Your clinic may already have an EMR, online forms, automated reminders, a phone system, digital fax, and several other tools. Yet staff still copy information between systems, chase missing details, answer the same questions, and rely on memory to keep work moving.
The immediate reaction is often to add another tool.
That is rarely the best first step.
Most clinics do not need to automate everything. They need to identify the small number of repeated tasks, unclear handoffs, and unreliable processes creating most of the daily drag.
The 80/20 idea is useful here as a practical question, not a mathematical promise:
Which small group of workflows is consuming a disproportionate amount of the team’s time and attention?
Once you can answer that question, you can make a better decision about whether the clinic needs a clearer process, better system configuration, staff training, an integration, controlled automation, or a different tool.
More software is not automatically an operational improvement
A new tool can make a good workflow faster. It can also make a poorly defined workflow harder to understand.
Imagine a clinic where referrals arrive through several channels. One person checks the fax inbox, another receives email attachments, and physicians sometimes send instructions through internal messages. Nobody owns the complete queue.
Adding an automated referral tool may help, but it does not answer the operational questions:
- What starts the referral workflow?
- Who confirms that the required information is present?
- Where is the current status recorded?
- Who follows up when something is missing?
- What counts as complete?
- Who handles exceptions?
If these questions remain unanswered, the clinic may automate part of the work while preserving the confusion around it.
In implementation work, I have seen clinics look for a new system when the more immediate problem was an undefined handoff. The useful first move was not selecting software. It was making the work visible, assigning ownership, and agreeing on what completion meant.
Step 1: Map what actually happens
Do not begin with the workflow described in a policy manual. Begin with what staff and clinicians actually do on a normal day.
Choose one workflow and write down:
- Trigger: What starts the work?
- Steps: What happens from beginning to end?
- Roles: Who touches the work?
- Systems: Where is information entered, viewed, or moved?
- Handoffs: When does responsibility pass to someone else?
- Exceptions: What changes when information is incomplete or the usual path fails?
- Finish line: How does the team know the work is complete?
The AHRQ Workflow Assessment for Health IT Toolkit similarly emphasizes understanding clinical and administrative workflow when planning or improving health-information technology.
Keep the first map simple. A page, whiteboard, or shared document is often enough. The purpose is not to create a perfect diagram. It is to expose where work waits, repeats, disappears, or depends on one person’s memory.
Step 2: Find the drag
Review the map with the people who perform the work. Look for patterns such as:
- the same information being entered more than once;
- staff switching repeatedly between systems;
- work waiting because ownership is unclear;
- messages being used as an informal task list;
- staff searching for the current status;
- predictable exceptions handled differently each time;
- one employee holding knowledge that is not documented;
- clinicians completing administrative steps that another role or system could support;
- work that is technically finished but not visibly closed.
The goal is not to criticize the team. Workarounds usually develop because people are trying to keep the clinic functioning. The map helps convert those workarounds into improvement opportunities.
Three workflow areas worth inspecting first
Every clinic is different, but three areas commonly reveal avoidable administrative friction.
1. Intake and check-in
Possible friction includes incomplete information, paper forms that must be retyped, repeated identity checks, or staff calling patients for details that could have been collected earlier.
Ask: Which information do we request more than once, and why?
The answer may lead to a clearer intake form, better use of an existing portal, a configuration change, or a different sequence. It does not automatically require new software.
2. Documentation and task completion
Possible friction includes unfinished notes, inconsistent templates, unclear responsibility for orders or referrals, and repeated corrections after the visit.
Ask: Where does the visit workflow remain open after the patient leaves?
This may reveal a need for template improvement, role clarification, training, EMR workflow optimization, or a controlled AI-scribe pilot. The intervention depends on the cause.
3. Follow-up, referrals, results, and patient communication
Possible friction includes multiple queues, missed status updates, repeated callbacks, and uncertainty about who closes the loop.
Ask: Can anyone on the team see what is waiting, who owns it, and what happens next?
If not, the first intervention may be shared workflow visibility and clear ownership rather than automation.
Step 3: Measure the current baseline
Without a baseline, almost any new process can feel faster during the first week. It is new, people are paying attention, and the pilot group may be unusually motivated.
Before changing the workflow, collect a small amount of usable information:
- number of items entering the workflow each week;
- approximate staff time per item;
- average waiting time or backlog;
- number of handoffs;
- frequency of rework or corrections;
- common exceptions;
- percentage completed within the expected time.
The measurement does not need to be perfect. It needs to be consistent enough to compare the old and new process.
A Doctors of BC Practice Support Program case study describes a clinic improvement cycle that included reviewing roles and communication, assessing EMR functionality, defining an aim, and comparing baseline and post-intervention measures. The sequence matters: understand the workflow, define the change, then measure whether it helped.
Step 4: Choose the right type of intervention
Not every workflow problem is an automation problem.
| What you observe | A reasonable first intervention |
|---|---|
| Different staff perform the same task differently | Define the process and train the team |
| The EMR already supports the desired step, but the team does not use it consistently | Improve configuration and adoption |
| Information must be copied between systems | Examine integration or redesign the workflow |
| A stable, repetitive task consumes predictable time | Consider a controlled automation pilot |
| Exceptions are frequent and require judgment | Keep human review and simplify the process first |
| Nobody knows who owns the next step | Clarify responsibility and the completion rule |
This distinction protects the clinic from purchasing technology that solves the wrong problem.
It also makes vendor conversations more useful. Instead of asking, “What can your platform automate?” the clinic can ask, “Can your platform support this specific workflow, under these conditions, with these exceptions?”
Step 5: Run a focused 30-day pilot
Choose one workflow with meaningful friction and a manageable scope.
Define:
- the exact workflow being tested;
- the people participating;
- the process or technology included;
- actions that remain outside the pilot;
- who owns the pilot;
- what happens when an exception occurs;
- the baseline measures;
- weekly review points;
- the decision date.
Avoid testing several workflows at once. If intake, documentation, scheduling, and referrals all change together, it becomes difficult to understand what improved and what created new problems.
A focused pilot also reduces disruption. Staff can learn one new process, identify practical exceptions, and improve it before the clinic expands the change.
Step 6: Keep, adjust, expand, or stop
At the end of the pilot, compare the results with the baseline.
Ask:
- Did the workflow require less staff effort?
- Did waiting time or backlog improve?
- Was there less rework?
- Did staff follow the new process consistently?
- Did the intervention create new exceptions or safety concerns?
- Is ownership clearer?
- Can the process continue without constant supervision?
The decision does not need to be “successful” or “failed.”
There are four useful outcomes:
- Keep: The workflow is more reliable and can continue.
- Adjust: The direction is useful, but the process needs refinement.
- Expand: The pilot is stable enough to extend to another team or workflow.
- Stop: The intervention does not fit the clinic or creates more work than it removes.
Stopping a poor-fit intervention is a valid result. The goal is not to automate more. The goal is to create a clinic workflow that people can follow and leaders can evaluate.
Start with one workflow
This week, choose one area where the team repeatedly feels friction.
Map the trigger, steps, roles, systems, handoffs, exceptions, and finish line. Then ask where the work repeats, waits, disappears, or lacks a clear owner.
That exercise may reveal a technology opportunity. It may also show that the clinic can improve the workflow using tools it already has.
If your team knows where the friction is felt but not where it starts, a Clinical Operations Assessment can help map the workflow, identify the most important bottlenecks, and establish a practical improvement roadmap.
