How your voice is transforming virtual medicine 

The human voice is an undervalued data source, but not for long: New technologies can identify moods and medical conditions from your voice through virtual communication. Voice-based tools are changing virtual patient care through the unique data and patient accessibility they provide, and are allowing medical care providers to improve diagnosis of medical conditions over the phone, enabling quicker and enhanced care for many patients.

_{New companies are researching into the physiological and psychological mechanisms underlying voice patterns through voice-based technology; the results could improve medical care (123rf / Viktor Gymria)}

The current state of medical care is predominantly dependent on direct patient-clinician interaction through doctors’ visits; patients must meet with medical professionals in person through routine appointments or in an emergency. However, the current system is insufficient when a patient needs immediate medical assessment but has limited access to care. The incorporation of voice-based diagnostic tools could provide these patients with near instantaneous medical attention.

Additionally, virtual tools could drastically expand the accessibility of medical care. In an emergency, patients receive urgent clinical attention and medical professionals benefit from accurate diagnostic data. Virtual patient communication during ambulance transport or at assisted-living facilities could better inform medical professionals and expedite urgent care. Virtual interactions will increase a patient's access to medical care, and increased interaction with care-providers could afford patients a better understanding of their medical conditions. Voice-based technologies may, therefore, improve medical care by providing additional information to both patients and care providers.

Voice stress analysis (VSA) is used in intelligence and criminal investigations to survey and analyse levels of stress in a person's voice. VSA functions by measuring frequency changes associated with physical or psychological stress – micro-muscle tremors – and is utilised as a "truth-verification technology" (Chapman & Stathis, 2012). This technology is used in legal and military situations to detect deception in informants' interviews (Hopkins, Ratley, Benincasa, & Grieco, 2005).

New companies have continued research into the physiological and psychological mechanisms underlying voice patterns through similar voice-based technology. Voice stress and frequency data have been combined with analytical software to produce powerful tools, which may enable medical analysis and diagnosis based on patients’ voices.

AudioProfiling, a Humboldt University spin-off, has developed voice-analysis algorithms based on ‘musical parameters (eg, volume, articulation, etc) as well as measures of voiced/unvoiced sounds. Research collaborations with Humboldt University and Charité Universätsmedizin Berlin have focused on medical and psychological disorder detection from the human voice. Other partnerships are working on voice-based detection of Attention Deficit Hyperactivity Disorder (ADHD) (Meditech GmbH) and various psychological states (University of Paderborn).

This software is noteworthy for accuracy in the absence of limited recordings, poor audio quality, and compatibility with telecommunication. AudioProfiling offers services ranging from medical detection of ADHD, to examining psychological personality traits.

Sonde Health, a by PureTech company, offers software designed to recognise: "Subtle, characteristic changes," in voice patterns in order to detect psychological conditions such as depression, as well as physical conditions like respiratory and cardiovascular disorders.

Sonde Health's voice-scanning technology is based upon research from MIT’s Lincoln Laboratories and is designed for patients’ smart devices, allowing health-related data to be continually collected and monitored, and to be accessible to healthcare providers. Currently, this software is in early-stage development, which relies on participant interaction. Ultimately, this technology may function as a continual analysis and detection tool.

Cogito has also designed voice-based software to detect changes in moods. The company was founded upon the findings of the MIT Human Dynamics Lab, and has since developed an artificial intelligence platform, which can interpret communication and psychological information, such as bipolar disorder and depression, from audio inputs. Partnerships with healthcare management programs and insurance organisations have provided further data for health-related detection. This voice-based technology may allow patients suffering from psychiatric conditions to monitor the condition through phone calls with medical care professionals.

It is clear, therefore, that these recent developments in voice-based software have increased the medical capability to collect health-related data from patients and provide broader medical access to patients.

These tools can be used to provide critical missing or supplemental information to medical care providers in cases that require urgent medical attention and future progress in developing voice-based software will continue to shape virtual medicine. The development of these voice-based technologies and virtual tools expands the opportunities for doctor-patient communication. Virtual medicine reshapes modern medical systems to enable better access to potentially limited emergency medical care.

Sarah Ayton; Carly Esteves & Ian Portelli