Thora-3Di™ Respiratory Rate Validation

Respiratory rate (RR) is a vital sign used to monitor clinical condition of a patient. Various devices using different techniques are available to measure RR. One technique, capnography, uses continuous monitoring of expired gases, via a nasal cannula to assess end tidal CO2 concentration. This long-established technique is frequently used for patients in intensive care. Both RR and the shape of the capnogram waveform, which has a wave pattern marked by alternating inspiratory and expiratory phases, are used to help monitor the patient. Clinician Over-scored End Tidal C02 (COSC) waveform (whereby an expert identifies and scores each breath on the waveform and counts the number of breaths per minute) is considered to be a "gold standard" for measuring RR.

Being able to record RR, without contact or interference with the patient is appealing because it requires minimal patient co-operation, enables measurements even during acute respiratory conditions, and may be more representative of "real life" physiology. Based on a principle originally described in the 1980s, a novel instrument that uses a completely non-contact system based on structured light plethysmography (SLP) has recently been designed by Cambridge University Hospitals Foundation Trust. The device has been refined with the development of the Thora-3Di by Pneumacare Ltd, Cambridge. A grid of visible light is projected onto the thoraco-abdominal wall and two digital video cameras record changes in the grid pattern due to breathing motion. A waveform is produced by the anterior excursion of the thoraco-abdominal wall (SLPVol) over time, and a numerical output of RR is provided . The present study aims to simultaneously measure tidal breathing with the Thora-3Di and a BCI Capnograph 9004 device to compare the Thora-3Di RR output against that of the gold standard clinician over-scored end-tidal C02 (COSC) and to assess equivalence of the two devices with the aim to establishing the validity of SLP for RR measurement.

Study Overview

• Status: Unknown
• Conditions: Pulmonary Disease
• Intervention / Treatment: Device: Structured Light Plethysmography (Thora-3Di); Device: BCI Capnograph 9004 (Smiths Medical)

Detailed Description

Objectives:

The primary objective is to demonstrate equivalence of the Thora-3Di™ SLP device and a Clinician Over-scored End Tidal C02 (COSC) derived from a Capnograph (BCI Capnograph 9004), Smiths Medical) for measurement of Respiratory Rate.

Endpoint:

The primary endpoint is the difference in the Respiratory Rate between the two devices.

Test Article:

Thora-3Di™, non-contact; measures Respiratory Rate by SLP from the SLPVol/time trace.

Comparator:

Clinician Over-scored End Tidal C02 derived from a Capnograph (BCI Capnograph 9004, Smiths Medical); measures RR from the end tidal CO2 (EtCO2) waveform.

Patient Enrolment and Procedures:

A maximum of 43 subjects will be enrolled across 2 sites to achieve 34 evaluable data sets of:

• 24 Patients with Respiratory disease/impairment (with a minimum of 7 and maximum of 17 to be enrolled at each site)
• 10 Normal subjects with no current or previous diagnosis of respiratory disease (minimum of 3 and a maximum of 7 to be enrolled at each site)

Sample Size:

A sample size of 34 subjects has been determined to yield at least 80% power when the MSD is 0.4. A previous study found that the MSD between the test device and the Embletta device was 0.31, thus we conservatively use the value of 0.4 to calculate power. In that study, the mean difference was -0.26 and the standard deviation of the differences was 0.49. We assume that 80% of the enrolled subjects will yield evaluable data. Specifically, an enrolled sample size of 43 subjects will yield 43*.80 = 34 evaluable subjects. This sample size yields at least 80% power when the true MSD is 0.4 or smaller.

The study will consist of one stage, which will consist of:

• Each participant will be measured simultaneously with both devices on 1 occasion for a duration of 5-9 minutes.
• The participant will be measured in the seated position.

Measurement Devices:

End Tidal C02 will be measured using a Capnograph (BCI Capnograph 9004, Smiths Medical). Expired gas will be collected via a nasal cannula with sampling line and moisture trap. Expired gas will be sampled at a flow rate of 150 mL/min ± 20 ml. The device has an average delay time of 1.8 seconds and a response time of 2.1 seconds which will be taken into account in the analysis. The BCI Capnograph 9004 has a display range of 0 - 100 mmHg and is accurate to ±2 mmHg, further information can be found in the device instruction manual. Capnograph waveforms will be recorded into a computer by utilising the analogue output of the capnograph. The analogue output will be sampled at 30 Hz using a 16 bit analogue to digital converter (Micro 1401-3, Cambridge Electronic Design, UK). The sampling rate of 30 Hz is selected since it is equal to the display update rate of the waveform on the device, and it is therefore closest to what clinicians commonly assess. Clinician over-scored capnographic RR will be derived from the capnogram CO2 waveform. The Capnography device will be calibrated according to manufacturer instructions. The range of Respiration rate of the Capnography devices are 0 -150 brpm.

Thoraco-abdominal movement will be measured using a Thora-3Di. The grid size used will be 14 x 10. Five minutes of continuous resting tidal breathing data will be acquired for each subject. The signals are sampled at 30 Hz and represent the anterior thoraco-abdominal breathing motion. The Thora-3Di software provides a numerical output of RR as mean brpm, calculated from the SLPVol/Time trace over a selected time frame.

The five minutes of continuous resting tidal breathing data will be collected simultaneously from both devices.

Data Transfer:

Capnographic outputs will be acquired and saved as data files for subsequent processing and review. The Thora-3Di™ software thoraco-abdominal signals will be exported to comma-separated value (CSV) files.

Before data analysis, all measurements will be assessed on their quality, and any protocol violators identified. The Thora-3Di™ device software detects the noise level of the output signal, and where the software flags a measurement as excessively noisy a measurement will not be obtained. For example, if a patient moves their upper body, this may prohibit a valid Thora-3Di™ measurement from being recorded. Likewise, interference with, or condensation within, the Capnography sampling tubing may produce artefacts or prevent a valid Capnography trace being recorded. Where paired samples cannot be obtained due to protocol violations, this will render the data unanalyzable so these subjects will be excluded from the analysis. Details of protocol violators but will be tabulated with reasons for exclusion listed. Any specific sub-population or disease state where the Thora-3Di™ did not work in as expected will be noted. Details of analyzable datasets (i.e. excluding protocol violators) will be tabulated

Data Analysis:

The primary endpoint analysis is to compare the output respiratory rate of Thora-3Di with that of a COSC measured over one randomly selected epoch of one minute for each subject. The 5 minute measurement period will span the range of intended use of the Thora-3Di™. A duration of 1 minute has been chosen as an appropriate comparison period for the analysis, as the number of breaths in one minute is the duration over which Respiratory Rate is clinically defined. Where the SLP tracing is orange or there is a significant change in baseline or a spike caused by body movement or such that would otherwise be excluded by labelling, those segments of the recording will not be used in the analysis. This is in accordance with intended use of the device.

To identify the randomly selected epoch for each subject, a uniform random number generator will be used to generate a random starting sample from which point a 1 minute epoch of data will be truncated for further analysis. This will be done using the rand function in MATLAB. The code will be:

starting sample = round((rand*4)*60*Fs), where rand is the uniform random number generator, it is multiplied by 4 to increase the default range of [0,1] to [0,4]. rand*4 essentially produces the time (in minutes) from which 1 minute of data will be truncated. The rest of the code translate the starting minute to a sample number, that is, multiplied by 60 to get the time in seconds, multiplied by Fs (sampling rate of the device) to get the sample and finally, since samples are always integers, round the number to get a the closest starting sample number.

This random number generation process will be repeated in order to allocate a random starting point to each of the paired analyzable datasets. To derive the SLP Respiratory Rate, the section corresponding to a 1 minute epoch for each sample (as identified on the 'Appraised Paired Sample' table) will be selected on SLPVol/Time trace of the Thora-3Di. The RR value for this period, which is a direct output from the software, will be recorded for this time section.

To derive the COSC RR, an expert clinician will be provided with the capnograph traces only and will verify each breath in the indicated 1 minute epoch of the capnograph output and calculate the RR for the time section for each subject.

To ensure no bias in the scoring, the clinician scoring the capnograph output, will have no reference to the Thora-3Di™ trace or SLP Respiratory Rate and likewise, the individual deriving SLP Respiratory Rate will have no reference to the capnograph output or the derived COSC Respiratory Rate.

To ensure no bias is introduced from the data collection stage to the data processing stage, the operator collecting the data have no involvement in the scoring of the outputs.

Hypotheses:

The following hypotheses will be tested with a significance level of =0.05:

• Null hypothesis: The true root mean squared difference (RMSD) between test and reference device RR is greater than or equal to 0.9 brpm.
• Alternative hypothesis: The true RMSD is less than 0.9 brpm.

The hypotheses above are equivalent to the following:

Null hypothesis: The true mean squared difference (MSD) between test and reference device RR is greater than or equal to 0.81 (breaths per minute)2.

Alternative hypothesis: The true MSD is less than 0.81 (breaths per minute)2. A RMSD< 0.90 brpm implies that the 95% limits of agreement are within 2 brpm (see Section 14.2). Pre-defined limits of agreement of 2 brpm were suggested by Smith et al. 2011 as clinically insignificant.

Analysis Populations:

The study population will comprise normal subjects, and subjects with other respiratory diseases. The full analysis population will comprise all patients who were entered into the study, excluding protocol violators, and provided one pair of measurements from the Thora-3Di device and the COSC. This population will be used as part of the primary analysis.

Poolability of Data Across Sites:

To assess the poolability of data across sites, the mean absolute differences between RR measured by Thora 3Di and the Capnograph will compared with a two-sample t-test. If the p-value of the test is greater than 0.10, poolability of data will be concluded.

Primary Endpoint Analysis and Acceptance Criteria:

Respiratory rate counts from the Thora 3Di and the COSC will be compared in a pair-wise manner. To test the hypothesis that the MSD is less than 0.81 brpm2, a Wald test will be done. The following test statistic will be calculated and compared to the threshold value of -1.645 from a standard normal distribution:

T=(MSD-0.81)/(υ(τ=0.81)) where υ(τ=0.81) is the estimated standard error of the MSD assuming the mean MSD is 0.81, as defined in Section 14.2. If T<-1.645, the null hypothesis will be rejected and we will conclude that the test device has sufficient accuracy.

Additionally, Bland-Altman plots will be produced and the 95% limits of agreement will be calculated. Specifically, the differences between test and reference device RR measurements will be plotted against the mean of the two RR measurements. Reference lines representing the 95% limits of agreement, defined by the mean difference ±1.96x standard deviation of the difference, will also be included on the plot. A scatter plot with the linear regression line and its slope and intercept will also be given.

The acceptance criteria for primary endpoint is +/- 2 brpm. This is equivalent to a MSD<0.81 brpm2.

Overall Analysis All subjects entered into the study will be accounted for by listing the numbers of patients entered, the numbers in each population, withdrawals and the reasons for withdrawal. Summary statistics for adverse incidents, demographics, clinical characteristics, Respiration Rates from the Thora 3Di device and the COSC will be produced. In general, categorical data will be summarized using frequency counts and percentages, and continuous data will be summarized using means, standard deviations, minimums, medians and maximums. These summaries will be inspected for any concerning patterns and further analysis may be performed as appropriate.

• Study Type: Interventional
• Enrollment (Anticipated): 43
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Rachel C Wilson, PhD; Phone Number: 07907 071978; Email: rachel.wilson@pneumacare.com
• Study Contact Backup: Name: Iles, MD

Study Locations

• United Kingdom
  • Cambridgeshire
    • Cambridge, Cambridgeshire, United Kingdom, CB2 0QQ
      • Recruiting
      • Cambridge University Hospitals Foundation Trust
      • Contact: Angelique Laubscher; Phone Number: 01223 811882; Email: alaubscher@cambridge-clinical-research.co.uk
      • Principal Investigator: Shuaib Nasser, MD
  • Suffolk
    • Stowmarket, Suffolk, United Kingdom, IP14 1NL
      • Recruiting
      • Stowhealth
      • Contact: Karen Norcott; Phone Number: 01223 811882; Email: knorcott@cambridge-clinical-research.co.uk
      • Principal Investigator: Neil Macey, MD

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years to 80 years (Adult, Older Adult)
• Accepts Healthy Volunteers: Yes
• Genders Eligible for Study: All

Description

Inclusion Criteria:

• Male or female,
• Age range 18-80 years,
• BMI range 18-40 kg/m2
• Patients with respiratory disease or breathing impairment (Patient Group) or normal subjects with no previous or current diagnosis of respiratory disease (Normals group)

Exclusion Criteria:

• Patient unable to sit in an upright position for required period of time
• Patients with significant co morbidities (assessed by the clinician at screening only):
• Significant unilateral lung disease e.g. pneumonectomy
• Chest wall or spinal deformity e.g. scoliosis
• Obstructive Sleep Apnoea, Apnoea hypopnoea index > 30 (if known)
• BMI>40
• Inability to consent/comply with trial protocol
• Presence of an acute disease process that might interfere with test performance (e.g. Nausea, vomiting, persistent coughing)

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Supportive Care
• Allocation: N/A
• Interventional Model: Single Group Assignment
• Masking: Single

Number of Arms

1

Arms and Interventions

Participant Group / Arm

Intervention / Treatment

Other: Respiratory Rate Measurement Comparision; To demonstrate equivalence of the Thora-3Di Structured Light Plethysmography device and a Clinician Over-scored End Tidal C02 (COSC)derived from a BCI Capnograph 9004 (Smiths Medical) for measurement of Respiratory Rate

Device: Structured Light Plethysmography (Thora-3Di); Respiratory Rate measured by Structured Light Plethysmography Thora-3Di and Clinician Over-scored End Tidal C02 (COSC) during 1 randomised minute of a 5 minute measurement period; Other Names: Thora-3Di Structured Light Plethysmography; Device: BCI Capnograph 9004 (Smiths Medical); Respiratory Rate measured by Structured Light Plethysmography Thora-3Di and Clinician Over-scored End Tidal C02 (COSC) during 1 randomised minute of a 5 minute measurement period; Other Names: Capnocheck (Smiths Medical)

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
The root mean squared difference (RMSD) in Respiratory Rate between the test (Thora-3Di) and reference Device (COSC)	The test device will be accepted as sufficiently accurate if the RMSD is significantly less than 0.9 brpm (significance level 0.05, power 0.8)	1 minute (randomly selected) of a 5 minute measurement period

Collaborators and Investigators

• Sponsor: Pneumacare Ltd
• Collaborators: Cambridge University Hospitals NHS Foundation Trust; University of Cambridge; Norfolk and Suffolk Primary and Community Care; Cambridge Clinical Research Ltd; Kelly Statistical Consulting
• Investigators: Principal Investigator: Richard Iles, MD, Cambridge University Hospitals Foundation Trust

Publications and helpful links

General Publications

• de Boer WH, Lasenby J, Cameron J, Wareham R, Ahmad S, Roach C, Hills W, Iles R. SLP: a zero-contact non-invasive method for pulmonary function testing. In: Labrosse F, Zwiggelaar R, Liu Y, Tiddeman B, eds, Proceedings of the British Machine Vision Conference. BMVA Press, 2010; pp 85.1-85.12
• Morgan MD, Gourlay AR, Denison DM. An optical method of studying the shape and movement of the chest wall in recumbent patients. Thorax. 1984 Feb;39(2):101-6. doi: 10.1136/thx.39.2.101.
• Bland JM. Sample size in guidelines trials. Fam Pract. 2000 Feb;17 Suppl 1:S17-20. doi: 10.1093/fampra/17.suppl_1.s17.
• Cretikos MA, Bellomo R, Hillman K, Chen J, Finfer S, Flabouris A. Respiratory rate: the neglected vital sign. Med J Aust. 2008 Jun 2;188(11):657-9. doi: 10.5694/j.1326-5377.2008.tb01825.x.
• Gaucher A, Frasca D, Mimoz O, Debaene B. Accuracy of respiratory rate monitoring by capnometry using the Capnomask(R) in extubated patients receiving supplemental oxygen after surgery. Br J Anaesth. 2012 Feb;108(2):316-20. doi: 10.1093/bja/aer383. Epub 2011 Dec 11.
• Jaffe MB. Infrared measurement of carbon dioxide in the human breath: "breathe-through" devices from Tyndall to the present day. Anesth Analg. 2008 Sep;107(3):890-904. doi: 10.1213/ane.0b013e31817ee3b3.
• Kodali BS. Capnography outside the operating rooms. Anesthesiology. 2013 Jan;118(1):192-201. doi: 10.1097/ALN.0b013e318278c8b6. No abstract available.
• Storm-Versloot MN, Verweij L, Lucas C, Ludikhuize J, Goslings JC, Legemate DA, Vermeulen H. Clinical relevance of routinely measured vital signs in hospitalized patients: a systematic review. J Nurs Scholarsh. 2014 Jan;46(1):39-49. doi: 10.1111/jnu.12048. Epub 2013 Oct 11.
• Yanagidate F, Dohi S. Modified nasal cannula for simultaneous oxygen delivery and end-tidal CO2 monitoring during spontaneous breathing. Eur J Anaesthesiol. 2006 Mar;23(3):257-60. doi: 10.1017/S0265021505002279.
• Box, Hunter and Hunter (1978). Statistics for experimenters. Wiley. p118. ISBN 0471093157
• Johnson NL, Kotz S, Balakrishnan N. 1970. Continuous Univariate Distributions, Volume 2, Wiley.
• Smith I, Mackay J, Fahrid N, Krucheck D. Respiratory rate measurement: a comparison of methods. British Journal of Healthcare Assistants 2011; 5: 18-23

Study record dates

Study Major Dates

• Study Start: February 1, 2015
• Primary Completion (Anticipated): May 1, 2015
• Study Completion (Anticipated): May 1, 2015

Study Registration Dates

• First Submitted: February 23, 2015
• First Submitted That Met QC Criteria: February 27, 2015
• First Posted (Estimate): March 2, 2015

Study Record Updates

• Last Update Posted (Estimate): March 6, 2015
• Last Update Submitted That Met QC Criteria: March 5, 2015
• Last Verified: March 1, 2015

More Information

Terms related to this study

• Keywords: Respiratory Rate; Capnography; Structured Light Plethysmography
• Additional Relevant MeSH Terms: Respiratory Tract Diseases; Lung Diseases
• Other Study ID Numbers: PMC-T3D-001