Pressure ulcers (PUs) are lesions of the skin and underlying soft tissues, caused by prolonged or elevated exposure to compressive and shearing forces, usually over a bony prominence. Coleman et al (2014) stated that the four main factors implicated in PU development were interface pressure, shear, friction and moisture. Indirect causal factors, such as compromised sensory perception, diabetes and poor nutrition often occur in frail or critically ill people (Coleman et al, 2014; VanGilder et al, 2017). Age alone is not a risk factor; rather, problems common in older people, such as hip fractures, faecal and urinary incontinence, smoking, dry skin, chronic systemic conditions and terminal illness, are associated with pressure ulceration (Hahnel et al, 2017).
PU symptoms include pain, exudate and odour, compromising all areas of patient functioning, consequently reducing quality of life (Gorecki et al, 2010; 2011; McGinnis et al, 2014). Patients with PUs are at a high risk of further complications, such as nosocomial infections and sepsis, and they experience longer length of hospital stays (Graves et al, 2005; Moore, 2013). Managing PUs places a significant financial burden on healthcare organisations, with the most recent figures it works out at between £1214 and £14 108 per patient per year (Dealey et al, 2012; Schuurman et al, 2012), costing the NHS more than £1.4 million every day (Guest et al, 2017). The most recent information on PUs in older patients shows they are associated with a fivefold increase in mortality: in-hospital mortality in this group is estimated to be 25–33%.
The National Institute for Health and Care Excellence (NICE, 2014) clinical guideline on PU prevention and management separates risk factors for developing a PU (limited mobility, significant loss of sensation, previous or current PU, nutritional deficiency, inability to reposition and significant cognitive impairment) from skin assessment (assessing skin integrity in areas of pressure, colour changes, variations in heat, firmness, moisture, pain or discomfort). The NICE quality standard on PUs (NICE, 2015) states that people admitted to hospital or a nursing home should have a PU risk assessment within 6 hours of admission.
Despite the guidance, variations remain between and within NHS hospitals and wards regarding how and when to conduct skin assessment. For example, the authors have observed that some clinical experts may emphasise the use of a risk assessment tool, while others may focus on frequent repositioning, pressure reduction equipment, specialist support services, decreasing friction, reducing shear and moisture, adequate nutrition and dressing selection.
The National Stop the Pressure Ulcer audit included 10 144 patients in England across 36 hospitals in 18 NHS trusts, identifying 917 patients (9.04% [95% CI 8.48–9.60%]) with one or more PUs (excluding moisture-associated skin damage) (Stephenson and Fletcher, 2020). However, incidence rates of PUs in hospitals of between 3.9% and 27.7% have been reported (Vanderwee et al, 2007; Tubaishat et al, 2018; Stephenson and Fletcher, 2020). Therefore, there is a clinical need for an objective physiological skin measurement to reduce the variation in care across the NHS and potentially enable earlier detection of PUs (before visual signs appear), leading to a reduction in incidence.
NICE (2020) has recommended a randomised controlled trial (RCT) should be carried out to address the uncertainties regarding the clinical benefit of the SEM Scanner 200, a prognostic medical device measuring subepidermal moisture, in preventing PUs (NICE, 2020). A number of alternative prognostic non-invasive techniques have been proposed including laser Doppler flowmetry, transcutaneous gas tension measurement, tissue reflectance spectrometry, near-infrared spectroscopy, skin temperature and active body pressure relief systems (Liu et al, 2015; Lee et al, 2019).
This article proposes a theoretical study protocol to assess a prognostic medical device for skin assessment relative to key practical and ethical considerations relating to conducting research in this area. The hypothesis of a proposed trial is that the use of a medical device that scans epidermal, dermal and subdermal tissue reduces the incidence of PUs through earlier detection of warning signs and the adoption of preventive measures, compared with standard care.
Methods and analysis
Theoretical study design
RCTs are considered the gold standard for demonstrating the efficacy and safety of new treatments. The theoretical study outline of a prospective, three-armed, multi-centre, stratified cluster randomised controlled superiority trial is detailed in Table 1. Each cluster is a hospital ward, allocated in a 1:1:1 ratio, which would enable multiple statistical comparisons to be made quantify the effect of the investigational medical device (IMD) separately from the effect of moving the patient, to take the measurement and the impact of the additional objective measure on clinical decision-making.
Table 1. Study protocol outline
| Study design | Interventional: a prospective, three-armed, multi-centre, stratified cluster randomised controlled superiority trialParallel assignment masking: double-blinded (patient and clinical staff) (intervention arm 1 only), outcome assessors blinded to all arms |
| Health condition(s) or problem(s) studied | Pressure ulcer prevention |
| Population | Inclusion criteria: adult patients (aged ≥18 years) who undergo risk assessment for developing a pressure ulcer during an admission to a recruiting NHS secondary care wardExclusion criteria: patients contraindicated to the use of the medical device |
| Intervention(s) | Active comparator 1: pressure ulcer detection medical device (blinded to staff and patient) and standard careActive comparator 2: pressure ulcer detection medical device (unblinded to staff and patient) and standard care |
| Comparator | Standard care |
| Outcomes | Primary: proportion of patients developing pressure ulcerSecondary: time to development of a new pressure ulcer (any category) during a hospital stay as determined by visual skin assessment using local procedures, calculated from time of admission to recruiting ward with a maximum length of stay of 60 days. Time frame; category of pressure ulcer; pressure ulcer treatment; length of hospital stay; documentation of whether pressure ulcer risk assessment device readings contributed to clinical decision making; detail of change in clinical behaviour; and adverse events associated with the pressure ulcer medical device |
| Setting | UK NHS hospital wardsWard inclusion criteria: ≥500 annual admissions, mean length of stay ≥3 daysExclusion criteria: paediatric, maternity, acute medical assessment or emergency wards |
| Target sample size | 16 200 across 33 wards |
Because of significant variation in standard care, hospital wards should be stratified based on baseline PU incidence rates, ward type, PU prevention standard care regimens and mean length of stay before cluster randomisation to ensure covariate balance across all arms. Cluster randomisation is a pragmatic approach that negatively impacts on study power; however, including more locations and hospital sites generates results that are likely to be more generalisable across the NHS.
Following computer-generated randomisation, wards should undergo a initiation visit to ensure the protocol and all elements of the study are fully understood and to confirm any additional support, documentation and oversight required. Training should be provided and disseminated to all staff involved in the study in line with the IMD instructions for use and competency evaluation. Patients who are admitted to a participating ward will undergo PU risk assessment as per arm allocation. Patients, staff and investigators cannot be blinded to arm allocation. Patients should remain enrolled within the study from the point of ward admission until ward discharge or death, whichever is earlier.
To measure the impact of a diagnostic or prognostic medical device on decision-making, both the patients and the clinicians should be blinded to the readings in one of the intervention arms. This could involve obscuring the device display or relying on device memory storage to capture the readings without involving the direct care team. Only outcome assessors will be blinded to all three arms. There may be circumstances where unblinding of the results from the IMD may be appropriate, for example, where an investigation into patient care requires information to be released or where there are concerns relating to patient risk or safety. Where possible, unblinding should not be released to staff taking measurements using the IMD.
Population
Because of patient-level risk of PU development, all adult patients (aged ≥18 years) who undergo risk assessment for developing a PU during an admission to an NHS secondary care ward should be included. To meet minimum recruitment targets and capture patients at the greatest risk of PU, wards with at least 500 annual admissions and a mean length of stay of ≥3 days should be considered. Patients contraindicated to the IMD should be excluded. Subgroups of interest include patients with history of PUs, younger age groups, comorbidities (eg diabetes) and ward type.
Intervention
The intervention was the use of an IMD to objectively measure the skin, combined with a standard care PU risk assessment. The IMD should be employed as per its instructions for use and contraindications. Using the IMD to make a physiological measurement of anatomical areas at most risk (the sacrum and heels) is likely to require the patient to be moved. As patients are repositioned to prevent PUs, this may confound the outcomes. Therefore, there are two interventional arms: standard care skin assessment plus IMD with the practitioner and patient blinded to its results; and standard care skin assessment plus IMD with results revealed to the both of them.
Comparator
The comparator is the current standard of care at individual ward level. The frequency of visual skin assessment varies by hospital and ward practice as well as because of the unpredictable nature of clinical working. Skin assessments should be performed as per local protocol to ensure patients are not unnecessarily put at increased risk of PUs. It may not always be possible or feasible to perform part or all of the protocol for the skin assessment, including the use of the IMD (for example, an inability to take measurements on broken skin). Where the protocol cannot be fully followed, justification should be clearly provided within the medical records.
Patients should continue to receive their ongoing NHS care as required on an individual basis, including accessing PU prevention adoption strategies where appropriate, irrespective of the PU risk assessment protocol randomised to the ward.
Outcomes
The primary outcome measure should be the proportional change in PU incidence. It should exclude moisture lesions (caused by excessive moisture, not pressure or shear) and diabetic foot ulcers (an arterial element of the disease process) as identified from medical records.
Secondary outcomes (Table 2) should include time to development of PU (any category) calculated from time of admission up to 60 days; time to PU diagnosis; category of PU at discharge or death; PU treatment; length of stay; visual assessment score; changes in clinical decision-making or care; and device-related adverse events. Staff should clearly document whether the IMD readings contributed to clinical decision making, including where preventive strategies were not adopted where indicated by the IMD.
Table 2. Secondary outcomes
| Measurement variable | Justification | Description |
|---|---|---|
| Time to development of new pressure ulcer (any category) | To measure reduction in hazard of pressure ulcer by using the risk assessment device | Determined by visual skin assessment using local procedures. Measured in days calculated from day of admission up to 60 days |
| Category of pressure ulcer | There is no evidence to suggest that use of a pressure ulcer risk assessment device reduces pressure ulcer staging (through earlier detection), so pressure ulcer staging will be recorded as a secondary outcome of interest following recommendations from the National Pressure Injury Advisory Panel | Each category will be validated independently by the nurse in charge. However, with consent, clinical photographs could be taken and uploaded on to the patient records. During analysis, all or a random sample of pressure ulcers identified through the study by at least one an independent assessor will be undertakenPressure ulcer categories include: category 1: non-blanchable erythema of intact skin; category 2: partial-thickness skin loss with exposed dermis; category 3: full-thickness skin loss; category 4: full-thickness tissue loss; category U: unstageable, full thickness skin or tissue loss; deep tissue injury |
| Pressure ulcer treatment | To determine whether any treatment needs to be introduced to prevent the occurrence of a pressure ulcer or further deterioration of an existing pressure ulcer. This is particularly significant in patients who have developed extreme categories of pressure ulcers: categories 3, 4 and U as well as deep tissue injuries | Management options include: nutritional supplements, hydration, high-specification foam mattress, therapeutic dressings, systemic antibiotics, specialist dressing, heel pressure offloading and surgical debridement (National Institute for Health and Care Excellence, 2014) |
| Length of ward stay | Patients who are hospitalised for a longer period are at a higher risk of developing a pressure ulcer. This includes patients who have been transferred between wards | Measured in days and hours from day of admission to the randomised ward to day of discharge back to community setting or death, whichever is earlier |
| Pressure ulcer risk assessment using validated tool (recorded as per local policies) | Using validated risk assessment scores is part of NHS standard practice; correlation with objective measurement from medical device should be analysed | The Braden scale is an example of a structured risk assessment tool and has six categories: sensory perception; moisture; activity; mobility; nutrition; and friction/shear (Bergstrom et al, 1987). Braden scores are in a range of 6–23, with a lower score indicating greater risk of pressure ulcers |
| Documentation of whether pressure ulcer risk assessment device readings contributed to clinical decision-making | To determine whether use of the device itself contributed to the change in clinical behaviour | This will be measured as either Yes or No, but a text box will be provided so the clinician can add information |
| Detail of change in clinical behaviour | To determine which method of pressure ulcer prevention was used or justification for why prevention was not used where considered appropriate | Additional prevention measures applied: positional changes, increased turning, heel elevation or high-specification foam mattress. Documentation relating to non-compliance with prevention measures |
| Time to moisture lesion diagnosis (if one occurred) | To calculate incidence of moisture lesions | Measured in days, with the day of admission to randomised ward being day 1 |
| Adverse events associated with the pressure ulcer risk assessment device | To determine the incidence of faults/errors occurring from the pressure ulcer risk assessment device | For example failed reading, value not recorded and device not available at the time needed, or patient movement not possible |
PU severity classified according to the visible depth of skin and tissue damage is subjective and may contribute to further variation. A method that could be used to mitigate this is using clinical photography to examine all or a random sample of PU identified through the study by at least one independent assessor, although this is likely to be a significant task depending on the incidence of PUs during the trial. Validated AI technology evaluating clinical photography could be considered if available.
Setting
The settings should be acute secondary NHS wards across the UK, including within tertiary centres and district general hospitals. Involvement of community settings, such as nursing homes or palliative care units, would introduce additional variation in standard care practices and matching baseline characteristics may not be feasible.
To estimate sample size, it was assumed that included wards would have a minimum of 500 annual admissions with a mean length of stay ≥3 days to ensure adequate recruitment with sufficient follow-up to evaluate primary and secondary outcomes. The incidence of PUs in paediatric and maternity patients is low, so these ward types should be excluded. Acute medical assessment or emergency wards are also unlikely to have a minimum length of stay for ≥3 days study inclusion, so should be excluded.
Because of the variation in PU incidence across the NHS, many patients would be have to be screened to reach a patient recruitment target. Multiple wards per hospital may be recruited to reduce burden on research staff coordinating the trial, with co-ordination of multiple NHS hospitals by a trial manager. To enable appropriate comparison of wards or hospitals in statistical analysis, baseline characteristics of setting should also be recorded (for example, turning regime, staff:patient ratio and staff rotation).
To ensure covariates are balanced across the three arms at baseline, all eligible wards should be recruited and stratified before randomisation. This may be difficult in practice and could be mitigated by widely publicising the trial through meetings with hospital ward managers, matrons and tissue viability nurses to generate interest in the study before recruitment. Wards could also be identified through National Institute for Health and Care Research (NIHR) Clinical Research Networks (CRNs) and the study could also be publicised across organisations such as the National Wound Care Strategy Programme, the Greater Manchester Pressure Ulcer Prevention and Management Network or the 1000 Elders patient and public involvement (PPI) group. Broad inclusion criteria should be applied to maximise the number of patients and wards eligible for recruitment and capture patients who are at greatest risk of developing a PU.
The study should not be restricted to a certain geographic area but could be conducted anywhere within the UK.
Additionally, if there are insufficient wards with adequate throughput of patients (eg <500 per annum), this could be reconsidered to widen the ward inclusion criteria, although research staff and oversight capacity and duration of the trial should be taken into consideration.
Sample size
There are large variances in PU rates between individual trusts, ranging between 3.9% and 27.7%, as reported by the National Stop the Pressure audit (Stephenson and Fletcher, 2020).
Using data from a single trust (Newcastle upon Tyne Hospitals NHS Foundation Trust), restricted to wards meeting the study eligibility criteria, the mean number of PUs was 17 per ward per year. Assuming a 33% reduction (ie the mean number of PUs reducing to 11.33 per year) then, to achieve 90% power (to enable two comparisons across the three arms), the trial would need to randomise nine wards per arm (27 wards in total). This is calculated assuming the number of PU per ward is Poisson distributed. The number of wards would increase to 28 wards per arm (84 in total) if assuming a 20% reduction in PU (ie reducing to 13.6 PUs per year). If the percentage reduction in PU incidence is unknown, an internal pilot phase could inform the subsequent randomised trial. Assuming a potential 20% drop-out of wards or patients, these assumptions lead to a recruitment target of 16 200 patients across 33 wards, with 500 patients recruited per ward over a 2-year period.
This large study size brings operational challenges that need to be considered, for example around managing quality assurance, having regular, up-to-date documentation from all recruited patients and wards, maintaining a high compliance rate and adherence, regularly reviewing performance, management and training of existing and new clinical staff and ensuring access to resources, including the IMD and any associated consumables. There are problems around implementing large trials in hospital settings, and trial budgets must have adequate investment.
The authors have observed clinical experts recognise the most significant barriers to recruitment or adherence to interventions are nursing shortages across the NHS and increasing pressures on wards. To alleviate the additional burden of the study, sufficient support should be available to enable study activities. If the study is eligible, registering it with the NIHR portfolio will enable access to CRN resources. It is anticipated that the arm allocation will become the locally adopted protocol for assessing a patient's risk of developing a PU, with most data collected for the study being routinely collected as per standard care. To relieve the burden on ward staff, data retrieval and collection tasks could be delegated to supporting research staff. For the two interventional arms, the time taken for staff to use the IMD where results are unblinded and the additional time to document any actions taken should be carefully considered alongside ward capacity and patient safety.
Interim analysis should be performed by a statistician on the primary endpoint when 50% of patients have been enrolled. Criteria for trial termination will include failure to recruit to time and target, failure to recruit sufficient sites or a national change in standard care in detecting or treating PU occurring during course of the trial.
Data collection plan
Data could be extracted from participant medical records (using a standardised data collection form or existing record system) and transcribed into the centralised online secure database. Potential confounders should be recorded from existing medical records, for example age, sex, height and weight (or BMI), ethnicity, comorbidities, history of PUs, presence of diabetic foot ulcers, calluses on heels, existing PU on admission, posture (supine, prone or side-lying) and approximate time spent in each position, nutrition (eg using the Malnutrition Universal Screening Tool), anatomical location of where the IMD was used and the use of additional medical devices (eg mechanical ventilation, which would also impact the ability to apply PU prevention strategies such as patient repositioning).
For patients who have been transferred from another ward, additional information would need to be collected from their previous ward(s). This would include: previous ward type; baseline PU incidence; mean length of stay; previous PU prevention regimen; and the reason for transfer.
Patient identifiers should be retained at each recruiting centre and only study identifiers entered into an electronic database. Given the repeated measurements per patient throughout their hospital stay, data collection would require additional staff time, which may be difficult in some hospital settings (ie intensive care), particularly if there are nursing shortages. Funding of staff time should be carefully considered to ensure the trial can be conducted in practice. If treated as an investigator-initiated trial, the lead NHS trust would act as data controller of information collected from all hospitals and would retain responsibility or ownership.
Missing data should be reviewed during interim analysis for each group. Given the pragmatic and practical application of the theoretical study, data made for the study in the clinical notes are likely to be incomplete. A minimum data set required for analysis should be considered based on the sample size. Data missing at random may be addressed with multiple imputation. Monitoring visits by trial managers at designated time points after each site opens should be conducted to ensure data collection is reliable and complete.
Analysis
All analysis should be conducted on an intention to treat basis. A per-protocol analysis should include all eligible patients while excluding major protocol deviations. Interim analysis at specified time points should be defined prospectively. Depending on the learning curve of the IMD, it may be appropriate to exclude results from an initial cohort of recruited patients to avoid skew in the results, with the impact explored during statistical analysis.
A negative binomial model to account for possible over dispersion, with the dependent variable being PU incidence, should be used to determine whether there is a significant difference between randomised arms. Furthermore, variables that were used for stratification (ward type, baseline PU incidence, mean length of stay and PU prevention regimen adopted) should be controlled for.
A negative binomial model could also be used to explore whether the PU risk assessment device reading is a significant predictor of PU outcome. Patient-level covariates (eg mean age, sex proportion, mean BMI, comorbidities, pre-existing PU and posture) should be adjusted within this model one at a time.
Suitable regression models should be used to compare the three interventions across the individual institutions, comparing the proportion of scanned patients with the proportion of those who develop a category 1, 2, 3 or 4 PU. Further analysis should compare the device reading and the type of PU treatments that are used and the anatomical location of the PU. Comparisons should also be made between the device readings and the length of hospital stay. Additional comparisons of PU incidence could be explored between institutions with similar baseline PU incidence, PU prevention strategy adopted and mean length of stay. Further subgroup analyses could be performed with the exclusion of data for patients who were admitted for one only night, have a pre-existing PU (from a transferring hospital or community setting etc), have been moved from another ward or who are at very low risk of developing a PU (given their Braden, Waterlow or PURPOSE-T score).
Data monitoring
A trial committee should have representation from the chief investigator, statistician, two or more clinical experts with specialist knowledge and experience in wound care or PUs and a patient representative with lived experience.
The trial committee will meet quarterly to discuss recruitment, data completeness and the interim and final analyses. The data monitoring committee will be independent from the sponsor and have no competing interests.
Ongoing monitoring of data entry and completeness should be managed by the central study team. Regular feedback to site research staff should be provided to maximise data quality.
Patient and public involvement
Input PPI representatives should be sought at the initial stages of protocol development to ensure outcome measures and consent processes align with patient priorities and trial acceptance. A study-specific PPI steering group should be formed to offer consultation and advice relating to the study protocol, development of study materials (including lay summaries and patient-facing documents), trial conduct or compliance reviews and dissemination. NIHR PPI resources should be used and appropriate reimbursement and recognition of PPI should be implemented.
Ethics
This proposed study design is regarded as clinical research covered by the UK Policy Framework for Health and Social Care Research and therefore submission for relevant regulatory approval (research ethics committee, Health Research Authority or Medicines and Healthcare products Regulatory Agency) would be required via the Integrated Research Application System.
This study design poses additional considerations relating to consent that should be addressed in the application for ethical approval. McRae et al (2011a) proposed that human research subjects involved in cluster-randomised trials are defined as those who are directly intervened on by investigators, interact with investigators, are intervened on through manipulation of their environment that may compromise their interests or whose identifiable private information is used to generate data. Therefore, the clinical staff delivering the trial interventions as well as the patients admitted to randomised wards are considered research subjects within this trial design so their informed consent should be carefully considered. McRae et al (2011b) also noted that there might be circumstances in which informed consent waivers may be appropriate: where the research involves no more than minimal risk to the subjects; the waiver will not adversely affect the rights and welfare of the subjects; the research could not be practicably carried out without the waiver; and, wherever appropriate, additional pertinent information is provided. In this case, the adoption of a new PU risk assessment protocol at ward level may meet the criteria for consideration of an informed consent waiver or alteration. Because of staff and patient turnover during the length of the trial, it is not feasible for all subjects to provide individual informed consent before randomisation and arm allocation. To achieve the minimum sample size and ensure protocol adherence (including avoiding mixed or delayed intervention resulting from to waiting for consent), it may be appropriate not to obtain informed consent from patients admitted to the randomised ward. An option may be for admitted patients to be offered the opportunity to have their data removed from the study, although this may result in selection bias and impact recruitment.
To maintain independence, no IMD manufacturer should be involved in the collection, analysis, write-up or publication of data derived from the study. Support from a company may be considered in the form of provision of devices, component parts, servicing or training, where there is a financial cost to the manufacturer; there would be no direct financial contribution.
Dissemination
The study should have a website informing stakeholders of recruitment status and include links to any published results (interim and final). The study results should be published in a peer-reviewed, freely available journal to ensure they can be reached by clinicians, commissioners and patients. The study team should consider providing full public access to the full protocol, anonymised participant-level dataset and statistical code. The PPI steering group should review a lay summary to be sent to all participants who consented to receive a copy of the study findings.
Discussion
Designing a randomised trial to ascertain whether a medical device reduces the incidence of PUs through earlier adoption of preventive measures is challenging. This is largely because there is substantial variation in PU prevention strategies across the NHS, with multiple factors affecting whether a patient develops a PU. To control for these factors, the authors propose a three-armed cluster RCT recruiting multiple wards in tertiary centres and district general hospitals. To detect an effect, target recruitment is approximately 33 wards involving 16 200 patients (assuming each ward has an annual throughput of 500 patients and a drop-out rate of 20%). This large sample size may not be practically feasible, particularly in terms of staff time required to screen, consent and collect data, especially in the current climate where there are staff shortages within the NHS.
Challenges/limitations and risk mitigation
Despite developing the ‘best case’ study protocol to demonstrate whether a prognostic medical device can reduce the incidence of PU, several limitations remain. Any results generated in a secondary care setting may not be fully generalisable to a primary care or community setting or to different healthcare institutions in different countries where standard care and baseline characteristics may vary.
Depending on the IMD, not all anatomical sites may be assessed. Uncertainty remains over whether the use of an IMD on specific locations would reduce a patient's overall risk of developing a PU.
Owing to the large size of this theoretical study and the number of clinical staff and patients involved, it may be difficult to monitor strict adherence to the protocol. The authors recommend additional research staff to be present on the study sites, who will collect data and relieve the burden on ward staff. Registering the study on the NIHR portfolio will enable access to CRN resources, which could be used to improve study compliance.
The authors have also considered that it may not be meaningful or feasible to obtain consent or assent from individual patients and individual ward staff to meet the minimum sample size, ensure protocol adherence and avoiding mixed or delayed intervention owing to waiting for consent. However, agreement or assent to be enrolled into a cluster can be made when a ward is approached for potential recruitment.
The theoretical trial design does not account for any qualitative data to be collected owing to the consent process and to reduce demands on the clinical staff delivering the trial. This may limit the conclusions reflective of patient or staff experience with the use of the IMD. Additional data collection outcomes could be included for all or a sample of patients who are able to consent and are able to self-complete qualitative questionnaires including Pressure Ulcer Quality of Life-Utility and Pressure Ulcer Quality of Life-Prevention questionnaires, as well as pain assessment via a visual analogue score.
Conclusion
There is a need and opportunity to improve the design and analysis of pragmatic trials in PU research. By describing these challenges and the actions taken to overcome them, the authors aim to improve the rigour of similar studies assessing the effectiveness of wound care in secondary care settings for future studies.
KEY POINTS
- A best case study design to support researchers should consider key practical and ethical issues in this field
- Research on devices to assess skin condition and the risk of pressure ulcer development before visual signs appear is hampered by variations in clinical practice and patient-level confounders
- Investigational medical devices may only assess specific anatomical locations; uncertainty over whether their use would reduce a patient's overall risk of developing a pressure ulcer.
- The findings from a study evaluating the impact of use of an interventional medical device to assess the risk of developing a pressure ulcer in a hospital may not reflect practice in other settings, such as primary or community care
CPD reflective questions
- Do the considerations outlined in this article sufficiently overcome the barriers to conducting this research trial?
- Should clinical trials involving medical devices in wound care be prioritised for research funding?
- If this device is shown to be effective, to what extent does this have potential to impact clinical practice?