
Available Jones, M.D.What if all patients could be seen on the day they call?
Posted Tuesday, Sept. 4, 2007, at 11:54 AM ETMarina Krakovsky was online Sept. 6 to chat with readers about this article. Read the transcript.
Why do we have to wait days, weeks, or even months for a doctor's appointment? Such delays have become so routine that they seem normal. But if some of the most popular restaurants can take same-day reservations, why should run-of-the-mill doctors routinely make patients wait and wait and wait?
In fact, they shouldn't. The challenge of reducing waiting times is a classic queuing problem in operations research. Professionals in all sorts of service industries, from restaurants and hotels to banks and department stores, have faced it in one form or another. Most of them handle the juggling of clients far better than physicians, despite the lower stakes. Mounting evidence shows that doctors can see patients quickly, too—even in perennially backlogged practices—and that when they do, they benefit themselves and the people they treat.
The road to reform is called "open access" (or "advanced access" or "same-day scheduling"). Dozens of papers have been published showing how practices around the country have done it, starting in 2000, when Dr. Mark Murray and colleague Catherine Tantau wrote about their experience of reducing the wait at a Kaiser Permanente clinic near Sacramento from 55 days to just one. Since then, clinics in Texas, Illinois, Minnesota, and elsewhere have similarly cut down patients' waiting times. The Palo Alto Medical Foundation offers same-day appointments, as does the Veterans Administration, the Mayo Clinic, and even some solo physicians.
When a patient calls in the morning asking to see a doctor who uses open access, the office offers an appointment for that same day. Why are there openings available? Well, the main reason most doctors defer today's work to some time in the future is that today's schedule is clogged with appointments made weeks ago. Doctors following the same-day scheduling model, on the other hand, are free today because they saw yesterday's patients yesterday. Using open access, doctors might still schedule some early-morning appointments in advance, for follow-up visits or for patients who actually prefer a future appointment. But the key is that they keep most of their time free for same-day visits and fill up their schedules as the day goes.
This sounds simple, and it flows easily once the system is up and running. But getting open access off the ground takes hard work. Doctors have to chip away at their backlogs, a task that typically takes several months of overtime. That's because during the transition, practices must meet their earlier obligations while at the same time offering same-day visits, to keep future dates clear. They also have to cope with fluctuating demand: If Monday is always the busiest day of the week, the office has to work longer hours on Mondays for open access to run smoothly.
Doctors also have to make sure their practices don't take on more patients than they can handle. The total number of patients in a practice, called the panel size, is crucial because it determines the demand for service on a typical day. Obviously, the larger the panel, the higher the number of expected daily appointments, and if demand outstrips supply, waits are inevitable. Yet many doctors have no idea whether their panel size is too large. They track only the patients they see, not the patients who wanted an appointment but didn't get one. That's a formula for underestimating actual demand for service.
Taking into account the total number of appointment requests is the first step to open access, but it doesn't do the trick on its own. It seems like common sense to balance the number of daily appointment slots with the average daily number of appointment requests. But a mathematical model built by operations researchers at Columbia University shows this intuition to be wrong. That's because demand varies from day to day, and not always predictably. If the average number of appointments is 20, for example, some days there may be 25 and other days only 15. Scheduling 20 slots every day won't work because extra service capacity can't be transferred from day to day: The unused slots from slower days cannot be recouped any more than empty airline seats can be sold after takeoff.
The only solution is to build in a margin of safety in the form of more appointment slots than an average day will ever use. That sounds like wasted capacity, but it's actually more efficient than filling up the appointment book in advance. That's because the further in advance patients make appointments, the likelier they are to miss them. A no-show rate of 30 percent is not uncommon. According to one study, many patients don't understand scheduling systems and find long waits insulting, so they think nothing of missing their appointment without calling to cancel. All these no-shows also add up to waste and lost revenue—the very problems traditional scheduling would seem to prevent. The strange upshot: By juggling too many patients, doctors lose income even as their backlog grows longer and longer.
But to dig themselves out, doctors have to take a leap of faith. Open access rejects the time-honored and seemingly sensible distinction between urgent and nonurgent visits, instead assuming that even routine checkups can be done on short notice. Some doctors who are sick of squeezing in very sick patients reserve a few slots each day for urgent appointments. Called the "carve out" approach, this works a little better, but it introduces new problems. There's less time for nonurgent needs, so patients have to wait even longer than in traditional practices if they're not in dire straits. And distinguishing between a problem that's urgent and one that can wait requires constant and costly triage. It also gives priority to the squeaky wheel, encouraging patients to squawk the loudest or stretch the truth about their condition. In the end, there's little gain: Most patients in carve-out as well as traditional practices can't see their doctor on the day they call.
That's not merely frustrating—long waits can worsen a patient's condition. The experience of Liz Contreras, a 28-year-old in Washington, D.C., is fairly typical. As she relates in her blog, Contreras waited three months for a skin cancer diagnosis because she couldn't get an earlier appointment—only to be told that she had a mole that was cancerous, and that she would have been in trouble if treatment had been withheld much longer. Research also shows that long waits make patients less likely to get necessary preventative care, such as immunizations and mammograms. Several studies have linked delayed access to higher death rates. Practices that have switched to open access, on the other hand, report higher measures of clinical care, greater patient satisfaction, and even lower costs.
And yet, doctors who see patients on the day they call remain the minority. In a recent study from the University of California at San Francisco, researchers phoned dermatologists around the country complaining of a changing mole, and found the average wait for an appointment to be about 38 days (and much longer in some regions). Other specialties may fare better, but primary care physicians—who for many patients are the gatekeepers to specialized care—keep patients waiting far longer than they need to. According to a recent Commonwealth Fund study in several countries of "sicker adults" (those with a history of serious illness or hospitalization), the wait time in the United States is one of the worst among peer nations. Only 30 percent of sicker American adults surveyed could see their doctor on the same day, far fewer than in the United Kingdom (45 percent), with its oft-maligned National Health Service. And almost one-quarter of Americans reported waiting six days or longer—the second-highest fraction of any nation surveyed.
Successful restaurants understand that long waits lead to no-shows and lost income, which is why many of the most popular places don't allow reservations until at most, say, a month in advance. In the restaurant business, deftly balancing supply and demand enables most places to take same-day reservations. Likewise, doctors should stop deferring for weeks and months what, with proper planning, can be done today. We can't wait.
Most Fast Food Restaurants Thrived During the Recession. Not Arby's.
Did Anti-Communists Really End Communism? Two Historians Say No.
Dear Farhad: How Does Facebook Know I'm Gay?
What Ever Happened to Hood Ornaments?
Are Doctors Allowed To Say They're Sorry?
Hitchens: Let's Not Get Sentimental About Communism












Remarks from the Fray:
I think this vastly oversimplifies the problems associated with getting an appointment to see a doctor. I am a child psychiatrist. There is an estimated need for about 12,000 child psychiatrist by 2020, but there will only be about 8,000 practitioners. I see the effects in my own practice on a daily basis, where the wait to see me is about 3 months, and I struggle to have a follow up of even an abbreviated time in 2 months. Same day scheduling would not be practical, as I need information from schools and therapists.
In addition, I find the restaurant example to fall apart (or perhaps to support my views?). Locally, most restaurants do not take reservations. When I show up at dinner time, there is often a wait of several hours, which discourages people from staying. Perhaps in urban settings with a greater variety of choices in both restaurants and physicians, this analogy holds better, and maybe medical access could be simplified. But for now, the restaurants don't do any better at getting me access to food than I do at providing access to treatment. In both cases, I think the solution is to have more providers, a task not easily accomplished when the training period for child psychiatry at least is 13 years of post-high school education.
--dberne
(To reply, click here.)
I am getting the impression while reading this article that the author has a somewhat simplified idea of how a medical office functions. Patients do not queue up for doctor visits the way one queues up for concert tickets or roller coaster rides. The vast majority of patients schedule appointments several days or even weeks in advance to accommodate their own busy schedules. They know that they have time on Thursday afternoons to see the doctor and they make an appointment accordingly. A smart patient does not call on Thursday morning to get seen the same day because they run the risk of there not being an opening and having to reschedule on a workday.
I think that the real reason waits in physician offices can be quite long is that it is often very difficult to predict how long a particular patient's visit will last and the schedule will be pushed back accordingly. A patient who comes in with an "ear infection" may take five minutes, twenty minutes, or even hours of attention. The patients scheduled afterward are going to be delayed. Also, a fair number of patient arrive late and are still accommodated. Open access scheduling will do nothing for that.
--enterdoc
(To reply, click here.)
It always amuses me when people try to make analogies between medicine and other kinds of commerce. It shows a profound lack of understanding how different the practice of medicine is from, say, the travel industry. I don't just show up at the airline counter and expect to fly that day.
If you show up to the Hilton and there are no rooms in the Inn, Hilton says, sorry about your luck. They are not legally obligated to house you, whereas doctors are (failing to do so is called abandoning the patient). If all that is available is a single and you want to house your family of 4, too bad, take it or leave it. Try shortchanging a patient because she is shoehorned into an inappropriately short office visit. Hilton Hotels don't shut down for several days because their is an a shortage of employees at the Marriot and everyone leaves to go help (read: hospital emergencies).
So the smart consumer makes a reservation (especially at high utilization time.) What the Hilton does is have (some) excess capacity (some of the time,) The walk in will end up paying top dollar if he gets into the hotel at all. Try that with Medicare patients and you'll be doing time in prison.
I agree that open access CAN work, but it works best for those practices where the needs are usually acute (pediatrics), predictable in time utilization (many primary care offices) and most of the visits do not be set up at an interval.If I know I need to recheck a cancer patient in 6 months, why not schedule it now, for six months from now?
--Stop-truth-decay
(To reply, click here.)
I know this one. I'm a rural solo practice doctor. I have patients call my cell phone to make appts. I see nearly all of my established patients same day. I have daily walk ins which are by definition same day. Also, I am a published researcher on certain mathematical simulations.
The article used the term 'doctors' only later breaking the term down into doctors seeing people for diseases whose course is days (primary care and colds) vs. months (dermatologists) and implicitly including others with very long courses (neurologists). The pattern of disease is not comparable.
For primary care, the instance of disease and patient visits follows an extremely non-gausian distribution. Infections are epidemic. Each individual 'flu' goes from not being present at all to hitting everyone. Weather causes chronic lung disease to flair in the entire susceptible population on the same day. Distribution of patients needing drug refills is mal-distributed by schedules of holidays and pay cycles.
As 'doctors' is an overly broad term, 'patient' is equivalently imprecise including as it does, those waiting for a signature and those waiting taking a complex history including not only looking up all of the patient's clinical data, but also the newest literature on her medications and conditions and their interactions. Setting aside the very small number of complex patients into the future, as a backlog, prevents a logjam for the entire office. The distribution of patients and number of hours in the day becomes less uniform and less predictable and less regimented as efficiency and frankly, medical care improves.
--drugdoc
(To reply, click here.)
(9/5)