Telemedicine Journal and e-Health
`Volume 6, Number 3, 2000
`Mary Ann Liebert, Inc.
`
`The Physiologic Cipher at Altitude: Telemedicine and
`Real-Time Monitoring of Climbers on Mount Everest
`
`RICHARD SATAVA, M.D., F.A.C.S.,1 PETER B. ANGOOD, M.D., F.A.C.S.,1
`BRETT HARNETT, B.S.,2 CHRISTIAN MACEDONIA, M.D.,3
`and RONALD MERRELL, M.D., F.A.C.S.2
`
`ABSTRACT
`
`Advanced wearable biosensors for vital-signs monitoring (physiologic cipher) are available
`to improve quality of healthcare in hospital, nursing home, and remote environments. The
`objective of this study was to determine reliability of vital-signs monitoring systems in ex-
`treme environments. Three climbers were monitored 24 hours while climbing through
`Khumbu Icefall. Data were transmitted to Everest Base Camp (elevation 17,800 feet) and re-
`transmitted to Yale University via telemedicine. Main outcome measures (location, heart rate,
`skin temperature, core body temperature, and activity level) all correlated through time-
`stamped identification. Two of three location devices functioned 100% of the time, and one
`device failed after initial acquisition of location 75% of the time. Vital-signs monitors func-
`tioned from 95%–100% of the time, with the exception of one climber whose heart-rate mon-
`itor functioned 78% of the time. Due to architecture of automatic polling and data acquisi-
`tion of biosensors, no climber was ever without a full set of data for more than 25 minutes.
`Climbers were monitored continuously in real-time from Mount Everest to Yale University
`for more than 45 minutes. Heart rate varied from 76 to 164 beats per minute, skin tempera-
`ture varied from 5 to 10°C, and core body temperature varied only 1–3°C. No direct correla-
`tion was observed among heart rate, activity level, and body temperature, though numerous
`periods suggested intense and arduous activity. Field testing in the extreme environment of
`Mount Everest demonstrated an ability to track in real time both vital signs and position of
`climbers. However, these systems must be more reliable and robust. As technology transi-
`tions to commercial products, benefits of remote monitoring will become available for rou-
`tine healthcare purposes.
`
`INTRODUCTION
`
`FIVE CLIMBERS DIED during their May 1996
`
`climb of the summit of Mount Everest. This
`tragedy brought into focus the extraordinary
`risks all individuals endure in remote and ex-
`
`treme environments. The accounts, popularized
`by Jon Krakauer in Into Thin Air1 and Brougton
`Coburn and David Beshears in Everest, Moun-
`tain Without Mercy,2 described in detail the hard-
`ships and circumstances leading to disaster. In-
`credibly in that episode, two climbers died
`
`1Department of Surgery, Yale University School of Medicine, New Haven, Connecticut.
`2National Aeronautics and Space Administration (NASA), Commercial Space Center for Medical Informatics and
`Technology Applications (CSC/MITA), Virginia Commonwealth University, Medical College of Virginia, Richmond,
`Virginia.
`3Department of Obstetrics and Gynecology, Uniformed Services University of Health Sciences, Bethesda, Maryland.
`
`303
`
`UA-1007.001
`
`

`

`304
`
`during a snowstorm within proximity of the
`safety of the camp. Another was left for dead
`but actually survived to walk into camp the fol-
`lowing morning with frostbite so severe that he
`lost his nose, right hand, and fingers. Had the
`expedition members known that two of their
`colleagues were fallen at a position just outside
`their tents, or had it been known that one was
`severely hypothermic but alive with barely de-
`tectable vital signs, the outcome may have been
`more positive. These three tragedies could have
`been avoided if current technology had been
`available to the expedition.
`A physiologic cipher is a noninvasive identi-
`fying tool to measure physiologic status of an in-
`dividual in real time, through monitoring of vi-
`tal signs, biochemical, and other parameters with
`sensors worn by the individual. When concerned
`with remote locations, such as expeditions or the
`battlefield, such monitoring also includes geolo-
`cation using the Department of Defense global
`positioning satellite (GPS) system.
`The military sector has been developing a
`number of wearable systems for location and
`vital-signs monitoring (VSM).3 These systems
`consist of three principal components:
`
`1. The vital sign sensors, including heart
`rate, temperature, respiratory rate, elec-
`trocardiogram (EKG), motion (accelerom-
`eters) and pulse oximeter, are currently
`worn at various sites on the body. They
`are typically strapped across the chest or
`wrist or swallowed in pill form.
`2. The GPS device is commercially available
`and accurate to within 0.75 meters longi-
`tude and 1.01 meters latitude.
`3. The telecommunications system, usually
`a radio frequency (RF) transmission sys-
`tem, is also commercially available but
`repackaged into a miniaturized wearable
`configuration.
`
`In remote environments, strategically placed
`transceivers are required to receive and trans-
`mit the signals from the wearable systems to a
`base station containing the receiver, signal
`processor, and computer workstation (or lap-
`top computer). Some systems, such as the Sar-
`cos Personnel Status Monitoring or PSM™ sys-
`tem (Sarcos, Inc., Salt Lake City, UT) also have
`
`SATAVA ET AL.
`
`a hand-held “medic unit” (Fig. 1) that permits
`a member of the expedition to be in the field
`and still monitor the vital signs and location of
`other individuals or members of the squad.
`During the Mount Everest climbing season in
`May 1999, an Everest Extreme Expedition was
`conducted by a team from Yale University
`School of Medicine in collaboration with the
`Yale University–NASA Commercial Space
`Center for Medical Informatics and Technology
`Applications, Millennium Healthcare Solu-
`tions, and The Explorers’ Club. The mission
`had three objectives: (1) to provide advanced
`medical support to the climbing expeditions
`from a base camp at 17,800 feet at Mount Ever-
`est Base Camp (EBC) from a telemedicine clinic,
`(2) to test an emerging VSM system for moni-
`toring a physiologic cipher and vital signs of
`climbers as they ascend toward the summit of
`Mount Everest, and (3) to assess the cardio-
`vascular adaptation to hypoxia at high altitude.
`This report concerns the VSM system and the
`medical implications of a physiologic cipher.
`
`FIG. 1. The medic unit from the Sarcos, Inc. Personnel
`Status Monitor (PSM) system being developed for the mil-
`itary. (Courtesy of Dr. Stephen Jacobsen, Ph.D., Sarcos,
`Inc., Salt Lake City, Utah.)
`
`UA-1007.002
`
`

`

`PHYSIOLOGIC CIPHER AT ALTITUDE
`
`MATERIALS AND METHODS
`
`The VSM system was custom designed by
`FitSense Technologies (FitSense Technologies,
`Wellesley, MA) in conjunction with require-
`ments by the Department of Defense for the
`“next generation soldier system.” Figure 2 il-
`lustrates the three modules which comprise the
`system:
`
`1. Non-invasive physiologic sensors to mea-
`sure vital and physical signs. This in-
`cludes: heart rate (accurate to 64 beats per
`minute), 3-lead EKG, accelerometer (or
`chest actigraph, for gross body motion
`and activity), a surface body temperature
`monitor (accurate to 60.01°C), and a core
`body temperature monitor (ingested pill,
`accurate to 60.04°C). The accelerometer is
`a standard threshold accelerometer with a
`sensitivity of 0.01 g. The core body tem-
`perature pill, similar to the one used by
`Senator John Glenn during his space shut-
`tle mission, is a silicon-covered jelly-bean
`size capsule with a temperature-sensitive
`quartz crystal oscillator in which the signal
`is inductively coupled to an RF transmitter
`to an external belt-worn transceiver as de-
`scribed by Mittal et al.4 The pill is swal-
`lowed and the core body temperature is
`
`305
`
`transmitted every minute to the receiver,
`which aggregates the data into a data
`logger.
`2. Accurate position tracking using the
`Global Positioning Satellite (GPS) system
`(Lassen SK-8, Trimple, Inc., San Jose, CA).
`3. Wearable, wireless communication sys-
`tem with radio frequency (RF) transmis-
`sion using the 928 mHz band. The receiv-
`ing station at EBC was a laptop computer
`with a proprietary graphical user interface
`from FitSense Technologies.
`
`The purpose of the VSM system is to provide
`a small, unobtrusive monitoring system that can
`continuously acquire and transmit vital signs
`and other bio-data in real time while the climbers
`were performing an ascent. The system is de-
`signed for medical research, industrial, and mil-
`itary applications where biodata are recorded
`while subjects perform normal activities. Cur-
`rently measured parameters (either directly mea-
`sured or calculated) include heart rate, core body
`temperature, surface body temperature (chest
`and wrist), chest actigraphy (gross body motion),
`sleep time, and geolocation.
`The VSM system was designed to be a mod-
`ular system which uses an open architecture.
`The data hub can dynamically receive and store
`data while reconfiguring as needed to accept
`
`FIG. 2. The vital-signs monitoring (VSM) system of Fitsense, Inc., demonstrating (right to left) the global position-
`ing satellite (GPS) module, the central processing hub, and the radio frequency (RF) transmitter. (Courtesy of
`Dr. Tom Blackadar, Ph.D., FitSense Technologies, Inc., Wellesley, Massachusetts.)
`
`UA-1007.003
`
`

`

`306
`
`new input, with a maximum of 16 sensors per
`person without RF degradation. The biosensor
`acquisition is transmitted to the hub on a Per-
`sonal Local Area Network (PLAN) with digital
`RF signals. The hub then stores the data or can
`transmit to a variety of long-haul options, such
`as CDPD cellular modem or other long haul ra-
`dio connections.
`Each datum is time-stamped and geo-
`stamped with the GPS sensor. Time-correlated
`data from multiple sensors are more robust than
`an isolated datum, facilitating error correction
`and datum fusion for superior physiological in-
`terpretation. The data may either be stored in
`the hub or telemetered for remote storage and
`analysis. For this expedition, the sensor was
`queried four times per minute until the datum
`was acquired, then sent to the hub and stored
`for the subsequent 5-minute transmission. The
`system hub runs for 10 days on two AA-size
`batteries storing data once per minute.
`The hub comes in three configurations, with
`integral long-haul radio and GPS, with integral
`GPS, and with wireless GPS and wireless long-
`haul radio. The hubs have a direct serial (wired)
`connection available to the PC for configura-
`tion and direct download. The hubs can also
`communicate through the supersensor. The su-
`persensor can be connected to a Palm Pilot or
`a personal computer running Windows 95 or
`Windows 98.
`The data from a climber were transmitted over
`a number of telecommunication links. The initial
`link was from the climbers using an RF of 918
`MHz. Because wireless systems typically require
`line-of-sight transmission, a device called a “re-
`peater” was positioned on a neighboring moun-
`tain (Pomori) to facilitate a vectored path by
`which the signal could be bounced. This proce-
`dure proved quite reliable except when a climber
`walked into a radio black area, which was occa-
`sionally encountered in the fractured glacier
`topology of the Khumbu Ice Fall. At Base Camp,
`data were received and compiled onto the lap-
`top. The software was programmed to collect ag-
`gregate ASCII (American Standard Code for In-
`formation Interchange) datasets from each of the
`monitoring devices at predetermined intervals.
`For the expedition, the intervals were every
`5 minutes. However, the hub is capable of up-
`dating every few seconds.
`
`SATAVA ET AL.
`
`From the laptop at Base Camp, the data were
`sent via satellite (Imarsat) using TCP/IP (trans-
`mission control protocol/Internet protocol)
`where they were grounded in Malaysia and
`routed to an Internet backbone and into a con-
`ference room at Yale University School of
`Medicine in New Haven, Connecticut (Fig. 3).
`Laplink (Traveling Software, Bothell, WA) was
`used to download the data from Base Camp to
`Yale University because of its intuitive interface
`and reliable transport method. At Base Camp
`the incremental ASCII datasets were stored in
`a directory that was indexed to Laplink and
`marked as “shareable.” The physicians at Yale
`University downloaded the data at 5-minute in-
`crements. Because the data sent were standard
`ASCII, the encoded data were plotted on a
`graphical user interface, designed by Fitsense
`Technologies, Inc., which displayed intuitively
`the ciphered information (Fig. 4). Access to in-
`dividual continuous display of vital signs (tem-
`perature and heart rate) was linked by simply
`clicking on the climber number, and the chrono-
`logical graph was plotted (Fig. 5).
`In addition to the vital signs, hundreds of im-
`ages were captured during the expedition us-
`ing digital photography, digital microscopy,
`ultrasound, retinal photography, and microcir-
`culation photography. Many of these images
`were sent to Yale University servers using file
`transfer protocol. During the daily videocon-
`ferences, images were stored in a browser and
`run in the background for quick reference.
`Some images, such as ultrasound, were trans-
`mitted real-time from EBC.
`The technical feasibility of the system was
`tested during two ascents from EBC to Camp 1
`(19,500 feet) through the arduous and treacher-
`ous Khumbu Ice Fall. This is an exceptionally
`difficult and dangerous climb, requiring scaling
`of 40- to 80-foot-high ice cliffs with as many as
`8–10 aluminum ladders lashed together (Fig. 6)
`and traversing over 1,000-foot-deep crevasses
`on aluminum ladders strapped together, using
`ropes for a hand railing (Fig. 7). The first trek
`involved three climbers (two Americans and
`one Nepalese mountain guide or Sherpa) who
`were monitored during the trek to Camp 1 and
`then on the forced return the following day be-
`fore an oncoming snowstorm. On the second
`trek, two Sherpas were monitored as they re-
`
`UA-1007.004
`
`

`

`PHYSIOLOGIC CIPHER AT ALTITUDE
`
`307
`307
`
`Repeater
`station on
`Kalapathar
`
`FIG. 3. Diagram of the telecommunications pathway from the individual climber back to the conference room at
`Yale University School of Medicine (author, BH).
`
`turned to Camp 1 after the snowstorm to retrieve
`equipment and supplies. The vital signs and po-
`sition were acquired every 5 minutes and
`archived and transmitted every 5 minutes. These
`data (Table 1) consisted of time stamps (using
`
`Greenwich Mean Time), GPS location, heart rate,
`activity status, skin temperature, and core body
`temperature. During the daily morning telemed-
`icine conference between Yale University and
`EBC on the day of the trek to Camp 1, vital signs
`
`Base
`Camp
`
`Camp 1
`
`Khumbu
`Ice fall
`
`FIG. 4. The laptop computer screen, showing the intuitive graphical interface. On the left is the terrain map of Mount
`Everest with the overlay of the climbers’ position, and on the right are the individual climber’s vital signs. (Courtesy
`of Dr. Tom Blackadar, Ph.D., FitSense Technologies, Inc., Wellesley, Massachusetts.)
`
`UA-1007.005
`
`

`

`308
`
`SATAVA ET AL.
`
`FIG. 5. Chronological graph of the vital signs of Climber 3. (Courtesy of Dr. Tom Blackadar, Ph.D., FitSense Tech-
`nologies, Inc., Wellesley, Massachusetts.)
`
`were retransmitted to Yale University in real
`time from the climbers, allowing physicians
`at Yale University to follow vital signs and lo-
`cation while the climbers were ascending
`through the icefall. Vital signs were updated
`every 5 minutes back at Yale University.
`
`During the climb into the Khumbu Icefall,
`more than 4,000 data points were logged for
`analysis.
`Institutional Review Board approval was ob-
`tained through the Yale University School of
`Medicine.
`
`FIG. 6. The lower face of the Khumbu Icefall, with aluminum ladders lashed together to allow for the ascent. (Cour-
`tesy of Rick Satava, Jr., Base Camp Manager, Boulder, Colorado.)
`
`UA-1007.006
`
`

`

`PHYSIOLOGIC CIPHER AT ALTITUDE
`
`309
`
`FIG. 7. Crossing a deep crevasse on the Khumbu Ice-fall, using ladders roped together. (Courtesy of James Williams,
`Climbing Team Leader, Jackson, Wyoming.)
`
`RESULTS
`
`Figure 4 illustrates the home screen of the lap-
`top computer that was the interface for view-
`ing the data from the climbers. On the left-hand
`side is a scale map of the terrain between EBC
`and Camp 1, demonstrating the trail and loca-
`tion of an individual climber. The large con-
`centration of data points in the upper left is at
`Camp 1, where the climbers established camp
`and remained overnight. One small excursion
`slightly beyond Camp 1 for purposes of photo-
`graphic documentation can be noted heading
`diagonally to the upper left corner. In addition,
`there is a large skew of one climber position due
`to malfunction (most likely a miscalculation of
`GPS signal capture). The inset in the lower left
`corner is a reference graphic representation of
`the vertical ascent usually taken to the summit,
`with the numbers representing the locations of
`the four camps. On the right side of the screen
`are the “thumbnail” graphics of the continuous
`vital-signs summaries of the three climbers (two
`being active at the time of the screen capture)
`along with their latest updated values. Clicking
`on any of the climber boxes takes you to the de-
`tailed individual vital-signs screen (Fig. 5),
`which provides specific overall information as
`
`well as the chronological, high-fidelity presen-
`tation of data points acquired in real time and
`plotted every 5 minutes. To provide as robust
`data acquisition as possible, standard strategies
`were employed such as continuous resampling
`and graceful degradation. Each sensor was
`polled 4 times per minute, and once the datum
`point for the 5-minute interval was acquired, it
`was sent to the hub where it was stored with
`all the other sensor data. At 5-minute intervals,
`the hub data “cache” was read and transmitted
`to the base station, as well as archived for later
`analysis. Missed data were stored as either a
`“0” value or a default datum point. The data
`were monitored in real time and stored on both
`the receiving computer and the wearable data
`logger. Whenever there was loss of transmis-
`sion, the data logger would continue to store
`the data and then retransmit the entire data
`from the last successful transmission.
`
`Reliability of data acquisition and transmission
`The GPS location functioned well for two of
`the three climbers. Climber 1 initially had an
`excellent functioning system. After a short
`time, however, the system became erratic in lo-
`cation acquisition and on the reascent totally
`
`UA-1007.007
`
`

`

`310
`
`SATAVA ET AL.
`
`TABLE 1. TYPICAL DATASET FROM CLIMBER 3 WHILE CLIMBING IN THE KHUMBU ICEFALL
`
`Data ID #
`
`Time (GMT)
`
`Latitude
`
`Longitude
`
`Heart rate
`
`Activity
`
`Skin temp
`
`Core temp
`
`3378
`3379
`3380
`3381
`3382
`3383
`3384
`3385
`3386
`3387
`3388
`3389
`3390
`3391
`3392
`3393
`3394
`3395
`3396
`3397
`3398
`3399
`3400
`3401
`3402
`3403
`3404
`3405
`3406
`3407
`3408
`3409
`
`12:17:46
`23:22:43
`23:22:43
`23:35:47
`23:40:35
`23:45:47
`23:50:47
`0:00:01
`0:00:01
`0:05:47
`0:10:47
`0:15:47
`0:20:47
`0:25:47
`0:30:47
`0:35:47
`0:40:47
`0:45:47
`0:50:47
`0:55:47
`1:00:41
`1:05:46
`1:10:46
`1:15:47
`1:20:47
`1:25:47
`1:30:45
`1:35:44
`1:40:37
`1:45:47
`1:50:47
`1:55:47
`
`28.007
`28.00663
`28.00658
`28.00725
`28.00704
`28.00716
`28.00625
`237.3914
`26.63363
`28.00522
`28.0039
`28.00399
`28.00293
`28.00216
`28.00199
`28.00154
`28.00136
`28.00012
`27.9993
`27.99964
`27.99952
`27.99786
`27.99752
`27.99665
`27.99625
`27.99513
`27.99513
`27.9941
`27.99466
`27.99372
`27.99415
`27.99523
`
`86.86044
`86.86144
`86.86138
`86.86089
`86.86126
`86.85998
`86.86282
`122.0376
`29.78837
`86.86506
`86.86366
`86.86554
`86.86737
`86.86763
`86.86862
`86.86896
`86.86934
`86.87052
`86.8708
`86.87203
`86.87135
`86.87095
`86.87148
`86.87207
`86.87303
`86.87352
`86.87316
`86.87302
`86.8732
`86.87372
`86.87329
`86.87547
`
`92
`96
`96
`84
`84
`148
`128
`128
`124
`100
`160
`168
`172
`172
`176
`172
`180
`176
`172
`172
`172
`172
`168
`156
`164
`168
`168
`152
`168
`136
`152
`156
`
`30
`19
`16
`27
`48
`44
`36
`37
`48
`18
`44
`49
`38
`30
`37
`35
`38
`32
`26
`31
`26
`31
`26
`37
`29
`33
`35
`32
`27
`21
`35
`39
`
`11.02
`16.24
`22.11
`24.11
`25.16
`25.85
`26.55
`27.68
`29.7
`31.19
`31.71
`32.36
`32.86
`33.05
`33.18
`33.25
`33.34
`33.16
`33.16
`33.13
`33.01
`32.94
`32.96
`32.74
`32.64
`32.46
`32.28
`31.93
`31.79
`31.75
`31.83
`31.94
`
`38.23
`37.85
`37.36
`37.12
`37.27
`37.95
`37.67
`37.61
`37.53
`37.45
`37.36
`37.53
`37.8
`37.95
`38.13
`38.21
`38.23
`38.26
`38.23
`38.26
`38.26
`38.15
`38.08
`38.02
`38.06
`38.26
`38.04
`38.08
`37.93
`37.93
`38.02
`37.95
`
`Time is Greenwich Mean Time (GMT). Heart rate is in beats per minute. Activity is in motions per minute. Skin
`and core temperatures are in degrees Celsius.
`
`failed. The two other systems continued to
`function perfectly well at the same time and
`place. The reliability of acquiring location on
`the two functioning GPS systems for continu-
`ous monitoring was 100%. In the third system,
`data acquisition was initially 75% (9 of 12 ac-
`curate locations) in the first hour; however, it
`rapidly degraded to 36% acquisition (18 of 50
`data points) in the next 4 hours, then quit al-
`together. The vital-signs data had a loss of
`transmission rate from 3% to 12%, with the ex-
`ception of one climber who must have had an
`improper affixing of the leads for the heart rate
`and activity monitor because of erratic loss of
`signal with only 56% data acquisition during
`the later descent portion of the trek. However,
`no vital-signs signals were lost for more than
`35 minutes or seven serial recordings. In all
`proper functioning monitors, no signal was
`lost for more than four consecutive readings or
`
`20 minutes. This occurred on only one occa-
`sion.
`
`Vital signs data
`The vital-signs monitors functioning from
`95%–100% of the time, with the exception of
`one climber whose heart rate monitor func-
`tioned 78% of the time. The heart rate varied
`from 76 beats per minute at rest to 176 at stren-
`uous exercise. The climbers had different base-
`line heart rates before the ascent, in the 100 to
`120 beats per minute range, which corre-
`sponded to their resting heart rate at the EBC
`after acclimatization. The Sherpa had a lower
`baseline heart rate of 86–100 beats per minute.
`All climbers experienced an increase in heart
`rate to the 150 beats per minute range, often
`(but not always) correlating to significant in-
`crease in the actigraph which indicated signif-
`
`UA-1007.008
`
`

`

`PHYSIOLOGIC CIPHER AT ALTITUDE
`
`icant exercise. The skin temperature sensor
`functioned extremely well, even during a time
`when there was loss of signal from the tem-
`perature pill (see below). The maximum vari-
`ability of skin temperature of 22.1°C to 34.3°C
`(usually 5–10°C over the duration of a climb,
`which was monitored over 6–9 hours) was
`much greater than the core temperature of
`36.7°C to 39.6°C, but it was usually within
`4–7°C of core body temperature. The one ex-
`ception was when one climber had improper
`position or an open overcoat at the beginning
`of the reascent, when there was abnormally
`low (11–21°C) readings. This was corrected,
`and it functioned well throughout the remain-
`der of the climb. There was no direct propor-
`tionality between the skin and core body tem-
`perature, and there were a few times when
`skin temperature changed in the opposite di-
`rection from core temperature. The core tem-
`perature pill was extremely accurate and fluc-
`tuated only 1–3°C over the duration of the
`climb. There was one climber in whom the pill
`suddenly stopped functioning in the middle of
`giving normal stable readings. It is not known
`if the pill simply stopped working or was pos-
`sibly excreted. An interesting phenomenon
`with the temperature pill is that it is possible
`to tell when the climber is taking a drink of liq-
`uid (either hot or cold) because there is a sud-
`den change (usually 1–3°C) in pill tempera-
`ture. Intermeasurement change was not
`greater than 60.5°C per 5-minute interval,
`even with strenuous activity. The correlation
`between heart rate, activity level, skin tem-
`perature, and core temperature was not con-
`sistent, though there were numerous intervals
`of 10–20 minutes when a sudden dramatic in-
`crease in heart rate was accompanied by a rise
`in actigraph level, skin temperature, and even
`a gradual and persistent increase in core body
`temperature. Unfortunately, the strenuous
`climb did not permit an event recorder to in-
`dicate what any individual climber was doing
`during the sudden change in vital signs.
`
`DISCUSSION
`
`In April 1999, a workshop of Home Care
`Technologies for the 21st Century sponsored by
`the National Science Foundation and the Cen-
`
`311
`
`ter for Devices and Radiologic Health of the
`Food and Drug Administration reported that it
`is “anticipated that health care will migrate to
`a more proactive, preventative model rather
`than reactive, episodic model utilized today
`[with]: Intelligent wearable sensors, trend
`analysis tools, predictive algorithms.5 There
`seems to be a consensus that wearable, wireless
`transmission of health data is commonplace.
`However, there are few reports in this area.
`The ambulatory monitoring of EKG signals
`can be traced to the Holter monitor. However,
`this wearable system stored the data in a belt-
`worn recorder for analysis 24 hours later. The
`concept of a wearable computer that could re-
`ceive and transmit information in real time was
`conceived by Steve Mann of the Massachusetts
`Institute of Technology Media Lab in the late
`1970s, and it developed into what is now rec-
`ognized as “wearable computing.”6 This
`brought to the forefront the possibility of con-
`tinuously monitoring and transmitting medical
`information on patients. In 1986 Shichiri et al.7
`attempted continuous monitoring of glucose
`concentration in five diabetic patients with a
`needle-type glucose sensor that transmitted the
`signals. This prompted the development of a
`wearable glucose monitoring systems by Fabi-
`etti et al.8 and Pfeiffer,9 but difficulties per-
`sisted with signal to noise and reliability of sen-
`sors. In 1991, Mittal et al.4 reported the first use
`of an ingested temperature pill that teleme-
`tered the signals to a belt-worn recorder and
`described its characteristics. This pill acquired
`core body temperature with an accuracy of
`0.04°C over a range of 30–50°C. The ingested
`temperature pill has had a number of applica-
`tions within the research community but has
`not been routinely employed in clinical prac-
`tice. Novel wearable approaches to monitor
`blood lactate by Meyerhoff et al.,10 ethanol by
`Swift et al.,11 and clotting of an arteriovenous
`fistula for dialysis by Shinzato et al.12 once
`again demonstrated the feasibility of such sys-
`tems. Due to several factors, however, includ-
`ing accuracy, power requirement, signal ac-
`quisition, and poor ergonomics, these systems
`have not achieved common clinical use. In
`1997, Richey et al.13 compared ambulatory
`monitored blood pressure measurements to
`traditional techniques in 216 patients. Ninety-
`one percent of the recordings were successfully
`
`UA-1007.009
`
`

`

`312
`
`acquired, and 81% of the measurements were
`interpretable. However, these systems are not
`user friendly. They are cumbersome and larger
`than desired for comfortable use. Their de-
`ployment has only been in feasibility studies or
`unique circumstances. Surprisingly, there have
`been no reports of VSM with remote transmis-
`sion of real-time signals using wireless wear-
`able systems. An interesting test was con-
`ducted on a commercial aircraft in July 1997 by
`Gandsas and Montgomery14 in which a com-
`mercial VSM system (Propaq 106, Protocol Sys-
`tems Inc., Beaverton, OR) connected an EKG on
`a volunteer to a cellular phone on an aircraft
`through a computer interface (with a 4,800
`baud rate). Vital signs were recorded using 3-
`lead EKG, and blood pressure, pulse, respira-
`tion, and oxygen saturation were monitored
`over the Internet “without any corruption of
`data with an average delay time of 15 sec.” The
`signals were transmitted during a flight at
`35,000 feet between Los Angeles and Chicago
`to receiving stations at Saddle Back Memorial
`Hospital (Laguna Hills, CA) and Santojanni
`Hospital (Buenos Aires, Argentina). This clearly
`demonstrated the feasibility (though rather
`cumbersome) of transmitting vital signs from
`remote sites without terrestrial connectivity.
`However, the signal acquisition systems were
`not wearable. There are no subsequent reports
`regarding further research of this type of sys-
`tem or whether the airline industry is using it.
`There are no documented reports of contin-
`uous real-time monitoring of vital signs on an
`ambulatory person in truly remote or haz-
`ardous conditions. Therefore, one of the goals
`of this expedition was to assess a physiologic
`cipher and to validate the technical feasibility
`of real-time monitoring of the vital signs and
`geolocation position of individual climbers in
`a real-world, extreme environment. The system
`was robust, fault tolerant (resampling when a
`GPS signal was not acquired or when a vital
`sign was not detected), and easily monitored
`through the graphical interface. The obvious
`physical stress of hypoxia and strenuous activ-
`ity were seen in the overall elevated baseline
`heart rate. As the climbers ascended, however,
`there were no specific event markers to corre-
`late level of exercise or effort with heart rate.
`The skin and core body temperatures showed
`
`SATAVA ET AL.
`
`different variations. As expected, skin temper-
`ature had a wider variation than did core tem-
`perature, and, at the outside air temperature
`and level of activity, the core body temperature
`was easily maintained and the skin tempera-
`ture was only modestly affected (usually 3–9°C
`less than core temperature).
`There were several episodes of lack of signal
`acquisition. However, the frequent sampling
`(every 15 seconds) provided adequate com-
`pensation when signal acquisition was mo-
`mentarily lost. The values obtained compared
`appropriately with the values derived from the
`person a few days before the climb at EBC. Data
`transmission from the individual to EBC and
`retransmission in real time to Yale University
`validated the concept of remote (or perhaps
`even global) monitoring of individuals. Al-
`though there was no methodology to verify
`that the GPS location of the climbers correlated
`with their exact location, their position at Camp
`1 was accurate. The weather was clear during
`the trek up to Camp 1 and slightly overcast on
`the return down. The effects of severe weather,
`such as a snowstorm, on the quality of trans-
`mission were not determined.
`This expedition exemplifies the extreme con-
`ditions for operating a VSM system over which
`there is nearly no control. There are numerous
`applications of VSM, such as prehospital and
`in-hospital settings, nursing homes, or even as-
`tronauts in space, but more so when the condi-
`tions are not nearly so extreme and the partic-
`ular environment allows for control of factors
`that would otherwise degrade the acquisition of
`signals. Most hospitals can secure an excellent
`communications system for monitoring vital
`signs by a wireless system. As more parameters
`are added to such a system through the imple-
`mentation of noninvasive sensors, the physio-
`logic cipher will become less of a mystery and
`more of a quality control asset.
`
`ACKNOWLEDGMENTS
`
`Support for this research was provided
`through grants from Olympus America, Inc.,
`Yale University–NASA Commercial Space
`Center for Medical Informatics and Technology
`
`UA-1007.010
`
`

`

`PHYSIOLOGIC CIPHER AT ALTITUDE
`
`Applications (NASA NCC5-191), and the Saint
`Charles Hospital in Port Jefferson, New York.
`A review of this project also appeared in the
`Yale Journal of Biology and Medicine (1999;
`72:19–27), and any similarities are uninten-
`tional. The authors received permission to pub-
`lish the current manuscript.
`
`REFERENCES
`
`1. Krakauer J. Into thin air (illustrated ed.). New York:
`Villiard Books, 1998:322–337.
`2. Coburn B, Beshears D. Everest: Mountain without
`mercy. Willard, OH: National Geographic Society
`Press, 1997:76–87.
`3. Satava, RM. Virtual reality and telepresence for mili-
`tary medicine. Comput Biol Med 1995;25(2):229–236.
`4. Mittal BB, Sathiaseelan V, Rademaker AW, Pierce
`MC, Johnson PM, Brand WN. Evaluation of an in-
`gestible telemetric temperature sensor for deep hy-
`perthermia applications. Int J Radiat Oncol Biol Phys
`1991;21(5):1353–1361.
`5. Dighe A, Warren S. Smart health care systems and the
`home of the future. In: Winters J, Herman W, eds. Work-
`shop on home care technologies for the 21st century. http://
`www.hctr.be.cua.edu/hctworkshop/HCTr_F.htm
`(Originally accessed June 23, 2000.)
`6. Mann S. An historical account of the “WearComp”
`and “WearCam” inventions developed for applica-
`tions in “Personal Imaging.” In: Bass L, Pentland A,
`eds. The first international symposium on wearable com-
`puters (Oct 1997). Los Alamitos, CA: IEEE Computer
`Society, 1997:66–73.
`7. Shichiri M, Asakawa N, Yamasaki Y, Kawamori R,
`Abe H. Telemetry glucose monitoring device with
`needle-type glucose sensor: A useful tool for b