Ask this question during an EHR outage at 9 a.m. on a Monday and the answer gets expensive by the minute. A medical practice runs on a small stack of critical software, the EHR, practice management and scheduling, often imaging and a patient portal, and “who supports it” is never one phone number. It is two parties with two distinct jobs, plus a coordination job that belongs to one of them and too often falls on your front desk instead.
Here is how the layers split, why the finger pointing happens, and what changes when one party owns the whole diagnosis.
Who owns what: the three layers of EHR support
Layer one: your EHR vendor owns the inside of the application. How the software behaves, what its error messages mean, how charting templates work, how claims flow to the clearinghouse, when a feature does not work as designed. You pay for this through your license and support agreement, and no IT company should pretend to replace it. A provider who claims deep expertise inside every EHR and practice management product is telling you something, and it is not reassuring. The honest claim is narrower and more useful: own everything the application depends on, and know how to work with the vendor.
Layer two: your IT partner owns the environment the application depends on. This is Braintek’s layer. For a medical practice it covers:
- The server the EHR lives on, or the internet path to the vendor’s cloud
- The workstations in exam rooms, at the front desk, and in the billing office
- The network connecting all of it, wired, wireless, and between locations
- Performance, because “the EHR is slow” almost always traces to this layer
- Document scanners, insurance card scanners, label printers, signature pads, and the rest of the peripherals clinical workflows lean on
- Backups of everything the vendor does not cover, actually tested, not assumed
- Access and accounts: who can sign in, from where, and prompt offboarding when someone leaves, which HIPAA cares about deeply
- Coordinating updates so a new version rolls out evenly instead of taking down half the practice
Layer three: the coordination gap between the first two. This is the layer nobody sells and every practice pays for. When a problem could plausibly live in either territory, someone has to own the diagnosis across the boundary, gather evidence, and get the right people talking to each other. If no one owns it, it defaults to your office manager, who now project manages an outage between two support desks while patients wait. The single most useful thing to settle before anything breaks is that your IT partner owns this layer, explicitly.
When those lines are drawn in advance, most problems resolve on the first call. When they are not, you get the failure mode every practice recognizes.
Why do the EHR vendor and the IT company blame each other?
Because most real problems can plausibly live in either territory, and neither party can see the other’s. The EHR freezes when loading charts. Is that the application, the server or cloud platform behind it, the internet connection, the wireless in that wing, or the one workstation in exam room three? The vendor’s support desk checks their platform and says it looks fine. The IT company reboots the workstation and says it looks fine. Both are telling the truth about their own layer, and the problem sits in the seam.
So the practice becomes the messenger. The office manager opens a ticket with the vendor, relays the vendor’s questions to the IT company, relays the answers back, waits on hold twice per round trip, and the morning schedule slides while nobody is technically wrong and nobody is accountable.
The fix is structural, not technical. One party owns the diagnosis, and it should be your IT partner, because the environment is where most problems start and because they can generate the evidence a vendor engineer actually responds to. “The EHR is slow, please advise” goes to the bottom of the vendor’s queue. “Charts load in two seconds on the front desk machine and thirty in exam room three, both on the same version, here is what differs between them” gets an engineer on the phone. If your current IT company hands you the vendor’s number and closes the ticket, that is the arrangement to fix, and it costs you every time something breaks. What that ownership looks like day to day is covered on our healthcare IT support page.
Hosted or on premises: where should the EHR live?
Both models are common, many practices run a mix, and the placement decides what the IT layer covers, not whether you need one.
On premises means the EHR and its database sit on a server in your office. Speed is local and predictable, and the day does not depend on your internet connection. The costs are ownership costs: the server ages, backups are your responsibility, patching and monitoring are constant, and every five to seven years there is a replacement project. On premises also complicates a second location and remote access for providers finishing charts from home, both of which have to be built securely rather than assumed.
Hosted or cloud means the vendor runs the platform and your practice reaches it over the internet. The aging server disappears, a second location gets dramatically simpler, and platform updates are largely the vendor’s problem. In exchange, every exam room now depends on your internet connection, your firewall, and your wireless, all day, every day. A practice on a cloud EHR should treat the internet connection like a clinical utility: sized for the load, monitored, and backed by a second path from a different carrier, because when the connection drops, the entire clinical day drops with it.
Either way, the environment layer remains. Workstations still have to run well in every exam room, scanners still have to scan, access still has to be controlled and logged, and the data the vendor does not back up, email, local documents, imaging, still needs protecting. We cover that split in detail in what a medical practice actually needs to back up. A practice weighing hosted against on premises should decide with someone at the table who does not earn more from either answer.
The exam room is where EHR support gets physical
EHR support conversations tend to happen in the abstract, servers, cloud, network, and then the actual failure is a signature pad in exam room two. Clinical workflows run through hardware: the workstation or cart in each exam room, the document scanner digesting referral paperwork, the insurance card scanner at check in, label and wristband printers, sometimes interfaces to lab or imaging equipment.
Two realities follow. First, peripherals fail more often than servers, and each one interrupts a specific step of patient flow, check in backs up, or charting stops in one room while the schedule keeps moving. Second, these devices sit exactly on the vendor seam: the scanner is IT’s hardware running the vendor’s integration software, so a failed scan is a coin flip between a driver problem and an application problem. This is precisely where the coordination layer earns its keep. Your IT partner should stock the answer to “is it the device or the software” as a routine diagnostic, not a week of back and forth.
Exam room hardware is also a quiet HIPAA surface: screens visible to patients, devices logged in and unattended, old workstations retired with drives full of cached patient data. The environment layer owns those exposures too, which is one reason the IT relationship in a practice carries HIPAA obligations of its own, including a signed BAA.
Uptime during clinic hours is the whole job
A law firm can absorb an hour of downtime by shuffling work. A clinic mostly cannot. Patients are physically present on a schedule, providers document in real time, and prescriptions, orders, and referrals flow through the system as the visit happens. When the EHR is unreachable, the practice either sees patients blind or stops seeing them.
That changes what good IT support looks like in three ways:
- Maintenance lives outside clinic hours. Updates, reboots, and network changes happen evenings or early mornings as scheduled changes, never as daytime surprises.
- Monitoring catches problems before the doors open. A failing drive, a saturated internet connection, or a switch throwing errors should page your IT partner at 6 a.m., not announce itself at 9.
- Downtime has a rehearsed procedure. Not every outage is preventable, so the practice needs a documented answer to “how do we keep seeing patients on paper, and how does that documentation get back into the system afterward.” An IT partner should help build that procedure, because it fails at the technical seams, printed schedules, offline access to key information, and the re-entry plan.
Response time matters here more than in most industries. Braintek typically answers in about a minute and starts working urgent issues within minutes, not hours, and while no honest provider guarantees a number, a provider whose typical response is measured in hours is the wrong fit for a clinic. Our guide to what IT response times actually mean breaks down how to benchmark this fairly.
What Braintek owns, and what your EHR vendor owns
The honest division, in one place:
Your EHR vendor owns: the application and its features, clinical templates and workflow configuration, the hosted platform if you are on their cloud, application bugs and fixes, claims and clearinghouse behavior inside the software, and application training.
Braintek owns: the server or the connectivity to the vendor’s cloud, every workstation and exam room device, the network and wireless, scanners and peripherals and their drivers, backups of everything outside the vendor’s platform, user accounts, access control, and offboarding, security across the environment, update coordination, and, critically, the diagnosis when something breaks and the vendor relationship during it. Your staff reports a symptom once; we work it to the vendor boundary and bring the vendor a narrowed question with evidence.
Nobody should ask your staff to own: relaying messages between support desks, deciding whose fault a problem is, or absorbing the environment layer informally because the office manager is “good with computers.”
Who does this for medical practices in Houston and DFW?
Braintek has supported Texas businesses since 2002 with local teams in Houston and DFW, including the servers, cloud connectivity, exam room workstations, and networks that EHR and practice management systems depend on. We own the environment layer, sign the BAA, coordinate with your EHR vendor so your staff never plays telephone between support desks, and schedule changes around clinic hours instead of through them. Fully managed support fits practices of roughly 10 to 50 people; larger groups with internal IT usually fit co-managed support, where your staff keeps the day to day and we cover security, projects, and escalations. Independent practices can start with our dedicated medical practice page, and if you are comparing providers, our guide to choosing an MSP for a medical practice lists the questions worth asking, including the ones about vendor coordination that most practices learn to ask one outage too late.
If the current arrangement leaves your practice refereeing between the EHR vendor and an IT company, that is the first thing worth fixing, and it is a structural fix, not a software change. Tell us what your practice runs and where it lives, and we will show you where the lines should be drawn.