Thursday, May 29, 2008

Open Acces: Part 3 Analyzing Open Access Scheduling

Today is the final installment of my series on open access scheduling. In part 1 I reviewed an implementation of Open Access Scheduling at Baylor College’s Family Practice Unit in Houston, Texas. In the second part of the series I discussed the theoretical relationship between wait time and total capacity. Today, I’ll take a deeper look at some of the management changes required to implement open access and undertake a more scholarly critique of the concept.

One of the first major articles about open access was from the primary care department at Kaiser Permanente in northern California as a means to decrease wait times in the clinic. The premise of open access is that today’s work is done today by offering all patients appointments for either the same or next day. Open access can function with zero wait time because most practices are in a state of equilibrium where the patient demand is roughly equal to the supply of provider time. The prerequisite for open-access, however, is careful management of resources and patient demands to maintain a uniform level of each with as little variation as possible.

At Baylor, Dr. Steinbauer and his team leveled the requests by patients for appointments using several techniques. Provider schedules were made accessible to the patients, the number of appointment types was decreased from 30-50 to 2-3 and allowances were made to triage/treat patients based on a phone assessment during flu season. This was combined with a patient education program.

On the practice management side, the team maintained the usual staffing levels, established rules for provider leave, developed a system for reminding patients of necessary appointments and reengineered the scheduling processes around health maintenance appointments (see last post for details on this change). The concept of better practice management to facilitate open access is not unique to the Baylor experience. In reports of open access implementation, management changes play a major role in the descriptions.

This raises the question of whether or not it is the scheduling or management changes that lowered wait times. In fact, a recent study from the UK used a single-blinded trial to compare 24 open access to 24 non-open access clinics. There was no significant difference between wait times in the two clinics. However, both had very low wait times for an appointment (1.00 vs 1.87 days). Whereas a Minnesota medical group conducted a longitudinal study between 1998 and 2002 comparing non-open access (1998-2000) to open access (2000-2002). They had a 38% increase in capacity (WRVU’s), PCP compensation rose 20% and costs fell 20% (cost per WRVU fell 13%). The context of the experiment, however, was observation during a period of organizational restructuring which happened to included changing to open access. These studies suggest that management rather than schedule alone is responsible for many of the benefits of open access.

Regardless of the cause, the overall effect seems to be a more efficient clinic. But if the same changes where made in a clinic that used block booking which would have greater capacity? There is no clear cut answer as of yet. Theoretically, offices with zero wait times will have a capacity lower than that of those with slightly longer wait (assuming there is some variation to the patient mix). But actual open access practices realize a capacity (total patient volume) near averages with standard scheduling types. It may be that better management makes up for the lack of a buffer of patients. When I asked Dr. Steinbauer if the total number of patients in the practice increased with the open access implementation he responded, “I didn't track the number of total patients in the practice. But with the increased availability of slots (and no increase in visits per patients; i.e.., we weren't "churning") I assume we got more patients.”

It may also be that that allowance for appointments’ up to 14 days out is enough of a buffer to reach 100% capacity. According to Dr. Steinbauer, “We tracked a metric called "percent capacity usage". Simply the number of patients seen over the available appointment slots expressed as a percent. We maintained appointment filling at 95-105% during and after the transition. There were more patients seen by the physicians because there were fewer long appointment slots, and therefore, more appointments available”.

It has been the experience of my own clinic that even when open access appointments are offered there is a tendency to choose 7-10 days to make allowances for work, car rides, etc…. My final assessment of open access is that it by itself it confers neither a net benefit nor loss in total patient capacity.

The effect on patient satisfaction with open access is unequivocal. Almost all of the reports have conducted patient satisfaction surveys which show a definite improvement. Interestingly, in a more recent survey of 10,821 patients in the UK found that the day of choice was more important than immediate access. This further supports the idea that a 14 day buffer may be beneficial by maximizing patient satisfaction and improving capacity.

The effect on follow-up is mixed. As anecdotally described by Dr. Steinbauer follow-up is less likely to occur because the patient leaves the clinic without an appointment. His group, however, did not track a follow-through metric. The number of appointments each patient used each year (2.7) did not change which is suggestive of similar follow-through pre and post implementation. Another way to examine the problem is realized access. That is how often a patient actually makes use of the open access system. In the words of Dr. Rohrer of the Mayo Clinic “…. other clinic characteristics may overcome the effects of open-access scheduling”. Dr. Rohrer’s group also found the realized access to care did not correlate to the use of open access scheduling. A third study by Dr. Sperl-Hillen et al., from Minnesota found better care in the management of diabetes (as characterized by better control of HbA1c and LDL) with open access but there was no correlation to wait time. This is suggestive of better realized access due to the effect of better management rather than schedule type. Finally, our own clinic sees follow through rates anywhere between 55%-95% depending on the procedure. However a clinic reminds patients of follow-up, the system should be rigorous so people don’t get forgot regardless of whether or not they leave the office with their next appointment.

In the end, I’m left to conclude that open access scheduling could well be an answer to some of the primary care access problems. It is a simple tool to effectively manage a primary care practice. Based on the evidence, it maintains capacity, nearly eliminates wait time and improves satisfaction. A well managed practice could have many of the same results using the benefits of block booking but open access is a straight forward means to achieve similar (or better) results.

For more information on open access take a look at Dr. Kishore Visvanathan’s site who has been posting his the results of his transition to open access as they occur.

Part 1: Open Access Scheduling: Interview with Dr. Jeffrey Steinbauer
Part 2: The Relationship between Wait & Capacity

No comments: