It took the mobile, unplugged world of smartphones to realize that widely ignored but remarkably rich medical data sources could be readily captured and processed. Two areas that are particularly being pursued are the analytics of breath and voice, with multiple potential uses for both.

The breath story is quite interesting. Our pattern of breathing — rate, depth and flow — appears to be an excellent metric for our mood and represents one of many useful metrics for quantifying anxiety and stress. The microphone of a smartphone can be used to detect how much air is moving and thereby capture a key metric of lung function, potentially providing a signal for an impending asthma attack before the individual has any symptoms. But the value of breath as substrate is far greater. While long known as a way to gauge the level of alcohol in the blood (and there are indeed smartphone breathalyzers for this function), the ability to detect and quantify organic chemicals, such as benzenes and alkenes, sets up the smartphone with a hardware attachment to potentially detect cancer. Surprisingly, it’s not just lung cancer, but a diverse array of cancers that ultimately may be detectable via such sensors.

This concept is an outgrowth of the appreciation of the keen sense of dogs to detect cancer by smell and now simulation of that via an “electronic nose.” The list of assays that might someday be performed via breath include acetone for diabetes, ammonia for kidney disease, carbon dioxide for liver disease, alkanes for transplant rejection, and nitric oxide for asthma. Many of these applications are currently being pursued. Imagine how this might someday take the place of expensive medical scans (such as CT, PET and nuclear imaging), which often expose individuals to considerable radiation — and only provide a one-off assessment.

Equally exciting is the processing of voice. Like breath, it is a powerful indicator of mood and anxiety by virtue of the tone, inflection, volume, rhythm and rapidity of speech. Some have suggested that voice is the most sensitive single metric for quantifying mood, but is expected to be integrated with many other components such as heart rate, blood pressure, galvanic skin response, facial expression, pattern of communication and movement. Beyond its being a window into emotion, voice has the distinct potential for diagnosis of medical conditions. In fact, one pilot study suggested that Parkinson’s disease could reliably be diagnosed simply be analyzing one’s voice.

It is striking to me that these essential human features largely have been missed until we embarked into the era of mobile medicine. While both require much further validation to go mainstream, they are illustrative of where we are headed when there is such a powerful microprocessor in the little device that we all carry and connect with each day.