The Case for Patient Flow Management
A 52-year-old woman with a history of major depression, posttraumatic stress disorder, and alcohol abuse was hospitalized for suicidality in March. After several weeks of inpatient treatment, the patient stabilized and was discharged back to her (outpatient) psychiatrist, a resident in the final year of training and due to graduate at the end of June. The patient saw this physician multiple times during April, May, and early June. At her last visit before the academic year-end transfer, the patient was not given a follow-up appointment because the clinic schedules for incoming residents, who would begin on July 1st, were neither finalized nor operational in the electronic scheduling system. Per existing protocol, the patient was asked to contact the clinic in July to set up an appointment with her new psychiatrist.
The patient did not call to schedule an appointment and was not prompted to do so. The incoming resident psychiatrist recognized this a month after starting (the resident had been given a brief sign-out by the outgoing resident that included this patient's tenuous condition) and contacted the patient to set up an initial visit. Because the resident's schedule was already booked through August, the patient was not seen until early September, at which point the patient stated that she felt better. She set up another appointment for later that month and told the resident that her primary care provider had given her sufficient medication refills.
Unfortunately, the patient did not make her second scheduled appointment. The patient's daughter notified the resident that the patient had died after driving her car into a tree. Autopsy results indicated alcohol and drug intoxication. While there was no way to be certain, a review of the case by the involved clinicians raised the possibility that the patient's 3-month hiatus (from last appointment with the outgoing resident in early June until the appointment with her new physician in early September) may have contributed to her demise.
In this case, the ambulatory training clinic had no mechanisms to ensure structured sign-out or priority scheduling of higher-risk patients, gaps that endangered patient safety. To reduce and eliminate this risk, several steps, including improved communications and checklists, must be implemented. Safety issues involving academic year-end transfer of outpatients from outgoing to incoming residents were addressed in a recent AHRQ WebM&M commentary. Our commentary will focus on a different issue—how scheduling systems and processes can impede access to outpatient care.
This case clearly illustrates how poor scheduling of patients can result in diminished quality of care. Timeliness is an extremely important determinant of patient safety—arguably as important as clinical quality of care itself. We and others have raised the controversial possibility that the majority of sentinel events (1) might be more attributable to poor management of patient flow (2-5)—particularly the timing of care delivery—than medical errors.
Unfortunately, in the world of modern health care, scheduling usually happens on a trial and error basis. Consider a typical primary care physician (PCP) or specialist office. A patient is scheduled in advance for an appointment at a particular time, often far in advance, assuming there is no clinical urgency. The physician knows this time as well and yet, a patient is frequently subjected to (sometimes excessive) waiting. Why?
Even though the patient in this case had elevated clinical risk, the patient could not secure a follow-up appointment at her last visit. Then, when contacted by the resident, the first available appointment was more than a month out. Problematic access and poor time management contributed to the death of this patient. Regrettably, the situation is widespread: Even with a proper care transfer, timely access to PCP or a specialist is often exceedingly difficult due to suboptimal scheduling in these offices. For example, in Boston, waiting time for an appointment with a dermatologist exceeds 2 months.(6)
What is the ultimate challenge in this situation? There are two main patient flows: patients who are scheduled for an appointment and those who are unscheduled but urgent, therefore requiring an appointment on the same day of their request or soon thereafter. Both patient flows compete for the same health care resources: physicians' and nurses' time, exam rooms, lab tests, etc. Let's consider three potential scheduling options and their effect on patient waiting times and office resources.
Option 1: Schedule patients back to back in order to serve as many as possible. An advantage of such a design, assuming it is administered correctly, is that patients have fairly predictable schedules, and they are assured they will be given care approximately when planned. It is also satisfactory to the physicians as they can anticipate their workload on a given day. Thus, this design benefits scheduled patients and physicians, and increases resource utilization and the office bottom line as long as patient non-attendance rates are low. Consequently, unscheduled patients (those who make an appointment and come on the same day) would be left waiting for an inordinate and unpredictable amount of time until there is a "hole" in the schedule. On the other hand, if unscheduled patients are given priority due to their purportedly higher acuity, then scheduled patients' waiting times are likely to become prohibitive.
Option 2: Full day open access ensures that patients are seen on demand assuming that office resources are sufficient (these resources can be estimated using the core component of operations research, queuing theory [7-9]). However, regardless of resource sufficiency, this scenario results in substantial underutilization. Given the unknown number of unscheduled patients coming to the office on a particular day, the schedule would have to hold open slots "just in case." This would have a significant adverse effect on the office bottom line.
Option 3: Combine open access and schedule implementation to improve the bottom line, while simultaneously controlling patient waiting times.(10) If both scheduled and unscheduled patient flows are combined throughout the entire day without separating them in time (designating a separate time interval, or separate resource for each category of patients), then it is almost certain that scheduled appointments will be delayed and/or the waiting time for unscheduled patients will be prohibitive. The right solution will be to find the least busy part of the day, when scheduled patient demand is lowest, and designate this time for unscheduled visits. How long this period of time should be depends on patient daily demand and available resources (physicians, nurses, etc.) for providing simultaneous care for these unscheduled patients. Once these resources are known, the amount of necessary time could be determined using queuing theory.(7) At one medical center, open access led to a drop in wait time for routine appointments in family medicine and pediatrics from 30 days to 1 day, and the percentage of patients seeing their own physician increased from 28% to 75%.(11)
Unfortunately, the deadly consequences of suboptimal access to care are not limited to PCP or specialists' offices. Situations in hospitals can be even worse, especially when hospital capacities and patient demand are misaligned. Multiple studies (3-5,12) have demonstrated that when hospital census is high, ambulance diversions and nurse overloading take place with a subsequent increase in patient morbidity and/or mortality. A recent study (4) demonstrated that patient-per-nurse staffing ratios caused by fluctuation in hospital census and/or admissions have a considerable impact on patient mortality.
"Fortunately" a significant part of such misalignment is also due to the poor scheduling of elective admissions.(13-15) In terms of scheduling, there is a strong analogy between elective/scheduled admissions and scheduled ambulatory visits as well as between urgent admissions and urgent/unscheduled ambulatory visits. The solution to this scheduling problem in hospital settings is very similar to the above Option 3: separating resources for both types of admissions.(14) A recent editorial (5) shows that eliminating artificial variability (fluctuations in patient scheduled admissions based on physicians' personal preferences rather than patient needs) is the only plausible way to reduce mortality associated with misaligned patient demand and hospital resources. The same effect exists in ambulatory settings. If, for example, many PCPs affiliated with the same hospitals would have limited office hours on the same week day and at the same time, artificial influxes of patients would take place in the emergency department.(16) In addition to improving patient safety and quality of care (17), this approach would substantially reduce ambulance diversions (16,18), reduce readmissions (19), and dramatically (20-22) reduce health care cost.
Operations research and its core component, queuing theory, as well as system analysis still remain "foreign" to the health care field. In the era of limited health care resources, these important disciplines should be introduced to medical students. Only last month, Centers for Medicare & Medicaid Services recognized the need for the health care delivery system to be introduced to these areas of management science by announcing a new training program for health care advisors.(23)
- Suboptimal patient appointment processes are a source of reduced patient access to care and consequently diminished quality of care and safety.
- Misaligned patient demand and health care capacity could be a significant source of medical errors and sentinel events.
- Operations management tools, particularly queuing theory, allows developing optimal schedules with maximized patient throughput and reduced waiting times.
- There is a growing necessity for including operations management in medical schools' curriculum.
Eugene Litvak, PhD President and CEO
Institute for Healthcare Optimization
Adjunct Professor, Department of Health Policy and Management
Harvard School of Public Health
Sarah A. Bernheim Project Coordinator
Institute for Healthcare Optimization
1. Sentinel Event. The Joint Commission. [Available at]
2. Litvak E. Healthcare and traffic management. Exec Healthc Mag. 2010;11:88-89, 91, 93-94. [Available at]
3. Litvak E, Buerhaus PI, Davidoff F, Long MC, McManus ML, Berwick DM. Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety. Jt Comm J Qual Patient Saf. 2005;31:330-338. [go to PubMed]
4. Needleman J, Buerhaus PI, Pankratz VS, Leibson CL, Stevens SR, Harris M. Nurse staffing and inpatient hospital mortality. N Engl J Med. 2011;364:1037-1045. [go to PubMed]
5. Litvak E, Laskowski-Jones L. Nurse staffing, hospital operations, care quality and common sense. Nursing. 2011;41:6-7. [go to PubMed]
6. Tsang MW, Resneck JS Jr. Even patients with changing moles face long dermatology appointment wait-times: a study of simulated patient calls to dermatologists. J Am Acad Dermatol. 2006;55:54-58. [go to PubMed]
7. Gross D, Harris CM. Fundamentals of Queueing Theory (Wiley Series in Probability and Statistics). 3rd ed. New York, NY: John Wiley and Sons, Inc.; 1998. ISBN: 9780471170839.
8. Butterfield S. A new Rx for crowded hospitals: math. Philadelphia, PA: ACP Hospitalist; 2007. [Available at]
9. Weber DO. Queue Fever, Parts 1 and 2: A little number crunching can show hospitals how many beds and staff members they really need. Chicago, IL: Hospitals and Health Networks; 2006. [Available at]
10. The CAHPS Improvement Guide. [Available at]
11. Open Access Scheduling for Routine and Urgent Appointments. Agency for Healthcare Research and Quality. [Available at]
12. Shen YC, Hsia RY. Association between ambulance diversion and survival among patients with acute myocardial infarction. JAMA. 2011;305:2440-2447. [go to PubMed]
13. Litvak E, ed. Managing Patient Flow in Hospitals: Strategies and Solutions. 2nd ed. Oakbrook, IL: Joint Commission Resources; 2009. ISBN: 9781599403724.
14. Litvak E. Optimizing patient flow by managing its variability. In: From Front Office to Front Line: Essential Issues for Health Care Leaders. Oakbrook Terrace, IL: Joint Commission Resources. 2005; 91-111. ISBN: 9780866889506.
15. Litvak E, Long MC. Cost and quality under managed care: irreconcilable differences? Am J Manag Care. 2000;3:305-312. [go to PubMed]
16. Litvak E, McManus ML, Cooper A. Root cause analysis of emergency department crowding and ambulance diversion in Massachusetts. Boston, MA; Boston University Program for the Management of Variability in Health Care Delivery, Report to the Massachusetts Department of Public Health; 2002. [Available at]
17. Mismanaged hospital operation: a neglected threat to reform. Health Affairs Blog. February 22, 2011. [Available at]
18. Litvak E, Long MC, Cooper AB, McManus ML. Emergency room diversion: causes and solutions. Acad Emerg Med. 2001;8:1108-1110. [go to PubMed]
19. Baker DR, Pronovost PJ, Morlock LL, Geocadin RG, Holzmueller CG. Patient flow variability and unplanned readmissions to an intensive care unit. Crit Care Med. 2009;37:2882-2887. [go to PubMed]
20. Health Affairs Blog. [go to PubMed]
21. Litvak E, Green Vaswani S, Long M, Prenney B. Managing variability in healthcare delivery. In: Young PL, Saunders RS, Olsen L, eds. The Healthcare Imperative: Lowering Costs and Improving Outcomes: Workshop Series Summary. Washington, DC; 2010. ISBN: 978-0-309-14433-9. [Available at]
22. Litvak E, Bisognano M. More patients, less payment: increasing hospital efficiency in the aftermath of health reform. Health Affairs. 2011;30:76-80. [go to PubMed]
23. Innovation Advisors Program. Center for Medicare & Medicaid Innovation. [Available at]