Prescription of prosthetic ankle‐foot mechanisms after major lower limb amputation

Objectives

This is a protocol for a Cochrane Review (intervention). The objectives are as follows:

To assess the effects of different prosthetic ankle‐foot mechanisms for improving health‐related quality of life, functional, and biomechanical outcomes in adult prosthesis users after major lower limb amputation.

Background

Description of the condition

Amputation

An amputation is the removal of a body part, generally as a result of disease or trauma, although this can also be related to congenital conditions. A major lower limb amputation is the removal of a body part through or above the ankle; toe and partial foot amputations are categorised as minor lower limb amputations. Following a major lower limb amputation, people typically receive a prosthesis to replace the missing limb with the goal of aiding mobility and thereby quality of life. People with a below‐knee or transtibial amputation will have a prosthetic ankle‐foot, while people with a through‐knee or transfemoral amputation will also have a prosthetic knee joint. Anyone with a hip disarticulation or hemipelvectomy will also have a prosthetic hip joint. Most major lower limb amputations result in some gait and balance abnormalities when using a prosthesis.

In higher‐income countries, most people who experience a major lower limb amputation are over 50 years of age and undergo the amputation because of diabetes mellitus or peripheral arterial disease, or both. In lower‐income countries, amputations are more commonly associated with trauma (e.g. land mines, road traffic accidents) and infection. Although the overall trend of the number of major amputations declined internationally between 2010 and 2014 [1], and declined in England between 2003 and 2013 [2], living with the condition long‐term still presents significant physical and mental challenges for people with limb loss and their caregivers, and financial burden for health services.

Categorisation of functional mobility

People who have had a major lower limb amputation are often categorised according to their functional status. A widely‐used system is the Medicare Functional Classification Levels (also known as K‐level classification) [3]. There are five functional K‐level categories (Table 1). Many prosthetic ankle‐feet have been designed to meet the complex needs of the person according to their K‐level classification. However, there is no consensus on the most suitable ankle‐foot mechanism for people according to their broader functional status. The Special Interest Group in Amputee Medicine (SIGAM) mobility grade is another method of categorising people according to their level of mobility, including: assistance to mobilise, walking distance, use of aids, and ability to negotiate difficult environments and adverse weather conditions [4]. The SIGAM scale includes six grades (Table 1). This Cochrane review will focus on people who use a prosthetic ankle‐foot following a major lower limb amputation (hereafter referred to as prosthesis users).

1. Descriptions of the K‐level classification and SIGAM mobility grade categories.

K‐level classification categories		SIGAM mobility grades
K0	Non‐prosthesis user. The person cannot walk with a prosthesis or a prosthesis would not improve their quality of life.	A	Non‐prosthesis user, or the person wears a prosthetic limb for cosmetic purpose only.
K1	The person has the ability, or potential, to use a prosthesis for transfers or to walk on level ground at a steady state. This is representative of a person who mostly uses a prosthesis to move around their home, or indoor environments, to perform basic activities with assistive walking aids.	B	Therapeutic limb wearer. The person wears a prosthesis for transfers only, to assist nursing or walking with the physical aid of another person during therapy.
K2	The person has the ability, or potential, to use a prosthesis to traverse low‐level environmental barriers such as kerbs, steps, stairs or uneven surfaces. They may require additional support using assistive walking devices. This is representative of a person who has limited mobility in the community.	C	The person can walk on level ground for up to 50 metres, with or without a walking aid, categorised as: a = frame b = crutches/sticks c = 1 crutch/stick d = no stick
K3	The person has the ability, or potential, to use a prosthesis for walking with variable cadence. This person can traverse most environmental barriers in the community and uses their prosthesis for work, or therapeutic or physical activities beyond simple walking.	D	The person can walk outdoors on level ground only, for more than 50 metres in good weather, with or without a walking aid categorised as: a = frame b = crutches/sticks c = 1 crutch/stick
		E	The person can walk for more than 50 metres without a walking aid, except occasionally for confidence or to improve confidence in difficult terrain or weather.
K4	The person uses a prosthesis beyond basic ambulation skills, without restriction, and engages in high‐impact activities or sports. This person has the highest level of functional mobility.	F	Normal or near normal gait.

Abbreviations: SIGAM: Special Interest Group in Amputee Medicine

Prosthetic limb provision

Prosthetic limb provision varies widely between low‐ and middle‐income versus high‐income countries, but also due to funding models (e.g. insurance‐based versus publicly‐funded models versus privately‐funded versus for military personnel). While there is clear evidence of the functionality of specific types of ankle‐foot components as described in the literature, their prescription in clinical practice can vary considerably and is frequently cost‐driven or based on clinicians’ experience. For example, the prescription of a prosthesis is primarily influenced by:

• estimation of a person’s functional mobility and their goals;

• level of amputation; and

• amputation aetiology, including general health and residual limb status within the context of prosthetic services’ budget.

It is common for people who have an amputation for vascular reasons, and have lower mobility levels (e.g. K1‐K2 versus K3‐K4), to receive less functional prosthetic components [5]. The 2004 Cochrane review that investigated the prescription of prosthetic ankle‐foot mechanisms concluded there was “insufficient evidence from high quality comparative studies for the overall superiority of any individual type of prosthetic ankle‐foot mechanism” [6]. In the past 20 years, there have been considerable advancements in prosthetic limb technologies that warrant further investigation.

Description of the intervention and how it might work

A prosthesis is an assistive technology that has been designed to enable function, aid mobility, and restore the appearance of the missing limb. All people who have had a major lower limb amputation require a prosthetic ankle‐foot. Some prosthetic ankle‐feet are ‘rigid’, allowing very limited movement; while some facilitate some ankle motion, including plantarflexion‐dorsiflexion, or inversion‐eversion, or both. Further devices use different technologies (e.g. energy storage and return, hydraulic, microprocessor) to help mimic more natural ankle motion.

Why it is important to do this review

Prosthetic limb prescription is often based on the clinical judgements made by prosthetists, physiotherapists, and physicians [7]. There is a lack of clear prescription protocols or criteria, and a lack of robust evidence on the clinical and cost‐effectiveness of the prosthetic ankle‐foot componentry. This means that the prescription of some, often more expensive, prosthetic limbs is hard to justify, even where a clinical benefit or patient preference, or both, is indicated. Consequently, people tend to be prescribed basic functional prostheses that may not be best fit for purpose. This often contributes to patient dissatisfaction with their prosthetic services and may lead to prosthesis abandonment. Moreover, the absence of evidence‐based guidelines can lead to health inequalities, even across the same funding model. It is important that clinicians strive for fair treatment that maximises people’s functional abilities and helps them to achieve their mobility goals.

We searched for different priority‐setting exercises established by international organisations. In 2019, a consensus‐building initiative was undertaken for limb loss stakeholders in Canada. The aim was to establish priority areas for building research capacity related to the broad field of amputation. Their results prioritised three themes:

• the creation of a national dataset;

• obtaining health economic data related to the socioeconomic cost of amputation for their healthcare system and patients; and

• improving strategies related to the measurement of health outcomes (i.e. development and validation) [8].

In 2021, the Vascular Society of Great Britain and Ireland conducted a James Lind Alliance priority‐setting process. Thirty people with a major lower limb amputation and healthcare professionals identified “What are the best ways to support rehabilitation following amputation?” and “How can we improve clinical outcomes for patients following major lower limb amputation?” as the 2nd and 3rd highest‐ranking priorities [9]. Prosthetic limb provision, appropriately prescribed according to the user’s functional needs and preferences, is integral to supporting rehabilitation and improving health outcomes for people following major lower limb amputation. To help achieve better equity of health outcomes and optimal prosthetic prescription, especially for long‐term health conditions such as a major lower limb amputation, it is necessary to undertake a systematic review of current prosthetic ankle‐feet technologies. This is an updated protocol to a Cochrane review first published in 2004 and previously updated to 2006 [6].

Objectives

To assess the effects of different prosthetic ankle‐foot mechanisms for improving health‐related quality of life, functional, and biomechanical outcomes in adult prosthesis users after major lower limb amputation.

Methods

Criteria for considering studies for this review

Types of studies

We will only include published and peer‐reviewed randomised controlled trials (RCTs), controlled clinical trials, crossover RCTs, and cluster‐RCTs. These may include pilot studies and feasibility trials. We will not exclude trials based on language restrictions. We will exclude conference proceedings, theses, opinion pieces, correspondence, and stand‐alone abstracts. We will also exclude adaptive and quasi‐RCTs, where trials may use flexible design methods to allocate people to trial arms, or where the randomisation method is not truly random and may be based on order of recruitment or medical record data.

Types of participants

We aim to include prosthesis users of both sexes over the age of 18 years, who have undergone major lower limb amputation, including ankle disarticulation (Syme's amputation), transtibial amputation, through‐knee amputation (such as Gritti‐Stokes, Nellis/van de Water; Mazet, Burgess, Youkey), transfemoral amputation, hip disarticulation, and hemipelvectomy. Indications for amputation may be vascular and diabetes‐related (e.g. vascular disease, peripheral arterial disease, neuropathy, aneurysm, blood clot, diabetic indications) or non‐vascular (e.g. road traffic accident, workplace accident, explosive device or other trauma, chronic regional pain syndrome, orthopaedic injury, congenital malformation, septicaemia, infection, cancer, tumour, elective). Vascular and diabetes‐related prosthesis users are more likely to experience significant multimorbidities (e.g. hypertension, diabetes, hypercholesterolaemia), whereas non‐vascular prosthesis users are less likely. Prosthesis users may have undergone previous major or minor lower limb surgeries, including salvage attempts, limb reconstruction, or revascularisation procedures. We will include individuals who use either a lower limb prosthesis with a socket system or an osseointegrated system, and who use it for walking indoors or outdoors, or both. Participants with varying activity levels, as classified by the K‐level classification system [3] or SIGAM [4], and residing in the community or in institutional care settings across diverse international regions, will be included.

Other information can be important to describe the group of prosthesis users; e.g. time since amputation and other prosthetic limb componentry. Although many studies will not report the (exact) time since amputation, we will try to capture these details. We will include studies with prosthesis users who have health comorbidities (including bilateral lower limb amputation) and present as much information as possible. We will attempt to report on the effects of the comorbidities where reported in the literature.

Our stop‐go criterion to include a study in our review will be reporting of level of amputation. If a study does not report the participants’ level of amputation, we will not consider it for inclusion. If a study only partially describes its participants’ characteristics, e.g. incomplete description of level of amputation, aetiology, or functional mobility level, we will contact the study authors to try to obtain additional data. We will include a study if more than 70% of participants are eligible based on our stop‐go criterion.

This Cochrane review will exclude studies involving children or individuals with partial foot or toe amputations, and adults who have undergone upper limb amputation only.

We will attempt to analyse prosthesis users using the following subgroups:

• level of amputation;

• functional mobility level (K‐classification level or SIGAM); and

• amputation aetiology: vascular and diabetes‐related or non‐vascular.

Types of interventions

We will include trials that compare commercially available prosthetic ankle‐feet. These include the solid ankle cushioned heel (SACH) foot, single axis foot, multi‐flex ankle (MFA) foot, multi‐axis foot, Seattle foot, FlexFoot, energy storing and release (ESAR) foot, carbon‐fibre ankle‐foot, passive ankle‐foot, hydraulic self‐aligning ankle‐foot, microprocessor ankle‐foot, powered or bionic ankle‐foot mechanisms. The brand, technology, or functionality of the ankle‐foot mechanism must be described. We will consider trials in which participants were randomised to receive one of the aforementioned prosthetic ankle‐feet. Trials will be included where two or more prosthetic ankle‐feet are compared, and the interventions were under the same conditions. We will consider trials for inclusion when they investigate overground walking (including climbing stairs, inclined walking, or ramp walking) or treadmill walking (level and inclined walking). We will exclude trials that only investigate sit‐to‐stand, stand‐to‐sit, transfers, standing, turning, and dynamic balance (e.g. perturbations, rhythmic weight‐shifting) activities.

Trials must present at least health‐related quality of life or functioning with a prothesis (including biomechanical gait outcomes) to compare prosthetic ankle‐feet and at two time points, in case of follow‐up.

We will exclude trials involving surgical amputation techniques, early prosthetic limb fitting (e.g. early walking aids or temporary prosthesis prior to the permanent prosthesis), prototypes, 3D printed ankle‐foot prostheses or prostheses without CE marking (i.e. a certification mark that indicates a product complies with European Union (EU) safety, health, and environmental protection standards), or computational modelling studies from this review.

Comparisons are based on the interventions grouped as follows.

• SACH foot

• Single‐axis foot

• Multi‐axis foot

• Dynamic‐response/ESAR foot

• Self‐aligning foot

• Microprocessor and bionic (computer‐controlled) foot

Outcome measures

We will consider five types of outcome measures: health‐related quality of life and functioning with a prosthesis, including clinical, biomechanical, and physiological measures. Whether a study reports one or more of the outcomes listed here will not be used as a criterion for including studies in this review, ensuring that the most important outcome measures are selected to answer the review’s objectives. We will use the World Health Organization (WHO) International Classification of Functioning, Disability and Health (ICF) to include studies that investigate outcome measures related to walking impairment, activity limitations, and participation restrictions.

Critical outcomes

• Health‐related quality of life outcomes that have been psychometrically tested with a lower limb amputee population. Eligible quality of life outcomes may include:

  • EuroQol Five‐Dimension Questionnaire (EQ‐5D) [10]

  • 12‐item Short‐Form Health Survey (SF‐12) [11]

  • 36‐item Short‐Form Health Survey (SF‐36) [12]

  • Patient‐Reported Outcomes Measurement Information System (PROMIS‐29) [13]

  • Orthotics Prosthetics Users' Survey (OPUS) Health Quality of Life Index [14]

  • WHO Quality of Life (WHOQOL‐BREF) [15]

Important outcomes

• Functioning with a prosthesis outcome measures related to the person’s ability to complete functional tasks. They may have been recommended by the International Society of Prosthetics and Orthotics (COMPASS) [16], and have been psychometrically tested with a lower limb amputee population. These measures can be self‐reported or measured objectively by a clinician or researcher in a pre‐defined standardised manner; e.g. distance walked (measured in metres) in 2 minutes (2‐minute walk test) [17]. Eligible functioning with a prosthesis outcomes may include:

  • Activity Balance Confidence (ABC) scale [18]

  • Amputee Mobility Predictor (AMP) [19]

  • Comprehensive High‐Level Activity Mobility Predictor (CHAMP) [20]

  • Figure 8 walking test (F8WT) [21]

  • Four Square Step Test [22]

  • Houghton scale [23]

  • Locomotor Capabilities Index – 5‐point scale (LCI–5) [24]

  • L‐test [25]

  • Narrowing Beam Walking Test (NBWT) [26]

  • Patient‐Specific Functional Scale (PSFS) [27]

  • Prosthesis Evaluation Questionnaire (PEQ scales) [28]

  • Prosthetic Limb Users Survey of Mobility (PLUS‐M) [29]

  • Prosthetic Mobility Questionnaire (PMQ) [30]

  • SIGAM Mobility Scale [4]

  • Stair and/or hill assessment index [31]

  • Timed walk tests (10 metres or 2 or 6 minutes are common) [32]

  • Timed Up and Go (TUG) test [33]

• Gait outcomes related to:

  • temporal‐spatial characteristics, such as speed (m/s), distance (m), stance or swing duration or both (% gait cycle), stride/step length (m), cadence (steps/m or strides/m)

  • joint kinematics, such as 2D or 3D joint angles and range of motion (ankle, knee, hip joint movements including plantarflexion/extension and dorsiflexion/flexion)

  • kinetics, such as ground reaction forces, centre of pressure, 2D or 3D joint moments and powers (ankle, knee, and hip joints) – either standardised to body weight or not

• Prosthetic satisfaction and comfort outcomes, such as:

  • Satisfaction with Prosthesis Questionnaire (SAT‐PRO) [34]

  • Socket Comfort Score [35]

  • Trinity Amputation and Prosthesis Experience Scales‐Revised [36]

• Energy expenditure as the gross metabolic energy cost of walking expressed in mL O2/kg/min or as energy expenditure in J·kg·s. Heart rate as measured in heart beats per minute (bpm).

We will include any study that fulfils our eligibility criteria and reports at least one of the outcomes of interest, which we have specified in this section. If we exclude studies on the basis of outcomes, we will take care to ascertain that relevant outcomes are not available because they have not been measured, rather than simply not reported, by contacting the study authors.

For the purpose of this review, we will extract data on each outcome for two time points that will correspond with the short‐term (baseline measurement) and long‐term effects (follow‐up measurement). If measures are taken on multiple occasions within a specified timeframe (short‐ and long‐term), we will prioritise outcomes measured initially at baseline to represent the short‐term effects. Equally, if measures are taken on multiple follow‐up occasions, we will prioritise outcomes measured at the latest time point to represent the long‐term effect. We will only combine data for similar time points within a certain timeframe: short term (up to 1 month) and long term (one month or longer).

We will extract continuous data, as reported in the different studies, up to the longest follow‐up period presented, but will only combine data for similar time points.

Search methods for identification of studies

Electronic searches

We will conduct a systematic search of the following databases for RCTs, controlled clinical trials, crossover‐RCTs, and cluster‐RCTs, without language restrictions.

• Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library via the Cochrane Register of Studies Online (CRSO)

• Cochrane Musculoskeletal Group Register

• MEDLINE (Ovid MEDLINE)

• Embase Ovid

• Cumulative Index to Nursing and Allied Health Literature (CINAHL Ultimate via EBSCOhost)

• AMED Ovid (Allied and Complimentary Medicine)

Detailed search strategies for the major databases are in Supplementary material 1. A summary of our participants, interventions, comparisons, and outcomes (PICO) criteria is presented in Supplementary material 2. We will conduct our search from April 2006 to the present, as the original Cochrane review was last updated to April 2006 [6]. Where appropriate, we will combine search strategies with adaptations of the Highly Sensitive Search Strategy designed by Cochrane for identifying RCTs and controlled clinical trials, as described in Chapter 4 of the Cochrane Handbook for Systematic Reviews of Interventions [37].

Searching other resources

We will search the following trial registries for ongoing studies.

• WHO International Clinical Trials Registry Platform (ICTRP) (https://trialsearch.who.int/)

• ClinicalTrials.gov (clinicaltrials.gov)

We will search for any relevant retraction notices and errata for information for studies selected for inclusion via the Retraction Watch Database.

We will also scan the reference lists of included studies.

Data collection and analysis

Selection of studies

We will follow the PRISMA 2020 reporting guidelines when selecting studies for this review [38]. Firstly, we will eliminate any duplicates from the records. Two review authors (NV and CJH) will then independently screen the titles and abstracts of all records. They will determine which studies are potentially eligible for inclusion and eliminate any studies that are clearly irrelevant. These review authors will be unaware of each other’s decision. If there is a disagreement between the two review authors, a third review author (of CH, FCDS, CTB, WvO, RFvE) will discuss the conflict with the review team to reach consensus. Next, we will retrieve the full‐text publications of potentially eligible records and the same review authors (NV and CJH) will assess these against the eligibility criteria independently. We will record all reasons for non‐eligibility using a hierarchy of exclusion reasons. The two review authors will resolve any disagreements through discussion; if a disagreement persists, then we will consult with a third review author (of CH, FCDS, CTB, WvO, RFvE) to arbitrate.

We will report a list of excluded studies. This covers all studies that may initially appear to meet the eligibility criteria, but on further inspection do not, and also those that do not meet all the criteria but are well known and likely to be thought relevant by some readers. We will give the primary reason(s) for exclusion.

We will use Rayyan software for study selection [39]. This is a free online screening and extraction platform that helps expedite the initial screening of abstracts and titles using a process of semi‐automation while incorporating a high level of usability, and supports blinded screening of search results. Rayyan is recommended to systematic review authors looking for suitable and easy‐to‐use tools to support title and abstract screening within healthcare research. This tool consistently demonstrated good alignment with user requirements [40].

We will use Google Translate to translate a non‐English language abstract or full‐text to determine eligibility if needed. If this translation does not provide us with enough information, we will seek out translation services. The study selection process will be illustrated using a PRISMA diagram [41], as outlined in Figure 1.

1.

Data extraction and management

We will use a data collection form to record the study characteristics and outcome data. We will pilot the data collection form with at least one included study. Two review authors (NV and CJH) will extract the study characteristics independently, as follows.

• Publication details: year, country, study authors

• Methods: study design, total duration of study, number of study centres and location, study setting, date of study, withdrawals

• Participants: number, mean age, age range, sex, level of amputation, mobility level of classification (e.g. K‐classification level or SIGAM level), cause of amputation (aetiology, e.g. vascular and diabetes‐related or non‐vascular), time since amputation, presence of multimorbidities, inclusion and exclusion criteria

• Interventions: type of ankle‐foot (including brand or functionality, or both e.g. SACH foot, ESAR foot, hydraulic self‐aligning ankle‐foot), type of walking environment (e.g. level walking, stair walking, inclined walking, overground versus treadmill), comparison ankle‐foot/feet, length of intervention

• Outcomes: primary and secondary outcomes specified and collected, time points reported (as described in Outcome measures)

• Notes: source of funding, notable conflicts of interest of trial authors

If there is a conflict of interest with any potentially eligible study (i.e. due to trial authorship), an other independent review author(s) (of CH, FCDS, CTB, WvO, or RFvE) will replace the original review author(s) [42]. In the process of data extraction and management, we will have to consider source of funding and conflicts of interest of the study authors, which may inform the exploration of directness and heterogeneity of study results, assessment of risk of bias within studies, and assessment of risk of bias in syntheses owing to missing results. In case of conflicts of interest, we will use the upcoming Tool for Addressing Conflicts of Interest in Trials (TACIT) .

We will contact study authors to confirm the eligibility of studies where it is unclear [41]. We also plan to contact study authors to request raw data for level of functional mobility, aetiology, or level of amputation, and for incomplete data for the primary outcome measure, if the descriptions of any study characteristics are incomplete. We aim to resolve any disagreement by discussion within the review team. Any outcome data not reported in a useable way will be listed in a ‘Characteristics of included studies’ table. We will calculate or convert the reported data into the required format for meta‐analysis when needed, following guidance from the Cochrane Handbook for Systematic Reviews of Interventions [43]. A third review author (of CH, FCDS, CTB, WvO, or RFvE) will resolve any disagreement between the review authors. One review author will import data into Review Manager (RevMan) software [44]. We will cross‐check data entry by comparing the outcome data presented in the systematic review with the data extraction form. One review author (of CH, FCDS, CTB, WvO, or RFvE) will spot‐check study characteristics for accuracy against the trial reports.

Risk of bias assessment in included studies

Two review authors (of NV, CH, FCDS, CTB, WvO, RFvE, CJH) will independently assess the risk of bias in each included study using the Cochrane risk of bias tool RoB 2 [45], as outlined in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions [46]. RoB 2 is structured into a fixed set of domains of bias, focusing on different aspects of trial design, conduct, and reporting. We will resolve any disagreements by discussion or by involving another of our review authors (of NV, CH, FCDS, CTB, WvO, RFvE, CJH). We intend that a senior author (NV, CH, RFvE, CJH) will assess a random subset of papers (10% to 20%) for risk of bias, as quality control.

If adequate information is unavailable from the publications, trial protocols, clinical study reports, or other sources, we will contact the study authors to request missing data on risk of bias items. We will assess the risk of bias of specific results of each trial, according to the following domains.

• Randomisation process

• Deviations from intended interventions

• Missing outcome data

• Measurement of the outcome

• Selection of the reported result

We will prioritise health‐related quality of life and functioning with prosthesis for the assessment of the risk of bias in each of the above domains. Following Cochrane’s guidelines, we will choose the effect of interest according to the following order of preference [46].

• The result corresponding to a full intention‐to‐treat (ITT) analysis

• The result corresponding to an analysis (sometimes described as a ‘modified intention‐to‐treat’ (mITT) analysis) that adheres to ITT principles except that participants with missing outcome data are excluded

• A result corresponding to an ‘as‐treated’ or naïve ‘per‐protocol’ analysis, or an analysis from which eligible trial participants were excluded

We will use the signalling questions in the RoB 2 tool and rate each domain as 'low risk of bias', 'some concerns', or 'high risk of bias'. We will consider the algorithm‐proposed judgements and use relevant quotes from studies to support our decisions. We will justify our judgements in the risk of bias tables, and provide reasons that do not follow the algorithm. We will summarise the risk of bias judgements for each of the domains listed across our studies.

When considering treatment effects, as part of the GRADE methodology, we will take into account the risk of bias for the studies that contribute to quality of life or prosthesis use outcome(s), or both. These domain‐level judgements will inform an overall risk of bias judgement or a single result in the form of the following.

• ‘Low risk’ if we judge all domains at ‘low risk’

• ‘Some concerns’ if we judge one to three domains at ‘some concerns’

• ‘High risk’ if we judge one or more domains as being ‘high risk’ or if we judge four domains as being ‘some concerns’

For cluster‐RCTs, we will use the dedicated version of RoB 2 [45, 47], as recommended in Chapter 23 of the Cochrane Handbook for Systematic Reviews of Interventions (Section 23.1.2 and Table 23.1.a) [48]. In addition to the domains listed above, RoB 2 for cluster‐RCTs covers bias arising from the timing of identification and recruitment of participants.

For crossover‐RCTs, we will use the dedicated version of RoB 2 [45, 49], as recommended in Chapter 23 of the Cochrane Handbook for Systematic Reviews of Interventions (Section 23.2.3 and Table 23.2.a) [48]. In addition to the domains listed above, RoB 2 for crossover‐RCTs covers bias arising from bias due to deviations from intended interventions.

We will summarise the risk of bias judgements across different studies for each outcome for each of the domains listed. The overall risk of bias for the result is the least favourable assessment across the domains of bias. We will exclude studies that present high risk of bias. The ‘overall’ risk of bias will be used to assess the certainty of the evidence.

We will use the RoB 2 Excel tool to manage the answers to the signalling questions and risk of bias judgements [49]. We will present these data as supplementary material in the Open Science Framework platform (osf.io).

This review will be conducted according to this published Cochrane protocol. We will report any deviations from the protocol in our review.

Measures of treatment effect

The types of outcome data that we will encounter are most likely to be dichotomous data, continuous data, or ordinal data. To compare these outcome data between two intervention groups (‘effect measures’), we can identify different ways [50], as outlined below.

• Dichotomous outcomes: we will compare between intervention groups using a risk ratio (RR), an odds ratio, a risk difference (RD), or a number needed to treat

• Continuous outcomes: we will compare between intervention groups using a mean difference (MD) or a standardised mean difference (SMD)

We will express dichotomous data as a RR with 95% confidence intervals (CIs). For continuous data using the same scale, we will use the MD and its 95% CI to estimate treatment effect.

If multiple instruments or scales have been used to measure the same outcome domain (e.g. quality of life), we will calculate the SMD with 95% CI. We will enter data presented as a scale with a consistent direction effect. We will use data reported at the longest follow‐up point for all outcomes.

We will interpret the size of effect if an established minimal clinically important difference is reported in the literature.

When studies are frequently reported in more than one publication, we will collate the information from the multiple reports. We will perform the following.

• Extract data from each report separately, then combine information across multiple data collection forms (in case of two or more detailed journal articles)

• Extract data from all reports directly into a single data collection form (in case of one journal article and multiple conference abstracts)

Unit of analysis issues

If trials have multiple comparator arms, or repeated outcome measurements at different time points, we will not combine data between arms unless the study authors presented the data in such a way. For trials with multiple comparator arms, we will first check whether the intervention groups are relevant to the systematic review and exclude those that are not. For repeated time measurements, we will choose the ones corresponding to the short‐term and long‐term measurements as defined in the primary outcomes.

In the event of cluster‐RCTs, we will follow the guidance from the Cochrane Handbook for Systematic Reviews of Interventions regarding cluster‐RCTs to adjust treatment effect estimates for clustering, using estimates of the average cluster size and intracluster correlation coefficient [51]. In the event of crossover‐RCTs, we will extract and report on the data at baseline and with the intervention foot/feet.

Dealing with missing data

If outcome data are missing, we will attempt to contact the study authors to request missing data. We will contact the study authors again after five weeks, and allow for a maximum of six weeks for a response, before we treat the data as missing. We will carefully evaluate important numerical data such as screened, randomly assigned participants, as well as intention‐to‐treat and as‐treated and per‐protocol populations in our risk of bias assessments. For this, we will investigate attrition rates (e.g. dropouts, losses to follow‐up, and withdrawals) and critically appraise issues concerning missing data and imputation methods (e.g. last observation carried forward). For our primary analyses, we will conduct available‐case analyses, considering these issues when assessing the risk of bias and the certainty of the evidence.

For studies in which the SD of the outcome is not available at follow‐up, or we cannot recreate it, we will standardise by the mean of the pooled baseline SD from studies that reported this information. As a sensitivity analysis, we will repeat the analyses after removing studies with missing data. We will address the potential impact of missing data on the findings of the review in our 'Discussion' section.

Reporting bias assessment

Reporting biases may occur when the dissemination of research findings is influenced by the nature and direction of results. If we include 10 studies or more, we will use funnel plots to assess small study biases [52]. If the funnel plots present with asymmetry, we will investigate the different possible reasons [53], such as true heterogeneity of effect with respect to study size, methodological design, and selective non‐reporting, and will seek statistical advice for their interpretation.

Synthesis methods

We only intend to undertake a meta‐analysis if we believe the PICOs are adequately similar for a clinically meaningful result. We will conduct the meta‐analysis via a random‐effects model, using Cochrane’s Q test to check for significant heterogeneity and calculating the I² statistic to quantify heterogeneity. When meta‐analysis is not possible, we will conduct alternative forms of synthesis, including the summary of effect estimates, the combination of P values, and vote counting based on the direction of effects, as described in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions [54]. In such a case, we will follow the Synthesis Without Meta‐analysis (SWiM) reporting guideline for results reporting [55].

We will consider the clinical, methodological, and statistical heterogeneity of the included studies to guide our decisions regarding pooling of the data. We will assess heterogeneity using Chi² and I², and use the below guidance for interpretation, as described in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions [56]:

• 0% to 40%: might not be important

• 30% to 60%: may represent moderate heterogeneity

• 50% to 90%: may represent substantial heterogeneity

• 75% to 100%: considerable heterogeneity

We will interpret the I² statistic in relation to the size and direction of effects and the strength of evidence for heterogeneity (e.g. P value from the Chi² test or CI for the I² statistic). In the event of substantial clinical or methodological heterogeneity, we will not report study results as the pooled effect estimate in a meta‐analysis.

Investigation of heterogeneity and subgroup analysis

We expect the following characteristics to introduce clinical diversity and will perform subgroup analyses using within‐study contrasts if we include at least 10 studies.

• Mobility/functional level: K‐classification level/SIGAM classification of K1 to K2 versus K3 to K4; SIGAM levels C‐D versus SIGAM levels E‐F

• Level of amputation: transtibial, transfemoral, through‐knee, hip disarticulation, hemipelvectomy

• Cause of amputation (aetiology): vascular and diabetes‐related versus non‐vascular

We will conduct subgroup analyses using the formal test for subgroup differences in RevMan [44]. We will make comparisons of the magnitude of the effects between the subgroups by evaluating the overlap of the CIs of the summary estimate. We will use the non‐overlap of the CIs to indicate statistical significance. If there are insufficient studies, or insufficient information about our subgroups, we will report this as a difference between the protocol and the review.

Equity‐related assessment

Prosthetic feet are interventions aimed at the general population. The main causes of amputation vary according to geographical location. In higher‐income countries, amputations are largely disease‐related; whereas in countries affected by war or conflict, traumatic amputations are more common.

It is likely that access to (or prescription of, or both) prosthetics technologies will vary according to high‐, middle‐, and low‐income countries [57]. We will report any relevant characteristics, as suggested by PROGRESS‐Plus (place of residence, race/ethnicity/culture/language, occupation, gender/sex, religion, education, socio‐economic status, social capital, age, sexual orientation and disability) [58], to determine if the populations included in trials are subject to health inequity in terms of the interventions. We will report on this descriptively in our review and present any differences in baseline risks in our population that might cause disadvantages in our summary of findings table.

Sensitivity analysis

We will perform sensitivity analyses of our primary outcome to assess the robustness of our conclusions. We will explore the influence of the below‐mentioned factors on effect sizes.

• We will restrict the analysis to studies with an overall low and moderate risk of bias

• We will repeat the analyses after removing studies with missing data (if we were unable to obtain these data from the trial authors)

• We will repeat the analyses using a fixed‐effect model

Certainty of the evidence assessment

We will create a summary of findings table(s), using GRADEpro GDT software [59], for the following comparison of two or more commercially available prosthetic ankle‐feet. Each summary of findings table will present the following outcomes. We will prioritise short‐term measurement comparisons.

Critical outcome

• Health‐related quality of life outcomes

Important outcomes

• Functioning with a prosthesis outcomes

• Gait outcomes (temporal‐spatial, joint kinematics, kinetics, joint moments, and powers)

• Prosthetic satisfaction and comfort outcomes

• Energy expenditure

Two review authors (NV, CJH) will use the GRADE approach, outlined in Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions [60], to assess the certainty of the evidence based on the studies that contribute data to the meta‐analyses for each outcome. According to the GRADE approach, evidence from RCTs is initially considered to be high‐certainty evidence, but can be downgraded by one or two levels based on limitations related to five considerations (risk of bias, inconsistency, imprecision, indirectness, and publication bias). We will downgrade the certainty of the evidence by one level if a GRADE consideration is serious and by two levels if very serious. For the risk of bias consideration, we will follow the guidance in Table 14.2.a of the Cochrane Handbook for Systematic Reviews of Interventions [60]. We will consider a partially contextualised approach when downgrading for imprecision [61]. We will resolve any disagreements by discussion or by involving our review team.

We will justify all decisions to downgrade or upgrade the certainty of evidence in the footnotes and provide comments to aid the reader's understanding of the review where necessary. We will consider whether there is additional outcome information that was not incorporated into the meta‐analyses, note this in the comments, and state if it supports or contradicts the information from the meta‐analyses.

If a meta‐analysis is not possible, we will present results in a narrative summary of findings table. We will provide key information about the best estimate of the magnitude of the effect in relative terms and absolute differences for each relevant comparison, the number of participants and studies addressing each outcome, and the rating of the overall confidence in effect estimates for each outcome, following guidance in Section 14.1 of the Cochrane Handbook for Systematic Reviews of Interventions [60].

We will formulate statements for the findings and certainty of the evidence using wording templates that combine the size and certainty of an effect to improve the clarity of communication [62].

Consumer involvement

Consumers will not participate directly in this Cochrane review; however, we will involve them in the development of the plain language summary. We will approach prosthesis user groups, both in the UK and in the Netherlands, to help us write the plain language summary so that the results will be of interest to prosthesis users. We will involve them once we have generated the findings of this systematic review.

Supporting Information

Supplementary materials are available with the online version of this article: 10.1002/14651858.CD016076.

Supplementary materials are published alongside the article and contain additional data and information that support or enhance the article. Supplementary materials may not be subject to the same editorial scrutiny as the content of the article and Cochrane has not copyedited, typeset or proofread these materials. The material in these sections has been supplied by the author(s) for publication under a Licence for Publication and the author(s) are solely responsible for the material. Cochrane accordingly gives no representations or warranties of any kind in relation to, and accepts no liability for any reliance on or use of, such material.

Supplementary material 1 Search strategies

Supplementary material 2 PICO summary of trial inclusion and exclusion criteria

New

Additional information

Contributions of authors

Protocol development: NV, CJH

Review of the protocol: CH, FCDS, CTB, WvO, RFvE

Co‐ordination of the review: NV, CJH

Final approval of this protocol version: all authors

Declarations of interest

Natalie Vanicek is the author of a potentially eligible study.

Chao Huang has no conflicts of interest.

Fiona C Davie‐Smith is the author of a potentially eligible study.

Cleveland T Barnett is the author of a potentially eligible study.

Wieneke van Oorschot has no conflicts of interest.

Rene F van Ee has no conflicts of interest.

Cheriel J Hofstad has no conflicts of interest.

Sources of support

Internal sources

• Internal sources, Other

  No sources of support provided

External sources

• External sources, Other

  No sources of support provided

Registration and protocol

This title was registered with Cochrane in November 2023. This protocol will supersede a 2004 Cochrane review [6].

Data, code and other materials

Data sharing is not applicable to this article as it is a protocol, so there were no datasets generated or analysed.