112 Patient Flow Solutions for Emergency Departments

This list should get us started…  Please add your ideas in the comment section at the end!

  1. Close your waiting room – bring patients straight inside to chairs if no beds open
  2. Use triage to enhance flow: triage = sorting and nothing else
  3. Limit the number of nurses at triage – 1 nurse can SORT 200 patients per day easily
  4. Have patients self triage
  5. Use on-line triage
  6. Use Bedside registration
  7. “Quick Reg” – limited registration; just enough to create a chart
  8. Have patients use self registration
  9. Offer pre-registration on line
  10. Post live wait times on-line to smooth patient volumes
  11. Use patient passports – patient education hand-outs at front door
  12. Educate the community to arrive in ‘slow’ hours (e.g., before 11am Tues – Fri)
  13. Educate community to avoid the surges on Sunday evening/all day Monday
  14. Limit 1 visitor per patient
  15. Encourage patients to bring med lists with them
  16. Encourage community MDs to send in referral notes
  17. Have on-call MDs
  18. Have flexible start and stop times for MDs
  19. Get MDs to take responsibility for flow in real time
  20. Have on call RNs
  21. Shorten nursing documentation (1-2 pages max)
  22. Use combined triage & nursing secondary assessment form
  23. Use physician scribes/navigators
  24. De-zone – move staff to where need is greatest; don’t leave a zone overstaffed
  25. Use advanced directives
  26. Use pre printed orders
  27. Measure and reward MD performance
  28. Measure consultant response times
  29. Insist on in-house consultant coverage for internal medicine, anesthesia, pediatrics…
  30. De-unionize – flow will improve
  31. Have nurses only do nursing tasks (carry out orders, give medications), not clerical work
  32. Track RN break times – insist on accountability
  33. Reward RN extra effort (staying late, skipping breaks, going the extra mile)
  34. Match RN staffing to patient volumes by hour
  35. Match MD staffing to patient volumes by hour
  36. Never allow MDs to go home if waits are long
  37. Staff extra MD and RN shifts on known high volume days (Mondays, holidays)
  38. Encourage MDs/RNs to work in teams and hand over readily
  39. Stagger RN shift changes
  40. Have dedicated ED X-Ray
  41. Have U/S (and tech) in the ED
  42. Use techs for lab draws and ECGs
  43. Stat labs
  44. Prioritize ED lab and DI
  45. Don’t batch
  46. Dedicated porters (RN/tech should porter if porters overwhelmed)
  47. Track DI and lab turn-around times
  48. Get a great EDIS (ED information system)
  49. Create meaningful alerts on EDIS to identify LOS, reassessments, etc
  50. Have a modern EMR linked to the EDIS
  51. Retire outdated EMRs – an old, slow EMR might be worse than none at all
  52. Consider a real time locating system (e.g., RFID)
  53. Consider EMR on tablets for each MD
  54. Computer terminal in each room
  55. Link ED EMRs with community EHRs
  56. Have forms available on-line
  57. Bypass ED for STEMI identified by EMS (straight to PCI)
  58. Eliminate phone calls for CT, etc
  59. Extend CT hours of operation
  60. Encourage the hospital to work on a 24-7 service model (at least a 7 day service model!)
  61. Do not schedule big surgical cases on Monday
  62. Track admits and discharges by time of day and day of week
  63. Eliminate day-day variations of admits/discharges
  64. Perform nurse handover on the ward; not by phone from the ED
  65. Get admitted patients straight up to the ward before a bed becomes available
  66. Use a visual bed management system for inpatient flow admitted patients leave promptly
  67. Use patient flow navigators
  68. Create robust medicine clinic follow-up clinics (next day)
  69. Do not allow consultants to ‘send patients to the ED’ and see them there
  70. Teach residents about quality and efficiency as paramount in their education
  71. Use PO instead of IM, and IM instead of IV treatments if possible
  72. Position EMS off-load in-front of the main nursing station – not hidden away where patients can languish
  73. Form psychiatric patients promptly as needed
  74. Do not perform an internal medicine ‘ward’ work-up in the ED
  75. Order all tests and treatments on the first touch
  76. Plan on disposition from the first encounter
  77. Have Multi-use rooms (eliminate bottle-necks)
  78. Establish procedures to sedate patients in any room
  79. Partner with volunteers – they can help a ton!
  80. Establish CDUs on in-patient wards – do consultations there
  81. Give every MD, RT and Consultants a phone to carry
  82. Do not scale down services over holidays when demand always goes up!
  83. Encourage same day, out-patient cardiac diagnostics and consultation
  84. Establish direct referrals to cardiology (not internal med, NP, cardiology, etc)
  85. Have everything needed for work in every area (don’t make staff walk to the ‘tube system’)
  86. Use pre-printed prescriptions
  87. Have the chief call in 2-3 times per day to monitor flow
  88. Create an internal, real-time ED surge plan
  89. Create a hospital wide surge plan and link it to the ED surge plan
  90. Give admin on call authority to move admitted patients out the ED
  91. Have back on-call to support internal medicine consults
  92. Do not allow surgeons to be on-call to the ED on their OR day
  93. Teach all nurses to apply splints and/or casts
  94. Use ‘just in time’ approach to patient movement – don’t stock-pile patients by loading rooms
  95. Avoid batching
  96. Assign patients to areas; not rooms
  97. Use overhead paging liberally – don’t walk around looking for patients
  98. Improve patient signage
  99. Use patient instruction sheets
  100. Use a re-assessment check-list so MDs aren’t called to reassess prematurely
  101. Build a minor treatment area (aka fast track)
  102. Get rid of as many stretchers as possible (limits holding admitted patients)
  103. Use exam tables where-ever possible
  104. Use some chairs instead of stretchers in the acute area for telemetry patients
  105. Get rid of walls – use curtains to divide most rooms
  106. Eliminate sequential processing
  107. Insist on parallel processing
  108. Look for bottlenecks – theory of constraints
  109. Learn queuing theory and how it applies to your department
  110. Learn and love LEAN
  111. Employ an unlimited capacity mindset – don’t limit flow for lack of ‘rooms’
  112. Adopt a ‘get it done NOW‘ attitude across the organization!

More ideas?  Questions?  Feedback?  Click on leave a reply or # replies below:

Physician Scheduling Extreme – Interview with Dr. Marko Duic

Imagine a schedule where you could work any day you wanted, see as many (or few) patients as you wanted, and take as many holidays as you wanted.

Sound unbelievable?

Dr. Marko Duic has honed a physician scheduling system that delivers MD choice, MD control and a perfect fit between physician speed and patient volumes.  He shares how he does it:

Welcome, Marko.  You’ve figured out a way to give physicians choice, control and as many holidays as they want.  Is that true?

For individual physicians, it’s true.  The only restriction is that a few physicians need to stick around to keep the emerg going–so not everyone can take off at the same time.  This might be an issue if everyone in the department wants to go to the same conference. But if an individual physician wants to go off for 3-6 months, to, say, cycle from Cairo to Capetown, it’s not a problem.

How long have you been doing this for?  

Since 2001

How do the physicians like it?  

They state that it’s a major reason why they’d never go to another ED to work.

How do patients like it – what results do you achieve?

The most important patient satisfaction correlate is their time to see the doctor (Physician Initial Assessment time).  Well, the two hospitals in Ontario that use this system are often number 1 and number 2 in PIA times, and one of them is definitely a leader in satisfaction among peers.

What do the nurses think of this system?

At first, not that pleased that we could muster up ANY number of physicians ANY time, and they would get stuck with a pile of orders.  They had to see that a pile of orders is better than a pile of unseen patients.  At least with the patients having been seen by MDs, nurses could be asked to do the most urgent orders first.   What the nurses like about his system is that there are always enough doctors to see the patients, so they never have to get abused by impatient patients.  There’s never a day when three slow physicians work back to back and the place explodes.

Could this system work anywhere; do you think you could teach others to do it?

Of course it could work anywhere.  It’s how patients would schedule doctors, if patients were allowed to schedule us. 

Most emergency departments schedule a fixed number of shifts every single day, but you don’t.  Why?

Each physician has a number of patients per shift that the physician is most comfortable seeing.  It could be 15, could be 30, could be 45, could be 60.  And each emergency department has a number of patients that they see any given day.  So for example, in one of our departments, we see 300 on Sundays and Mondays and 270 the other five days.  So I have to schedule enough physicians that their combined capacity to see patients adds up to the number of patients I’m expecting.  So if all the fast docs go on a conference together, the slow remaining ones have to be scheduled in larger numbers–maybe 9 or 10 of them in a given day to see that many patients.  If the slow guys go on a conference, I might only need five or six of the remaining fast docs to see the patients.  If everyone’s in town, and I alternate fast and slow docs, I might need 7 or 8 physicians.  If the physicians are different speeds, how can you expect to have the same number of them per day?

So, how do you determine exactly how fast each MD works?  

I have stats, but they don’t really work that well.  So I do it by trial & error repetition and intuition.

But what if more patients arrive on a particular day?  

Shifts start when patient waits get up to a certain level.  Physicians call each other to figure out when the next one needs to show up.  If the day’s busier, it becomes evident in the conversations.  Physicians come in early, and stay late, and if needed, call an extra physician.

And what if you need another physician to help?

We use our on-call funds to pay one of them to show up.  We call in turn, alphabetically, and change the order by one physician each time, so everyone gets a chance.

Who decides if more MDs are needed on any given day?

The physicians who are working in the department at the time.  If they risk running over the target patient waiting time, they call extra help in.

What if the physicians working that day do NOT call for extra help?

Then the times go over, this is a disaster, and they need to explain why they don’t.

What are your thoughts on the provincial Hospital On-Call funding system?

It’s good to have money to pay physicians to come in to serve patients.  Especially in the ED, where volumes and acuities are unpredictable, and where timely access to care is what’s held out to the public in the name EMERGENCY DEPARTMENT that’s posted on the door.

Are there times when you have scheduled too many MDs on one day?  What happens then?

They either shorten all the shifts, or cancel one of them, or both.  They come to an agreement that suits them all.

Okay, let’s focus on the mechanics of schedule creation.  How do physicians ‘pick their shifts’?

They submit a selection form that shows me when they want to work, when they can work, when they would prefer not to work, and when they can’t work.  I use all the physicians’ forms to give everyone a schedule that’s almost entirely made up of shifts they want or can do.

Is it completely different every month?

Basically, yes.  Some patterns repeat–some people like nights; other people like Wednesday mornings; some people can never work Friday evenings.  But overall, it’s different every month.

What happens if there are too many shifts requested by the group?

Everyone gets a bit less than they asked for.

What happens if you can’t provide enough coverage to meet the expected volumes on a day?

Short term, everyone works a bit more than they would like.  Long term, hire more people.  But if it’s one day, and no one wants to work then, then there’s a lottery.

Can MDs take holidays?

Yes, any time, for any length of time.  We’ve had people take full-year sabbaticals or 4 to 6 months LOAs and come back to a full shift roster.  This system is totally flexible.  20 doctors can cover the absence of 1-2 docs with minimal disruption for a long time.

What if everyone wants to take holidays at the same time?

Then we close the department.  No, seriously, there’s a max of about 1/3 of the department that can be on holidays for a longer period, or 1/2 the department for a few days, or 2/3 for a day or two, and it can still have full staffing.

What about Christmas, New Years and summer vacation – how do you handle those holiday requests?

In whatever way EPs want, but overall, some people always want to work them, and if there are not enough, then there’s always the record of who did it last year and the year before–those people get first dibs on taking them off this year.

Wow – this seems too good to be true, but it looks like the results prove it works.  Do you have any final comments?

Try it, you’ll like it.  More importantly, the physicians will love it and insist on always doing it this way.  Patients will love it too.

Fix Patient Flow Bottlenecks – Forever

Patient flow seems impossible to fix in most hospitals.

Nearly every step in a patient’s journey can stop flow.  Even the most LEAN-ed processes require fanatical vigilance to guarantee patients don’t get stopped as they move through our hospitals.

Water Pouring from Bottle

Eli Goldratt popularized ‘bottleneck’ in his classic ‘The Goal‘ – a long narrative presentation of the Theory of Constraints.  A worthy read.

Bottlenecks decrease flow.

They cause havoc by showing up unplanned.  In industry, bottlenecks turn up as critical processing steps whose maximum speed determines the maximum speed of a whole production line.

In healthcare, bottlenecks change all the time.  In fact, a dozen different things could be the bottleneck – the rate limiting step – over 1 day in the ED.  During the morning, it’s lack of staff; during the afternoon, it’s lack of beds.  Despite this, most folks like to blame their favourite bottleneck:

 “If we only had more beds. If only we didn’t have admitted patients!”

“If only we didn’t have such slow docs!”

“If only we had more nurses (or faster nurses…)!”

“If only we had more space!”

If only…

Picking one bottleneck – usually out of our control – removes responsibility to address all the bottlenecks within our control.

In recent posts, we’ve been talking about closing the waiting room, bringing patients straight into the ED and cycling them from chairs to exam tables.  These steps bypass the ‘bed block’ excuse most EDs use for making patients wait for hours in the waiting room.  Once patients get ushered straight in, dozens of new bottlenecks show up: nursing shift change, waiting for porters, DI/Lab back-up, not enough MDs, patient reassessment delays…running out of patient gowns…

There’s only one way to guarantee a bottleneck will never slow flow in your ED.

Parallel processing

Like putting together an IKEA cabinet, some things need to be done in order.  Sequential processing means doing things one after the other – like a long train of boxcars.  However, insisting on sequential processing creates hundreds of potential bottlenecks.  Parallel processing unhitches the boxcars and lets them all run on separate tracks at the same time.  That way, if any process stops, all the others can continue.

Simple enough…

But, medicine loves process.  Many of the sacred cows in emergency medicine are core to sequential processing:

Triage followed by

Secondary Assessment followed by

MD assessments followed by

Lab and DI followed by

Portering…

Reassessments…

Discharge

Most providers fight to keep care in a general order.  It’s what they’re used to.

A parallel approach looks like this:

Triage (sorting only…more in another post) followed by

Everything else

Easy, right?  It’s not.

If you truly adopt this thinking, you might have MDs assess and discharge a patient before the patient even sees an RN.  Or, you might have Lab/DI and discharge planning involved before an MD gets to a patient.  This requires huge flexibility for providers who’ve been bound by historical process.

But…but…what happens first?

After triage, everything else gets done ‘as soon as possible’.  Whoever can get to the patient first, gets started on their part of the process even if it means they can’t complete it before another provider arrives.  Get it done, now!  Sequential steps can never be tolerated as an excuse for making patients wait.

What bottlenecks are holding you back?  What’s holding you back from adopting a fully parallel approach?  Share your thoughts by clicking on ‘leave a reply’ or # replies below.