The purpose of the AcRIS study is to obtain data to characterize the relationship between symptoms and voice features for (reverse transcription polymerase chain reaction (RT-PCR) confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), influenza virus, or Respiratory Syncytial Virus (RSV) positive participants with acute viral respiratory illness. This data will be used as the basis to build voice and symptom algorithm(s) for detection and monitoring of these illnesses. This would benefit vaccine development across several key disease areas, including SARS-CoV-2, influenza virus and RSV. The study also models concepts of more efficient "flexible" clinical trials involving not only voice capture, but also web-based participant recruitment, enhanced participant engagement, and remote sample collection that could make future clinical studies more efficient. The clinical data obtained in this observational study could provide the documentation of the technology's performance needed to enable its deployment in future interventional studies.
Each subject will be required to stay in the study for 6 weeks. If the participant tests positive for any of the three viruses at swab #1 or swab #2, they will continue the study until the end of Week 8. Participants will record acute respiratory illnesses symptoms and voice data daily for up to a maximum of 8 weeks in both the well state and, should they become ill, the sick state, utilizing the Electronic diary on their Mobile application. Once enrolled, the participant will start recording symptoms and voice in the Electronic diary, with daily time commitment to this portion of the study expected to be 2-4 minutes. Two nasal self-swab collection kits will be ordered for delivery to the participant once they are enrolled in the study. The participant will be asked to self-swab when the test kit arrives (swab #1). The kit, including the specimen, will be returned to the central lab for RT-PCR SARS-CoV2/Influenza/RSV RT-PCR testing. The participant is expected to complete 3 phonemes and 5 lines of reading each day, in addition to score the self-reported symptoms in the Electronic diary. If participants become sick (self-report) with new or increased symptoms of acute respiratory illness symptoms, they will be asked to self-swab (swab #2) and return the sample for central SARS-CoV-2/Influenza/RSV RT-PCR testing. If the participant does not develop any new or increased symptoms between swab #1 and end of Week 6, they will obtain a self-swab (swab #2) at Day 42. If the participant tests positive for any of the three viruses at swab #1 or swab #2, they will continue the study until end of week 8. If they test negative for the three viruses at swab #1 and swab #2, they will exit the study at approximately the end of week 6 when the test results are returned. The results of the RT-PCR testing will be shared with participants.
Study Type
OBSERVATIONAL
Enrollment
9,151
nasal swab
Ochsner Clinic Foundation
New Orleans, Louisiana, United States
Change From Baseline in Self-Reported Symptom Scores From Well-to-Sick State Through Day 56
Symptoms (fever, cough, difficult breathing, fatigue, runny nose, stuffy/blocked nose, sore throat, loss of taste/smell, chills, muscle pain, diarrhea, vomiting, headache, nausea, rigors, wheezing) recorded at least once daily in e-diary for 8 weeks \& rated 0:none to 4:severe for fever \& 0:none to 7:severe for other symptoms. Total symptom score=sum of all symptom scores in a recording session \& mean of daily total symptom score=average across available sessions for each day, range=0 to 109, higher value=more severe total symptoms. Baseline=average of values up to 7 days before 1st occurrence of new/increased symptoms. If no data in 7 days, average of endpoint values for closest 3 days prior to 7 days used as baseline. Linear mixed effect model used for analysis for change from baseline in each symptom versus day during well to sick period, with slope (without intercept) included as fixed effects and random effects. Fixed effect of slope, as population level estimate, was reported.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Features (AHH_Max Phonation Time, EE_Jitter Local Absolute, MM_Jitter Local Absolute) Values From Well-to-Sick State Through Day 56
Participants recorded voice features such as pitch, jitter once daily in an electronic diary for 8 weeks. Voice assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Max Phonation Time: duration that sound is held for. Jitter Local Absolute: average absolute difference between consecutive periods. Indicates how unsteady the pitch is across neighboring glottal pulses. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor in addition to values mentioned. Individual exponential factors have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Feature (Cepstral Peak Prominence, Harmonicity, MFCC Mean, MFCC Std, SNR, Shimmer Local dB, Spectral Flatness, Third Octave Band, and VLHR) Values From Well-to-Sick State Through Day 56
Harmonicity, flatness, shimmer recorded and assessed by phonemes: 'eee' and 'mmm' for 4 seconds, "ahh" sustained and 5-sentence reading passage. Cepstral Peak Prominence: voice quality measure. Harmonicity: degree of periodicity in signal. mel frequency cepstral coefficients (MFCC) mean: quantifies shape of spectrum. MFCC std: quantifies variation in spectral shape over time. Signal-to-noise ratio (SNR): how loud signal is compared to background. Shimmer Local dB: average absolute difference between amplitudes of consecutive periods (how unsteady sound intensity is across neighboring glottal pulses). Spectral Flatness: quantifies how tone-like a signal is based on spectral distribution. Third Octave Band: energy in 200 Hz third octave band relative to total. VLHR: degree of nasality. Linear mixed effect model used for analysis. Fixed effect of slope, as population level estimate is reported. Individual exponential factors for data presented have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Feature (Coefficient of Variation, Mel Frequency Cepstral Coefficients (MFCC) 1st Order Delta, MFCC 2nd Order Delta) Values From Well-to-Sick State Through Day 56
Voice features were recorded once daily in an electronic diary for 8 weeks. Assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Coefficient of Variation of F0: measures variation in pitch over time. MFCC 1st order delta: time-averaged estimate of first derivate of the MFCCs. MFCC 2nd order delta: time-averaged estimate of the second derivative of the MFCCs. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor in addition to values mentioned. Individual exponential factors have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Features (EE_Entropy, MM_Entropy) Values From Well-to-Sick State Through Day 56
Participants recorded voice features such as entropy once daily in an electronic diary for 8 weeks. Voice assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Entropy: Shannon entropy of the spectral distribution. Quantifies tonailty similar to spectral flatness. "EE\_Entropy" and "MM\_Entropy" refer to this same measure computed on the "eee" and "mmm" phonemes, respectively. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor in addition to values mentioned. Individual exponential factors have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Features (Formant and Formant Bandwidth) Values From Well-to-Sick State Through Day 56
Participants recorded voice features once daily in an electronic diary for 8 weeks. Voice assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Formant: frequency at which the vocal tract produces an acoustic resonance. Formant Bandwidth: the spectral width of the acoustic resonance. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor in addition to values mentioned. Individual exponential factors have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Features (EE_Voiced Frames, MM_Voiced Frames) Values From Well-to-Sick State Through Day 56
Participants recorded voice features once daily in an electronic diary for 8 weeks. Voice assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. The voiced frames rate indicates how much of the sound is voiced. EE and MM refer to the sounds on which the measure is computed. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor in addition to values mentioned. Individual exponential factors have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Features (EE_Jitter Local, MM_Jitter Local) Values From Well-to-Sick State Through Day 56
Voice features such as jitter were recorded once daily in an electronic diary for 8 weeks. Voice assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Jitter Local: average absolute difference between consecutive periods divided by the average period as a percentage. Indicates how unsteady the pitch is across neighboring glottal pulses. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor in addition to values mentioned. Individual exponential factors have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Features (EE_Shimmer Local, MM_Shimmer Local) Values From Well-to-Sick State Through Day 56
Voice features such as shimmer were once daily in an electronic diary for 8 weeks. Voice assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Shimmer Local: average absolute difference between the amplitudes of consecutive periods divided by the average amplitude as a percentage. Indicates how unsteady the sound intensity is across neighboring glottal pulses. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor in addition to values mentioned. Individual exponential factors have been mentioned in respective row title.
Time frame: Baseline up to Day 56
Change From Baseline in Voice Features (READ_Speaking Rate) Values From Well-to-Sick State Through Day 56
Participants recorded voice features once daily in an electronic diary for 8 weeks. Voice assessments were done by 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Baseline was average of endpoint values up to 7 days before first occurrence of new/increased symptoms. If participant had no data in 7 days, average of endpoint values for closest 3 days prior to these 7 days was used as baseline. Linear mixed effect model was used for analysis for change from baseline in each voice feature versus day during the well to sick period, with slope (without intercept) included as fixed effects and random effects. The fixed effect of the slope, as the population level estimate, was reported. The data given has exponential factor (i.e. LSM\*10\^-2, SE\*10\^-3) in addition to values mentioned.
Time frame: Baseline up to Day 56
Percentage of Compliant Days in Total Days of Symptoms and Total Days of Voice Recordings Entered in the Electronic Diary
In this outcome measure, percentage of compliant days in total days of self-reported symptoms and total days of voice recordings as entered in the electronic diary were reported. Compliant days were calculated for the following and expressed as percentage: symptom compliance = number of compliant days of self-reported symptoms divided by the total number of days in the study. Voice compliance (expressed as percentage) = number of compliant days of completed voice recordings divided by the total number of days in the study. Partial voice compliance (expressed as percentage) = number of compliant days of partially completed voice recordings divided by the total number of days in the study. Symptom and voice compliance (expressed as percentage) = number of compliant days of self-reported symptoms and completed voice recordings divided by the total number of days in the study. Participants who experienced technical issues due to study operational errors were excluded.
Time frame: Day 1 up to Day 56
Percentage of Quality Voice Recordings
In this outcome measure, percentage was calculated over the number of days that the participant was able to complete the task for each recording. Participants were required to record their daily voice in electronic diary and the assessment included 2 phonemes: 'eee' and 'mmm' for 4 seconds (minimum 3 seconds), 1 phoneme: "ahh" sustained (as long as possible) and a 5-sentence reading passage. Quality recordings were defined as recordings exceeding signal-to-noise (SNR) and duration thresholds (i.e. SNR \>= 20 decibels \[dB\] that have duration \>= 3 seconds for phoneme and \>= 10 seconds for reading tasks). In the case of multiple electronic diary sessions per day, the session with the highest quality voice recordings for that day was used in this analysis.
Time frame: Day 1 up to Day 56
Percentage of Participants With a Positive Self-Swab Result for SARS-CoV-2 and/or Influenza and/or RSV
In this outcome measure, percentage of participants with valid self-swab 1 and 2 results positive for SARS-COV-2, Flu A, Flu B, RSV and SARS-CoV-2 and/or Influenza (Flu A/Flu B) and/or RSV were reported.
Time frame: Day 2 up to Day 43
Percentage of Participants Who Administered Self-swab 1 and Self-swab 2
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.
In this outcome measure, percentage of participants who administered self-swab 1, self-swab 2 and self-swab 1 and 2 excluding participants who experienced technical issues due to study operational errors were reported.
Time frame: Day 2 up to Day 43
Number of Days Between Reporting Symptoms and Recording Swab 2 in the Electronic-Diary
In this outcome measure, number of days from symptom onset (reporting new or increased symptoms in the electronic diary) to the day of recording Swab 2 day in the electronic diary averaged across participants who administered self-swab 2 was reported.
Time frame: Day 2 Up to Day 43
Percentage of Participants Reporting Symptoms in the Electronic Diary Who Have a Self-swab Collected at or Around Symptom Onset
In this outcome measure, 1) participants (par) who administered (adm) self-swab 2 (SS2) before Week 6 (W6) divided by the number of participants who reported new or increased (inc) symptoms (sym) in the electronic diary (e-diary) before W6 and 2) number of participants who reported new or increased symptoms in the e-diary before Week 6 divided by the number of participants who administered self-swab 2 before Week 6 were reported in percentage.
Time frame: Day 2 Up to Day 43
Percentage of Participants Who Administered Self-swabs With Valid Results
In this outcome measure, percentage of participants with valid (positive or negative) self-swab results for SARS-CoV-2, Flu A, Flu b and RSV were reported.
Time frame: Day 2 Up to Day 43