The 12 KPIs Every Medical Practice Should Track in 2026

The twelve numbers this article walks through. Seven have a credible 2025–2026 benchmark; for the other five the honest answer is that none exists, so you measure against your own trend.

For each one below you'll get the plain-English definition, a credible 2025–2026 benchmark where one actually exists — and an honest flag where it doesn't, so you don't anchor to a number someone invented for a blog post. That distinction matters more than it sounds: half the "industry benchmarks" repeated online have no primary source at all. (Running a dental practice? The numbers differ — see the 10 dental practice KPIs.)

Why KPIs matter more in 2026 than they did five years ago

Running a practice on instinct used to be survivable. It's getting harder, because the margin for error is thinner than it's been in a generation.

Costs are climbing, independent practices are consolidating, and Medicare physician payment has fallen 33% in real terms between 2001 and 2025 (AMA). At the same time, payers are denying more: the initial claim-denial rate rose to 11.81% in 2024 across the 2,100 hospitals and 300,000 physicians on one vendor's analytics platform (Kodiak Solutions), which is the largest such dataset published and still one vendor's client book. In that environment, the practices that thrive aren't the ones with the most patients — they're the ones who can see where the money and time are leaking, and close the gap. You can't fix what you can't measure.

1. Net collection rate

The single most important revenue number in the practice: of the money you were contractually allowed to collect, how much did you actually get?

A healthy practice collects 95–99%; the best clear 99%+ (AAFP). Anything under 95% means earned money is quietly leaking out through underpayments, write-offs and claims nobody worked. This is the number to fix first, because a one-point improvement drops straight to the bottom line. Here's why the gross rate hides that leak — and how to plug it.

2. Days in accounts receivable (A/R)

How long, on average, it takes to get paid — total A/R divided by average daily charges. Keep it under 50 days at a minimum; 30–40 is where well-run practices live (AAFP). Days in A/R is an early-warning gauge: when it drifts up, billing is falling behind or denials are piling up — usually a month or two before you feel it in cash flow. Here's what a good days-in-A/R benchmark really is, and how to cut it.

3. Claim denial rate

The share of claims payers reject on first submission. AAFP puts the industry average at 5–10% and calls under 5% more desirable. Note what it measures: the dollar amount denied in a period over the dollar amount submitted, not a count of first-pass rejections. The trend is the concern: the industry-wide initial-denial rate climbed to 11.81% in 2024, across a dataset of 2,100+ hospitals and 300,000+ physicians (Kodiak Solutions). Every denial is revenue you have to re-work or lose outright — track the rate and the top denial reasons, because a handful of codes usually drive most of it.

4. Revenue per visit & per patient

Net revenue divided by encounters (per visit) and by unique patients (per patient). Honesty flag: there's no reliable public benchmark here — the tidy per-visit dollar figures circulating on vendor blogs aren't sourced to anything, and real numbers vary enormously by specialty (MGMA's DataDive is the actual source of record). So don't chase someone else's number. Track your own trend: is each visit — and each patient relationship — worth more this year than last? That single line tells you whether your payer mix and service mix are drifting in the right direction. We break down per-visit versus per-patient, the gross-vs-net trap, and the honest Medicare and MEPS anchors in revenue per visit and per patient.

5. Patient retention rate

Of the patients who saw you this year, how many come back? Define your window (say, seen in the last 18–24 months) and apply it consistently. Honesty flag: there is no credible healthcare-specific benchmark — ignore the "75% / 85%" figures repeated on marketing sites. What is widely cited, across industries and hedged by the people citing it, is the economics: a 5% lift in retention can raise profits 25–95% (Reichheld, Bain / Harvard Business Review), and keeping a patient costs a fraction of winning a new one. It's the most under-watched number in most practices — which is exactly why it belongs on the dashboard. We go deep on how to measure and lift patient retention in its own guide.

6. New-patient rate

New patients as a share of the total, tracked month over month. No industry "healthy growth" benchmark exists, so measure against your own trailing average and your local market. One nuance worth knowing: rising appointment wait times (see #11) mean new-patient demand often already exists — it's being throttled by access, not by too little marketing spend. Before you buy more leads, check whether you can even see the ones you have.

7. Patient lifetime value (PLV)

What a patient relationship is worth over its life — the number that connects retention to revenue.

There's no standard benchmark figure (be wary of the "$3,000 per patient" numbers online — unsourced). The value is in the formula itself: it makes a "cheap" acquisition channel that only ever produces one-visit patients look like the expensive one it really is, and it turns a small retention gain into a visible dollar amount. See how to calculate PLV for a clinic — the margin version, and why no benchmark exists.

8. Patient satisfaction

Measured rigorously with CG-CAHPS (the AHRQ-maintained standard for ambulatory experience) or quickly with an internal "how likely are you to recommend us?" score. A useful reference point: Press Ganey put medical practices at 84.1 out of 100 on likelihood-to-recommend for 2023, a five-year high at the time and since superseded. Skip the made-up "average NPS" figures; there's no authoritative one. Trend your own score over time and — more importantly — read the written comments, because that's where the fixable problems hide. Our full guide on patient satisfaction surveys covers CG-CAHPS vs NPS, why neither has a real benchmark, and what the research actually says about satisfaction and quality.

9. No-show rate

Missed appointments divided by scheduled ones. MGMA's single-specialty aggregate, which excludes multispecialty groups, hit 6.81% in 2023, near the 7% it recorded pre-pandemic. I have seen 5–7% called healthy often enough to repeat it, and I cannot source it to anyone; double digits is real money walking out the door — and it's fixable. We break down exactly how to measure it and seven evidence-based ways to cut it in our full no-show guide. And ignore the "$150 billion a year" figure you'll see quoted everywhere — it's folklore with no primary source. Size the cost from your revenue per visit instead.

10. Provider & room utilization

How full your providers, exam rooms or chairs actually are: time used divided by time available. The commonly cited target of 80–89% is not an MGMA measurement. It appears in an article MGMA hosts, written by a scheduling vendor and footnoted to that vendor's own whitepaper, from 2020, about exam rooms rather than providers. The 20-point shortfall in the same piece is health-system executives self-reporting a number the article elsewhere says most of them cannot see. Under-utilization is invisible on a P&L — it never shows up as a cost — but it's capacity you've already paid for in salaries, rent and equipment. Lifting a provider from 65% to 80% is often the cheapest revenue growth available to a practice. We break down how to measure provider, room and chair utilization — and where no honest benchmark exists — in its own guide.

11. Patient access (third next available appointment)

How long a patient waits to be seen — measured as the third next available appointment (the IHI standard; using the third slot filters out flukey last-minute openings). The ideal is same-day. Reality is moving the other way: the average new-patient wait reached 31 days in 2025, up 48% since 2004 (AMN Healthcare). Slow access quietly caps new-patient growth and sends people to whoever can see them sooner — which is why this metric belongs next to your marketing numbers, not buried in operations. We untangle the three different things people call "wait time" — and the honest truth that there's no national in-office benchmark — in how to measure patient wait times.

12. Cancellation rate

Cancellations (with notice) divided by scheduled appointments — the close cousin of no-shows, and often more recoverable because you get warning. There's no honest industry benchmark, so track it against your own baseline. The move that matters isn't lowering it to zero; it's pairing it with a waitlist so a cancelled slot becomes a filled one instead of lost revenue. A cancellation you can backfill barely costs you anything.

The point isn't the list — it's seeing it in one place

Twelve numbers sounds like a lot, until you realise you already have almost all of them. They're sitting in your EMR, your practice-management and billing system, and your scheduling tool right now. The problem was never collecting them — it's that they live on twelve different screens and nobody looks at all of them together.

A number you check once a quarter, buried in a PDF, doesn't change how anyone behaves. The same number on a dashboard you open every Monday morning does. That's the whole game: not more data, but the data you already have, current, on one screen — so a leak becomes obvious the week it starts, not the quarter after. And it's less work to set up than most owners expect — turning your existing exports into a five-page dashboard takes about 15 minutes.

Frequently asked questions

WRITTEN BY
Olha · clinic data analyst

I build the reporting our managers open every morning at a multi-branch medical clinic — and package it so other practices don't have to start from scratch.

Benchmarks are drawn from the sources below and are illustrative — your practice's numbers will vary by specialty, payer mix and market. Where no credible benchmark exists, that's stated plainly. Lucid Vitals is not affiliated with Microsoft.

35ARTICLES WRITTEN