Edge Rewrite
everything Source code View original
// HTMLRewriter · presentation

This page was redesigned at the edge.

Cloudflare fetched the original article and streamed it through HTMLRewriter to apply an entirely new visual system without rebuilding the source page.

// request.cf · coarse context

A page that knows where it met you.

Only coarse request metadata is shown. This demo does not display or persist visitor IP addresses.

Country
US
Cloudflare location
CMH
Connection
HTTP/2
Language
Not provided

Ray ID: a22b430d9b15d319

Jump to content

Draft:Objective cough monitoring

From Wikipedia, the free encyclopedia

Objective cough monitoring is the automated, quantitative measurement of cough frequency using acoustic or vibration based sensing technologies.

Unlike subjective methods such as patient reported outcomes (PROs) and diaries that rely on patient perspectives, objective measures capture cough as a physiological signal. These approaches detect and quantify cough frequency, rate (coughs per hour), temporal patterns, and bout structure over time [1]. Currently used cough monitoring technologies include ambulatory audio recording systems, smartphone based acoustic detection, and wearable devices with on device AI processing. The latest technologies enable continuous monitoring in real world settings, supporting longitudinal disease tracking.

Cough is the most common respiratory complaint worldwide, with chronic cough affecting approximately 10% of adults globally [2]. It is a burdensome symptom observed across respiratory and non respiratory conditions, and also serves as a clinically relevant biomarker providing diagnostic and prognostic insight and correlating with disease severity and outcomes [1]. Despite its prevalence and clinical importance, clinical evaluation often relies on patient recall, which is systematically inaccurate and prone to bias. This limitation underscores the need for more reliable objective measurement approaches that can quantify cough frequency, intensity, and diurnal patterns over days or weeks.

Technologies for cough measurement have evolved across three generations: early clinician observation and patient diaries, ambulatory audio recording systems with human annotation, and continuous monitoring using wearable devices with on device AI.[3][4]. These systems are now used as endpoints in clinical trials, particularly in chronic cough drug development, and have been deployed in multicenter studies for quantifying chronic cough across various indications [5]

These cough monitoring devices fall within the class of sensor based digital health technologies. The Digital Medicine Society's V3+ framework evaluates such technologies across verification, analytical validation, clinical validation, and usability validation [6]. As of April 2026, no system has received specific U.S. Food and Drug Administration clearance for cough counting.

Analytical validation of these systems compares algorithm output against expert manual cough counts, with reported event level sensitivity typically high in adult systems; a 2026 meta-analysis pooling 16 studies reported sensitivity of 89 percent and specificity of 99 percent, although sensitivity falls substantially in pediatric and high noise settings. Objective cough frequency has been used as a primary or secondary endpoint in chronic cough drug trials, most prominently the gefapixant Phase 3 trials, where 24 hour cough frequency was the primary endpoint. Regulatory status differs by region: the European Medicines Agency accepted objective cough frequency as a primary endpoint and authorized gefapixant in 2023, whereas a U.S. advisory committee found the same evidence did not demonstrate clinically meaningful benefit. Because monitors that record environmental audio may capture intelligible speech, privacy considerations including consent, data minimization, and on device processing are central to study design under HIPAA and the General Data Protection Regulation. Open questions remain around the absence of standardization, limited pediatric validation, degraded performance in noisy environments, and the weak to moderate correlation between objective cough counts and patient reported outcomes.

Background and clinical significance

[edit]

Cough as a clinical measure

[edit]

Cough is a common burdensome feature associated with several respiratory diseases, such as asthma, chronic obstructive pulmonary disease (COPD), bronchiectasis, interstitial lung disease (ILD), and tuberculosis. It is also observed in non respiratory disorders such as gastroesophageal reflux disease (GERD) and lung malignancies.[7][8]. Beyond being a symptom, chronic cough (CC) also has clinical relevance as a biomarker, reflecting underlying disease processes. Variations in cough pattern and characteristics differ across etiologies, providing diagnostic and prognostic insight. Higher symptom burden and exacerbations are often correlated to poorer outcomes in conditions such as asthma and idiopathic pulmonary fibrosis (IPF) [7]

Cough counts and patterns offer additional clinical value, capturing circadian and longitudinal trends that reflect disease severity and progression.[1]. Incorporating objective measures such as cough duration, frequency, intensity, diurnal patterns, and other characteristics (e.g., wet or dry) further enhances phenotyping (grouping patients by observable disease characteristics) and diagnostic precision. Cough frequency also carries pharmacodynamic relevance, enabling detection of dose response relationships and early treatment effects compared to patient reported outcomes (PROs). A 2022 validation study proposed that a 30% or greater reduction in 24 hour objective cough frequency may be more clinically meaningful than changes in subjective scores [9]. Cough frequency data also allow patient stratification and subgroup identification in clinical trials [10]

Subjective cough severity and objective cough frequency are correlated but non identical constructs. Patients with similar cough counts per hour can differ in cough intensity and perception. Several studies show that subjective scores exhibit only moderate correlation with objective cough frequency.[11]. Use of objective measures therefore adds value by providing more precise and reproducible assessment for research on drug mechanisms, treatment evaluation, and disease monitoring [12][13]

Epidemiology

[edit]

CC is one of the most common reasons for respiratory consultations, representing a substantial clinical and public health burden. Globally, the prevalence of CC is estimated at approximately 10% in adults, with variations across geographies. Prevalence ranges from 10% to 20% in Europe, America, and Australia; 2% in Africa; 18% in Oceania; and between 4% and 18% in Asia.[14][2]. These differences reflect heterogeneity in study design, population characteristics, and definitions of CC. Demographically, CC is more frequently reported in middle aged women, possibly due to increased cough reflex sensitivity. Prevalence also increases with age, with the greatest burden observed in older adults [15]. In children, prevalence estimates range widely (10% to 20%), with community based studies showing 10% experiencing cough lasting more than three weeks [15]

The most common causes of CC in adults include upper airway cough syndrome, asthma, non asthmatic eosinophilic bronchitis, and gastroesophageal or laryngopharyngeal reflux.[16]. An estimated 0.5% to 3% of adults report refractory chronic cough (RCC), a condition where cough persists despite extensive guideline recommended evaluation and treatment of underlying conditions. An additional 12% to 40% of CC cases are classified as unexplained chronic cough (UCC), where no clear cause is identified despite thorough investigation [17]

CC also imposes a substantial burden on health related quality of life, extending beyond its clinical implications. Evidence from questionnaires such as the Leicester Cough Questionnaire (LCQ) shows that patients frequently experience complications such as sleep disruption, chest discomfort, and stress urinary incontinence (particularly in women), as well as psychological distress, anxiety, and social withdrawal [2]. This underscores the burden of CC and supports the need for systematic measurement approaches for clinical assessment and diagnostic decision making.

The measurement problem

[edit]

Despite its clinical relevance, accurate assessment of cough remains methodologically challenging. In practice, evaluation often relies on patient recall, which is systematically inaccurate and limits reliability [1][5]. Objective methods such as 24 hour cough monitoring have improved measurement but do not fully address underlying challenges.

Cough exhibits substantial variability within individuals. A single 24 hour measurement may poorly represent habitual cough rate. Longitudinal studies have shown substantial within day and day to day variability for each subject with persistent cough recorded over 30 days.[18]. Measurement is further influenced by context. Cough frequency recorded in clinical settings may be lower than in free living environments, as the awareness of being monitored can alter natural cough behavior [5]. Some studies have also reported screen day inflation, where cough counts at trial screening are elevated relative to run in values, requiring placebo run in periods for patient selection [19]. A scoping review on continuous cough monitoring found wide variability in the relationship between perceived severity and measured frequency, driven by differences in perception, tolerance, intensity, and context [5]

Cough is therefore a dynamic physiological response that cannot be fully captured by any single measure, reinforcing the need for multiple approaches to more reliably capture real world patterns while maintaining accuracy and clinical relevance.

History

[edit]

Subjective methods and early instruments

[edit]

The assessment of cough severity is central to evaluating treatment response, encompassing frequency, intensity, and impact on quality of life. Before the advent of objective monitoring, clinical evaluation relied on clinician observation, patient diaries, and simple questionnaires. Over time, more structured and validated PRO methods were developed to improve consistency, accuracy, and clinical relevance.

The Leicester Cough Questionnaire (LCQ) is widely used to assess cough related quality of life across physical, psychological, and social domains, demonstrating reliability, repeatability, and responsiveness.[20]. The Hull Airway Reflux Questionnaire (HARQ) focuses on symptoms related to airway reflux and has utility in identifying reflux associated cough mechanisms [21]. Other tools include the Cough Severity Index (CSI) for upper airway symptoms, the Cough Symptom Score (CSS) that captures daily daytime and nighttime symptom patterns, and visual analogue scales (VAS), a linear scoring method for measuring perceived severity [22][23]

These instruments are practical, well validated, and remain integral to both clinical practice and as research endpoints in clinical trials. Regulatory frameworks require both PROs and objective measures to demonstrate physiological change and meaningful clinical benefit to patients.

Ambulatory audio recording (1990s to 2000s)

[edit]

Advances in digital technology during the 1990s and 2000s enabled the development of ambulatory cough monitoring systems. Improvements in audio recording, data storage, microphone miniaturization, and battery life made it feasible to capture continuous, high quality cough data over extended periods [4]. Two of the most widely used systems are the Leicester Cough Monitor (LCM), developed by Birring et al. at the University of Leicester, and the VitaloJAK system, developed through collaboration between Vitalograph Ltd. and the University Hospital of South Manchester.

The LCM is a lightweight system with a portable digital recorder and a lapel microphone with an automated algorithm capable of detecting cough sounds while filtering background noise. It has demonstrated high sensitivity and specificity [4]. The LCM established key methodological principles, including ambulatory recording, algorithm based detection, and human validated calibration.

The VitaloJAK system uses a combination of a lapel microphone and contact microphone attached to the upper sternum with a specially designed ambulatory recording device worn in a belt bag. Its software algorithm compresses audio recordings by removing non cough segments. This also requires trained operators to review audio visual displays to identify cough events.[4][24]

Both these technologies have been employed in clinical trials to monitor cough across RCC, COPD, asthma, and pulmonary fibrosis. In the gefapixant trials, they enabled reliable measurement of cough frequency and underpinned its use as a clinically meaningful endpoint.[4]

Smartphone and app based detection (2010s)

[edit]

The 2010s marked a shift toward smartphone based cough monitoring, enabled by advances in machine learning [24]. The widespread availability of smartphones, combined with improvements in processing power and embedded sensors, created opportunities for continuous, real world cough monitoring using built in microphones [4][24][25]. This marked a transition from semi automated systems requiring manual review to automated approaches capable of real time detection and analysis.

However, smartphone based monitoring introduces challenges. Inconsistent microphone placement, ambient noise interference, battery drain, and privacy concerns related to continuous audio recording and data transmission reduced adherence [1]. These challenges emphasized the need for more user friendly and privacy conscious technologies.

Wearable continuous monitoring (2020s)

[edit]

The 2020s saw the development of wearable continuous cough monitoring, driven by the convergence of miniaturized sensors, low power edge AI (machine learning that runs directly on the device rather than in the cloud), and wrist worn form factors. These systems support continuous, real world monitoring over extended periods, typically days to weeks.

The AI pipeline involves two stages: detection of cough like acoustic signals, followed by classification to distinguish true cough events from background noise. Machine learning models are designed to achieve high sensitivity while minimizing false detections, learning complex cough signatures and differentiating them from similar sounds.[3]

Several wearable platforms have been developed and deployed in multicenter clinical trials across indications such as COPD, asthma, IPF, and RCC/UCC [5]. These systems incorporate high performance microphones with cloud or on device processing, allowing continuous monitoring while reducing data burden. Some systems transmit only cough timestamps rather than audio, addressing privacy concerns [5]. Wearable cough monitoring enables longitudinal tracking over days to weeks, generating insights into cough patterns and disease activity.

Technology and methods

[edit]

Acoustic detection

[edit]

Cough detection algorithms are predominantly based on acoustic detection principles. The process involves two main steps. In the first step, relevant acoustic events are selected from a continuous audio stream via detection of short, high amplitude, transient sounds, which are further processed to produce time frequency or spectral representations such as mel spectrograms and cepstral coefficients (compact numerical summaries of a sound's frequency content that make patterns easier for an algorithm to classify).[4][24]. In the second step, a classifier categorizes events as coughing or noncoughing [4][24]. Earlier research focused mainly on feature extraction and statistical classifiers, whereas more recent efforts make use of neural networks [24]. Performance is affected by acoustic properties of the environment, including the presence of speech, throat clearing, laughing, body movement, and ambient noise [4][24]. Data collected in controlled settings cannot be directly extrapolated to free living application [4][24]

Ambulatory recording systems

[edit]

The first clinically validated generation of objective cough monitoring used extended sound recordings in ambulatory conditions, typically for 24 hours, followed by algorithm assisted reduction and manual annotation.[4][26][27][28]. Sounds from the environment are collected using a microphone, followed by exclusion of segments where there is little probability of a cough event, resulting in shorter files for annotation [4][28]. Such semiautomated systems provided a means for prolonged cough counting and for producing evidence in research studies and clinical trials [4][26][27][28]. Limitations of this methodology include dependence on reviewer time, limited scalability, reliance on audio data storage, associated privacy issues, and restriction to single day observation periods rather than extended multi week monitoring [4][28]

Continuous wearable monitoring

[edit]

A more recent generation of systems focuses on fully automated classification and extends the monitoring period beyond a single day long recording.[29][30][31][32]. Several implementations have utilized wrist worn devices, miniature chest worn sensors, contactless bedside devices, and smartphones [4][24][29][30][31][32]. A common design principle across these systems is processing at or near the sensor level, transmitting timestamps or derived cough event information rather than storing intelligible audio [29][30][31]. This architecture reduces privacy risk and makes repeated, longitudinal monitoring more practical [29][31]. These systems provide longitudinal output data such as hourly cough frequency, diurnal or nocturnal distribution, and day to day variation in cough burden, which are difficult to characterize from a single 24 hour recording alone [29][30][31]

Emerging approaches

[edit]

Other sensor modalities are under investigation, including accelerometric, strain sensor, and mechanoacoustic technologies, in which sensors detect cough related vibrations transmitted through the body surface rather than relying on airborne sound.[24][33][34]. Potential advantages include greater immunity to environmental noise and reduced privacy concern, since no intelligible voice recordings are stored [33][34]. These technologies remain at a relatively early stage compared with acoustic systems, and many have been tested only in pilot or single center studies [24][33][34]

Analytical validation

[edit]

Validation frameworks

[edit]

The Digital Medicine Society's V3+ framework defines four validation stages for sensor based digital health technologies: verification (whether the hardware generates the intended signal), usability validation (whether the technology functions effectively in the intended context), analytical validation (whether the algorithm output agrees with an appropriate reference standard), and clinical validation (whether the resulting measure is clinically meaningful).[6]. Failure can occur independently at each stage, making the framework particularly relevant to cough monitoring applications [6]

Ground truth methodology

[edit]

The accuracy of any cough monitor depends on the quality of the ground truth used for comparison.[35][29]. Manual scoring of coughs from audio recordings remains the accepted reference standard [35][29]. Earlier studies demonstrated excellent agreement between manually counted coughs from digital recordings and video recordings in well controlled settings [35]. In more recent free living studies, stricter criteria have been adopted, including two blinded expert listeners, predefined annotation units such as the cough second, adjudication of disagreements by a third listener, and explicit reporting of inter listener agreement [29]. Inter listener agreement effectively sets a practical upper bound on algorithmic performance, given that no reference standard surpasses expert human annotation [35][29]

Performance metrics

[edit]

Several complementary metrics are used to assess cough monitor accuracy.[4][30][36]. At the event level, sensitivity and specificity are the most widely reported [4][27][30][36]. Published figures vary substantially across systems and populations. The Leicester Cough Monitor reported sensitivity of 91% and specificity of 99% [27]. An automated nocturnal system (Albus Home) achieved sensitivity of 94.8% and specificity of 100.0% [30]. A pediatric smartphone algorithm validated in hospitalized children reported sensitivity of 47.6% with specificity of 99.96%, illustrating that high specificity does not ensure detection of all cough events [37]. A multicenter study of a wrist worn continuous monitor reported sensitivity of 90.4%, positive predictive value of 87.5%, and a false positive rate of 1.03 per hour [29]. Semiautomated systems such as VitaloJAK function primarily by compressing audio files for subsequent manual analysis rather than as fully autonomous cough counters [28]

At the count or rate level, agreement metrics provide a more informative assessment than correlation alone.[4][30][36]. The intraclass correlation coefficient (ICC) measures consistency between two methods accounting for both systematic and random differences, and published studies have reported high ICC values for validated cough monitors [4][30]. Bland-Altman analysis quantifies bias and limits of agreement; for example, one nocturnal system showed a mean hourly bias of −0.6 ± 2.0 coughs per hour [30], while a multicenter wrist worn monitor study reported a bias of 0.23 with limits of agreement from −3.7 to 4.8 [29]. Pearson correlation, by contrast, measures association rather than agreement: two methods can be highly correlated while one systematically overestimates the other, which is why agreement metrics are preferred [35][29]

Heterogeneity of reported metrics across studies remains a limitation, as it complicates direct cross study comparisons.[4][36]. A 2026 systematic review and meta-analysis of 16 studies comprising 52,612 cough events reported pooled sensitivity of 89% and pooled specificity of 99% [36]

Real world versus laboratory validation

[edit]

Validation results are strongest when recording conditions are controlled and background noise is minimal.[24][36]. Performance declines in free living contexts where cough detection must contend with speech, movement, and ambient sounds [24][29][30][31][37][32]. The degree of deterioration varies across systems and study conditions [24][36]. For example, a pediatric smartphone study reported sensitivity of 47.6% in hospitalized children [37], while a nocturnal ward system achieved sensitivity of 72% on device and 82% on computer with specificity of 99% on both platforms [32]. A multicenter wrist worn monitor study reported sensitivity of 90.4% during everyday routine activity [29], and a small chest worn sensor achieved sensitivity of 88.5% for daytime and 84.2% for nighttime outpatient monitoring [31]. Across studies, accuracy is typically lower in real world settings than in laboratory testing, but real world results better reflect intended clinical use [4][24][36]

Patient reported outcomes

[edit]

Objective cough measurement complements rather than replaces PRO measures.[4][38][39][40][21][41][42][9][1]. The LCQ is a 19 item instrument evaluating cough related quality of life across physical, psychological, and social dimensions [38]. An improvement in total LCQ score of at least 1.3 points is considered clinically meaningful [39]. The CSS, first described by Hsu et al., is a brief subjective tool with separate daytime and nighttime components used to assess cough severity [40]. The HARQ consists of 14 items assessing symptoms associated with cough hypersensitivity and airway reflux [21]. The Cough Specific Quality of Life Questionnaire (CQLQ) is a 28 item instrument measuring cough specific quality of life [41]. The cough severity VAS measures cough severity on a 100 mm scale ranging from "no cough" to "worst cough," and a reduction of at least 30 mm has been proposed as clinically meaningful in chronic cough studies [42]

These scales do not measure the same construct as an objective count of cough episodes.[4][9][1]. Objective frequency measurement counts occurrences, whereas PROs measure perceived severity, burden, and the extent to which symptoms affect sleep, mood, social interactions, and daily activities [4][9][1]. The relationship between objective frequency and PROs is usually statistically significant but only low to moderate [9][1]. A 2022 validation study found statistically significant but modest correlations between objective 24 hour cough frequency and cough specific PROs (Spearman rho 0.30 to 0.58) [9]. A 2025 scoping review reported median correlation coefficients of 0.42 for cough severity and −0.49 for cough related quality of life [1]. Both objective and subjective measures are therefore used in modern chronic cough trials [4][9][1]

Clinical trial applications

[edit]

Cough frequency as an endpoint

[edit]

As a quantitative trial endpoint, objective cough frequency allows for observable cough counts over a defined period.[43], can be measured with validated monitoring systems [44], and is less subjective than patient reported measures such as diaries or symptom scales [43]

Objective cough frequency has been used as both primary and secondary endpoints in clinical trials. In the gefapixant Phase 3 trials (COUGH-1 and COUGH-2), 24 hour cough frequency was the primary endpoint and cough frequency during waking hours was a secondary endpoint.[45]

Respiratory conditions where objective cough frequency has been used as an endpoint include acute cough.[9], asthma [9], COPD [9], and idiopathic pulmonary fibrosis (IPF) [46]

Baseline measurement and variability

[edit]

Observation induced behavioral changes may occur when subjects know they are being observed, a phenomenon known as the Hawthorne effect.[47]. This effect may occur in the presence of visible recording devices, contributing to the placebo context in cough trials [48]

Trials in chronic cough commonly enroll patients with severe symptoms and exclude milder cases, which can introduce regression toward the mean.[48]. This occurs because, over time, extreme values tend to move toward the average [48]

Day to day variability in cough frequency limits the predictability of cough patterns from a single day of monitoring. One study investigated cough variability over 30 days in 97 individuals with persistent cough using a continuous cough monitoring system.[18]. The study reported substantial within subject and day to day variability in cough frequency, with single day predictability ranging from 30% to 95% across subjects [18]

Design variability across trials

[edit]

During metric selection, trial designers may consider whether to analyze waking hours, 24 hour, and nocturnal cough independently, as there is evidence that daytime cough frequency (08:00 to 22:00) is greater than overnight frequency (average 25 vs. 10 coughs per hour).[27]

To reduce day to day variability, some trials have incorporated monitoring periods longer than 24 hours. In one study, when monitoring was performed for only 24 hours, the error rate (how much the short recording differed from the longer term cough rate) was 47%, but fell to 14% when monitoring was extended to 240 hours.[49]. Researchers have proposed that 7 days may be a reliable monitoring period for establishing a baseline in chronic cough; using a full week also allows for exploring cyclical activities that may influence coughing frequency [5]

Data handling is also a relevant design consideration, as evidence suggests past monitoring studies have lost between 4% and 40% of cough monitoring data.[50]. Causes of missing data include non device related factors, device failure, uninterpretable recordings, occupational issues, and patient discomfort [50]

Applications by indication

[edit]

Refractory and unexplained chronic cough

[edit]

The European Respiratory Society (ERS) defines refractory chronic cough (RCC) as a cough that persists despite optimal treatment of underlying conditions.[51]. The same guideline defines unexplained chronic cough (UCC) as a cough that persists despite a thorough diagnostic evaluation that yields no confirmed cause [51]. Relevant to both conditions is cough hypersensitivity syndrome (CHS), an overarching diagnosis for adults who cough in response to low levels of thermal, chemical, or mechanical stimulation [51]

COUGH-1 and COUGH-2 were the first Phase 3 randomized controlled trials evaluating treatments for chronic cough.[45]. The trials reported a reduction in 24 hour objective cough frequency versus placebo, with an 18.5% reduction at week 12 in COUGH-1 (p = 0.041) and a 14.6% reduction at week 24 in COUGH-2 (p = 0.031) [45]. COUGH-1 showed no reduction in awake cough frequency versus placebo that reached statistical significance, while COUGH-2 showed a 15.79% reduction (p = 0.022) [45]

COPD and chronic bronchitis

[edit]

Chronic cough and sputum production are prevalent symptoms in patients with COPD and may constitute risk factors for exacerbation.[52]. A prospective exploratory monitoring study with 28 COPD patients showed that objective cough frequency predicted 45% of cough exacerbations at an average of 3.4 days before clinical onset [53]

Idiopathic pulmonary fibrosis

[edit]

According to a state of the art review, cough in IPF is an independent predictor of disease progression and may predict time to death or lung transplantation.[7]

The PAciFy Cough Phase 2 randomized crossover trial evaluated low dose morphine for the treatment of cough in IPF and found that morphine reduced objective awake cough frequency by 39.4% compared with placebo (p = 0.0005).[54]

Asthma

[edit]

One study showed that 24 hour ambulatory cough frequency correlates with Asthma Control Questionnaire (ACQ) scores but is statistically independent of airflow obstruction in multivariate analyses across 89 subjects.[55]. Objective measurements of cough therefore provide information about asthma control that is not fully captured by ACQ or Global Initiative for Asthma (GINA) questionnaires [55]

Bronchiectasis

[edit]

Cough is a common symptom among bronchiectasis patients, with estimates ranging from 82% to 98%.[7]. The ERS identifies cough changes as one of the core diagnostic criteria for acute exacerbation [56]

A prospective study of adults with bronchiectasis using ambulatory objective cough monitoring reported that cough frequency was higher in patients with sputum producing disease compared to those without.[57]

Tuberculosis

[edit]

The World Health Organization (WHO) guideline on tuberculosis recommends that symptom screening should include cough along with fever, night sweats, weight loss, or poor weight gain.[58]

Continuous cough monitoring has been proposed as a digital biomarker (an objective physiological indicator captured by a digital device) for TB diagnosis and treatment response monitoring in resource limited settings, where noninvasive, affordable tools may help reduce overtreatment and improve detection when routine symptom based screening misses cases.[59]. A prospective study of 22 adults with pulmonary TB found that continuously monitored cough frequency fell during treatment, from a median of 11.0 coughs per hour in week 1 to 1.0 by week 26 [60]. The study concluded that community based continuous cough monitoring is feasible in low resource settings, but adherence remains a challenge, as do access to internet and electricity [60]

Respiratory viral surveillance

[edit]

One prospective cohort study investigated whether acoustic surveillance of cough using artificial intelligence can be used for respiratory disease surveillance.[61]. The study collected 62,000 cough events from 616 participants and found that cough frequency increased along with respiratory related medical consultations at the individual level, while aggregated counts showed a weak correlation with population level COVID-19 incidence [61]

Other indications

[edit]

Reflux related mechanisms have been studied in chronic cough trials using 24 hour objective cough frequency as a primary endpoint; for example, one study evaluated lesogaberan in RCC while assessing gastroesophageal reflux at screening.[62]. Objective cough monitoring has also been used as a dependent outcome in patients with pulmonary sarcoidosis [63]. Of 77 studies assessing cough counting identified in one scoping review, 15 included a pediatric sample with respiratory disorders [1]

Digital health and consumer applications

[edit]

Consumer cough tracking

[edit]

Smartphone based cough apps have been used to monitor cough frequency at home.[5]. These apps use artificial intelligence enabled algorithms to recognize cough events from ambient sound [18]. Beyond smartphones, consumer hearables such as earbuds with active noise cancellation microphones have been proposed for on device cough detection [64]. Consumer cough tracking apps are generally not validated to clinical grade accuracy and are not regulated as medical devices in the United States [5][50]

Remote patient monitoring

[edit]

Continuous cough monitoring has been used to study the biology and therapy of diseases associated with cough, including RCC and UCC, COPD, bronchiectasis, congestive heart failure, and gastroesophageal reflux.[5]. Continuous monitoring has also been used to analyze aspects of cough occurring in day to day outpatient settings, such as variability in daily cough frequency, diurnal and episodic patterns, and the relationship between subjective and objective cough measurements [5]

Digital therapeutics

[edit]

A meta-analysis of 12 studies found behavioral cough suppression therapy (BCST) to be effective in improving RCC and UCC symptoms.[65]. Objective cough frequency was reported by 3 of the 12 included studies, with results showing a reduction in cough frequency among patients with chronic cough compared to those receiving verbal education or before receiving BCST (p < 0.001) [65]

Population level surveillance

[edit]

One observational surveillance study used an environmental sensor platform to collect audio, thermal imaging, and radar data and found that cough counts correlated with positive COVID-19 cases.[66]. The study proposed using aggregated cough counts and other data to predict COVID-19 related hospital visit metrics in a hospital waiting room [66]

Regulatory landscape

[edit]

United States

[edit]

The regulatory classification of cough monitoring technologies depends on whether the software performs an independent medical function or serves only as a data acquisition tool. Software that autonomously classifies audio and outputs cough counts may meet the definition of software as a medical device (SaMD) as described by the International Medical Device Regulators Forum.[67]. Because no cough counting device has been previously marketed, there is no predicate to support a 510(k) submission (the 510(k) route clears a device by showing it is substantially equivalent to an existing legally marketed device, called a predicate). The relevant pathway is the FDA's De Novo authorization pathway, which allows devices without a predicate to be classified as Class I or Class II [67]

As of 2026, no cough monitoring system has received FDA clearance, approval, or De Novo authorization for cough counting.[5][68]. The VitaloJAK (Vitalograph Ltd) and Strados RESP biosensor (Strados Labs) are cleared only to acquire and store sound, not to count coughs [5]. In the November 2023 briefing document for gefapixant (NDA 215010), the FDA noted that validation data were insufficient to confirm that the VitaloJAK accurately measures cough frequency [68]

The FDA's Digital Health Center of Excellence issued final guidance in December 2023. The guidance recommends that digital health technologies used in clinical investigations be shown to be fit for purpose, with validation scaled to the intended use.[69]

Europe

[edit]

The European Union Medical Device Regulation (EU 2017/745) classifies standalone medical device software under Rule 11 of Annex VIII.[70] (the classification rule that assigns a risk class to software intended to provide information used for medical decisions). Software that provides diagnostic or therapeutic information falls under Class IIa at minimum. Commercial placement requires conformity assessment by a Notified Body (an independent organization designated to assess whether a device meets EU regulatory requirements) and CE marking [70]

The European Medicines Agency accepted objective 24 hour cough frequency as the primary efficacy endpoint in the gefapixant (Lyfnua) marketing authorization application; EU authorization was granted on 15 September 2023.[71]. In contrast, the FDA's Pulmonary Allergy Drugs Advisory Committee reviewed the same clinical data and voted 12 to 1 in November 2023 that the evidence did not demonstrate clinically meaningful benefit [68]

Japan

[edit]

Gefapixant received approval from Japan's Pharmaceuticals and Medical Devices Agency for refractory and unexplained chronic cough, supported by a Japan specific Phase 3 trial and the global COUGH-1 and COUGH-2 trials.[72]. A 2025 review of regulatory considerations for digital endpoints in clinical trials identified PMDA among the major health authorities that sponsors should consult when implementing digital health technology derived endpoints in global development programs [73]

Fit for purpose acceptance

[edit]

Use of a digital health technology as a clinical trial endpoint does not require the device to hold marketing authorization as a medical device.[69]. Validation evidence has been submitted within drug applications, as occurred in the gefapixant submissions [68][71], and the FDA's Drug Development Tool Qualification Program provides a pathway through which qualified tools can be used by any sponsor without revalidation [69]. The Digital Medicine Society's V3 framework (2020) evaluates biometric monitoring technologies across three stages: verification, analytical validation, and clinical validation [74]. A 2025 update (V3+) added usability validation as a fourth stage [6]

Privacy and data governance

[edit]

Audio recording and privacy risk

[edit]

Ambulatory cough monitors that continuously record environmental audio capture content beyond cough sounds. An analysis of 120 hours of VitaloJAK recordings identified 21.6 hours of intelligible speech alongside 3,376 cough events.[75]. Under HIPAA, audio recordings that contain individually identifiable health information are classified as electronic protected health information when created or maintained by covered entities [76]. Under the General Data Protection Regulation (GDPR), health data are a special category of personal data requiring explicit consent and enhanced protections under Article 9 [77]. The data minimization principle (Article 5(1)(c)) raises questions about whether continuous audio recording that captures large volumes of noncough data is justified by the monitoring purpose [77]. Informed consent in clinical trials using audio monitoring have been required to address the continuous nature of the recording and the incidental capture of third party speech [76][77]. Consent language must also specify data retention periods and the right to erasure [76][77]. Because ambulatory recordings contain both health data and identifiable third party conversations, a breach would expose not only participant information but also the speech of unconsenting individuals [75][77]

On device processing as a privacy architecture

[edit]

An alternative design approach processes audio locally on the wearable device using embedded machine learning classifiers. Under this approach, only metadata such as timestamps and cough counts are transmitted; raw audio is discarded after classification.[78]. This reduces the volume of protected data leaving the device. It also aligns with GDPR data minimization and privacy by design requirements (Article 25) [77][78]

Clinical trial data standards

[edit]

Digital health technologies generating data for regulatory submissions must comply with electronic records requirements. In the United States, 21 CFR Part 11 governs electronic records and electronic signatures in regulated activities. Its requirements cover system validation, audit trails, and access controls.[79]. The regulation's definition of electronic records explicitly covers audio data [79]. The European equivalent, EudraLex Volume 4 Annex 11, sets parallel requirements for computerized systems [80]. The Clinical Data Interchange Standards Consortium (CDISC) maintains submission data standards required by the FDA and PMDA, but no cough specific standard currently exists [81]

Limitations and open questions

[edit]

Lack of standardization. No universal standard exists for how cough monitoring studies should be conducted: not for monitoring duration, not for metric definitions or minimum wear time, and not for endpoint reporting.[5][82]. A 2022 review documented the absence of standardized approaches for cough counting, including unresolved debate over whether individual cough events or cough epochs are the appropriate unit of measurement [82]. A 2025 American Thoracic Society research statement confirmed that no professional respiratory society has published guidance to standardize home based respiratory monitoring [83]

Regulatory clearance gap. No cough monitoring system has received FDA clearance specifically for cough counting as of 2026, and the FDA has questioned the validation evidence supporting existing cough counting systems [5][68]. This creates uncertainty about the evidentiary threshold regulators will require for primary endpoint use in drug approval decisions.

Nocturnal monitoring gaps. Wearable cough monitors require periodic charging, creating data gaps during overnight use.[5]. Battery life on some platforms is limited to 12 to 18 hours, which may not cover a full 24 hour period [5]

Pediatric validation deficit. Validated cough monitoring systems for children are limited. A 2022 study built a pediatric cough detection algorithm that achieved sensitivity of 47.6%, far below the 91 to 99% reported in adult systems, a gap attributed to greater acoustic variability across age groups.[37]. A scoping review of 108 cough sound studies found that only a minority focused on pediatric populations [84]

Environmental noise robustness. Algorithm performance in cough detection degrades in high noise environments. Noise reduction methods were specifically evaluated in only 8.3% of studies in a 2024 scoping review, and validation conducted in controlled acoustic settings may overestimate performance in free living conditions.[84]

Optimal metric uncertainty. It is not established which cough metric is most clinically meaningful, whether total 24 hour count, waking hourly rate, bout frequency, or a composite measure.[82][20]. Different metrics may suit different indications, and no minimal clinically important difference has been established for any objective cough frequency measure [82][20]

PRO objective dissociation. Objective cough frequency reduction does not reliably predict patient reported improvement. In the gefapixant Phase 3 trials, statistically significant reductions in 24 hour cough frequency did not translate to advisory committee consensus on clinically meaningful benefit; the FDA's Pulmonary Allergy Drugs Advisory Committee voted 12 to 1 that the data were insufficient.[68]. A 2025 scoping review found only moderate correlations between objective cough counts and patient reported outcomes [1]. Median Spearman correlations were 0.42 for visual analogue scale scores and −0.49 for the Leicester Cough Questionnaire [1]

Acoustic phenotyping. Whether acoustic features of individual coughs carry diagnostic information beyond frequency is unresolved. Among 38 studies examining cough sound diagnostics in a 2024 scoping review, only 21.1% compared results against a diagnostic gold standard.[84]. Acoustic cough biomarkers have not been translated into validated screening or diagnostic tools [84]

See also

[edit]

References

[edit]
  1. ^ a b c d e f g h i j k l m n Zimmer, AJ; Das, R; Lopez, PE; et al. (2025). "Objective cough counting in clinical practice and public health: a scoping review". The Lancet Digital Health. 7 (11) 100908. doi:10.1016/j.landig.2025.100908. PMID 41274833.
  2. ^ a b c Hu, X; Zhang, K; Liu, T; et al. (2025). "Chronic cough: A review and prospects". Medicine. 104 (41) e45162. doi:10.1097/MD.0000000000045162. PMC 12517813. PMID 41088639.
  3. ^ a b Ijaz, A; et al. (2022). "Towards using cough for respiratory disease diagnosis by leveraging Artificial Intelligence: A survey". Informatics in Medicine Unlocked. 29 (5) 100832. doi:10.1016/j.imu.2021.100832.
  4. ^ a b c d e f g h i j k l m n o p q r s t u v w x y Hall, JI; Lozano, M; Estrada-Petrocelli, L; et al. (2020). "The present and future of cough counting tools". Journal of Thoracic Disease. 12 (9): 5207–5223. doi:10.21037/jtd-2020-icc-003. PMC 7578475. PMID 33145097.
  5. ^ a b c d e f g h i j k l m n o p q Galvosas, M; Small, PM (2025). "The value of continuous cough monitoring: a narrative review". Journal of Thoracic Disease. 17 (11): 10571–10583. doi:10.21037/jtd-2025-876. PMC 12688508. PMID 41376907.
  6. ^ a b c d Bakker, JP; Barge, R; Centra, J; et al. (2025). "V3+ extends the V3 framework to ensure user-centricity and scalability of sensor-based digital health technologies". npj Digital Medicine. 8 (1) 51. doi:10.1038/s41746-024-01322-2. PMC 11760348. PMID 39856145.
  7. ^ a b c d Hirons, B; Rhatigan, K; Kesavan, H; et al. (2023). "Cough in chronic lung disease: a state of the art review". Journal of Thoracic Disease. 15 (10): 5823–5844. doi:10.21037/jtd-22-1776. PMC 10636467. PMID 37969279.
  8. ^ Wu, J; Ma, Y; Chen, Y (2022). "GERD-related chronic cough: possible mechanism, diagnosis and treatment". Frontiers in Physiology. 13 1005404. doi:10.3389/fphys.2022.1005404. PMC 9630749. PMID 36338479.
  9. ^ a b c d e f g h i j Schelfhout, J; Nguyen, AM; Birring, SS; et al. (2022). "Validation and meaningful change thresholds for an objective cough frequency measurement in chronic cough". Lung. 200 (6): 717–724. doi:10.1007/s00408-022-00587-2. PMC 9675653. PMID 36348054.
  10. ^ Ghrabli, S; Elgendi, M; Menon, C (2024). "Identifying unique spectral fingerprints in cough sounds for diagnosing respiratory ailments". Scientific Reports. 14 (1): 593. Bibcode:2024NatSR..14..593G. doi:10.1038/s41598-023-50371-2. PMC 10770161. PMID 38182601.
  11. ^ Decalmer, SC; Webster, D; Kelsall, AA; et al. (2006). "Chronic cough: how do cough reflex sensitivity and subjective assessments correlate with objective cough counts during ambulatory monitoring?". Thorax. 62 (4): 329–334. doi:10.1136/thx.2006.067413. PMC 2092471. PMID 17101736.
  12. ^ Turner, RD; Birring, SS (2023). "Measuring cough: what really matters?". Journal of Thoracic Disease. 15 (4): 2288–2299. doi:10.21037/jtd-23-230. PMC 10183488. PMID 37197542.
  13. ^ Holt, K; Dockry, R; Mitchell, J; et al. (2023). "Objective measurement of cough". Barcelona Respiratory Network Reviews. 9 (4): 102–120. doi:10.23866/BRNRev:2023-M0099.
  14. ^ Yang, X; Hu, M; Huang, K (2025). "Demographics, risk factors and prevalence of chronic cough in Asian general adult population: a narrative review". Journal of Thoracic Disease. 17 (5): 3419–3432. doi:10.21037/jtd-24-248. PMC 12170037. PMID 40529737.
  15. ^ a b Das, S; Goldin, J; Alhajjaj, MS (2026). "Cough: evaluation and management". Nelson Pediatric Symptom-Based Diagnosis: Common Diseases and their Mimics. pp. 27–53. doi:10.1016/B978-0-323-76174-1.00003-1.
  16. ^ Sonoda, K; Nayak, R (2024). "Chronic cough: evaluation and management". American Family Physician. 110 (2): 167–173. PMID 39172674.
  17. ^ Bali, V; et al. (2024). "Patient-reported experiences with refractory or unexplained chronic cough: a qualitative analysis". Therapeutic Advances in Respiratory Disease. 18 17534666241236025: 17534666241236024. doi:10.1177/17534666241236025. PMC 10953008. PMID 38501735.
  18. ^ a b c d Chung, KF; Chaccour, C; Jover, L; et al. (2024). "Longitudinal cough frequency monitoring in persistent coughers: daily variability and predictability". Lung. 202 (5): 561–568. doi:10.1007/s00408-024-00734-x. PMC 11427503. PMID 39085518.
  19. ^ Smith, JA; et al. (2025). "Camlipixant in refractory chronic cough: a phase 2b, randomized, placebo-controlled trial (SOOTHE)". American Journal of Respiratory and Critical Care Medicine. 211 (6): 1038–1048. doi:10.1164/rccm.202409-1752OC. PMC 12180140. PMID 40043302.
  20. ^ a b c Spinou, A; Birring, SS (2014). "An update on measurement and monitoring of cough: what are the important study endpoints?". Journal of Thoracic Disease. 6 (Suppl 7): S728–S734. doi:10.3978/j.issn.2072-1439.2014.10.08. PMC 4222923. PMID 25383207.
  21. ^ a b c Morice, AH; Faruqi, S; Wright, CE; et al. (2011). "Cough hypersensitivity syndrome: a distinct clinical entity". Lung. 189 (1): 73–79. doi:10.1007/s00408-010-9272-1. PMID 21240613.
  22. ^ Shembel, AC; Rosen, CA; Zullo, TG; et al. (2013). "Development and validation of the cough severity index: a severity index for chronic cough related to the upper airway". Laryngoscope. 123 (8): 1931–1936. doi:10.1002/lary.23916. PMID 23737389.
  23. ^ Wang, Z; Wang, M; Wen, S; et al. (2019). "Types and applications of cough-related questionnaires". Journal of Thoracic Disease. 11 (10): 4379–4388. doi:10.21037/jtd.2019.09.62. PMC 6837954. PMID 31737324.
  24. ^ a b c d e f g h i j k l m n o Serrurier, A; Neuschaefer-Rube, C; Rohrig, R (2022). "Past and trends in cough sound acquisition, automatic detection and automatic classification: a comparative review". Sensors. 22 (8): 2896. Bibcode:2022Senso..22.2896S. doi:10.3390/s22082896. PMC 9027375. PMID 35458885.
  25. ^ Kvapilova, L; et al. (2019). "Continuous sound collection using smartphones and machine learning to measure cough". Digital Biomarkers. 3 (3): 166–175. doi:10.1159/000504666. PMC 7011715. PMID 32095775.
  26. ^ a b Matos, S; Birring, SS; Pavord, ID; et al. (2007). "An automated system for 24-h monitoring of cough frequency: the Leicester cough monitor". IEEE Transactions on Biomedical Engineering. 54 (8): 1472–1479. Bibcode:2007ITBE...54.1472M. doi:10.1109/TBME.2007.900811. PMID 17694868.
  27. ^ a b c d e Birring, SS; Fleming, T; Matos, S; et al. (2008). "The Leicester Cough Monitor: preliminary validation of an automated cough detection system in chronic cough". European Respiratory Journal. 31 (5): 1013–1018. doi:10.1183/09031936.00057407. PMID 18184683.
  28. ^ a b c d e Smith, JA; Holt, K; Dockry, R; et al. (2021). "Performance of a digital signal processing algorithm for the accurate quantification of cough frequency". European Respiratory Journal. 58 (2): 2004271. doi:10.1183/13993003.04271-2020. PMID 33875494.
  29. ^ a b c d e f g h i j k l m n Chaccour, C; Sanchez-Olivieri, I; Siegel, S; et al. (2025). "Validation and accuracy of the Hyfe cough monitoring system: a multicenter clinical study". Scientific Reports. 15 (1) 880. Bibcode:2025NatSR..15..880C. doi:10.1038/s41598-025-85341-3. PMC 11704278. PMID 39762316.
  30. ^ a b c d e f g h i j k Do, W; Russell, R; Wheeler, C; et al. (2022). "Performance of cough monitoring by Albus Home, a contactless and automated system for nocturnal respiratory monitoring at home". ERJ Open Research. 8 (4): 00265–2022. doi:10.1183/23120541.00265-2022. PMC 9720547. PMID 36478916.
  31. ^ a b c d e f g Kuhn, M; Nalbant, E; Kohlbrenner, D; et al. (2023). "Validation of a small cough detector". ERJ Open Research. 9 (1): 00279–2022. doi:10.1183/23120541.00279-2022. PMC 9868968. PMID 36699651.
  32. ^ a b c d Barata, F; Cleres, D; Tinschert, P; et al. (2023). "Nighttime continuous contactless smartphone-based cough monitoring for the ward: validation study". JMIR Formative Research. 7 e38439. doi:10.2196/38439. PMC 9989914. PMID 36655551.
  33. ^ a b c Otoshi, T; Nagano, T; Izumi, S; et al. (2021). "A novel automatic cough frequency monitoring system combining a triaxial accelerometer and a stretchable strain sensor". Scientific Reports. 11 (1) 9973. Bibcode:2021NatSR..11.9973O. doi:10.1038/s41598-021-89457-0. PMC 8113562. PMID 33976286.
  34. ^ a b c Tzavelis, A; Palla, J; Mathur, R; et al. (2024). "Development of a miniaturized mechanoacoustic sensor for continuous, objective cough detection, characterization and physiologic monitoring in children with cystic fibrosis". IEEE Journal of Biomedical and Health Informatics. 28 (10): 5941–5952. Bibcode:2024IJBHI..28.5941T. doi:10.1109/JBHI.2024.3415479. PMC 11653413. PMID 38885105.
  35. ^ a b c d e Smith, JA; Earis, JE; Woodcock, AA (2006). "Establishing a gold standard for manual cough counting: video versus digital audio recordings". Cough. 2: 6. doi:10.1186/1745-9974-2-6. PMC 1557531. PMID 16887019.
  36. ^ a b c d e f g h Rodrigues, ALC; Passafaro, LSR; dos Santos, DO; et al. (2026). "Diagnostic accuracy of cough frequency monitors: a systematic review and meta-analysis". Respirology. 31 (5): 459–470. doi:10.1002/resp.70229. PMC 13125381. PMID 41797231.
  37. ^ a b c d Kruizinga, MD; et al. (2022). "Development and technical validation of a smartphone-based pediatric cough detection algorithm". Pediatric Pulmonology. 57 (3): 761–767. doi:10.1002/ppul.25801. PMC 9306830. PMID 34964557.
  38. ^ a b Birring, SS; Prudon, B; Carr, AJ; et al. (2003). "Development of a symptom specific health status measure for patients with chronic cough: Leicester Cough Questionnaire (LCQ)". Thorax. 58 (4): 339–343. doi:10.1136/thorax.58.4.339. PMC 1746649. PMID 12668799.
  39. ^ a b Raj, AA; Pavord, DI; Birring, SS (2009). "Clinical cough IV: what is the minimal important difference for the Leicester Cough Questionnaire?". In Chung, KF; Widdicombe, J (eds.). Pharmacology and Therapeutics of Cough. Handbook of Experimental Pharmacology. Vol. 187. Berlin: Springer. pp. 311–320. doi:10.1007/978-3-540-79842-2_16. ISBN 978-3-540-79841-5. PMID 18825348.
  40. ^ a b Hsu, JY; Stone, RA; Logan-Sinclair, RB; et al. (1994). "Coughing frequency in patients with persistent cough: assessment using a 24 hour ambulatory recorder". European Respiratory Journal. 7 (7): 1246–1253. doi:10.1183/09031936.94.07071246. PMID 7925902.
  41. ^ a b French, CT; Irwin, RS; Fletcher, KE; et al. (2002). "Evaluation of a cough-specific quality-of-life questionnaire". Chest. 121 (4): 1123–1131. doi:10.1378/chest.121.4.1123. PMID 11948042.
  42. ^ a b Nguyen, AM; Bacci, ED; Vernon, M; et al. (2021). "Validation of a visual analog scale for assessing cough severity in patients with chronic cough". Therapeutic Advances in Respiratory Disease. 15 17534666211049743. doi:10.1177/17534666211049743. PMC 8552382. PMID 34697975.
  43. ^ a b Kum, E; Guyatt, GH; Devji, T; et al. (2021). "Cough symptom severity in patients with refractory or unexplained chronic cough: a systematic survey and conceptual framework". European Respiratory Review. 30 (162): 210104. doi:10.1183/16000617.0104-2021. PMC 9518000. PMID 34261745.
  44. ^ Birring, SS (2015). "How best to measure cough clinically". Current Opinion in Pharmacology. 22: 37–40. doi:10.1016/j.coph.2015.03.003. PMID 25819594.
  45. ^ a b c d McGarvey, LP; Birring, SS; Morice, AH; et al. (2022). "Efficacy and safety of gefapixant, a P2X3 receptor antagonist, in refractory chronic cough and unexplained chronic cough (COUGH-1 and COUGH-2): results from two double-blind, randomised, parallel-group, placebo-controlled, phase 3 trials". The Lancet. 399 (10328): 909–923. doi:10.1016/S0140-6736(21)02348-5. PMID 35248186.
  46. ^ Birring, SS; Wijsenbeek, MS; Agrawal, S; et al. (2017). "A novel formulation of inhaled sodium cromoglicate (PA101) in idiopathic pulmonary fibrosis and chronic cough: a randomised, double-blind, proof-of-concept, phase 2 trial". The Lancet Respiratory Medicine. 5 (10): 806–815. doi:10.1016/S2213-2600(17)30310-7. PMID 28923239.
  47. ^ Wickstrom, G; Bendix, T (2000). "The 'Hawthorne effect' — what did the original Hawthorne studies actually show?". Scandinavian Journal of Work, Environment and Health. 26 (4): 363–367. doi:10.5271/sjweh.555.
  48. ^ a b c Zhang, M; Zhang, B; Morice, AH (2024). "Decoding the impact of the placebo response in clinical trials for chronic cough". ERJ Open Research. 10 (5): 00335–2024. doi:10.1183/23120541.00335-2024. PMC 11514000. PMID 39469270.
  49. ^ Gabaldon-Figueira, JC; Keen, E; Rudd, M; et al. (2022). "Longitudinal passive cough monitoring and its implications for detecting changes in clinical status". ERJ Open Research. 8 (2): 00001–2022. doi:10.1183/23120541.00001-2022. PMC 9108969. PMID 35586452.
  50. ^ a b c Witjaksono, LE; Schulte, M; Holland, AE; et al. (2025). "Cough monitoring systems in adults with chronic respiratory diseases: a systematic review". European Respiratory Review. 34 (175): 230212. doi:10.1183/16000617.0212-2023. PMC 11880901. PMID 40044185.
  51. ^ a b c Morice, AH; Millqvist, E; Bieksiene, K; et al. (2020). "ERS guidelines on the diagnosis and treatment of chronic cough in adults and children". European Respiratory Journal. 55 (1): 1901136. doi:10.1183/13993003.01136-2019. PMC 6942543. PMID 31515408.
  52. ^ Miravitlles, M (2011). "Cough and sputum production as risk factors for poor outcomes in patients with COPD". Respiratory Medicine. 105 (8): 1118–1128. doi:10.1016/j.rmed.2011.02.003. PMID 21353517.
  53. ^ Crooks, MG; den Brinker, AC; Thackray-Nocera, S; et al. (2021). "Domiciliary cough monitoring for the prediction of COPD exacerbations". Lung. 199 (2): 131–137. doi:10.1007/s00408-021-00435-9. PMC 8053154. PMID 33829322.
  54. ^ Wu, Z; Spencer, LG; Banya, W; et al. (2024). "Morphine for treatment of cough in idiopathic pulmonary fibrosis (PACIFY COUGH): a prospective, multicentre, randomised, double-blind, placebo-controlled, two-way crossover trial". The Lancet Respiratory Medicine. 12 (4): 273–280. doi:10.1016/S2213-2600(23)00432-0. PMID 38237620.
  55. ^ a b Marsden, PA; Satia, I; Ibrahim, B; et al. (2016). "Objective cough frequency, airway inflammation, and disease control in asthma". Chest. 149 (6): 1460–1466. doi:10.1016/j.chest.2016.02.676. PMID 26973014.
  56. ^ Chalmers, JD; Haworth, CS; Flume, P; et al. (2025). "European Respiratory Society clinical practice guideline for the management of adult bronchiectasis". European Respiratory Journal. 66 (6). doi:10.1183/13993003.01126-2025. PMID 41016738.
  57. ^ Spinou, A; Lee, KK; Sinha, A; et al. (2017). "The objective assessment of cough frequency in bronchiectasis". Lung. 195 (5): 575–585. doi:10.1007/s00408-017-0038-x. PMC 5599483. PMID 28707107.
  58. ^ World Health Organization (2021). WHO consolidated guidelines on tuberculosis. Module 2: Screening: systematic screening for tuberculosis disease. Geneva: WHO. ISBN 978-92-4-002267-6.
  59. ^ Huddart, S; Asege, L; Jaganath, D; et al. (2023). "Continuous cough monitoring: a novel digital biomarker for TB diagnosis and treatment response monitoring". International Journal of Tuberculosis and Lung Disease. 27 (3): 221–222. doi:10.5588/ijtld.22.0511. PMC 9983626. PMID 36855045.
  60. ^ a b Raberahona, M; Zimmer, A; Rakotoarivelo, RA; et al. (2025). "Continuous digital cough monitoring during 6-month pulmonary tuberculosis treatment". ERJ Open Research. 11 (2): 00655–2024. doi:10.1183/23120541.00655-2024. PMC 11931562. PMID 40129545.
  61. ^ a b Gabaldon-Figueira, JC; Keen, E; Gimenez, G; et al. (2022). "Acoustic surveillance of cough for detecting respiratory disease using artificial intelligence". ERJ Open Research. 8 (2): 00053–2022. doi:10.1183/23120541.00053-2022. PMC 9149391. PMID 35651361.
  62. ^ Badri, H; Gibbard, C; Denton, D; et al. (2022). "A double-blind randomised placebo-controlled trial investigating the effects of lesogaberan on the objective cough frequency and capsaicin-evoked coughs in patients with refractory chronic cough". ERJ Open Research. 8 (1). doi:10.1183/23120541.00546-2021. PMC 8918934. PMID 35295236.
  63. ^ Sinha, A; Lee, KK; Rafferty, GF; et al. (2016). "Predictors of objective cough frequency in pulmonary sarcoidosis". European Respiratory Journal. 47 (5): 1461–1471. doi:10.1183/13993003.01369-2015. PMID 26846840.
  64. ^ Wang, Y; Zhang, X; Chakalasiya, JM; et al. (2022). "HearCough: enabling continuous cough event detection on edge computing hearables". Methods. 205: 53–62. doi:10.1016/j.ymeth.2022.05.002. PMID 35569734.
  65. ^ a b Yi, B; Wang, S; Xu, X; et al. (2024). "Efficacy of behavioral cough suppression therapy for refractory chronic cough or unexplained chronic cough: a meta-analysis of randomized controlled trials". Therapeutic Advances in Respiratory Disease. 18 17534666241305952. doi:10.1177/17534666241305952. PMC 11663276. PMID 39707881.
  66. ^ a b Al Hossain, F; Tonmoy, MTH; Nuvvula, S; et al. (2024). "Syndromic surveillance of population-level COVID-19 burden with cough monitoring in a hospital emergency waiting room". Frontiers in Public Health. 12 1279392. Bibcode:2024FrPH...1279392A. doi:10.3389/fpubh.2024.1279392. PMC 11007176. PMID 38605877.
  67. ^ a b U.S. Food and Drug Administration (23 July 2025). "Software as a Medical Device (SaMD)". Digital Health Center of Excellence. Retrieved March 2026. {{cite web}}: Check date values in: |access-date= (help)
  68. ^ a b c d e f U.S. Food and Drug Administration (November 17, 2023). "FDA Briefing Document: NDA 215010, Gefapixant". Pulmonary-Allergy Drugs Advisory Committee.
  69. ^ a b c U.S. Food and Drug Administration (December 2023). "Digital Health Technologies for Remote Data Acquisition in Clinical Investigations: Guidance for Industry, Investigators, and Other Stakeholders". Food and Drug Administration (Final guidance).
  70. ^ a b "Regulation (EU) 2017/745 of the European Parliament and of the Council of 5 April 2017 on medical devices". Official Journal of the European Union. L 117, 5.5.2017, Annex VIII, Rule 11; Articles 20, 52, 62.
  71. ^ a b Committee for Medicinal Products for Human Use (CHMP) (20 July 2023). "Assessment report: Lyfnua (gefapixant)" (PDF). European Medicines Agency. EMA/357321/2023 Corr.1.
  72. ^ Niimi, A; Sagara, H; Kikuchi, M; et al. (2022). "A phase 3, randomized, double-blind, clinical study to evaluate the long-term safety and efficacy of gefapixant in Japanese adult participants with refractory or unexplained chronic cough". Allergology International. 71 (4): 498–504. doi:10.1016/j.alit.2022.05.006. PMID 35752582.
  73. ^ Erdemli, G; Murphy, T; Walinsky, S (2025). "Regulatory considerations for successful implementation of digital endpoints in clinical trials for drug development". npj Digital Medicine. 8 (1) 142. doi:10.1038/s41746-025-01513-5. PMC 11882986. PMID 40044879.
  74. ^ Goldsack, JC; Coravos, A; Bakker, JP; et al. (2020). "Verification, analytical validation, and clinical validation (V3): the foundation of determining fit-for-purpose for Biometric Monitoring Technologies (BioMeTs)". npj Digital Medicine. 3 55. doi:10.1038/s41746-020-0260-4. PMC 7156507. PMID 32337371.
  75. ^ a b Barry, SJE; et al. (2022). "A speech obfuscation system to preserve data privacy in 24-hour ambulatory cough monitoring". IEEE Transactions on Biomedical Engineering. doi:10.1109/TBME.2021.3137237 (inactive 16 July 2026).{{cite journal}}: CS1 maint: DOI inactive as of July 2026 (link)
  76. ^ a b c U.S. Department of Health and Human Services (June 13, 2022). "Guidance on How the HIPAA Rules Permit Covered Health Care Providers and Health Plans to Use Remote Communication Technologies for Audio-Only Telehealth". Office for Civil Rights.
  77. ^ a b c d e f "Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 (General Data Protection Regulation)". Official Journal of the European Union. L 119, 4.5.2016, Articles 5, 9, 25.
  78. ^ a b Albini, S; Orlandic, L; Dan, J; et al. (2025). "Cough-E: A multimodal, privacy-preserving cough detection algorithm for the edge". IEEE Journal of Biomedical and Health Informatics. PP: 1–15. doi:10.1109/JBHI.2025.3577507. hdl:20.500.11824/2009. PMID 40478704.
  79. ^ a b "21 CFR Part 11, Electronic Records; Electronic Signatures". U.S. Code of Federal Regulations. §11.3(b)(6).
  80. ^ European Commission (June 30, 2011). "EudraLex Volume 4, Annex 11: Computerised Systems" (PDF) (Revision 1).
  81. ^ Clinical Data Interchange Standards Consortium (10 June 2024). "Study Data Tabulation Model (SDTM)".
  82. ^ a b c d Zhang, M; Sykes, DL; Brindle, K; et al. (2022). "Chronic cough — the limitation and advances in assessment techniques". Journal of Thoracic Disease. 14 (12): 5097–5119. doi:10.21037/jtd-22-874. PMC 9840016. PMID 36647459.
  83. ^ Khor, YH; Poberezhets, V; Buhr, RG; et al. (2025). "An Official American Thoracic Society Research Statement: Home-Based Monitoring for Chronic Lung Disease". American Journal of Respiratory and Critical Care Medicine. 211 (2): 174–193. doi:10.1164/rccm.202410-2080ST. PMC 11812536. PMID 39585746.
  84. ^ a b c d Hegde, S; et al. (2024). "Cough sounds in screening and diagnostics: a scoping review". Laryngoscope. 134 (3): 1023–1031. doi:10.1002/lary.31042. PMC 10915103. PMID 37672667.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Draft:Objective_cough_monitoring&oldid=1365927950"