Danger, Will Robinson!!! or injury prevention with sensors and algorithms

Health is determined by many factors including:

• Behavior (Physical Activity, Eating habits, Tobacco or substance abuse, responsible sexual choices, etc)
• Mental Health 
• Injury and Violence 
• Environmental Quality 
• Preventative measures such as immunization 
• Access to Health Care

All these factors are quantifiable, predictable and preventable. Injuries are most likely to be  perceived as “accidents” and “acts of fate”,  but they depend on the same determinants as other health factors: individual behavior, social and physical environment.  The likelihood of injuries -  unintentional ones and those caused by acts of violence - can be computed from physical location (estimations for USA¹ are a good example), gene-environment interactions², prior medical history, and physical traits³.

There are many ways to prevent injuries - just say "no" to risky behaviors, wear preventative gear while playing sports or fall-optimized shoes for elderly, watch out for others engaged in similar activities... Yet, sometimes we forget to watch, don't have access to histories of others or get diverted.  Would a body sensor or a gadget with smart software be able to warn us about potential accidents ahead to help prevent accidents?  What would it need to measure?

Software and devices automatically detecting and reporting accidents already exist:

Halo Monitoring's fall detection system, for example,  consists of a chest strap and belt clip with motionOnStar or mbrace that "intelligently integrate the driver, the vehicle and the environment" - capabilities such as this will be provided in the area of next-generation health management systems like Aurametrix.
sensors, heart rate and skin temperature monitors. Although the system detects falls only after they happen, a study showed that just the fact of wearing it increases alertness of seniors and reduces the number of falls. Although fall detection systems are not as advanced as telematics for cars - like
Unprecedented accumulation of data  - such as snapshots of driving behavior or 1.2 million person-years of hip fracture observations (Kanis JA) allows development of smarter software able to predict injuries.  Logistic regression models (Kononen et al., 2011) predict seriousness of auto accidents,  first-principles mathematical models (such as AHAAH for the ear) connect forces with injuries, neural net and other data mining approaches foretell which juvenile offenders are likely to return to crime (source of  "intentional injuries"), or allow to calculate risk of fractures based on milk intake, personal history of accidents and body mass index. Self-quantifiers such as René Ghosh are able to figure out how to use their own data to predict future injuries. Using simple math (Riegel equation bringing all running logs on to a comparable level) and trend analysis, he tied his accidents to wanes following waxes in running performance. Researchers keep refining the variables predicting injuries.  Swanenburg et al., for example, predicted multiple falls for those with a history of multiple falls (odds ratio, 5.6) and use of multiple medications (odds ratio, 2.3). And there is another simple measurement of standing position helpful in prediction. Frequent fallers, indeed, have a narrower stance width than non-fallers.


In the always-connected smart-sensor-equipped future, things such as Intel's magic carpet - picking up the weight, angle and pressure of steps - will be a commodity. Gene tests predicting injuries will be integrated with data coming from our carpets, clothing, footwear and location information. And this may be sooner than you think.


References

1. Centers for Disease Control and Prevention (CDC), National Center for Injury Prevention and Control. Web-based Injury Statistics Query and Reporting System (WISQARS); 2010 Mar 4 Available from: http://www.cdc.gov/injury/wisqars/index.html

2. Husted JA, Ahmed R, Chow EW, Brzustowicz LM, Bassett AS. Childhood trauma and genetic factors in familial schizophrenia associated with the NOS1AP gene. Schizophr Res. 2010 Aug;121(1-3):187-92. PMID: 20541371

3. Swanenburg J, de Bruin ED, Uebelhart D, & Mulder T (2010). Falls prediction in elderly people: a 1-year prospective study. Gait & posture, 31 (3), 317-21 PMID: 20047833

4. Kononen DW, Flannagan CA, Wang SC. Identification and validation of a logistic regression model for predicting serious injuries associated with motor vehicle crashes. Accid Anal Prev. 2011 Jan;43(1):112-22. PMID: 2109430

5. Price GR. Predicting mechanical damage to the organ of Corti. Hear Res. 2007 Apr;226(1-2):5-13. Epub 2006 Sep 15.PMID: 16978813