Home Patient Flow Getting rid of what doesn’t work

Getting rid of what doesn’t work

November 2014

Published in the November 2014 issue of Today’s Hospitalist

Editor’s note: This is the second in a series of articles on how to improve patient flow. The full series covers three themes:

1)Are you fixing patient-flow problems or causing them?
2)Getting rid of what doesn’t work
3)”Managing up” to improve patient flow

IN MY LAST ARTICLE, I discussed the changing paradigm in hospital reimbursement and how patient flow will play a key role in determining whether hospitals succeed in this new economic model.

One big reason our hospitalist group “Adult Inpatient Medicine Service (AIMS), which is part of the Presbyterian Medical Group in Albuquerque “had to solve our patient-flow problems is because New Mexico is No. 50 in the nation in terms of population covered by commercial insurance.

At the same time, our state is No. 2 in terms of the percentage of population covered by Medicaid, which is now under capitation. We needed to innovate to survive, which may be where other health systems find themselves as reimbursement gets squeezed.

To improve both patient flow and quality, many innovations we put in place realigned hospitalist workflow. To do so, we needed to challenge several “sacred cows” that most of us in the industry hold about hospital operations and group management.

That wasn’t easy. But it’s what we had to do to unclog the hospital and accomplish our main goal: providing quality health care that was financially viable.

The wRVU trap
The first “sacred cow” we took on was the whole wRVU and productivity model. Hospitalists love higher salaries but are overwhelmed by the census they need to chase for high wRVU-based compensation.

A higher census creates throughput problems as hospitalists struggling with their patient load discharge less efficiently. In 2010, our administration pushed us to adopt a wRVU model. We pointed out that we could meet those wRVU expectations, but only by sacrificing patient flow. We argued that revenue generated from productivity pales in comparison to the money the hospital could save by reducing average length of stay through more efficient discharging and better throughput.

Productivity drives another sacred cow we took on: level-loading. In many groups (particularly those with unit-based care, like ours), rounding teams insist on having the same number of patients.

But we do not. We distribute overnight swing-shift admissions equally among our rounding teams. Then, we let individual hospitalists’ census float, fluctuating from day to day like currency. (The one exception is the first day a hospitalist comes on service, when we give him or her 12 patients.)

Perverse incentives to not discharge
Why do away with level-loading? In our system, hospitalists who appropriately discharge patients are rewarded with a lower census the next day. But in level-loaded programs, those same physicians are punished by being given more new patients. And in a wRVU system, physicians with lower census are likewise dinged for efficient discharges due to lower wRVU production.

If you take level-loading, add hard caps on individual doctors’ census and mix in a wRVU productivity model, you’ve created the ideal but perverse incentive not to discharge. In such systems, hospitalists are rewarded not for efficient discharges, but for continuing to accumulate wRVUs “which their income is tied to “from patients with whom they are very familiar.

We do offer an almost immediate incentive if a rounding team or swing-shift physician performs extra admissions beyond their allotted admissions per shift on days when admission volume is heavy. Hospitalists receive this pay in their next paycheck, not when their wRVUs are reconciled at the end of the year. This allows us to flex up quickly for high volume, and it rewards physicians almost instantly for extra work.

Matching schedules to work
The next dogma we replaced was the hospitalist schedule. Instead of a strict week-on/week-off, we aligned our schedule to maximize throughput.

We started by breaking down FTE requirements into a number of hours, not a number of shifts. Our FTEs work 1,856 hours every year; with a typical rounding shift of 10 hours, that equals roughly a week-on/week-off load. Our shifts usually range between 10 and 12 hours, and those working nights earn an hourly differential.

Because we believe continuity of care is critical for throughput, doctors working rounding shifts have to work a minimum of five days in a row. Non-rounding shifts don’t have that constraint.

This allows us to add or delete shifts to meet volume variations throughout the year. And being hourly-based, physicians can craft very flexible schedules. While some do the traditional week-on/week-off, others work different blocks. And doctors who want additional time off can work any number of days in a row, if they need to.

We also schedule shifts to the work, challenging another sacrosanct practice among hospitals everywhere: running a sharply reduced staff on weekends. We believe that cutting the number of weekend shifts grinds patient flow to a halt, so our weekend schedule is exactly the same as on weekdays.

That decision put pressure on our ancillary services to match our schedule. Our colleagues in care coordination increased their weekend staffing model, guaranteeing a care coordinator on each unit-based floor every weekend.

Another scheduling innovation: Our two swing shifts run 4:30 p.m.-2:30 a.m. and 5 p.m.-3 a.m. Some hospitalist groups end their swing shifts around midnight because that’s more convenient for doctors. But our swing-shift schedules instead match our peak ED volume.

Discharge before noon?
Then there’s “discharge before noon,” an innovation many administrators are now pushing. Discharging patients before noon may be an appropriate goal, but it needs to be part of an integrated throughput process and not just a quick gimmick.

If not properly planned and incentivized, the push to discharge before noon can actually decrease patient flow. After all, if we discharged all patients before noon, everyone in the hospital would have to stop treating patients in the morning to gear up for discharges.

In fact, staggering discharges promotes discharge efficiency. And why incentivize a team to discharge a patient at 10 a.m. when that discharge could have happened at 4 p.m. the day before? We are piloting an early discharge program, but we have integrated that initiative into our unit-based multidisciplinary rounds so it is not a stand-alone process. Working with the hospitalists, our care coordinators identify patients who would be ideal for early discharge, such as those going to a SNF the next day.

Then there’s this innovation that hospitalist groups swear by: having dedicated rounders and admitters. We have dedicated admitting teams in the late afternoon and at night, but we also want to create “surface area,” the ability to see large numbers of patients quickly. That’s why we have each rounding team perform up to two admissions between 8 a.m. and 4:30 p.m.

Relying only on dedicated admitters creates admission bottlenecks; if 10 patients are waiting to be admitted and two dedicated admitters each process one patient an hour, two of those patients won’t be admitted for five hours. Instead, with 17 rounding teams, we can theoretically admit 34 patients within two hours. Our AIMS record? Eighty-five admissions with a history and physical and orders in 24 hours.

What are the “sacred cows” that you never question in your group? Have you really thought about why you are tied to wRVUs and certain schedules?

David J. Yu, MD, MBA, is the medical director of Adult Inpatient Medicine Services with Presbyterian Medical Group at Presbyterian Hospital in Albuquerque, N.M. You can send him questions or comments to dyu3@phs.org.