.2021.07.061. [DOI] [PubMed] [Google Scholar]
- 18.Sully B.G.O., Julious S.A., Nicholl J. An investigation of the impact of futility analysis in publicly funded trials. Trials. 2014;15(1):61. doi: 10.1186/1745-6215-15-61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Berry D.A. Adaptive clinical trials in oncology. Nat. Rev. Clin. Oncol. 2011;9(4):199–207. doi: 10.1038/nrclinonc.2011.165. [DOI] [PubMed] [Google Scholar]
- 20.Lawler PR, Goligher EC, Berger JS, Neal MD, McVerry BJ, Nicolau JC, Gong MN, Carrier M, Rosenson RS, Reynolds HR, Turgeon AF, Escobedo J, Huang DT, Bradbury CA, Houston BL, Kornblith LZ, Kumar A, Kahn SR, Cushman M, McQuilten Z, Slutsky AS, Kim KS, Gordon AC, Kirwan BA, Brooks MM, Higgins AM, Lewis RJ, Lorenzi E, Berry SM, Berry LR, Aday AW, Al-Beidh F, Annane D, Arabi YM, Aryal D, Baumann Kreuziger L, Beane A, Bhimani Z, Bihari S, Billett HH, Bond L, Bonten M, Brunkhorst F, Buxton M, Buzgau A, Castellucci LA, Chekuri S, Chen JT, Cheng AC, Chkhikvadze T, Coiffard B, Costantini TW, de Brouwer S, Derde LPG, Detry MA, Duggal A, Džavík V, Effron MB, Estcourt LJ, Everett BM, Fergusson DA, Fitzgerald M, Fowler RA, Galanaud JP, Galen BT, Gandotra S, García-Madrona S, Girard TD, Godoy LC, Goodman AL, Goossens H, Green C, Greenstein YY, Gross PL, Hamburg NM, Haniffa R, Hanna G, Hanna N, Hegde SM, Hendrickson CM, Hite RD, Hindenburg AA, Hope AA, Horowitz JM, Horvat CM, Hudock K, Hunt BJ, Husain M, Hyzy RC, Iyer VN, Jacobson JR, Jayakumar D, Keller NM, Khan A, Kim Y, Kindzelski AL, King AJ, Knudson MM, Kornblith AE, Krishnan V, Kutcher ME, Laffan MA, Lamontagne F, Le Gal G, Leeper CM, Leifer ES, Lim G, Lima FG, Linstrum K, Litton E, Lopez-Sendon J, Lopez-Sendon Moreno JL, Lother SA, Malhotra S, Marcos M, Saud Marinez A, Marshall JC, Marten N, Matthay MA, McAuley DF, McDonald EG, McGlothlin A, McGuinness SP, Middeldorp S, Montgomery SK, Moore SC, Morillo Guerrero R, Mouncey PR, Murthy S, Nair GB, Nair R, Nichol AD, Nunez-Garcia B, Pandey A, Park PK, Parke RL, Parker JC, Parnia S, Paul JD, Pérez González YS, Pompilio M, Prekker ME, Quigley JG, Rost NS, Rowan K, Santos FO, Santos M, Olombrada Santos M, Satterwhite L, Saunders CT, Schutgens REG, Seymour CW, Siegal DM, Silva DG, Jr., Shankar-Hari M, Sheehan JP, Singhal AB, Solvason D, Stanworth SJ, Tritschler T, Turner AM, van Bentum-Puijk W, van de Veerdonk FL, van Diepen S, Vazquez-Grande G, Wahid L, Wareham V, Wells BJ, Widmer RJ, Wilson JG, Yuriditsky E, Zampieri FG, Angus DC, McArthur CJ, Webb SA, Farkouh ME, Hochman JS, ATTACC Investigators, ACTIV-4a Investigators, REMAP-CAP Investigators Zarychanski R. Therapeutic anticoagulation with heparin in noncritically ill patients with Covid-19. N. Engl. J. Med. 2021;385(9):790–802. doi: 10.1056/NEJMoa2105911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lawler P.R., Hochman J.S., Zarychanski R. What are adaptive platform clinical trials and what role may they have in cardiovascular medicine? Circulation. 2022;145(9):629–632. doi: 10.1161/CIRCULATIONAHA.121.058113. [DOI] [PubMed] [Google Scholar]
- 22.Mehta C., Gao P., Bhatt D.L., Harrington R.A., Skerjanec S., Ware J.H. Optimizing trial design. Circulation. 2009;119(4):597–605. doi: 10.1161/CIRCULATIONAHA.108.809707. [DOI] [PubMed] [Google Scholar]
- 23.Zhang J., Saju C. A systematic review of randomised controlled trials with adaptive and traditional group sequential designs – applications in cardiovascular clinical trials. BMC Med. Res. Methodol. 2023;23(1):200. doi: 10.1186/s12874-023-02024-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lanini S., Ioannidis J.P.A., Vairo F., Pletschette M., Portella G., Di Bari V., et al. Non-inferiority versus superiority trial design for new antibiotics in an era of high antimicrobial resistance: the case for post-marketing, adaptive randomised controlled trials. Lancet Infect. Dis. 2019;19(12):e444–e451. doi: 10.1016/S1473-3099(19)30284-1. [DOI] [PubMed] [Google Scholar]
- 25.Stallard N., Hampson L., Benda N., Brannath W., Burnett T., Friede T., et al. Efficient adaptive designs for clinical trials of interventions for COVID-19. Stat. Biopharm. Res. 2020;12(4):483–497. doi: 10.1080/19466315.2020.1790415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Denholm J.T., Venkatesh B., Davis J., Bowen A.C., Hammond N.E., Jha V., et al. ASCOT ADAPT study of COVID-19 therapeutics in hospitalised patients: an international multicentre adaptive platform trial. Trials. 2022;23(1):1014. doi: 10.1186/s13063-022-06929-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Huskins W.C., Fowler V.G., Jr., Evans S. Adaptive designs for clinical trials: application to healthcare epidemiology research. Clin. Infect. Dis. 2017;66(7):1140–1146. doi: 10.1093/cid/cix907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.US Food and Drug Administration . Adaptive Designs for Medical Device Clinical Studies: Guidance for Industry and Food and Drug Administration Staff, U.S. Department of Health and Human Services, Food and Drug Administration, Center for Devices and Radiological Health (CDRH), Center for Biologics Evaluation and Research (CBER) Silver Spring: Food and Drug Administration; 2016. p. 45. [Google Scholar]
- 29.Zapf A., Stark M., Gerke O., Ehret C., Benda N., Bossuyt P., et al. Adaptive trial designs in diagnostic accuracy research. Stat. Med. 2020;39(5):591–601. doi: 10.1002/sim.8430. [DOI] [PubMed] [Google Scholar]
- 30.Ben-Eltriki M., Rafiq A., Paul A., Prabhu D., Afolabi M.O.S., Baslhaw R., et al. Adaptive designs in clinical trials: a systematic review-part I. BMC Med. Res. Methodol. 2024;24(1):229. doi: 10.1186/s12874-024-02272-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Stockley R.C., Graham I.S. The importance of embracing complexity in rehabilitation. J. Eval. Clin. Pract. 2023;29(4):657–661. doi: 10.1111/jep.13715. [DOI] [PubMed] [Google Scholar]
- 32.Page M.J., McKenzie J.E., Bossuyt P.M., Boutron I., Hoffmann T.C., Mulrow C.D., et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372 doi: 10.1136/bmj.n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ursprung S., Mossop H., Gallagher F.A., Sala E., Skells R., Sipple J.A.N., et al. The WIRE study a phase II, multi-arm, multi-centre, non-randomised window-of-opportunity clinical trial platform using a Bayesian adaptive design for proof-of-mechanism of novel treatment strategies in operable renal cell cancer - a study protocol. BMC Cancer. 2021;21(1):1238. doi: 10.1186/s12885-021-08965-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Waldron-Lynch F., Kareclas P., Irons K., Walker N.M., Mander A., Wicker L.S., et al. Rationale and study design of the adaptive study of IL-2 dose on regulatory T cells in type 1 diabetes (DILT1D): a non-randomised, open label, adaptive dose finding trial. BMJ Open. 2014;4(6) doi: 10.1136/bmjopen-2014-005559. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Englert S., Kieser M. Adaptive designs for single-arm phase II trials in oncology. Pharm. Stat. 2012;11(3):241–249. doi: 10.1002/pst.541. [DOI] [PubMed] [Google Scholar]
- 36.Gotte H., Kirchner M., Krisam J., Allignol A., Lamy F.X., Schuler A., et al. An adaptive design for early clinical development including interim decision for single-arm trial with external controls or randomized trial. Pharm. Stat. 2022;21(3):625–640. doi: 10.1002/pst.2190. [DOI] [PubMed] [Google Scholar]
- 37.Bofill Roig M., Burgwinkel C., Garczarek U., Koenig F., Posch M., Nguyen Q., et al. On the use of non-concurrent controls in platform trials: a scoping review. Trials. 2023;24(1):408. doi: 10.1186/s13063-023-07398-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Magaret A., Angus D.C., Adhikari N.K.J., Banura P., Kissoon N., Lawler J.V., et al. Design of a multi-arm randomized clinical trial with no control arm. Contemp. Clin. Trials. 2016;46:12–17. doi: 10.1016/j.cct.2015.11.003. [DOI] [PubMed] [Google Scholar]
- 39.Kidwell K.M., Almirall D. Sequential, multiple assignment, randomized trial designs. JAMA. 2023;329(4):336–337. doi: 10.1001/jama.2022.24324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Lewis R., Gómez Álvarez C.B., Rayman M., Lanham-New S., Woolf A., Mobasheri A. Strategies for optimising musculoskeletal health in the 21st century. BMC Muscoskelet. Disord. 2019;20(1):164. doi: 10.1186/s12891-019-2510-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.McConaghy K., Klika A.K., Apte S.S., Erdemir A., Derwin K., Piuzzi N.S. A call to action for musculoskeletal research funding: the growing economic and disease burden of musculoskeletal conditions in the United States is not reflected in musculoskeletal research funding. J Bone Joint Surg Am. 2023;105(6):492–498. doi: 10.2106/JBJS.22.00693. [DOI] [PubMed] [Google Scholar]
- 42.Nguyen A.T., Aris I.M., Snyder B.D., Harris M.B., Kang J.D., Murray M., et al. Musculoskeletal health: an ecological study assessing disease burden and research funding. Lancet Reg. Health, Am. 2024;29 doi: 10.1016/j.lana.2023.100661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Wilson N., Biggs K., Bowden S., Brown J., Dimairo M., Flight L., et al. Costs and staffing resource requirements for adaptive clinical trials: quantitative and qualitative results from the costing adaptive trials project. BMC Med. 2021;19(1):251. doi: 10.1186/s12916-021-02124-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Zhang Q., Dimairo M., Julious S.A., Lewis J., Yu Z. Reporting and communication of sample size calculations in adaptive clinical trials: a review of trial protocols and grant applications. BMC Med. Res. Methodol. 2024;24(1):216. doi: 10.1186/s12874-024-02339-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Silveira A.M.C., Comachio J., Mesa-Castrillon C., Roberts K., Beckenkamp P., Ho E.K., et al. My back exercise app: an automated exercise intervention supported by educational notifications, a sleep programme and diet advice to improve function in people with chronic non-specific low back pain – protocol for a superiority, adaptive multi-arm multi-stage randomised controlled trial. BMJ Open. 2025;15(7) doi: 10.1136/bmjopen-2024-098324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Herbert E., Julious S.A., Goodacre S. Progression criteria in trials with an internal pilot: an audit of publicly funded randomised controlled trials. Trials. 2019;20(1):493. doi: 10.1186/s13063-019-3578-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Meurer W.J., Legocki L., Mawocha S., Frederiksen S.M., Guetterman T.C., Barsan W., et al. Attitudes and opinions regarding confirmatory adaptive clinical trials: a mixed methods analysis from the adaptive designs accelerating promising trials into treatments (ADAPT-IT) project. Trials. 2016;17(1):373. doi: 10.1186/s13063-016-1493-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Dimairo M., Julious S.A., Todd S., Nicholl J.P., Boote J. Cross-sector surveys assessing perceptions of key stakeholders towards barriers, concerns and facilitators to the appropriate use of adaptive designs in confirmatory trials. Trials. 2015;16(1):585. doi: 10.1186/s13063-015-1119-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Morgan C.C., Huyck S., Jenkins M., Chen L., Bedding A., Coffey C.S., et al. Adaptive design: results of 2012 survey on perception and use. Therapeutic Innovation & Regulatory Science. 2014;48(4):473–481. doi: 10.1177/2168479014522468. [DOI] [PubMed] [Google Scholar]
- 50.Hampson L.V., Jennison C. Group sequential tests for delayed responses (with discussion) J. Roy. Stat. Soc. B. 2013;75(1):3–54. [Google Scholar]
- 51.Briel M., Elger B.S., McLennan S., Schandelmaier S., von Elm E., Satalkar P. Exploring reasons for recruitment failure in clinical trials: a qualitative study with clinical trial stakeholders in Switzerland, Germany, and Canada. Trials. 2021;22(1):844. doi: 10.1186/s13063-021-05818-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Jenkins M., Stone A., Jennison C. An adaptive seamless phase II/III design for oncology trials with subpopulation selection using correlated survival endpoints. Pharm. Stat. 2011;10(4):347–356. doi: 10.1002/pst.472. [DOI] [PubMed] [Google Scholar]
- 53.Miron L., Gonçalves R.S., Musen M.A. Obstacles to the reuse of study metadata in clinicalTrials.gov. Sci. Data. 2020;7(1):443. doi: 10.1038/s41597-020-00780-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Hatfield I., Allison A., Flight L., Julious S.A., Dimairo M. Adaptive designs undertaken in clinical research: a review of registered clinical trials. Trials. 2016;17(1):150. doi: 10.1186/s13063-016-1273-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Dimairo M., Pallmann P., Wason J., Todd S., Jaki T., Julious S.A., et al. The adaptive designs CONSORT extension (ACE) statement: a checklist with explanation and elaboration guideline for reporting randomised trials that use an adaptive design. Br. Med. J. 2020;369 doi: 10.1136/bmj.m115. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data supporting the findings of this study are available within the article and its supplementary materials or from the corresponding author upon reasonable request.