On the evening of July 30, 1945, a U.S. battleship traveling from Guam to the Philippines was spotted in the Philippine Sea by a Japanese submarine crew. Six torpedoes sped across the black water with devastating effect.
Three hundred sailors died immediately as the stricken ship sank. Another 900 men were left floating in an oil slick in the shark-infested Pacific Ocean for four days.1
The events that led to the sinking represent a classic tale of systems failures. They ring familiar to any physician who has closely examined modern medical error.
The communication breakdown that prevented the ship’s captain from being aware of submarine activity on the same route four days before is no different than communication breakdowns in complex hospital systems. The bureaucratic decision by a remote administrator to withhold the safety measure of an escort to prevent such an attack on the grounds that it “lacked necessity” likely resonates with physicians who have struggled with getting authorization for care.
Why did those sailors remain in the water four days before rescue? Despite being only a two-hour flight from the nearest base, they were not recovered—or even missed—for what must have seemed an endless amount of time.
In the end, slightly more than 300 men were alive when they were spotted by a plane that happened to fly past. The rest succumbed to dehydration, exposure, and sharks. The failure of the Navy to rescue the sailors offers lessons to the clinician trying to improve transitions of care more than a half a century later.
The Feedback Loop
Hospital discharge is a complex process initiated by physician orders on charts, prescription pads, and patient instructions. Most often, the things we assume will be done out of our direct view are carried out satisfactorily. However, any hospitalist can easily recount stories of tests or follow-up that didn’t happen as ordered. Patients who fall through the cracks at discharge—like the stranded sailors of the USS Indianapolis—are, in part, the story of a simple omitted step: the feedback loop.
A feedback loop occurs when the results or consequences of an action are returned as an input loop to the initiating step in order to modify subsequent actions. This fundamental engineering concept can keep complex systems on course. The feedback loop allows the lack of completion of a portion of a process to be recognized—and corrective measures taken—before additional harm occurs.
In Guam, the island base from which the USS Indianapolis departed, the marker indicating the ship on the plotting board was removed when the ship left. Later, the Philippine port of Leyte failed to note that the ship didn’t arrive. Policy at the time was that all ships that left port were presumed to have arrived at their destination unless a call indicating trouble was received. The junior officer who noticed that the ship hadn’t arrived assumed there had been an order to divert to an alternate port. The Navy had no feedback mechanism to communicate between the two ports and raise an alarm when a ship did not arrive. In fact, a Navy directive discouraged communicating the arrival of combat ships as a matter of military secrecy. As a result, no rescue mission was launched—and the sharks began to arrive.
A closer examination of usual hospitalist discharge practice reveals too much similarity for comfort. Similar to how the USS Indianapolis was removed from the plotting board on Guam, discharged patients are removed from the hospitalist’s census list. Follow-up becomes the responsibility of the patient and primary care physician.
Communication with the primary care provider is recognized as a best practice for discharge, but research suggests direct communication occurs less than 20% of the time.2 These dismal statistic suggests the “port of arrival” is unaware our patients are expected in a significant percentage of discharges. Outpatient physicians, like the junior port director in Leyte, may assume that patients who do not call or arrive for appointments have been readmitted, seen by another physician, or otherwise diverted.
Even a superficial review reveals significant deficits in the feedback provided by our current discharge practice. When patients don’t arrive at follow-up appointments, most hospitalists lack any ability to recognize this failure of their transition plan. The assumption in most hospitalist groups is that patients who leave the hospital will achieve follow-up as directed. This “presumption of success” is ill-founded and may expose the patient to potential harm—and the physician to liability.
The case of Shirk v. Kelsey offers parallels to discharge situations hospitalists commonly encounter.3 In this malpractice case, a procedure was unsuccessful in terminating a patient’s pregnancy. The performing physician left follow-up of the pathology results to the patient’s usual outpatient obstetrical provider. The patient’s lack of follow-up was not recognized by the discharging obstetrician. This provider—not the practitioner with whom follow-up was intended—was found liable.
The American Medical Association’s code of ethics states: “Once having undertaken a case, the physician should not neglect the patient.” Hospitalists form physician-patient relationships with hospitalized patients that usually terminate on discharge. Our duty to not abandon or neglect a patient diminishes significantly after discharge, when we are no longer responsible for ongoing hospital care or exchanging information with patients on a daily basis.
But our duty does not disappear. Certainly, the responsibility falls to the hospitalist to be aware of and ensure a follow-up plan for important results such as pathology reports that return after discharge.
Allegations of improper post-discharge communication or failure to pass along critical results that become available once the patient has left the hospital, are common in medical malpractice claims.
Most are settled out of court, and many do not find the physician liable for malpractice. However, legal consequences are far less relevant than the safety and quality of care compromised when patients are failed by a system that lacks feedback loops to ensure safe transitions.
The ultimate goal of medicine is to improve the health and quality of life of our patients. Whether or not a lawsuit results, we need to recognize the commonality of patients who have their care compromised, delayed, or mismanaged because of our inability to recognize a foundering transition plan. Instead of looking at this as a failure of individual physicians to communicate, the problem needs to be addressed by creating effective, reliable systems.
Providers discharging patients with follow-up needs should have a mechanism to identify those at highest risk for problems with transition. For these patients, follow-up with a post-discharge telephone call may be an effective feedback step.
One study looking at post-discharge phone contact found that 20% of patients had not filled new discharge prescriptions. Another post-discharge study revealed that a quarter of patients had medication questions that required clarification.4,5 Other research indicates more than one in 10 patients had new or worsening symptoms in the first five days after leaving the hospital.6 Despite these symptoms, 39% of these patients did not have a follow-up appointment established.
An integrated informatics system that prompts hospitalists and primary care physicians when patients do not arrive at expected follow-up or when test results return after discharge would be optimal. But a simple phone call to identify problems can be effective. Some hospitalist groups have incorporated routine post-discharge telephone contact into their practice—but most have not. Research identifying which patients would benefit is needed to allow targeted use of resources.7
While it’s understood that not all patient discharges will go smoothly, just as not all battleships will arrive at port without incident, there is frequently an opportunity to recover when things begin to go awry. A change in the common attitude that hospitalist responsibility ends when the patient leaves the hospital is necessary.
An element of the solution lies in the creation of feedback loops to identify patients who are not obtaining follow-up as expected. This step requires a commitment of resources—something our fragmented medical system, with location-based reimbursement, does not provide incentives for.
Creation of a feedback loop may be as complex as integrated medical informatics systems, or as simple as a follow-up phone call, but it is incumbent on each hospitalist to examine the environment in which they practice and ensure this vital element of a safe and reliable system is being addressed. TH
Drs. Cumbler and Egan are assistant professors in the Section of Hospital Medicine at the University of Colorado at Denver.
- Stanton D. In harms way: the sinking of the USS Indianapolis and the extraordinary story of its survivors. New York, NY: Henry Holt and Company LLC; 2001.
- Kripalani S, LeFevre F, Phillips F, Williams M, Basaviah P, Baker D. Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA. 2007;297(8):831-841.
- Alpers A. Key legal principles for hospitalists. Am J Med. 1999;111(9):5-9.
- Dudas V, Bookwalter T, Kerr KM, Pantilat SZ. The impact of follow-up telephone calls to patients after hospitalization. Am J Med. 2001;111:26S-30S.
- Boockvar K, LaCorte H, Giambanco V, Fridman B, Siu A. Medication reconciliation for reducing drug
- discrepancy adverse events. Am J Geriatr Pharmacother. 2006;4:236-243.
- Epstein K, Juarez E, Loya K, Gorman MJ, Singer A. Frequency of new or worsening symptoms in the posthospitalization period. JHM. 2007;2:58-68.
- Mistiaen P, Poot E. Telephone follow-up, initiated by a hospital-based health professional, for postdischarge problems in patients discharged from hospital to home. Cochrane Database of Systemic Reviews 2006; 4. Article No.:CD004510. DOI:10.1002/14651858.CD004510.pub3.