TL;DR
- 23.0% average outpatient no-show rate across 105 studies, against 6.81% in US single-specialty practice
- 27% of practice leaders name no-shows their top patient-access priority for 2026, ahead of scheduling and phone access
- 43% of health system executives say they lose more than 10% of revenue to patient leakage, and 19% lose more than 20%
- 1 in 5 patients switched provider in a year, and 90% of switchers said the practice was hard to do business with
- 63% would switch doctors over poor communication
- 75% of US health systems now use at least one AI application, up from 59%
Roughly a quarter of booked outpatient appointments do not happen.
The pooled figure across 105 studies is 23.0% (Dantas et al., Health Policy, 2018). US single-specialty practices report a much tighter 6.81% (MGMA, 2025), and the gap between those two numbers is the whole subject of this page. No-shows are not one problem with one rate. They are the visible end of a journey that leaks in four separate places, and most patient engagement software only addresses the last one.
This is written for the revenue side of healthcare. Acquisition, enquiry response, conversion, recovery and retention. Not clinical care, not records, not the parts of the stack that touch a diagnosis.
Where the journey actually leaks
Four leaks, in the order a patient meets them. Most software addresses the fourth.
Enquiry response
The first leak happens before anyone is a patient. An enquiry arrives, by phone, form or message, and how fast it is answered decides whether that person books anywhere.
We are deliberately not quoting a speed-to-lead multiple here. The famous ones come from B2B sales studies, get borrowed into healthcare content and are presented as clinical-sector findings, which they are not. What is defensible is the direction: 22% of practice leaders named phone access a top 2026 priority, alongside scheduling at 24% (MGMA, 2025). The front door is understaffed and everyone in the sector knows it.
The no-show
23.0% globally, 6.81% in US single-specialty practice (Dantas et al. 2018, MGMA 2025). The spread reflects specialty, payer mix and population more than it reflects reminder quality.
Reminders are close to universal now, which is precisely why sending more of them changes little. 42% of practices charge no-show fees, and those that do report 25% improvement against 16% for those that do not (MGMA, 2025). That is a real difference and a blunt instrument, since a fee deters the missed appointment and the next booking together.
The number worth having is what a no-show costs, and here the honest answer is that the sector's favourite figure does not hold. A frequently cited $196 per no-show comes from a VA-only study (Kheirkhah et al., BMC Health Services Research, 2016) and travels well beyond the setting it measured. The $150 billion a year figure circulating everywhere traces to a vendor opinion piece with no method. Neither belongs in a business case.
Referral and patient leakage
The quietest leak and the largest. 43% of health system executives say patient leakage costs them more than 10% of revenue, 19% say more than 20%, and 87% call it a priority (Sage Growth Partners and Fibroblast, 2018).
Leakage is a coordination failure. A referral is made, nobody confirms it landed, and the patient completes their care somewhere else entirely. The organisation finds out at the end of the quarter.
Retention
1 in 5 patients switched provider in a year, 90% of them said the practice was hard to do business with, and 70% cited access specifically (Accenture, 8,000 respondents, via AHA, 2024). Separately, 63% say they would switch doctors over poor communication (Deloitte, 2025).
Read those together. People are not leaving over clinical quality. They are leaving over the experience of trying to get seen, which is an operations problem wearing a loyalty problem's clothes.
What patients are already doing without you
Two shifts happened faster than most engagement roadmaps have caught up with, and both change what good looks like.
Patients arrive informed. 52% now use AI to research conditions, 54% to check side effects, and 42% bring what they found to the appointment. 74% say they trust AI for health information while 78% still want a doctor to validate it (Wolters Kluwer, 2026). The person booking is not a blank slate any more, and treating them like one reads as condescension.
Health systems are moving too. 75% now run at least one AI application, up from 59%, and half run three or more (Wolters Kluwer, 2026). Adoption is no longer the differentiator. What it is pointed at is.
That whack-a-mole framing is exactly right, and it describes the failure mode of buying a separate tool for each leak.
Why the software category misses this
The patient engagement software SERP is dominated by intake and portal tools, and they are good at what they do. Digital forms, payment collection, survey capture, chronic-care follow-up.
What almost none of them address is the funnel before the appointment exists and the recovery after one is missed. The closest competitor works the phone line for a single clinic and reports strong results on new-patient capture, but the framing stops at the front desk. Nobody in the top results treats patient acquisition and retention as one continuous journey across channels and locations.
That is the open ground, and it is the same shape as the leaks above. Four disconnected problems, four disconnected tools, one patient who experiences all of it as a single bad impression.
What can and cannot be automated here
This section matters more in healthcare than anywhere else, so it is worth being exact.
Appointment reminders, scheduling and logistics are treatment communications under HIPAA and do not require separate authorisation. Promoting a practice's own services is marketing and does require written authorisation. Content in an SMS or a chat thread stays logistical, never clinical, so no results, no diagnosis, no test findings. TCPA consent is captured at scheduling and opt-outs are honoured immediately.
Zigment operates inside the first category. We do not touch clinical decisions, we do not diagnose, and we are not a records system. What an agent does here is answer an enquiry fast, hold the thread across channels, confirm and re-confirm bookings, and pick up the patient who stopped responding.
Being specific about that boundary is not a disclaimer. It is the reason the deployment is straightforward.
What recovery looks like when the journey holds together
The pattern that works is not more reminders. It is continuity, so that a stalled patient is picked up in the channel they already use with the context of where they stopped.
De-identified, from our own healthcare deployments. A fertility clinic network running click-to-WhatsApp acquisition with real-time routing into the CRM saw a 5x increase in conversion rate and a 40% reduction in cost per acquisition, engaging enquiries 24 hours a day across 12 or more languages.
Language matters more in Indian healthcare than most acquisition plans assume, and it is the same gap we see across sectors. A journey that only works in English loses the majority of an Indian funnel before the clinical question is ever asked.
What to look for in patient engagement software
Four criteria separate tools that move revenue from tools that digitise a form.
Does it work before the appointment exists? Most patient engagement software activates once someone is already a patient. The enquiry that never converted is invisible to it.
Does context survive the channel switch? A patient who asked about availability by web chat and then replies to an SMS should not start over. If the tool cannot carry that thread, your staff will carry it by hand.
Does it handle recovery, not just reminders? Sending a reminder is table stakes. Picking up someone who missed, rescheduling them in the thread and confirming it is where the recovered revenue is.
Does it speak your patients' languages? In India this decides the majority of the funnel. In the US it decides a meaningful minority, and both get treated as an afterthought.
Any patient engagement software that answers no to the first two is an intake tool, which may still be worth buying. It is just not going to change acquisition.
Where to start
Measure the four leaks separately before buying anything, because they have different owners and different fixes, and no patient engagement software fixes all four out of the box.
Time from enquiry to first meaningful reply. No-show rate by specialty and by booking channel. Referral completion rate, which most organisations cannot produce on request and which is usually the largest number on this page. And switching rate with a reason attached.
Three of those four are operations metrics that happen to determine revenue. The fourth tells you how much the other three already cost.
If you want to see where your own patient journeys stall and what recovering them is worth, talk to our team.