Awareness training assumes your team catches the fake. Good programs plan for the morning someone doesn’t. What follows is what to do after you click on a phishing link at a medical practice: the machinery that contains it layer by layer, and the four steps that run if it gets past all of them.
Key takeaways
Training lowers how often a click happens. It doesn’t change what has to happen the moment one does, and that second part is where most practices are thin.
- Layered security means a fence, a front door, and an alarm system rather than a single wall. Each layer catches what the one before it missed, so none of them has to be perfect.
- The stack proves its value after the click, not before it. A compromised email account should surface as a blocked sign-in and a ticket within minutes.
- Two doors sit outside the inbox entirely: your identity layer, and the vendors who already hold your patient data.
Start where the training ends

Most phishing guidance stops at the click. Spot the sender, hover the link, don’t enter your password. Careful people still click on a distracted morning, because that layer gets tested every single day and eventually loses. Your front desk should know the basics cold, and we cover them in phishing detection for busy clinical staff.
So the question that decides whether a practice has a breach isn’t “will someone click?” Someone will. It’s “what is already in place the second they do?” Everything below assumes the email got through and the click already happened.
What the attacker is really reaching for
You get the phishing email. It looks fine. You click it to fill out your information, and just like that an attacker has the keys to the castle. They can log into your account and your environment. And here’s the quiet part. That risky sign-in from halfway around the world can go unnoticed for a week, because the attacker sets up rules in your inbox to hide their tracks. Nothing looks wrong while they work.
What they’re after isn’t the mailbox. It’s what the mailbox opens onto: protected health information, and your whole contact list, so they can turn around and run the same attack on your colleagues, your patients, and everyone you refer to. Each fake gets more convincing, because now they know your relationships. That’s the social engineering chain, and in healthcare it ends in a breach, with the cost, the notifications, and the hit to patient trust that come with it.
Hold onto that gap between the click and the discovery. A week of undetected access is the difference between an incident and a catastrophe, and the layers underneath the inbox exist to close it down to minutes.
Why one control is never enough

Here’s how I’d frame the whole approach. No single control has to be flawless, because the one behind it is there to catch what it missed. A phishing attack is the clearest place to watch that work, since you can trace one click through the whole stack and see each layer get its chance to stop it.
Trained staff alone vs. staff plus layered controls
Same click, two practices. What changed isn’t the person who fell for the email. It’s what was standing behind them when they did.
| When a staff member has clicked a phishing link | Training only | Training plus layered controls |
| Time to detection | Days to a week of silent access while inbox rules hide the intruder | Minutes. The impossible sign-in is blocked or the anomaly flags the account |
| Patient data exposure | Full reach: records, contact list, referral relationships | Contained to the compromised account before it moves laterally |
| Stolen credentials from a dark-web dump | Work. Nothing behind the login checks who is using them | Dead end. Identity and anomaly layers reject the impossible session |
| What the click becomes | A breach: notifications, cost, lost patient trust | A blocked sign-in and a ticket |
| When someone clicks again | Same outcome, every time | Same non-event, every time. The layers don’t get tired |
The right column doesn’t describe a better-trained staff. It describes an outcome that no longer depends on the click.
How a single click gets intercepted
The first layer keeps the email out. Good mail filtering stops most phishing from ever reaching an inbox. But some always gets through, and someone always clicks, so there’s a second layer waiting.
The second layer is identity. When someone signs in from your office and then, two minutes later, from another continent, that’s impossible, and the system blocks it. You physically can’t be in two places at once, and the sign-in knows it. Most practices have never looked at this layer, which is a shame, because it’s what turns a stolen password into a dead end.
The third layer is anomaly detection. If the first two somehow miss it, another system is watching the odd patterns around your environment, the small correlations that don’t add up. It can interrogate what’s happening and shut the account down. Because it watches behavior rather than credentials, it’s what closes the week-long gap.
That’s the whole design, and it’s why we don’t put our faith in one tool, including the trained person reading the email. A click that would have been a catastrophe becomes a blocked sign-in and a ticket. It’s the same defense in depth thinking that protects a practice at the organization level, applied to a single click.
What each layer is actually called, and how to check you have it
Everything above describes the layers by what they do. Here’s what they’re named on an invoice or a service summary, and the one question that tells you whether your practice actually has each one. You don’t need to know how any of it is configured. You need to be able to ask, and to recognize a real answer.
| Layer | What it’s called | The question that confirms it |
| Keep the email out | Email filtering, link protection, attachment sandboxing | Are links checked at the moment someone clicks them, or only when the message is delivered? |
| Identity | Multi-factor authentication, conditional access, impossible travel and risky sign-in policies | What happens when one of our accounts signs in from a country we don’t operate in? |
| Behavior | Anomaly detection and managed detection and response | Who sees that alert at 2am, and what are they authorized to do without calling us first? |
| Exposed logins | Dark web credential monitoring, external attack surface monitoring | When was our domain last checked against known breach dumps, and who got told? |
| Vendors | Third-party risk assessment | Which of our vendors hold patient data, and when was each of them last reviewed? |
Two of those deserve a note. Implementing multi-factor authentication is where most practices stop, and MFA on its own still leaves gaps that conditional access closes, so “we have MFA” is a starting answer rather than a finished one. And if the answer to any of the five is a name rather than a process, that’s the gap. A tool nobody watches is a line item.
If you can’t get clean answers to all five, that’s what a managed security program is for.
The two doors outside the inbox
The layers above all sit around email and identity. Two more sit outside it, and practices tend to forget them because nothing about them ever shows up in an inbox.
The first is your own exposed logins. Your staff’s emails and passwords may already be sitting on the dark web from some other company’s breach. An attacker who buys a working password doesn’t need to phish anyone. So part of a real program is watching those spaces for your exposed credentials, and watching your internet-facing systems for the openings attackers probe first.
The second is third-party risk. The vendors and partners who touch your systems or your patient data are part of your attack surface, and if you hand patient data to someone and they get breached, that lands on you. So a real security program includes evaluating how your third parties protect the data you share with them. In practice that means knowing which vendors hold what, asking them for evidence rather than assurances, and revisiting it on a schedule instead of at renewal.
The same door, a different disaster: ransomware
Everything above follows the branch where an account takeover stays quiet and an attacker reads your data. There’s a second branch off the exact same click. Instead of sitting inside your account, an attacker encrypts everything and holds it hostage. The EHR, the files, the schedule, all of it locks up with a ransom demand attached. For a practice, that means doors closed and patients turned away until it’s resolved.
The entry point is identical and so is most of the defense. The same layers that stop a quiet takeover, keeping the email out, catching the impossible sign-in, spotting the anomaly, are what keep a single click from becoming a practice-wide lockup. The one addition specific to this branch is backups you can truly restore from, tested on a real restore rather than a green checkmark in a console. Whether a practice should ever pay is its own decision, and one worth settling while nobody is staring at a demand.
What to do after you click on a phishing link: the first thirty minutes

Assume every layer above is in place and a click still slips through the gaps. What separates a scare from a breach now is what you decided in advance. Phishing incident response is only fast when it’s already written down, and the practices that come through cleanly are the ones who wrote these four steps before they needed them.
- Isolate the account. Cut off the compromised login immediately so the attacker loses access while you work. Speed here is what limits how far they reach.
- Reset access. Change the password on that system and anywhere the same credentials were reused, and revoke active sessions so a reset alone can’t be walked around.
- Assess the reach. Determine what the account could touch: which records, which systems, which contacts. This is what tells you whether you have an incident or a reportable breach.
- Notify on a plan you already wrote. If patient data was exposed, move into the breach notification process you drafted in advance, not one you improvise during the worst hour of the month. If you’re unclear on where security ends and the reporting obligation begins, that line is worth reading up on before you need it: HIPAA compliance vs cybersecurity.
The layers reduce the odds you ever run this. Written down and assigned to named people, it takes thirty minutes. Improvised, it takes the rest of the week.
Questions we get about this
An employee clicked a phishing link. What happens now?
That depends entirely on what was already running behind them. With layers in place, the sign-in gets blocked or flagged and you work a ticket. Without them, an attacker may be reading mail quietly for a week. Either way, run the four steps above and don’t wait to be certain before you isolate the account.
We already train our staff to spot phishing. Isn’t that the defense?
Training is one layer of it. It lowers how often a click happens but doesn’t change what happens when one does. A trained team plus a stolen password from the dark web still gets an attacker to your front door, and the identity and anomaly layers are what decide whether they get through it.
We have spam filtering. Isn’t that enough?
No. Filtering is the first layer, and some phishing always gets through it. That’s why identity checks, anomaly detection, and a monitored environment sit behind it, so the mail that slips past the filter still has three more things to get past.
What really happens if an attack gets through anyway?
A break-in becomes a contained incident instead of a catastrophe. The practices that recover fastest tend to be the ones who had a written response plan sitting in a shared folder, with names next to each step, before anything happened. It’s the same discipline behind the four things every medical practice has to protect.