✕
In a hospital, the on-call schedule is somebody's job. There is a department, a system it lives in, and a person whose title contains the word scheduling.
In an independent practice of six physicians, it is a task that lands on whoever is willing. Usually the practice manager. Sometimes the senior partner. Occasionally whoever happens to be good with spreadsheets.
That difference shapes everything about what independent practices actually need, and it is why most scheduling software sold to them is built for a problem they do not have.
Whoever maintains the rota is doing it between other work. They are not going to learn a complex tool, and they are not going to run an optimisation engine over a group of six people.
Building the month is not the hard part. Independent rotas are small enough that a competent person can lay one out in an afternoon, and most of it repeats last month.
The maintenance is the hard part, and it never stops. Somebody swaps. A partner adds a conference in March. Two people want the same holiday. A nurse practitioner joins the call rotation in June.
Software that is impressive at generating a schedule and awkward at changing one has solved the wrong half of the problem. When a vendor demonstrates, watch the edit, not the build.
Almost every difficult conversation about on-call in a small group is a fairness conversation. Not because clinicians are unreasonable, but because the arithmetic is unforgiving. In a group of six, one extra holiday weekend is a meaningful share of somebody's year. In a group of three, it is enormous.
Hospital scheduling systems try to solve this with rules engines and preference weightings. That machinery exists because a large department cannot negotiate face to face. An independent group can, and mostly does. What it needs is not an arbitrator but a shared set of numbers nobody disputes.
The complication is that nights, weekends and holidays are three different currencies, and they are not interchangeable. Somebody who takes more weeknights than anybody else may still, quite reasonably, feel hard done by if they also drew two of the four major holidays. A running total per person, per category, per year settles more arguments than any generator ever will. If a tool cannot show you that table, it is not helping with the part that is actually difficult.
Vendor demonstrations show one clinician per night. Real independent rotas are messier than that, and the mess is not exotic — it is ordinary practice.
Coverage is often split within a day: two partners divide a weekend at midday on Saturday. Nurse practitioners and physician assistants take first call with a physician behind them, which is two people covering at once in different roles. A practice with three sites may cover them separately on weekdays and jointly at weekends. Holiday blocks follow rules that have nothing to do with the weekly pattern. Some groups cross-cover with a neighbouring independent practice, so the person on call is not always on your own payroll.
There is also a distinction most systems miss entirely: being unavailable is not the same as not being on call. A partner who is away should not appear as an escalation option, a second-call backup, or a name anybody can select — and that is different from simply not holding the pager tonight.
When software cannot express these, the workaround is always the same: somebody keeps the real schedule separately and translates it into whatever the system will accept. Translation is where the errors live, and the translator is the person doing this as their second job.
It is worth stating plainly, because software design usually assumes the opposite: swaps are not exceptions. They are how the schedule stays true to reality between the day it is published and the day it is lived.
Hospital tools tend to impose a formal workflow — request, approve, publish — because in a department of eighty, an unapproved swap is a governance problem. In a group of six, the same workflow is friction that people route around. The swap gets agreed in a corridor or over text in ninety seconds, and then somebody is supposed to go and record it. That last step is where independent practices lose the schedule — not because anybody is careless, but because the agreement has already happened and recording it feels like paperwork after the fact.
So the useful question is not whether a system supports swaps. Everything supports swaps. It is how many steps sit between two clinicians agreeing and everything downstream knowing — and whether the person who made the change can make it themselves, from a phone, on a Friday evening.
This is the distinction that matters most, and it is easy to miss because both things are called the schedule.
A PDF on the noticeboard is a publication. So is a monthly email to the partners, a shared calendar, a printed sheet at the front desk, the copy your answering service typed into their system, the roster loaded into your messaging app. Each is a snapshot of what was true when it was made.
Every publication is a copy, and every copy is a way to be wrong. Change the schedule and you have not changed the copies; you have created work. Most of the time somebody does that work. The gap is the times they do not, or do it late, or tell one downstream system and forget the other.
The alternative is that the schedule is not a document at all. It is a live source, and everything that needs to know who is covering reads it at the moment it needs to know — the operators answering your phones, the app your clinicians carry, the referring office trying to reach whoever is on tonight. Change who is covering and routing changes with it, because there was never a second version to keep in step.
That is a different piece of software from a calendar, and it is worth being clear which one you are being sold.
Five questions, in rough order of how often they expose something.
Build my messiest week in front of me. Two partners splitting Saturday, a nurse practitioner on first call, one site covered separately. If it cannot be built live, it cannot be built.
Show me a swap on a Friday at four. Who can make it, from what device, and how many steps. Then ask what has changed downstream by the time they have finished.
Show me the totals. Nights, weekends and holidays per person for the year to date. This is the report the group will actually argue over.
Who else reads this schedule? If the answering service holds its own copy, or the messaging app holds its own copy, count the copies. That number is your exposure.
What does a sender select? If somebody sending a message has to pick a clinician by name, they are consulting the schedule in their head, and the schedule in somebody's head is the one that goes stale fastest.
MatchMD has been doing this since 1996, and the design follows from one decision: there is one live on-call schedule, and it is the routing engine rather than a document about the routing.
Our own live operators answer your phones and dispatch from that schedule, so there is no separate copy at the answering service to keep in step. Senders address a role or a team — the covering physician, the on-call team — and the system resolves who that is at the moment of sending, so nobody needs to know tonight's roster to reach the right person. Providers return calls to patients through the app, so personal mobile numbers stay private. Every message carries its own sent, delivered, read and replied timestamps, so the record of who was reached and when is a by-product of the work rather than something to reconstruct later.
It is worth saying what this is not for. If your group is small, your rota genuinely stable and your after-hours volume low, a shared calendar and a good answering service may be all you need, and we will say so. The case for changing anything starts when somebody is maintaining the same schedule in two places and hoping the two agree.
Bring the messiest week you cover and we will build it in front of you.
Book a demo