The Unison Score: Moving Beyond Adoption Rates to Measure True Provider Engagement

When a new electronic health record module or clinical decision support tool goes live, the first question asked is almost always: What's the adoption rate? It's a clean number, easy to report upward. But anyone who has watched a provider log in just enough to satisfy a mandate, then immediately revert to paper or a shadow system, knows that adoption rates can be a comfortable lie. The Unison Score offers a way to see past the login count and measure what actually matters: whether providers are using the tool in a way that changes their work for the better.

This guide is for clinical informaticists, implementation managers, and quality leaders who have felt the gap between what adoption dashboards show and what's happening on the ground. We'll define the Unison Score, show how it works, and walk through the trade-offs of moving to a more nuanced measurement system.

Where the Adoption Rate Deceives

Adoption rates typically measure a binary event: a user logged in at least once during a reporting period, or they didn't. That's useful for tracking whether a system is being accessed at all, but it tells you nothing about the quality of that access. A provider who clicks one checkbox and closes the application counts the same as one who uses the system to document a full encounter, check drug interactions, and review patient history.

We've seen teams celebrate 90 percent adoption rates only to discover, through shadow audits, that half of those logins were performed by medical assistants or residents logging in on behalf of the attending. The attending never touched the system. The adoption number looked great; the actual engagement was near zero.

The problem with binary measurement

Binary metrics create perverse incentives. When adoption is the key performance indicator, organizations push for sign-ins, often by any means necessary. Auto-login scripts, mandatory but meaningless clicks, and badge-swiping rituals all inflate the number without improving care. Worse, these behaviors train providers that the system is something to be gamed, not a tool to be used.

Why engagement matters more

True engagement means a provider actively integrates the tool into their clinical reasoning. They use it to inform decisions, catch errors, and save time over the long run. Engagement correlates with better outcomes, fewer workarounds, and lower burnout. The Unison Score attempts to capture this multidimensional reality.

What the Unison Score Actually Measures

The Unison Score is a composite metric that combines four weighted dimensions: frequency, depth, sentiment, and integration. Each dimension is scored on a 0–100 scale, and the final score is a weighted average. The exact weights depend on the context—a system meant for quick reference should weight frequency differently than one meant for complex order sets.

Frequency (25–40 percent weight)

This isn't just whether the provider logged in, but how often per shift or per patient encounter. We recommend measuring active sessions (where the provider interacts with the system for more than 30 seconds) rather than raw logins. A provider who opens the system once per patient encounter scores higher than one who opens it once per shift.

Depth (30–50 percent weight)

Depth measures which features are used. A provider who only views lab results has shallow depth. One who also documents, orders, uses clinical decision support, and reviews medication histories has deep engagement. Depth can be scored by counting distinct feature categories used per session, normalized against the roles available to that provider type.

Sentiment (10–20 percent weight)

Sentiment comes from brief, anonymized pulse surveys embedded in the workflow. A single question like How useful was this tool for your last patient? with a 1–5 scale, delivered after a key action, gives real-time feedback. Sentiment scores are averaged over a rolling 30-day window. Low sentiment can flag a tool that works technically but frustrates clinicians in practice.

Integration (15–25 percent weight)

Integration measures how seamlessly the tool fits into the existing workflow. This is assessed through observation sampling or time-motion studies: does the provider have to switch screens, re-enter data, or perform extra steps? A high integration score means the tool reduces cognitive load; a low score means it adds friction, regardless of how often it's opened.

Building Your Own Unison Score: A Practical Framework

Implementing the Unison Score doesn't require a data science team, but it does require thoughtful design. Here are the steps we recommend based on work with several health systems.

Step 1: Define the use case and provider roles

Not every tool needs the same score. A medication reconciliation module might weight depth and integration heavily, while a communication platform might weight frequency and sentiment. Map out the primary use case and the roles involved (physicians, nurses, pharmacists) before setting weights.

Step 2: Instrument the system for granular logging

Work with your IT team to log not just logins but specific actions: which screens were viewed, which buttons were clicked, how long the session lasted, and whether the session ended with a completed task. Most modern EHRs have audit trails that can be repurposed for this. If you need to build custom logging, focus on the actions that matter most for depth and integration.

Step 3: Deploy lightweight sentiment collection

Avoid long surveys. One question, delivered at the point of use, gets far higher response rates. Use a 1–5 scale and keep the question consistent so you can track trends. For example: How well did this tool support your decision for this patient? A response below 3 triggers a follow-up (optional, open-ended) to capture context.

Step 4: Calculate and visualize the composite

Normalize each dimension to a 0–100 scale. For frequency, define a baseline (e.g., expected uses per shift) and cap at 150 percent of baseline to avoid outliers. For depth, use the number of distinct feature categories used out of the total available for that role. For sentiment, multiply the average score by 20 (so 3.5 becomes 70). For integration, use a score derived from time-motion observations (e.g., percentage of clicks that are non-redundant). Then apply your weights and sum.

Visualize the score as a single number with a spider chart showing the four dimensions. A high overall score with a low depth dimension suggests the tool is used often but shallowly—a warning sign.

Step 5: Validate with qualitative rounds

The Unison Score is a starting point, not an endpoint. Once you have scores, walk the floor. Talk to high-scoring providers: what makes the tool work for them? Talk to low-scoring ones: what's blocking them? Use the score to identify outliers and then investigate. The numbers tell you where to look; the conversations tell you what to fix.

Common Missteps When Implementing the Unison Score

Even well-designed composite metrics can go wrong. Here are the patterns we've seen teams fall into—and how to avoid them.

Overweighting sentiment

It's tempting to weigh sentiment heavily because it's the only dimension that directly asks the user. But sentiment can be volatile. A provider who had a bad day with a different system may vent on your survey. A provider who just learned the tool may rate it low out of frustration, even if the tool is well-designed. Keep sentiment at 20 percent or less, and always look at trends over 30 days rather than single responses.

Ignoring integration because it's hard to measure

Integration is the dimension most teams skip because it requires direct observation. But if you skip it, you lose the dimension that best predicts long-term sustainability. A tool that scores well on frequency, depth, and sentiment but poorly on integration will eventually see all three decline as frustration builds. Start with a small sample (10–15 observations per provider role) and expand as you refine your observation protocol.

Comparing scores across different tools

The Unison Score is designed for a specific tool and role combination. Comparing the score of a lab results viewer (low depth, high frequency) to a clinical note template (high depth, moderate frequency) is meaningless. Always compare the same tool over time, or different tools for the same role. Cross-tool comparisons only work if you normalize for the expected depth and frequency of each tool's purpose.

Chasing a target score

Once a score becomes a target, it loses its value. Teams may start coaching providers to click more features or to give higher sentiment responses. The Unison Score is a diagnostic, not a performance goal. Use it to identify areas for improvement, then set separate, specific goals (e.g., increase use of medication reconciliation feature by 20 percent) that will naturally improve the score.

When Simpler Metrics Are Better

The Unison Score is not always the right answer. In some situations, a simple adoption rate or a single-question net promoter score is more appropriate and less burdensome.

Early pilot phases

During a small-scale pilot with 10–20 users, the overhead of building and maintaining a composite score isn't justified. You can learn more from direct observation and conversation. Save the Unison Score for when the tool is deployed to a larger population (50+ users) and you need a systematic way to track engagement across sites or shifts.

Tools with a single, narrow function

A tool that does one thing—like a drug reference lookup—doesn't need depth or integration dimensions. Frequency and sentiment may be enough. Applying the full Unison Score would add complexity without insight. In these cases, we recommend a simplified version: frequency plus a single satisfaction question.

Organizations without analytics support

If your IT team can't provide granular usage logs, or if you lack the capacity to do time-motion observations, building the Unison Score will be frustrating and possibly inaccurate. In that situation, focus on sentiment alone (easy to collect) and supplement with periodic shadowing. A crude but honest measure is better than a sophisticated one built on bad data.

When you need to compare across many tools quickly

If you're evaluating a portfolio of dozens of tools and need a quick triage metric, the Unison Score is too labor-intensive. Use a simpler composite like the System Usability Scale (SUS) or a single adoption threshold (e.g., used at least once per week). Reserve the Unison Score for the top-priority tools that warrant deeper investigation.

Frequently Asked Questions

How often should we calculate the Unison Score?

Monthly is a good cadence for most tools. Weekly is too noisy for sentiment trends; quarterly may miss early warning signs of disengagement. Monthly gives you enough data points to smooth out daily variation while staying responsive.

Can the Unison Score be automated?

Partially. Frequency and depth can be pulled from audit logs automatically. Sentiment requires the survey component, which can be automated but needs provider participation. Integration is the hardest to automate—it typically requires manual observation, though some teams are experimenting with screen recording and clickstream analysis to infer integration. We expect full automation to become feasible within a few years as analytics platforms mature.

What's a good Unison Score?

There's no universal benchmark because the score is tool- and context-specific. However, we've observed that tools scoring above 70 tend to have sustained engagement and positive user feedback. Scores between 50 and 70 indicate moderate engagement with room for improvement. Below 50, the tool is likely underused or actively causing frustration. Track your own baseline and aim for a 10-point improvement over 6 months.

How do we handle providers who refuse to participate in sentiment surveys?

Keep the survey optional and extremely brief. If response rates fall below 30 percent, the sentiment dimension becomes unreliable. In that case, either increase the weight of other dimensions or drop sentiment until you can improve response rates. Some teams offer small incentives (e.g., a coffee card drawing) for completing the survey, but be careful not to create a sense of coercion.

Does the Unison Score work for non-clinical tools?

Yes, with adjustments. For administrative tools, replace clinical depth with feature coverage (e.g., used all modules available to that role). Integration might measure how many clicks it takes to complete a common task. The principles are the same; only the specific feature categories change.

Next Moves: From Score to Action

The Unison Score is only valuable if it leads to change. Here are three concrete actions to take after your first score report.

First, identify the lowest-scoring dimension. If it's depth, run a training session focused on underused features. If it's integration, schedule a workflow walk-through with a provider to find friction points. If it's sentiment, hold a listening session to understand pain points.

Second, share the scores transparently with providers. When providers see that their feedback (sentiment) affects the score, and that the organization is acting on it, trust increases. Share the spider charts at staff meetings and ask for ideas on improvement.

Third, revisit your weights after three months. The first set of weights is an educated guess. After collecting data, you may find that frequency correlates poorly with outcomes, and you should shift weight to depth. The Unison Score should evolve as your understanding deepens.

Moving beyond adoption rates is not about replacing one number with another. It's about committing to a more honest, more useful picture of how providers actually interact with the tools meant to help them. The Unison Score is a framework for that honesty—not perfect, but far better than the comfortable lie of a login count.