Analytics for Your Health

Oct. 27, 2008
Intelligent video can streamline a hospital's surveillance efforts

Healthcare facilities have a growing responsibility to patients, employees and the community to provide a safe and secure environment. In addition to general perimeter security around main buildings, associated labs and parking structures, hospitals and other healthcare facilities are continuously exploring better ways to monitor patients. And hospitals, as well as research facilities and university laboratories, must protect hazardous waste and bio-chemical hazards that can be compromised if they fall into the wrong hands. Those same facilities also most protect against the theft of expensive diagnostic equipment, labware, pharmaceuticals and other high value material.

The result is a need for a more pervasive and effective video surveillance system to mitigate the risks and address the increasing liability and associated insurance costs impacting healthcare facilities.

To meet those needs, hospital facilities are investing more and more in surveillance equipment, making the healthcare industry one of the fastest growing markets for the deployment of security cameras and video management systems. According to a recent IMS Research market report, the healthcare industry is expected to generate an average 15-percent increase in CCTV and video surveillance equipment expenditures through 2010, which is higher than the projections for other heavy security equipment buyers such as retail (12.2 percent) and banking (10.7 percent).

The healthcare industry, like many other markets, is now looking to video analytics as the only effective means to deploy a constant-use, proactive solution. Intelligent video software provides detection, classification (human, vehicle, other), and tracking of objects within the field of view of a camera and compares object behavior against various rules modeled to identify various conditions.

Typical video analytics behaviors can be directly mapped to real applications considered by the healthcare industry. The best way to assess the applicability of video analytics to healthcare use is to look at a hospital/medical center security landscape from the macro (perimeter) to the micro (patient) level.

Perimeter Security

One of the more basic security monitoring functions for a hospital is basic perimeter security to identify access to grounds or buildings in a suspicious manner, or attempted access to buildings via various emergency, service and otherwise non-monitored access points. To establish a sound secure perimeter, adding more security personnel or physical barriers may be a costly and infeasible solution. Video analytics, however, provide a more cost effective and reliable solution. In the simplest implementation, this involves the creation of a “video tripwire” around the entire perimeter, or at various locations where hospital access must be monitored. The alerts from such a tripwire can be centrally monitored and used to provide real-time notifications to roving security forces of a potential threat, making that mobile force more effective, while not requiring constantly-manned video screens.

Video analytics can also look for potentially threatening behavior, such as loitering, around the exterior of a hospital or near monitored areas such as loading docks and hazardous waste collection containers.

Hospital Monitoring

Operations within the hospital represent significant challenges addressable by video analytics. Any secure area can be monitored for unauthorized access. Those same entry events can be used for one-to-one comparison to access control system entry events to spot illegal “tailgating” into controlled areas. Normal behavior such as people accessing hospital supplies or controlled pharmaceuticals can be recorded as “events of interest” to verify and monitor who is accessing these materials in case trends of abuse or unauthorized access require investigation.

Video analytics excel not only with tracking moving objects but also for monitoring stationary targets. Suspicious objects left inside the hospital, or potentially hazardous objects left in passageways or blocking fire exits, can be automatically detected. Conversely, high-value objects, such as laboratory equipment, can be monitored for sudden removal and possible theft.

Patient Monitoring

Hospitals are always looking for new methods to monitor in-room patient health and welfare. Video presents two issues: First, a camera in every patient room creates an infeasible situation for manual video monitoring; second, video immediately raises privacy issues. Video analytics can solve both issues.

There are several patient monitoring behaviors effectively performed by video analytics. Patient in-bed movement is of particular concern, especially for intensive care unit (ICU) patients. By defining areas surrounding a patient, video analytics can detect various reaching movements outside of the bed, tampering with monitoring or IV equipment, or unauthorized exiting of the bed. Real-time alerts to such movement enables hospital staff to respond faster to the potentially damaging or fatal activity, and proves the hospital's rapid response against possible malpractice suites.

The concept of using video for patient monitoring raises the red flag of privacy. Video analytics solves the problem because the software, in itself, does not stream video. The algorithms essentially take in video images and output data in the form of events and alerts. There is no video out. The analytics watch the activity instead and output data only when something of interest takes place.

When movement in and out of the bed is authorized, video analytics can alert to other conditions. Did lights go on or off? Did the patient fall down while walking to the bathroom? Did the patient exit the room? Intelligent video channels also have the ability to determine occupancy in a room at any time, which is useful for determining unauthorized visits during certain times of the day (intelligent video rules can be scheduled to be active for certain times of the day) or too many visitors in a room. Alternatively, knowing the occupancy in a patient's room throughout the day can be a useful tool to verify scheduled room visits by hospital staff.

Other Applications

What if the patient is not in a medical facility? Unassisted living for the elderly and home-based healthcare needs are growing, and the industry is in need of better home-based patient monitoring solutions. Nurse programs provide sporadic visits, but not 24-hour coverage. There are some infrared ( IR) -based sensor solutions to track patient behavior and movements, but false detections often impact the performance of these systems. Video intelligence can be used to detect human presence and movement, access to the kitchen to verify the patient is eating, regular bathroom visits, and slip-and-falls.

Leading analytic providers enable multiple rules to be actively monitored simultaneously, equipping hospital security directors with ultimate flexibility to model their security and safety policy. If the threat condition changes, the security director can literally at the push of a button activate a new “rule group” across all cameras and extend the perimeter and engage more stringent video analytic rules to meet the heightened threat level.

Bob Cutting is the director of product management for ObjectVideo, an intelligent video provider.