The Case

A 62-year-old woman with end-stage liver disease was hospitalized for recurrent variceal bleeding. On admission, she reported allergies to a number of medications as well as a food allergy to eggs. The patient was adamant about appropriate documentation of her allergies, especially her food allergy because "scrambled eggs almost killed me." Consequently, her medication and food allergies were clearly displayed on her medical chart as well as her wristband.

On hospital day 1, she underwent successful banding of a bleeding varix. Post-procedure, she developed some mild hepatic encephalopathy and was treated with lactulose. She was slightly disoriented but alert, and her diet was advanced—the diet order at the time was "low-salt diet."

She remained clinically stable until the morning of hospital day 2, when she had acute onset of tachypnea with audible wheezing and hypoxia. She recovered quickly with administration of continuous albuterol, hydrocortisone, and antihistamines. At the time of the event, one of the clinicians noticed that her breakfast tray, sitting by her bedside, included a plate of half-eaten bacon and eggs. She did not recall eating the eggs, probably because of her encephalopathy. She had no long-lasting complications from the allergic reaction.

The Commentary

Clinically significant allergic reactions to food in adults are relatively uncommon. Although population-based surveys of adults have yielded self-report rates of up to 20%, objective testing by food challenge confirmed a prevalence of only 1.8%.(1) In children, rates of parent-reported and confirmed allergies are higher, at 28% and 6%, respectively.(2) Severe allergies are much more common in young children, in whom the rate of severe allergy to the ovomucoid-ovalbumin allergens of hens' eggs has been estimated at 2.6% (3) and true anaphylaxis to peanuts at 0.5%.(4) Tolerance to allergenic foods usually develops with immunologic maturation during the preschool and school-age years. In older children and adults, peanuts and tree nuts (eg, walnuts, almonds, and pistachios) are the predominate causes of true anaphylaxis. In a case series of 32 fatal allergic food reactions reported to the national registry of the American Academy of Allergy, Asthma, and Immunology, 20 fatalities were from peanut ingestion, 10 from tree nut ingestion, and 1 each from fish and milk ingestion.(5) No deaths from egg consumption were reported.

As an example of an adverse event owing to a lapse in communication, this case is typical. An engaged patient and careful providers documented the patient's allergy in the appropriate places, but the information was lost on its way to the effector end of a care delivery system, in this case the production and distribution of meals. We do not know the methods by which food allergy information is communicated to food services personnel in the institution reporting this case, but practices are variable, and the institutional attention committed is likely to lag in comparison to drug allergies.

A common approach to the most severe allergens is exclusion of the offending item from the environment entirely. As peanuts are the predominant food cause of true anaphylaxis in adults, many providers of public accommodations (6), including hospitals and airlines, have elected to remove peanuts entirely as a food ingredient. The US Department of Transportation mandates that airlines that do serve peanuts offer a "peanut-free buffer zone" to air travelers who request it.(7) The well-established trend toward latex-free hospital environments is analogous. However, it is impractical to consider complete removal of potential allergens from hospital food services, in part because of the breadth of possible allergens, and in part because some of the more rare but well-documented allergens, such as eggs and fish, are staples in Western diets. Apart from excluding peanuts, manual methods to reduce the risk of delivering food to which a patient is allergic commonly include posting the information at the patient's door and bedside, or on patients themselves (eg, through the use of special wristbands), and requiring food services personnel to check the patient's allergies against the tray contents at the time of delivery. Much of this was in place at the reporting institution, however, and the case exemplifies the fallibility of these approaches.

In the health care setting, an allergic reaction to food is a variation of the more common hazard of allergic reactions to medications. Systematic prevention of the administration of drugs to which a patient is known to be allergic has been an early and sustained focus of institutional, regulatory, and commercial attention.(8) In comparison to many patient safety concerns, the problem is conceptually straightforward, and technological solutions, chiefly computerized provider order entry (CPOE) systems (9) with or without barcode medication administration (10,11), have been the primary approach. In the seminal study by Bates and colleagues, CPOE reduced orders for medications for which the patient was known to be allergic by 56%.(8)

Technological approaches to prevent delivery of foods to which a patient is allergic are considerably less well studied, perhaps owing to the comparatively low rate of clinically significant food allergies in adults. Many hospital food services departments remain paper based, and communication of allergy information from that collected by the health care providers (physicians, nurses, etc.) through to the arrival of the food at the patient's bedside relies on a complex network of communication and error-prone human practices.

For food services departments with robust information systems, however, the same principals established for preventing adverse drug reactions using CPOE and barcoded medication administration are immediately applicable. For allergy information recorded in a CPOE system to reliably trigger automated drug- or food-allergy checking, the allergies must be encoded in a way that the decision support logic can understand. Typically, a lexicon of choices is presented from which the documenting clinician must choose. The lexicon is often the formulary of available medications, and food items may or may not be listed. Allergy to eggs has historically received special attention, given the contraindication to influenza vaccination in patients with documented egg allergy (12), and so eggs (as well as latex and shellfish) are commonly included as choices. Modern CPOE systems usually receive their compendia of drug and allergy information by subscription from a commercial knowledge vendor, and these vendors are increasingly including a range of food items and other common non-drug allergens as selections.

Advanced food services automation software provides the health care industry–standard data interface (the Health Level 7 [HL7] Version 2 messaging standard [13]) that permits communication between the CPOE system and the food services software. If a physician documents a food allergy in the CPOE system, the food services information system receives the information via the HL7 interface and uses it to appropriately exclude hazardous items from the patient's menu and planned meals. The allergy information persists regardless of variation in diet orders over the course of the stay.

Unlike medications, which have few ingredients and are readily barcoded at the unit-dose level, prepared foods may have many ingredients, the offending allergen may not be obvious as an ingredient, and food ready for consumption is not easily barcoded. Modern food-service systems provide printed electronic labels for application to the tray for delivery, which include patient identifiers, the patient's food allergies, a room number, and the details of what is (or should be) on the tray. The patient identifiers on the label are manually checked for a match with the identifiers at the point of delivery. Such printed labels could easily include a barcode, permitting computerized verification that the right tray has been delivered to the right patient. However, this remains a step removed from the true closed loop of barcoded medication administration, as it is difficult to guarantee each printed label has been applied to the correct tray (unlike preprinted medication packets) or to assure that the assemblage of food items on the tray are as described on the label.

Fortunately, the patient in this case didn't suffer any long-term consequences from her allergic reaction. Although imperfect, adaptation of the techniques for preventing adverse medication events can significantly reduce the risk of patient exposure to non-medication allergens in the hospital, such as allergenic foods. Removal of peanuts from the hospital environment mitigates much of the risk, but a systematic approach to the range of allergens can help prevent the remainder of these rare but serious events.

Russ Cucina, MD, MS Assistant Professor of Medicine Hospitalist Group & Medical Informatics University of California, San Francisco

References

1. Young E, Stoneham MD, Petruckevitch A, Barton J, Rona R. A population study of food intolerance. Lancet. 1994;343:1127-1130. [go to PubMed]

2. Bock SA. Prospective appraisal of complaints of adverse reactions to foods in children during the first 3 years of life. Pediatrics. 1987;79:683-688. [go to PubMed]

3. Eggesbo M, Botten G, Halvorsen R, Magnus P. The prevalence of allergy to egg: a population-based study in young children. Allergy. 2001;56:403-411. [go to PubMed]

4. Sicherer SH, Munoz-Furlong A, Burks AW, Sampson HA. Prevalence of peanut and tree nut allergy in the US determined by a random digit dial telephone survey. J Allergy Clin Immunol. 1999;103:559-562. [go to PubMed]

5. Bock SA, Munoz-Furlong A, Sampson HA. Fatalities due to anaphylactic reactions to foods. J Allergy Clin Immunol. 2001;107:191-193. [go to PubMed]

6. Hartocollis A. Nothing's safe: some schools ban peanut butter as allergy threat. The New York Times. September 23, 1998.

7. Enforcing the Civil Rights of Air Travelers with Disabilities: Recommendations for the Department of Transportation and Congress. Washington, DC: National Council on Disability; February 26, 1999.

8. Bates DW, Cullen DJ, Laird N, et al. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group. JAMA. 1995;274:29-34. [go to PubMed]

9. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280:1311-1316. [go to PubMed]

10. Johnson CL, Carlson RA, Tucker CL, Willette C. Using BCMA software to improve patient safety in Veterans Administration Medical Centers. J Healthc Inf Manag. 2002;16:46-51. [go to PubMed]

11. Wright AA, Katz IT. Bar coding for patient safety. N Engl J Med. 2005;353:329-331. [go to PubMed]

12. Zeiger RS. Current issues with influenza vaccination in egg allergy. J Allergy Clin Immonol. 2002;110:834-840. [go to PubMed]

13. Health Level 7. Available at: http://www.hl7.org. Accessed June 13, 2006.