3D Telemedicine Outperforms 2D Telemedicine: A Randomized Crossover Trial in Reconstructive Plastic Surgery

Abstract

Background:

A key barrier to adoption of telemedicine is patient preference for in-person consultations. To address this, an immersive 3-dimensional (3D) telemedicine system was codeveloped with patients to improve the realism, quality, and patient experience of remote consultations, and the benefits were assessed in this randomized trial.

Methods:

Eighty patients were recruited from the Canniesburn Plastic Surgery Unit, UK to a randomized crossover trial of 3D versus 2-dimensional (2D) telemedicine consultations in 2022 and 2023. The primary outcome was presence using the Presence scale, which measured how closely a remote consultation resembled an in-person consultation. Secondary outcomes included the validated Telehealth Usability Questionnaire (TUQ), the Mental Effort Rating Scale, and satisfaction scales.

Results:

The Presence score for 3D telemedicine was significantly higher than that for 2D telemedicine (P < 0.001). Secondary outcomes also were superior for 3D telemedicine, including TUQ score (P < 0.001), mental effort (P < 0.001), and satisfaction (P < 0.001). Age, deprivation, education, sex, and technology familiarity were not associated with any outcome measures, indicating inclusivity of the 3D technology. Subjective interviews indicated that 3D telemedicine, by virtue of allowing annotation and drawing on the patients’ own 3D model, aided understanding of complex surgery, allowing a more personalized medicine approach in the consent process.

Conclusions:

3D telemedicine improves the realism, interaction quality, and experience of remote consultations relative to 2D telemedicine, with significantly higher presence and TUQ scores. Three-dimensional telemedicine more closely mimics reality, which may help overcome barriers to adoption related to patients’ preference for in-person consultations.

The COVID-19 pandemic led to an acceleration in innovations focusing on remote medical consultations,¹ providing the impetus for our research group to codevelop the world’s first real-time 3-dimensional (3D) telemedicine system, in conjunction with patient and clinician stakeholders, using Microsoft’s Holoportation technology.² This was developed to address one of the key barriers to adoption of telemedicine—patients’ preference for in-person consultations—with the goal of using 3D technology to improve the realism, immersion, and experience of remote consultations to more closely align with features of an in-person visit.³ Although our preliminary data suggested that there is value in the use of a real-time 3D system—with improvements in realism, quality, and satisfaction over 2-dimensional (2D) telemedicine systems—conclusive trial data are lacking.

DEVELOPING NOVEL DIGITAL HEALTH TECHNOLOGIES WITHIN A STANDARDIZED FRAMEWORK

No real-time, 360-degree 3D telemedicine systems are commercially available. Previous 3D systems have been predominantly used in research settings only, without participatory patient development, clinical validation, or real-world use.^4,5 We sought to address these shortcomings in the development process using the Virtual Reality Clinical Outcomes Research Experts (VR CORE) international guidelines, which aim to improve the methodologic quality of digital health trials.⁶ A cornerstone of these guidelines is the inclusion of participatory development to place patients at the heart of the development process, as “lack of patient involvement, poor requirement definitions, and nonadaptation to user feedback are some of the common factors that explain failures of digital interventions.”⁶ Participatory development is the first of 3 phases, followed by early clinical trials and randomized clinical trials (RCTs).

3D TELEMEDICINE MAY IMPROVE THE REALISM, QUALITY, AND IMMERSION OF REMOTE CONSULTATIONS

An international collaboration among the United Kingdom, Ghana, and the United States (Microsoft Corporation) initially sought to develop and deploy a low-cost 3D telemedicine system to increase access to specialized reconstructive surgical care in lower- to middle-income countries.⁷ The first phase of the VR CORE guidelines, including participatory codevelopment, was completed in 2021, with multiple iterations resulting in significant improvements in patient and clinician usability, satisfaction, and quality.² This was followed by a second phase, consisting of several studies centering on safety, clinical concordance with in-person consultations, protocol optimization, study design, and outcome instrument evaluation.² The final phase—RCT testing—aimed to provide evidence of clinical efficacy and benefits of the 3D telemedicine system in comparison with 2D telemedicine in a cohort of reconstructive plastic surgery patients.

One of the fundamental goals of telemedicine is to improve realism and immersion, to more closely mimic an in-person consultation. This trial therefore aimed to assess whether 3D telemedicine outperforms 2D telemedicine in this regard, using psychometric properties of realism and immersion deemed “presence.” This was supported by the Telehealth Usability Questionnaire (TUQ),⁸ one of the most widely used validated outcome instruments in telemedicine research,^9–11 which examines usefulness, ease of use, effectiveness, reliability, interaction quality, and satisfaction, providing additional insights into the clinical efficacy of 3D versus 2D telemedicine systems. The hypothesis of this trial was that 3D telemedicine provides a more realistic and immersive consultation (measured by presence) than 2D telemedicine, which translates to improved overall clinical quality (measured by TUQ).

PATIENTS AND METHODS

Ethics

This study was approved by National Health Service Greater Glasgow and Clyde Research and Innovation (approval no. INGN22SG035). The trial is registered at ClinicalTrials.gov (identification no. NCT05227235). The project was monitored by biannual governance reviews. Participants consented in writing, with patient data controlled by National Health Service Greater Glasgow and Clyde. CONSORT (Consolidated Standards of Reporting Trials) guidelines were followed for reporting of randomized trials.¹²

Study Design

Prestudy Protocol Optimization

Focus group and codesign with patients was conducted from 2019 through 2021, to allow protocol optimization and testing of outcome instruments before the RCT as per VR CORE guidelines.² This consisted of 3 studies: a clinician feedback study (n = 23), a patient feedback study (n = 26), and a cohort study focusing on safety and reliability (n = 40). “Lose,” “keep,” or “change” feedback prompts were used to guide development. Codevelopment is covered in detail in our previously published participatory development study² (Fig. 1).

Fig. 1.

Illustration of the 3 phases of the VR CORE guidelines on digital health trials. The project began in 2020 with the VR phase 1 participatory codevelopment phase. This was followed by VR phase 2, incorporating early clinical trials focusing on safety, reliability, and clinical concordance, and establishing parameters for the following RCT. VR phase 3 is the RCT comparing 3D with 2D telemedicine.

Randomized Crossover Design

A randomized 2-group crossover trial design randomized patients to either 3D or 2D telemedicine consultation first. As the included plastic surgery patients were heterogenous (with differing operations and anatomical sites), a crossover design accounted for confounding better than a between-subjects design, with patients acting as their own controls. Patients were assessed using the 2 systems by the same clinician, with standardization of the consultation and examination with each system (recorded separately). A washout period was not mandatory for tests of perception (such as presence) as opposed to tests of knowledge, as has been used in other digital research trials.¹³ Nonetheless, carryover effects were analyzed in this study. To further address potential concerns about carryover effects, the study design also included a prespecified sensitivity analysis that examined between-subjects analysis for the first telemedicine system used by participants. (See Figure, Supplemental Digital Content 1, which shows the study design [within-subjects randomized crossover design]. Primary analyses used a randomized crossover design, https://links.lww.com/PRS/I452. See Figure, Supplemental Digital Content 2, which shows the sensitivity analysis [between-subjects randomized design]. This removes the crossover component and analyses the first telemedicine system to which participants were randomized, https://links.lww.com/PRS/I453.) Because the interventions involved telemedicine, it was not possible to blind participants or investigators. Data analysis was performed by investigators blinded to intervention allocation.

Telemedicine Systems

The 3D system consisted of an array of 10 Azure Kinect cameras that fused depth and video output to create a 360-degree 3D model of the patient.² This was linked to a separate “viewer” room where the patient could be viewed remotely over the existing hospital network, using an interface that is intuitive and easy to use, with the 3D patient model manipulated with a mouse in a similar manner to 3D computed tomography image manipulation (Figs. 2 and 3). (See Video [online], which shows the 3D telemedicine system in use. The clinician is “co-present” with the patient in the same virtual clinic space. The patient’s 3D model can be annotated and drawn on. This helps the clinician illustrate complex 3D reconstructive surgery and facilitates patient understanding. The patient can be viewed in 360 degrees in real time, facilitating body part assessment, and, if desired, supporting remote physiotherapy.) The 2D system used a Logitech Brio 4K webcam running Microsoft Teams. Both telemedicine systems were set up in the same clinic room to minimize variability. The clinician “viewer” room was set up in a remote clinic room in the same hospital (West Glasgow Ambulatory Care Hospital).

Fig. 2.

Schematic of the 3D telemedicine system. Ten Kinect cameras are connected to Fusion and Render PCs, which create a fused 3D spatial model that is overlaid in real time with RGB camera output. This processing is performed locally at the remote clinic, and the output can be delivered through a standard Internet connection to the clinician viewing room. In this trial, the viewing room was in a remote location separate from the patient clinic room, but within the same hospital (West Glasgow Ambulatory Care Hospital).

Fig. 3.

The 3D telemedicine viewing screen. The patient can be seen in 3 dimensions in real-time and viewed in 360 degrees. The system also allows annotation of the 3D model to illustrate surgical procedures and anatomy for the patient. The clinician is visible in the same clinical space as the patient and can point to areas of interest using their hands.

Video. This video shows the 3D telemedicine system in use. The clinician is “co-present” with the patient in the same virtual clinic space. The patient’s 3D model can be annotated and drawn on. This helps the clinician illustrate complex 3D reconstructive surgery and facilitates patient understanding. The patient can be viewed in 360 degrees in real time, facilitating body part assessment, and, if desired, supporting remote physiotherapy.

Download video file ^{(47.9MB, mp4)}

Outcome Instruments

Multiple outcome instruments were tested during the prestudy participatory development phase.² Presence was deemed to be the primary research outcome, as it indicates how closely a remote consultation mimics an in-person consultation. Many scales measuring presence were too lengthy and confusing for patients,^14,15 and not suitable for literacy levels of the general UK population.¹⁶ The validated single-item presence question “To which extent did you feel present in the virtual clinic, as if you were really there?” was clear and simple, and resonated strongly with patients.¹⁷ The main secondary outcome measure was the TUQ, a validated 21-item questionnaire measuring subdomains of usefulness, ease of use, effectiveness, reliability, and satisfaction, giving a metric of the overall quality of the telemedicine system.⁸ Additional secondary outcome measures included the Mental Effort Rating Scale, a 9-point Likert scale that assesses extraneous cognitive load (ease of understanding),¹⁸ and satisfaction, measured with a 0 to 100 visual analog scale.¹⁹

Outcome Measures

The primary outcome measure was the single-item Presence scale score. Secondary outcome measures included TUQ, Mental Effort Rating Scale, and satisfaction scores.

Study Hypotheses

We hypothesized that 3D telemedicine outperforms 2D telemedicine for presence, providing improved realism and immersion closer to an in-person consultation, which may translate into improvements in the overall quality and experience as evaluated by the TUQ score.

Statistical Analysis

Sample sizes were calculated using prestudy data from a pilot trial.² For the primary outcome (Presence scale score), assuming an ⍺ of 0.05 and 80% power, a sample size of 9 participants per group was needed. For the secondary outcome (TUQ), a sample size of 40 participants per group was needed. The study was powered to allow assessment of primary and secondary outcomes for a randomized 2-group crossover and between-subjects sensitivity analysis. The total sample size was 80 patients. Ancillary analyses included effect sizes for outcome measures, descriptive analyses of TUQ subdomains, correlation of equality and diversity measures with outcome measures, and inductive thematic analysis of subjective patient interviews.

Randomization was provided by a university statistical service using 1:1 computer-generated allocation with a block size of unknown length. Allocation concealment was with a remote, password-protected website (Sealed Envelope Ltd.). Assignment was performed by a research assistant after participants completed study consent. Statistical analysis was performed by an independent statistical service using Minitab (version 18) at 5% significance. Outcome measures were tested for normality (Anderson-Darling tests) before appropriate parametric and nonparametric tests. Crossover effects were tested with the Mann-Whitney U test. There were no missing data in this study.

Patient Identification

Patients were recruited consecutively from clinics at Canniesburn Plastic Surgery Unit, Glasgow. Inclusion criteria included age 16 through 95 years, having undergone a plastic surgery procedure (free/pedicled flap, skin graft, major cancer, trauma, or burn surgery), and capacity to consent. Exclusion criteria included registered blindness or deafness, having undergone a minor procedure (primary closure or minor local flap), or perineal reconstruction (ie, a sensitive area of examination).

Code Availability

The 3D telemedicine code has been open sourced by Microsoft and is available on request.

RESULTS

Participants

Eighty patients were recruited from the Canniesburn Plastic Surgery Unit from December of 2022 through October of 2023 (Fig. 4). The mean age was 51 years; 35% of participants were female; body parts undergoing surgery included the upper limb (15%), lower limb (45%), head and neck (10%), or thoracoabdominal region (30%) (Table 1).

Fig. 4.

Recruitment flowchart.

Table 1.

Patient Demographics^a

Demographics	Values (Total n = 80)
Sex, F/M	28 (35)/52 (65)
Age, yrs	51 (18 to 78)
Deprivation level by postal codeb
1	10 (12.5)
2	13 (16.3)
3	5 (6.3)
4	7 (8.8)
5	8 (10)
6	3 (3.8)
7	10 (12.5)
8	8 (10)
9	10 (12.5)
10	8 (10)
Education level (highest level reached)
None	0
Primary school	0
High school	30 (37.5)
College	24 (30)
University	20 (25)
Postgraduate	6 (7.5)
Diagnosis
Cancer reconstruction	68 (85)
Benign tumor	5 (6.3)
Trauma	7 (8.8)
Body part examined
Upper limb	12 (15) (5 arm, 4 hand, 3 shoulder)
Lower limb	36 (45) (30 leg, 6 thigh)
Head and neck	8 (10) (5 face, 3 neck)
Thoraco-abdominal	24 (30) (13 chest, 5 back, 3 groin, 3 abdomen)
Technologic familiarity (can independently use video calls)	64/80 (80)

^aData are presented as n (%) or mean (range).

^bDecile 1 through 10, where 1 indicates greatest deprivation.

Descriptive Statistics

Differences between groups in presence, satisfaction, mental effort, and TUQ were not normally distributed, and comparisons therefore used Wilcoxon tests (Table 2). Mann-Whitney tests found no evidence of significant carryover effects for any outcome.

Table 2.

Summary Outcome Data^a

Group	No.	Minimum	Q1	Median	Q3	Maximum
Presence
3D	80	34.00	70.00	83.00	90.00	100.00
2D	80	12.00	44.50	61.00	79.50	100.00
Satisfaction
3D	80	50.00	78.00	90.00	96.00	100.00
2D	80	2.00	50.00	66.00	80.00	100.00
Mental effort
3D	80	1.000	1.000	1.000	2.000	6.000
2D	80	1.000	1.000	2.000	3.000	8.000
TUQ
3D	80	72.00	116.50	128.00	136.00	147.00
2D	80	44.00	99.00	112.00	122.00	145.00

^aMedian and interquartile range are presented for nonnormally distributed data.

Primary Outcome

The median presence score for 3D telemedicine was significantly higher (better) than that for 2D telemedicine (P < 0.001; difference 17; 95% CI, 12, 24).

Secondary Outcomes

The median TUQ score for 3D telemedicine was significantly higher (better) than for 2D telemedicine (P < 0.001; difference 13.5; 95% CI, 10.5, 17.5). The median Mental Effort Rating Scale score for 3D telemedicine was significantly lower (better) than for 2D telemedicine (P < 0.001; difference −0.5; 95% CI, −1, 0]). The median satisfaction score for 3D telemedicine was significantly higher (better) than for 2D telemedicine (P < 0.001; difference 20; 95% CI, 14, 24) (Fig. 5).

Fig. 5.

Outcome measures comparing 3D with 2D telemedicine. Median values of primary and secondary outcome measures are shown. For the mental effort rating scale, a lower score is better; for the other outcome measures, a higher score is better.

Between-Subjects Sensitivity Analyses of 3D versus Traditional Telemedicine

This study used only outcome data from the first telemedicine system that participants used, thereby excluding the crossover component. For all outcome measures, 3D telemedicine scored significantly better than 2D telemedicine. (See Table, Supplemental Digital Content 3, which shows between-subjects analyses of primary and secondary outcomes. Only data from the first system were included [n = 40]; data from the second system used by participants were not included, https://links.lww.com/PRS/I454.)

Ancillary Analyses

Subdomain Analyses of TUQ

Descriptive statistics are presented for 6 subdomains of TUQ; 3D telemedicine was rated better than 2D telemedicine for 4 of 6 domains, with no overlapping 95% confidence intervals (CIs) for usefulness, interface, interaction quality, or satisfaction/future use (Table 3 and Fig. 6).

Table 3.

TUQ Subdomains^a

TUQ Subdomain	3D Telemedicine				2D Telemedicine
	Median	Q1	Q3	95% CIb	Median	Q1	Q3	95% CIb
Usefulness	19	17	21	18, 20	16	13	18.75	15, 17
Ease of use/learnability	20	17.25	21	18, 21	18	13	19	16, 19
Interface	27	24	28	25, 27	22	18.25	24	21, 23
Interaction quality	27	24	28	26, 27	24	21	26	23, 25
Reliability	13	7	16	11, 14	11	7	13	8, 12
Satisfaction/future use	27	25	28	26, 28	23	19.25	26	22, 24

^aP values were not calculated to reduce multiple comparisons. Descriptive statistics present the median, interquartile range (Q1, first quartile; Q3, third quartile), and 95% CIs (with exact CI for the median calculated at 96.7% for all subdomains), with 4 out of 6 domains showing no overlap in 95% CI in favor of 3D over 2D telemedicine.

^bExact CI of median 96.7%.

Fig. 6.

TUQ subdomains. Box plot of median (central line), 25th and 75th percentile (box), and 5th and 95th percentile (whiskers). *Subdomains with nonoverlapping 95% CI of the medians (exact CI, 96.7%).

Effect Sizes

Effect sizes of 3D compared with 2D telemedicine included presence (Cohen d = 0.82), TUQ (d = 0.84), and satisfaction (d = 0.93).

Effect of Clinician Conducting Consultation

Consultations were equally distributed among consultants (n = 26), nurse specialists (n = 27), and surgical trainees (n = 27), with no difference in median presence scores among clinicians (3D telemedicine, Kruskal-Wallis P = 0.60; 2D telemedicine, P = 0.17).

Association of Equality and Diversity Measures with Outcomes

Subgroup analyses of age, deprivation level, education level, sex, and technology familiarity did not statistically correlate with any outcome measures for 3D telemedicine. (See Table, Supplemental Digital Content 4, which shows a univariate analysis of equality and diversity variables with 3D telemedicine outcome measures. Linear regression was used for age and deprivation level, correlation for education level, and logistic regression for sex and technological familiarity. P values are presented without corrections for multiple testing, https://links.lww.com/PRS/I455.)

Subjective Interview

Standardized patient interviews were conducted on completion of both telemedicine consultations. Inductive thematic analysis was performed using the Braun and Clarke approach,²⁰ involving data familiarization, code generation, combining codes into themes, reviewing, determining, and reporting. (See Table, Supplemental Digital Content 5, which shows a thematic analysis of subjective patient interviews. Inductive thematic analysis was performed using the Braun and Clarke 6-step approach, including data familiarization, code generation, combining codes into themes, reviewing, determining, and reporting, https://links.lww.com/PRS/I456.) Patients found that body part positioning was easier and more comfortable with the 3D system, as it avoided the need to maintain uncomfortable poses for prolonged periods of time. The 3D system also improved visualization (with its 360-degree view), ease of manipulation of the picture, and, most notably, the opportunity for the clinician to annotate the 3D images in real time. This allowed patients to visualize surgery on their own body, allowing for personalization of the consultation. This may translate to additional benefits pertaining to informed consent and understanding of complex operations, as patients better understood their surgical operation after the 3D consultation; 3D telemedicine also afforded an increased feeling of realism similar to that of an in-person consultation, with patients considering it a viable future alternative.

Harms

No harms were noted.

DISCUSSION

Holographic communication and the concept of a being co-present—as if the clinician and patient were actually together in person—has captured the public imagination through science fiction and popular media over the decades. Although intuitively it would seem that applying this concept in 3D telemedicine may provide significant benefits over conventional 2D video calls, particularly in a complex spatial surgical specialty such as plastic surgery, these benefits have been purely speculative. Previous studies have been conducted in nonclinical research settings or without true 360-degree, real-time 3D coverage.^21,22 The current study provides the first compelling evidence from a randomized trial of improved realism, interaction quality, and experience with 3D telemedicine over current standard of care 2D telemedicine.

One of the fundamental goals of telemedicine is to bring the remote consultation closer to the experience of a face-to-face consultation, which may be assessed through measures of presence or realism. This therefore formed the primary outcome measure in this trial, with the single item Presence scale asking the question “To which extent did you feel present in the virtual clinic, as if you were really there?” This strongly favored the 3D system (P < 0.001), with a large effect size (Cohen d = 0.82), compared with 2D telemedicine. Presence is clinically relevant, as it strongly correlates with satisfaction of the clinical consultation,² and is associated with improved task-related human performance.²³

The TUQ,⁸ one of the most widely used and accepted outcome instruments in telemedicine research,^9–11 provides an overall metric of the clinical quality of the telemedicine system. This favored 3D telemedicine in the current study (P < 0.001), with a large effect size (Cohen d = 0.84). Subdomain analyses of the TUQ score allowed more nuanced insights into the benefits of 3D telemedicine, with significant improvements over 2D telemedicine seen with usefulness, interface, interaction quality, satisfaction, and potential for future use. Together, this points not only to improved quality of consultation with 3D telemedicine, but also to a more immersive interaction and experience that increases patient satisfaction and, therefore, the likelihood of future patient adoption. These are key drivers in the implementation of novel technologies and quality in health care systems.²⁴ In addition, the Mental Effort Rating Scale was used to evaluate patients’ extraneous cognitive load (or ease of comprehension),¹⁸ and indicated that 3D visualization decreases the mental effort required by patients to comprehend complex spatial surgery.

This randomized trial focused on psychosocial, quality, and usability metrics. How these translate into tangible clinical benefits in diagnostic accuracy, health care outcomes, and cost savings was not the focus of this trial, although we have previously demonstrated in a comparative cohort study that diagnostic accuracy and concordance of 3D telemedicine with an in-person consultation (95%) exceed those seen with conventional 2D telemedicine (estimated at 91% to 92%).^25,26 The further potential benefits of 3D over 2D telemedicine noted in this trial are numerous, and directly address many issues relating to barriers to adoption. These include barriers to adoption relating to difficulties with camera and body part positioning, with the 3D system making it easier and more comfortable for patients to position themselves, particularly for awkward areas, such as the legs or back.^3,27 3D visualization also appeared to positively affect patients’ understanding of complex 3D concepts, with the ability to draw on the patient’s 3D model allowing the clinician to more accurately explain an operation on the actual patient’s body. This moves toward a more personalized medicine approach rather than generic explanations using diagrams or leaflets,¹³ and may provide tangible benefits in critical areas, such as consent. The 3D system was developed with patient input and has a simple, intuitive interface.²⁸ Diversity and equality measures demonstrated that the system was inclusive regardless of educational level, deprivation level, technology use level, age, and sex. Linear regression showed no differences in ease of use (through TUQ scores) with increasing age. Usability has previously been noted as one of the main barriers to adoption in older individuals.³

The 3D system was initially developed to increase access to remote reconstructive care in lower- to middle-income countries, on a philanthropic rather than commercial basis. Off-the-shelf components were therefore used to provide a relatively low-cost system (approximately $8000 in 2025), with a significantly lower cost than any other real-time 3D imaging system. The project has been implemented in our concurrent test bed in Ghana,⁷ and is already in use on a real-world basis to facilitate international discussions on the care of patients needing complex reconstruction, and surgical planning of cases before overseas surgical visits. To enable South-South development, Microsoft open-sourced the code and partnered with a technology company in Ghana to promote local ownership, development, and health framework adoption.

Limitations of this trial include selection of a specific patient group for whom 3D visualization is particularly relevant: those undergoing reconstructive plastic surgery.²⁹ As such, these data may not be generalizable across general medicine. However, the 3D system may prove of merit in fields such as orthopedics, physiotherapy, neurologic conditions, and psychiatry, where nonverbal bodily cues and body movements are of key importance. Although the primary outcome was a single-item scale, this was supported by secondary outcomes from the multiple-item, validated TUQ questionnaire.

The data presented here provide the first evidence from a randomized trial that 3D telemedicine provides a significantly more immersive, realistic remote consultation, which more closely mimics an in-person consultation than conventional 2D telemedicine. The validated TUQ showed additional nuanced benefits in subdomains of usefulness, interaction quality, satisfaction, and likelihood of future use. Together, this suggests that 3D visualization may help overcome one of the key barriers to adoption of telemedicine—patients’ preference for in-person consultations—and thereby increase acceptability of remote consultations as a viable future alternative.

DISCLOSURE

The authors have no financial disclosures to declare.

ACKNOWLEDGMENTS

The authors thank Ruchi Lalwani, Jonny Johnson, Amber Hoak, and David Tittsworth. This project was supported by the West of Scotland Innovations Hub. This study was funded by the NHS Greater Glasgow and Clyde Endowments Fund; the Chief Scientist Office, Scotland; and the Canniesburn Research Trust, Glasgow.

Supplementary Material

prs-157-0933-s002.pdf ^{(57.3KB, pdf)}

prs-157-0933-s003.pdf ^{(55.9KB, pdf)}

prs-157-0933-s004.pdf ^{(38.2KB, pdf)}

prs-157-0933-s005.pdf ^{(37.9KB, pdf)}

prs-157-0933-s006.pdf ^{(34KB, pdf)}