Web site surveys can be amusing, and one on Mobile Health Data's site
today caught my attention. Here's the set up (it's not really a
question, is it?):

Enabling
medical devices, such as infusion pumps, electrocardiograph machines
and glucometers, to wirelessly send data to a patient's medical record
or to a physician:

Why wireless? Well, unless you're in the OR or ICU with
gorked out patients and dedicated devices, or want to put a hub or switch in every patient
room for hardwired Ethernet connections, you're going to go wireless.
Wireless best fits the inherently mobile nature of hospital care: medical devices
connected to patients, care givers and support staff who constantly
circulate, and patients who transit through (too many) care units and
departments during their stay.

When thinking of integrating medical devices, the first two I think of
are spot vital signs monitors and continuous patient monitors. Patient
vital signs are taken several times daily. The problems with manually
documenting vitals are legend; indecipherable handwriting, data in the
wrong chart, vital signs written on scraps of paper (hands, scrub
suites, etc.) that get forgotten, and the sometimes considerable lag
between readings and when they get into the chart. This is a problem;
now add an EMR -- now everything must be written down and then typed
into the electronic chart, and cries of "double entry" and "extra
steps" arise. So, spot vital signs monitors probably represent the
greatest amount of data that makes it into the patient's chart.

Continuous patient monitoring is being used more and more outside the
ICU and Telemetry. One driver is the increased emphasis on pain
management. When therapeutic intensity is increased (in this case increased pain
meds), there must be a related increase in surveillance. Reducing nurse
to patient ratios is one way to do this, as is adding ECG and SpO2
monitoring. Hospitals that have foregone increased surveillance have experienced sentinel events (PFD file). Another driver is what's called the variable acuity care
model (aka, flexible monitoring, universal bed, flex bed). This care
model keeps the patient on their medical service unit and varies the
nurse-to-patient ratio and related therapies and surveillance
(monitoring) as the patient's condition changes. Typical ranges of
acuity go up to high dependency or Telemetry units, just short of the
ICU. Hospitals are using this approach to relieve patient flow
bottlenecks in the ICU and Step-down or Telemetry units. Finally, some
hospitals are making wireless patient monitors available in units
outside the ICU simply to improve patient safety.

Wireless diagnostic devices, especially the two mentioned in the
survey, are great examples. Wireless enablement of these devices gets data into the chart much quicker, without the risk of data
entry errors.

Sending medical device data to physicians is a non-starter; an alert
about that out of range lab test yes, the 15 arrhythmia alarms the
patient had in the last 45 minutes, no. The unsung hero in all this is
the person directly responsible for the patients care and safety 24/7,
the caregiver. Communications at the point-of-care is incredibly
chaotic and disjointed. Many life critical alarms are still audible
alarms annunciated solely at the bedside by the device itself. Much of
the coordination of care (especially when patients start to crash) is
done with unreliable pagers, overhead pages and phone calls to
roles ("on-call respiratory therapist") rather than individuals.
The amount of meaningful automation and technology to support
caregivers is miniscule compared to that provided for patient
accounting, medical records and diagnostic departments. The adoption of
an EMR on its own will make little difference. Is it surpising that
past and current IT investments have had little impact on patient
safety?

Here are the survey choices (the results when I took the survey are in parentheses):

  • Still too expensive to implement in routine clinical care (5%)
  • Doesn't offer a secure enough mode of transmission for clinical data (10%)
  • Could help save time and improve patient safety (50%)
  • Could also be used to help monitor patients from their homes (35%)

In the scheme of things, the cost to integrate medical devices is but a
rounding error in measuring the cost of an EMR. However, there are
several hidden costs. Some vendors insist hospitals provide separate
network infrastructures for their medical devices. Medical devices
communicating "life critical" data are a new application for most
hospital IT shops, but the SLAs required for an EMR are virtually
identical to those required for medical devices. Getting IT to the
place where they offer up-time, latency and support coverage for "life
critical" applications is part of good EMR implementation planning.

If people bought cars like most hospitals buy medical devices, we'd all be driving much
older cars. Consequently, when looking to integrate devices many
hospitals are told that the older devices or system they have must be
upgraded to support HL7 integration. Ca-ching. Another hidden cost is
the proliferation of point solutions. If you're not careful, you can
end up with multiple infrastructures, HL7 interfaces, and numerous
gadgets for your nurses. Fortunately there are enterprise solutions to
these challenges, they're just not widely known.

Protection of ePHI
(electronic protected health information, i.e., patient identifiable
data) is far from insurmountable. Although HIPAA rules for both privacy
and security apply to medical devices,
most hospitals are not in compliance. Here's a gut check question: has
your facility had all of your medical device vendors complete a Manufacturer Disclosure Statement
for medical device security? Ever heard of such a thing? If the answer
to either question is "no" you've got some homework to do. (Let me know
if you need any help...)

It was heartening to see that the majority of survey-takers recognize the benefits of medical device integration.

Finally, home monitoring is a completely different kind of application
from EMR integration. Rather than caregivers as users, home monitoring
must be patient friendly. The available networks and their technical
characteristics are considerably different from hospital networks.
There are also different parameters that must be monitored that are
managed differently in the hospital (e.g., the weight of a patient with
chronic heart failure). Finally, home care applications are chronic
disease monitoring systems that provide patient feedback, patient
monitoring and alerts (to both patients and physicians). If there are
any systems that are sold and used in both inpatient hospital and home
care I'd like to know about them.

Medical device integration is a critical (and often overlooked) part of EMR planning. To be
successful, any plan must take into account many more
considerations beyond getting an HL7 feed into the EMR. Multiple
stakeholders include nursing and biomedical engineering must be engaged. Putting
together a successful plan requires going across traditional hospital
silos, and an in depth understanding of point-of-care workflows, medical device connectivity and
device vendor offerings and product strategies. In this area, there is
no such thing as a "one shot" plan; multiple phases over a period of
years will be required for you to reach your ultimate goal. And great
care must be taken to avoid "throw away" purchases and to ensure that
caregiver productivity is enhanced rather than degraded. If you'd like
a little help, let me know.

Take the survey yourself, and let me know what the results were when you took it.