The 1999 publication of the Institute of Medicine's (IOM) Report on patient safety entitled To Err is Human focused attention on the problem of errors in health care.¹ Directors of pediatric residency programs should be involved in the initiatives for improving patient safety that ensue from the IOM report. This article will provide program directors and other pediatric leaders with an overview of the patient safety problem, discuss a framework for addressing safety, and describe actions pediatric program directors can take to improve patient safety.

The problem of patient safety in pediatrics

In its report, the IOM defined an error as “the failure of a planned action to be completed as intended or the use of a wrong plan to achieve an aim.”¹ According to the IOM report, as many as 98,000 Americans are killed each year because of errors in health care.¹ Although these figures have been disputed, concerns about patient safety have already had an impact on resident training. For example, safety concerns have stimulated legislative action attempting to change resident work hours.²

Errors occur in pediatric care as well. A JAMA study suggested that pediatric patients may be at increased risk of medical error during inpatient hospitalization compared with adult patients.³ In addition, hospitalized preterm infants may be at an increased risk of mortality during hospital shift changes.⁴ The public views safety as the foundation of quality, and it responds not only to such analyses but also to the narratives of patients injured by errors. Such narratives can be compelling when they relate to pediatric patients.⁵

Of course, safety has long been a concern of pediatric residency programs. Safety has frequently been viewed as an individual professional responsibility. The “hotbox” conference (where a resident explains his/her management of a clinical problem in front of senior physicians), morning report, and board certification and renewal all support the concept that professional development is a key contributor to quality of care and safety. James Reason suggests that there are 3 types of errors: skill-based slips, rule-based mistakes, and knowledge-based mistakes.⁶ Skill-based slips are actions that “deviate from current intention due to execution failures and/or storage failures.”⁶ For example, putting a carton of milk in the freezer as you jump up from the breakfast table in a rush for work is an example of a slip. Rule-based errors occur when an incorrect rule is applied to a decision. You may have a rule, “when I put away groceries, the milk goes in the refrigerator.” This rule leads to unnecessary refrigeration when applied to “milk-in-a-box,” milk in a container that does not require refrigeration. Knowledge-based errors occur when we do not have readily retrievable information upon which to base a decision.⁶ For example, when microwave ovens first appeared on the market, you might have heated food in an aluminum container because you were unaware that metal ignites in a microwave.

Training programs typically focus on improving the sophistication of our rules and the extent of our knowledge. Even in the unlikely event that a person had perfect rules and knowledge, they would still be vulnerable to slips. In other words, even experts make slips. For example, a recent Joint Commission Sentinel Event Alert discusses inadvertent medication errors made by experienced health care professionals resulting from confusing look-alike or sound-alike drug names.⁷ Traditional medical education view slips that damage patients as an occasion for discipline or shame, but slips seldom become the focus of systematic study.

New approaches to patient safety can complement traditional approaches by focusing on the systems in which physicians and other professionals operate.⁸ Such error management approaches require significant organizational forethought. Error specialists would describe such forethought as working at the blunt end of a problem. This shifts the focus from the sharp end of the problem, where the actual interaction between a medical professional and a patient occurs, to issues such as organizational culture, management decisions, information technology deployment, and training that can precede these interactions by years (Figure).

Figure. This graph schematically portrays the fact that most opportunities to decrease errors occur before the actual “error event.” These opportunities are greatest at the blunt end of the system, removed temporally and perhaps geographically from the error-generating act that occurs at the sharp end of the system. Latent errors occur at the blunt end of a delivery system, where they lay dormant until circumstances arise that follow it to facilitate an actual mistake at the sharp end of the care system.

Click on Image to view full size

In the case of medication errors, focusing on the blunt end of the medication process often leads organizations to consider the manner in which therapeutic intentions are communicated between members of the care team. Traditionally, medication delivery has been managed using oral and written systems of communication without embedded safety systems. On an individual nursing unit, the handwritten medical administration record is an example of a blunt-end system that tolerates errors.

Recent interest in computerized physician order entry (CPOE) stems from the premise that a blunt-end investment in CPOE can decrease errors in the care delivery process by improving, for example, communication between physicians, pharmacists, and nurses at the sharp end of the system. Such systems might also minimize slips that occur during the ordering process by checking dose calculations.⁹ Similarly, information systems might decrease the occurrence of rule-based and knowledge-based errors by enhancing the choice of rules and the knowledge available at the sharp end of the care delivery system (where an individual physician orders a specific dose of medication for a particular patient).^10,11 For example, a CPOE using a knowledge base and simple logic function can check for complicated drug interactions. Such a system therefore has the potential to enhance the information brought to bear at the sharp end of clinical problems (Figure).

Many of these error reduction strategies are just now being tested in medicine. But past success in the aviation and nuclear industries suggests that improvement in the safety of care provided to patients can be enhanced through such systematic approaches.^12,13

Unique opportunities

Pediatrics as a discipline has significant need for and experience with error reducing technologies. Perhaps the most obvious example is the use of portable calculators, a decision aid used universally by pediatric house staff to decrease drug dosing errors. Pediatrics is a medical specialty that is comfortable managing the blunt end of complex problems. This is precisely what we attempt to accomplish with our injury prevention, anticipatory guidance, and immunization programs. Pediatricians commonly think about prevention even when they are dealing with acute problems. For example, when they are faced with a motor vehicle accident victim in the emergency department, it is not uncommon for pediatric clinicians to wonder if the child was appropriately restrained at the time of the accident. The concept of a cascade of poor decisions leading to catastrophe is therefore very familiar to pediatricians. Program directors need to apply the same mindset when they are confronted with errors in their care delivery systems.

What we can do

In one form or another, the safety issue is likely to remain an important subject in health care. Because of the special vulnerability of children, the safety of pediatric care is likely to be scrutinized more than the safety of adult health care.⁵ Groups such as the IOM and the Joint Commission on Accreditation of Healthcare Organizations have sounded calls to action around the issue of medical safety. Thus, even among the unenthusiastic and the skeptical, prudence would support the following course of action:

1. Educate: Program directors and pediatric faculty should educate themselves about safety issues and introduce modern error reduction approaches into their residency training programs. In addition, they should encourage residents to participate actively in the patient safety initiatives undertaken by their hospital partners.
2. Report errors: To anticipate errors, physicians need to understand the types of errors that occur in their care environment, even if these errors rarely generate bad outcomes. The best method for accomplishing this level of understanding is to develop a system for reporting errors.¹⁴ It is helpful to focus on errors that do not actually harm patients, because these are the errors that professionals seem most willing to discuss. Residency programs should adopt a nonpejorative approach to error reporting so that residents will provide information that can help institutions better understand their error problems.
3. Fix systems, not people: Work on the blunt end of problems. When a resident makes a mistake, move upstream to find aspects of the delivery system that contributed to the resident's error. This will help prevent another resident from making the same mistake in the future. Residents are early warning detectors for system problems, because they tend to have a first-hand understanding of how clinical systems work. Use their knowledge as a starting point to improve your delivery system.

These 3 suggestions may appear overly simplistic. However, each calls for significant changes in teacher/student and clinician/administrator relationships within our academic institutions. The fundamental shift demanded by an error prevention perspective is not to perceive residents as a cause of errors, but as an important resource to help prevent them.¹⁴

• Christi Napper is Director of Research and Grants Administration at Children's Hospital of Alabama, Birmingham.
  James B. Battles is Senior Service Fellow in Patient Safety for the Agency for Healthcare Research and Quality (AHRQ).
  Crayton Fargason is Associate Professor at University of Alabama at Birmingham, and Medical Director and VP of Clinical Affairs at Children's Hospital of Alabama, Birmingham.
   
• The opinions expressed herein by the author(s) do not necessarily reflect the official endorsement of The Association of Medical School Pediatric Department Chairs, Inc. (J Pediatr 2003;142:359-60)
   
• Copyright © 2003 Mosby Inc. All rights reserved.
   
• 0022-3476/2003/$30.00 + 0
   
• doi:10.1067/mpd.2003.99