Acoustic Epidemiology and Syndromic Surveillance — How they can apply in Respiratory Health
Acoustics is a science that studies sound waves and their effects on people, animals, and buildings. Epidemiology aims to link the causes and risk factors of health-related events. These elements frequently relate to diseases.
Acoustic epidemiology is the synergistic intersection of these two fields to use the knowledge about sound waves to gain insight into causes and patterns of disease in human populations.
Following in the footsteps of epidemiological efforts and goals, acoustic epidemiology is concerned with using body sound data to improve disease surveillance capabilities for COVID-19 and any other applicable diseases of the future.
What is Acoustic Epidemiology?
Acoustic epidemiology is a field that studies bodily sounds, such as coughs and breath sounds, to better identify determinants and distribution of disease. It consists of analyzing sounds either by themselves or in conjunction with other data points as a clinical, objective epidemiological tool.
Current technology can already record sounds using more powerful computing hardware and more accurate sensors than ever before. These sounds can then be analyzed independently or in conjunction with additional information.
In the long run, there will be many more devices that catch and interpret more sounds, comparing them to one another or preexisting databases. These could include, for example, coughing, sneezing, snoring, breathing, other sounds, and data collected via basic sensors.
This advancement indicates that more data will be available for analysis and a high-level understanding of health and its dynamics, resulting in machine learning and artificial intelligence having a greater chance of becoming more sensitive and specific.
AI diagnostics and their applications
Artificial intelligence (AI) is making progress in diagnostics. As a result, diagnostics are becoming less expensive, faster, and more accurate. In addition, it provides many medical facilities with access to high-quality diagnostics that they might not otherwise have.
The use of AI in the diagnostic process to assist medical practitioners might be highly beneficial to the healthcare industry and the overall health of individuals.
Artificial intelligence integration into current technology infrastructure accelerates finding critical medical data from various sources customized to the patient’s needs and treatment procedure.
Current practical applications, such as healthcare and illness diagnosis, are focused on a single task and are being developed with machine learning.
AI in robotic surgery can track surgeons’ motions and patterns in the most successful cases and use that data to help with robot control in subsequent procedures. In addition, robot-assisted surgery has dramatically reduced surgical complications, minimized discomfort, and achieved shorter recovery periods.
This added efficiency allows performing more sophisticated surgeries with fewer adverse effects, blood loss, and pain. Post-surgery recuperation is also faster and smoother.
Other few examples of AI in action include:
- Management of clinical health data
- Drug development
- Disease diagnosis
Advantages of AI diagnostics
AI-powered diagnostics are a new trend that has been introduced to the medical field and can be an excellent help for healthcare professionals. AI helps doctors predict the course of a disease and suggest possible treatments. It can also help doctors with complex cases that they have not seen before.
Doctors will not be replaced by AI diagnostics anytime soon, but they will become more efficient. Doctors will have more time with their patients and less time on paperwork, which is one of the most tedious parts of their job.
The advantages of AI diagnostics are:
- Personalized treatment
- A better understanding of patient’s health
- Clinical Decision Making
- Improved healthcare accessibility
- Rapid diagnosis/Early diagnosis
- Provide real-time data
For example, consistent findings can improve performance and cost-efficiency in a healthcare context. In addition, patients will benefit from this automation since practitioners would be more available. In other words, doctors will see more patients and diagnose them sooner. Above all, this can reduce severe cases in disease progression while also improving treatment.
Compared to manual diagnosis, analyzing imaging results and biochemical and physiological reports becomes more accessible and accurate, with the help of sophisticated AI models. Once it is created, a machine learning algorithm can run and quickly display useful diagnostic interpretations of the data.
Having the ability to detect sounds in acoustic epidemiology and syndromic surveillance could assist in uncovering cough variations and anomalies faster than traditional reporting.
AI-assisted tasks that eliminate the requirement for a medical specialist can benefit both patients and healthcare providers.
Clinical Relevance
Syndromic surveillance of symptoms such as fever and cough in emergency rooms is a well-established way to understand the distribution of diseases. Tracking cough frequency in a manner that preserves privacy across time and space is a novel approach that can provide valuable, real-time data on a population’s health.
“Acoustic surveillance” is the term we use to describe this. Because epidemiology is a population-based field of research, acoustic disease surveillance data are significant on a large scale and have far-reaching ramifications for society.
Measurements and Deviations
The ability to analyze respiratory sounds and identify variations from baseline is a powerful epidemiologic tool, and a sound surveillance system must also be quick and comprehensive. In addition, it should ideally track symptoms as they appear rather than when treatment is required because it inevitably comes later and only applies to a small minority of the population.
In acoustic epidemiology, sound-based surveillance could be a low-cost and scalable alternative. For example, we can create a baseline profile of all coughing sounds that occur in a specific place, such as an office or school campus.
Daily coughing and sneezing levels can serve as useful criteria. In addition, we can use this benchmark to detect any variation from the standard baseline and use it as an early warning system to take action sooner.
Tracking cough frequency over area and time can produce valuable, real-time, actionable data on a population’s health.
Syndromic Surveillance
Syndromic surveillance collects health-related data as soon as illness appears to offer an overall population-based awareness of disease spread in an area. For example, sick people may exhibit behavioral patterns, symptoms, indicators, or laboratory findings that we can follow through various data sources before the laboratory confirmation of an infectious disease.
Symptom and preliminary diagnostic information and rapid data collection methods are used in syndromic surveillance to offer information for public health action. Examples of syndromic surveillance include:
- Absenteeism.
- Over-the-counter and prescription pharmaceutical sales.
- Poison control reports.
- Emergency medical service ambulance data.
Traditionally, patient contacts with healthcare services have been the main pathway for data acquisition. Currently, the field is undergoing technological disruption: non-health-care syndromic surveillance data, such as social media or internet search data, is increasingly being studied.
Acoustic syndromic Surveillance
Cough detection for respiratory disease syndromic surveillance is a simple way to detect early outbreaks and disease surveillance in the context of the ongoing COVID-19 and future pandemics with respiratory symptoms.
Because topographic clustering is crucial in disease outbreaks, it is crucial to consider patterns in symptoms over time and space. Essentially, it would use data sources that are reliable and interoperable so that healthcare professionals could make valuable comparisons across different locations.
Acoustic surveillance of cough entails the following:
- Taking in sounds.
- Distinguishing coughs from non-coughs.
- Combining and analyzing data to find anomalous tendencies.
- Taking public health action if cough frequency increases dramatically in a given location.
Coughs would be tracked and learned from using a sophisticated sound monitoring system. A feedback loop would include the same data that triggered the anomalous warning. With time, the simple alarm system might factor in parameters like cough distribution throughout the day.
Bias in Syndromic Surveillance
Often, biases creep in sources of syndromic surveillance in the following ways:
- Healthcare access and geographic reach bias the population from where the data originated.
- Although faster than ever, syndromic surveillance might not be as fast as disease transmission. E.g., an individual may be infected and contagious for several days before being detected.
- Furthermore, it only applies to places where data is abundant and timely. We cannot carry out Syndromic surveillance efficiently without emergency rooms, digital school logs, or centralized laboratory test registers.
Acoustic Epidemiology and New Technologies
The world of acoustic epidemiology has received a considerable boost in recent years with the advent of new technologies such as smartphones, wearables, and voice-activated assistants.
Saving both time and resources is advantageous for all clinical trials. However, for those who utilize cough as an endpoint, the most significant dramatic benefits would come from long-term, continuous, AI-powered cough counting software delivered via inconspicuous smartphones, wearable devices, and Internet-of-Things items.
Smartphones
Artificial intelligence techniques combined with mobile device technologies can offer novel ways to clinical research. For example, several researchers have underlined the benefits of employing smartphones and their notification functionalities to promote participant retention and adherence to treatment. AI software could improve these efforts by applying dynamic drop-out risk forecasts.
It’s no coincidence that companies have seized this unique opportunity to improve individual and public health through the use of smartphones. Hyfe is one such firm that has developed an app that can transform the detection and management of respiratory disorders.
Researchers used the Hyfe app to track respiratory sounds in over 800 study participants in a 2020–2021 acoustic epidemiology study in Navarra, Spain. The study’s goal was to assess the ability of population-based digital cough surveillance to predict the incidence of respiratory diseases at the population level in Navarra and individual determinants of platform uptake.
This was the first population-based syndromic surveillance project to use passive cough monitoring at scale, spanning almost a year and generating tens of thousands of cough sounds across 9 years of person-time. It showed that cough monitoring can detect changes in cough frequencies at individual and community levels and that aggregated cough data was temporally associated with the incidence of COVID-19 in the community.
Conclusion
Everyone, without exception, can use the Hyfe App for diagnostics. It can minimize the strain on healthcare staff, empower individuals to take control of their health, and assist loved ones in caring for one another.
Wearables and smartphone health tracking apps provide essential health information by using the power of data — a lot of it! Get in touch with Hyfe to learn more about the Hyfe app to track your patients’ cough.
Visit the wiki page to learn more on acoustic epidemiology!
