Preventing medication errors in long-term care: results and evaluation of a large scale web-based error reporting system.

OBJECTIVE: To describe the implementation and evaluation of a web-based medication error reporting system. DESIGN: Evaluation study. SETTING: Long-term care. PARTICIPANTS: 25 nursing homes in the US state of North Carolina. INTERVENTION: Detailed information about all medication errors occurring in a facility during a 1 year period was entered into a web-based reporting system. An evaluation survey was conducted to assess usability and the potential for the system to prevent errors. MAIN OUTCOME MEASURES: Number and specific characteristics of medication errors reported. A survey evaluating ease of use of the system and whether the participants thought it would help improve medication safety. RESULTS: 23 (92%) sites entered 631 error reports for 2731 discrete error instances when weighted by the number of times the errors were repeated. 51 (8%) errors were classified as having a serious patient impact requiring monitoring/intervention or worse. The most common errors were dose omission (203, 32%), overdose (91, 14%), underdose (43, 7%), wrong patient (38, 6%), wrong product (38, 6%), and wrong strength (38, 6%). Errors most commonly occurred during medication administration (296, 47%) and were attributed to basic human error (402, 48%). Seven drugs were implicated in a third (175, 28%) of all errors: lorazepam, oxycodone, warfarin, furosemide, hydrocodone, insulin and fentanyl. 20 sites (86% of respondents) completed the evaluation survey and participants found the system easy to use and thought it would increase accuracy of reporting and improve patient safety. CONCLUSIONS: The web-based medication error reporting system was easy to use, with strong indications that it would be a valuable tool for preventing future errors.

Syndromic surveillance for emerging infections in office practice using billing data.

PURPOSE: We wanted to evaluate the feasibility of conducting syndromic surveillance in a primary care office using billing data. METHODS: A 1-year study was conducted in a primary care practice; comparison data were obtained from emergency department records of visits by county residents. Within the practice, a computer program converted billing data into de-identified daily summaries of International Classification of Diseases, Ninth Revision (ICD-9) codes by sex and age-group; and a staff member generated daily summaries and e-mailed them to the analysis team. For both the practice and the emergency departments, infection-related syndromes and practice-specific thresholds were calculated using the category 1 syndrome codes and an analytical method based upon the Early Aberration Reporting System of the Centers for Disease Control and Prevention. RESULTS: A mean of 253 ICD-9 codes per day was reported. The most frequently recorded syndromes were respiratory illness, gastrointestinal illness, and fever. Syndromes most commonly exceeding the threshold of 2 standard deviations for the practice were lymphadenitis, rash, and fever. Generating a daily summary took 1 to 2 minutes; the program was written by the software vendor for a fee of dollar 1,500. During the 2003-2004 influenza season, trend line patterns of the emergency department visits reflected a pattern consistent with that of the state, whereas the trend line in primary case practice cases was less consistent, reflecting the variation expected in data from a single clinic. Still, spikes of activity that occurred in the practice before the emergency department suggest the practice may have seen patients with influenza earlier. CONCLUSIONS: This preliminary study showed the feasibility of implementing syndromic surveillance in an office setting at a low cost and with minimal staff effort. Although many implementation issues remain, further development of syndromic surveillance systems should include primary care offices.

Types of medication errors in North Carolina nursing homes: a target for quality improvement.

BACKGROUND: Medication errors are an important problem in nursing homes, but little is known about the types of medications involved in errors in this setting. Gaining a better understanding of the types of medications commonly involved in medication errors in nursing homes would be an important step toward quality improvement. OBJECTIVES: This study sought to describe the types of medication errors most frequently reported to a statewide repository by North Carolina nursing homes. We also examined whether nursing homes reporting an error involving a drug on the updated Beers list of medications considered potentially inappropriate for use in the elderly were likely to report a greater number of medication errors or more harmful medication errors compared with nursing homes that did not report such an error. METHODS: Medication errors were defined as preventable events that had the potential to cause/lead to or actually caused/led to inappropriate medication use or patient harm. We analyzed summary reports of medication errors submitted to the State of North Carolina by licensed nursing homes for the 9-month period from January 1, 2004, to September 30, 2004, using a Web-based reporting system. Drugs commonly involved in medication errors were summarized for all nursing homes in the state. Errors involving medications on the updated Beers list also were identified. Nursing homes were profiled and compared according to the type of medication error and whether the error reached the patient and/or caused harm. RESULTS: Among the 384 licensed nursing homes included in our analysis, 9272 medication errors were reported. The specific medication involved was documented for 5986 of these errors. The medications most commonly involved in an error were lorazepam (457 errors [8%]), warfarin (349 [6%]), insulin (332 [6%]), hydrocodone and hydrocodone combinations (233 [4%]), furosemide (173 [3%]), and the fentanyl patch (150 [3%]). The medication errors disproportionately included central nervous system agents (16%) and analgesics (11%). Medications considered potentially inappropriate in the elderly were frequently involved in the reported errors (10% of all reported errors), with the greatest number of such errors associated with lorazepam (457 [8%]), alprazolam (130 [2%]), and digoxin (74 [1%]). Nursing homes reporting potentially inappropriate drugs among their 10 most common medication errors also reported a significantly greater mean number of errors compared with nursing homes that did not report such errors (26.9 vs 17.6, respectively; P < 0.001), as well as a significantly greater number of errors that reached the patient (6148 vs 1393; P < 0.001). CONCLUSION: Based on the experience in North Carolina, quality improvement efforts in nursing homes should focus on the medications commonly involved in errors and should continue to discourage or closely monitor the use of medications considered potentially inappropriate in the elderly.