In 1996, Ochsner Clinic Foundation in
New Orleans shared a problem with many hospitals and clinics.
"All of our patient information was wherever the mainframe terminal or
PC was sitting," said Kurt Induni, Ochsner’s network services manager.
"The information wasn’t anywhere near the patients."
As a result, nurses and other health care professionals shuttled back and forth
between patients and terminals to gather and enter information. "Who knows
how many errors came from that situation?" Induni wonders now.
In 1996, however, the head of Ochsner’s emergency department agreed that wireless
local area networks (WLANs), then a relatively new technology, could solve that
"In the Emergency department, it allowed admissions clerks to get the
patient in a bed before they got all the information like insurance information,"
Induni said. "We finally got support for the vision of getting the information
systems and the patient closer together."
Now, WLANs are deployed on all patient care floors in the eleven-story hospital,
the new emergency and critical care facility and 43 clinics. However, while
Ochsner’s use of WLANs is a success, it wasn’t easy getting there, according
Expect the Unexpected
In the first of three stages of implementation, Induni said he used the initial
iteration of 802.11, which provided, at best, 2 Mbs throughput. The slow throughput
was an issue, but the real problems were the ones he and his team didn’t anticipate.
"Initially, we used stainless steel surgical equipment carts to roll the
laptops," he said. "They had a three-quarter inch lip all the way
around. It took a while, but we learned that our spotty coverage was related
to the cart — it was causing interference."
The most perplexing early problem, though, was the sporadic dropping of connections.
Induni said he and his staff examined a number of solutions and grasped at more
than a few straws.
"I was sitting outside one day looking at the (Mississippi) river, wracking
my brain about the problem and a boat went by and I saw its radar dish,"
Induni said laughing. "I wondered if that wasn’t the problem."
It wasn’t. Eventually, he and his staff figured out the interference was caused
by the location of the access point antennas, which they had hidden from view
above the ceiling tiles. The metal infrastructure of the ceiling caused the
Then, there was the problem of dead spots that changed from floor to floor
and from time to time — even though the hospital floors all had the same layout.
"Finally, it came to me that hospitals are extremely mobile environments,"
Induni said. "Patients move, big food carts roll down the hallways and
they stop. There’s a lot of stainless steel equipment." The ever-changing
dead spots were sporadic because of that fluidity.
As a result, Induni and his team decided that it was time for phase two, which
meant running a professional site survey and deploying equipment based on the
then-new 802.11b standard.
Stages Two and Three
Unfortunately, Induni said, "the first survey didn’t help as much as we
had hoped because of the nature of the hospital. The environment the person
surveyed changed the minute he walked out the door."
So Induni and his staff continued to brain out proper placement of access points
as they expanded the WLAN to more parts of the facility and migrated to the
Aironet 340 series of 802.11b equipment.
Eventually, they understood that to ensure reliable coverage they had to install
more access points than they logically needed. They did so and the second stage
of the deployment was finished in 2000. By then, however, the size and nature
of the network had created yet another major problem.
"Our IP segmentation was that each floor was its own LAN segment,"
Induni said. "So an IP address for the eleventh floor wouldn’t work on
the other floors. That’s obviously no big deal for wired, but wireless devices
tend to roam."
Induni said he couldn’t use Dynamic Host Configuration Protocol (DHCP)
the mainframe required a fixed IP address on each device. That problem, along
with the construction of a three-story building for the emergency department
and intensive care unit, led to the third stage of the institution’s WLAN deployment.
To solve the segmentation problem, Induni said his network engineers started
using new features in their Cisco equipment such as auto channel selection.
That, combined with what Induni called "super-netting" — joining
four network segments into one big one — enable as many as 1,000 devices to
access a single LAN segment. That solution, along with some re-segmenting of
the LAN, now enables mobile users to move between floors, he said.
Currently, there are about 165 access points in both of Ochsner’s main buildings
accessed by between 500 and 600 WLAN-enabled laptops, Induni said. In addition,
the remote clinics also are WLAN-enabled, he said.
The original need — getting patient data closer to where the care is given
— has been met and that, Induni said, is the true return-on-investment. "It’s
been an investment in better patient care. It’s not about saving money; it’s
about doing our job better."
Induni acknowledged that deployment issues remain. Near the top of his list
are the inter-related problems of security and the Health Insurance
Portability and Accountability Act (HIPAA).
At this writing, the federal government was about to release a wide-ranging
set of HIPAA regulations mandating that specific types of security be implemented
by organizations that handle patient medical records. It wasn’t certain what
those regulations would require, but based on previous drafts of the rules,
it was certain they’d be complex. Induni said that he would need to take a broad
approach to security issues.
"We need to deal with it (security) as an enterprise, not just look at
a single (WLAN) system," he said. "If we approach it as a single system,
we’ll be here until 2020 figuring it out." He won’t have that long — the
federal government will give institutions such as Ochsner about two years from
the time the rules are released to comply.
Another issue, but one that is less pressing, is how to handle the ever-changing
set of wireless standards. Ochsner is standardized on 802.11b even though 802.11a
products with more bandwidth are available with 802.11g expected later this
"I don’t have any major bandwidth issues, so I’ve made the decision to
hold off on migrating," he said. When he does switch, it will probably
be to 802.11g because of its compatibility with 802.11b since the two standards
operate in the same frequency range. "I’m interested in ‘g’ because it’s
compatible. I could deploy it and not lose my current investment."
While Induni said he’s satisfied that the institution’s WLAN, he understands
that challenges remain.
"I’m watching carefully," he said.