Wearable devices are transforming healthcare by converting the body’s analog signals—like skin temperature and heart rate—into digital signals powering health insights. By analyzing this data, we can spot early signs of stress, illness, and even conditions like hypothyroidism—before symptoms appear.

The future of healthcare is personal, and it starts with your data - on your wrist or finger.

All signals generated by our body are analog (continuous with varying strength and frequency over time) representing information as a continuous wave. These analog signals are converted into digital (discrete and represented as binary value) using ADC (Analog-Digital converters) for processing, feature extraction, and modelling. Our smart devices (watches, rings etc., called as “edge devices” ) have miniature ADC components to measure and record these signals from our body. 

This is an amazing ballet of signal processing, miniaturization, Real-Time OS (RTOS),  Sensor Fusion, Feature extraction and Intelligent Learning models. Information power of this information is being harnessed for prediction to improve Quality of Life and longevity.

Two measurements that any good wearable device (WD) measure are – Skin Temperature and Heart Rate.

Let’s understand Skin Temperature:

Homeostasis is the ability of a living organism to maintain a stable internal environment in relation to a variable external environment. In other words, human body maintains a dynamic equilibrium of the  internal environment or ‘milieu interieur’  despite any change in the external environment. This is achieved at cellular level by regulating physiological processes such as fluid balance, blood sugar levels and body temperature. 

Human body has a core temperature (temperature of internal organs) of 37 deg C and remains stable, typically within a range of +/- 0.8 deg C.

Measuring core temperature has uses in daily wellness, preventive and recovery phases of health and in sports – Fever (rise in core temperature) is usually the first sign of an oncoming infection, used in recovery monitoring post-ailment, ICU, sleep quality analysis, hydration status and others.

Pulmonary artery temperature (temperature of blood returning from the body to the heart) is considered gold-standard. However, this is an invasive process and used only in critical care or in major surgery. There are relatively less invasive and reliable methods such as Esophageal temperature probe, rectal temperature, but are still uncomfortable and impractical for daily use. Skin Temperature measurement steps in to solve the problem at scale using wearable devices (WD).

Skin temperature (lower than core temperature due to larger surface area) measurement starts with the concept of an Open Thermodynamic System. Skin is the medium of exchange of energy and matter from human body to its environment . This exchange is by mechanisms of evaporation, conduction, convention etc.

When the core temperature reaches critical level, results in increased blood flow to skin to dissipate energy (heat) to environment.

When the core temperature drops below critical level, body works to increase rate of heat production (by means of muscle tension and shivering) with equivalent drop in blood flow (to reduce heat dissipation) to skin to reach equilibrium.

This change in temperature (analog signals) are picked up by Thermopiles or Infrared (IR) sensors (all warm objects emit IR) used in WD’s (smart rings, smart watches etc.) The IR sensors are combined with data from accelerometers, HR and ambient temperature sensors to derive context to the measure.

Moving on to Heart Rate:

Heart pumps blood - More the blood flow, faster is the heart pump. This principle along with light absorption by blood is the basis of measuring heart rate.

A green source of light from LED’s is directed onto the skin using a WD. This light is absorbed and reflected by the blood flowing in the capillaries. By measuring intensity of absorption and/or reflection using photo detectors, a pulse waveform can be recorded. The peaks of the waveform (analog) is used to derive heartrate (digital).

Multitude derived measures of HR such as HRV (Heart rate Variability – measured as time intervals between successive heart beats), reflect the balance of autonomic nervous system.

“Anxiety” is another important derived metric to evaluate emotional and physiological state of an individual. By combining measures of skin temperature and HR derivatives, it is possible to predict anxiety state. 

When body's metabolism slows, it results in less heat generation. This causes the skin temperature to fall below 33.5°C (typically between 32.5 and 33.0°C). This drop in skin temperature can also be due to other factors such as hormones, drugs and nervous system signals (“fight” or “flight” response). For contextual state, additional observations such as HRV, Sleep Quality and Quantity can be included. 

By analyzing the data points of skin temperature, HRV, Sleep Quality and Quantity, “anxiety” state is derived.

WD are classified as medical devices, and not as “gold-standard” of measure. However, it is widely accepted by research papers that by tracking data from WD, it is possible to predict a near-future state of health condition.

Let us evaluate Thyroid health.

Thyroid health reflects metabolic health, cardiovascular function, mood and cognition. Blood panel measures T3, T4 and TSH. Any imbalance of thyroid hormones directly influence metabolic rate and peripheral blood flow, deviations in baseline skin temperature (either a consistent drop in hypothyroidism or an increase in hyperthyroidism). These measures if measured, recorded and analysed can provide early indications of thyroid dysfunction.

Let us consider hypothyroidism – This condition is characterized by drop in temperature (core and skin), drop in HRV (normal: 70 to 90 ms | Drops to 48 to 55 ms), increased stress on the body (measured by - disturbances to respiratory rate, sleep latency, sleep quality) and anxiety state of body.

[Research indicates that hypothyroid patients have a 1 °C lower resting skin temperature than healthy persons, particularly noticeable during cold exposure. This minor change, is accompanied by decreases in HRV and slight elevations in resting heart rate.] 

Combining these observations – Skin Temperature, HRV and Sleep metrics over a period of time (say 6 months), gives a unique health-print of an individual. 

Let’s consider an example: 

Meet Mr. A, a 45-year-old with no major medical history and consistent physical activity. He regularly tracks his thyroid hormone levels in blood tests and uses WHOOP to monitor physiological parameters. It is also assumed that Mr. A is under no medication, no comorbidity, no supplements and has no hospitalization incidents for the last 12 months

Below are his thyroid readings.

Below are some of the physiological parameters (screenshots from WHOOP app) for our consideration.

HRV Trend – 6 months. Units: ms.

Respiratory Rate Trend – 6 months

Sleep Performance – represented in %

Measured as (Sleep obtained ÷ Sleep need) × 100

Restorative Sleep

Measured as = (Slow Wave Sleep (SWS) + Rapid Eye Movement (REM)) / Total Sleep Time

Skin Temperature Trend – 6 months

The above features are summarized below for easy of interpretation.

If we run an Intelligence learning model on this disparate data sources, we can help understand and answer below scenarios.

1. Is it possible to help an individual and a healthcare professional by providing an overall health score of an organ or a system which is derived from the disparate data sources ?
   
   a. Eg: Can a system assign a score for Thyroid health of 6.8 (on a scale of 1 to 10, 1 being low health and 10 being optimal health) by analysing T3, T4 and TSH readings along with data from WD ?
   
   

2. Early intervention by healthcare professionals for personalized care and treatment.
   
   a. Mr. A is on the lower end of T3 and T4 values; with the TSH level being on mid to high level of the normal range? Is there a potential risk of him becoming sub-clinically hypothyroid in the next 90 days, if untreated?
   
   b. Will Mr. A benefit from magnesium supplement to support his improvement in sleep performance, with focus to improve his thyroid health?
   
   

3. Identify patterns and correlations not apparent to human expertise.
   
   a. Why is Mr. A’s HRV dropping in the current month? Is his sleep performance impacting his HRV? How is this impact contributing to his anxiety level in his daily life?
   
   

4. Continuous refinement of insights based on new data.
   
   a. If Mr. A improves his sleep performance, what are the T3 and TSH test results he can achieve in the next 90 days?

The future of predictive and preventive healthcare, driving personalized care and intervention starts right on our wrist or finger where we carry our physiological monitoring device.

As our personal health monitoring devices get smarter, so will our ability to stay healthier, longer.

References:

1. https://pubmed.ncbi.nlm.nih.gov/35657799/

2. https://pubmed.ncbi.nlm.nih.gov/36009529/

3. https://pubmed.ncbi.nlm.nih.gov/31752113/