| Sumario: | In this study, computer-based surveillance techniques were used to identify hazards and adverse events associated with medical devices in a 520-bed teaching hospital that was accustomed to using computer monitoring to prevent and detect adverse drug events. 'Adverse events' were defined as incidents involving harm to patients, and events that caused no harm to patients were considered 'hazards.' This hospital used an electronic medical record, which consisted of an integrated clinical database and frame-based medical decision system with over 1300 readily accessed computer terminals to record all clinical information, including data from the pharmacy, billing, laboratory, electrocardiography, medical records, digital radiology, as well as hazards and adverse events reported by hospital personnel. Using information gathered from this electronic medical record and extensive interviews with caretakers, a database was created concerned with common medical devices and the problems associated with them. The computer was configured to identify or 'flag' reports of possible hazards or adverse events involving medical devices and to recognize International Classification of Diseases, 9th edition (ICD-9) codes that included medical devices in their descriptions. In addition, a telemetry checklist was developed to identify problems with cardiac catheters that were routinely reported by the nursing personnel. Over a 5-week period, telemetry technicians marked all false alarms and missed ectopy on the checklist. Finally, questions were added to the patient postdischarge telephone satisfaction survey to garner patient perspectives on errors involving medical devices. The study took place from January through September 2000 and included all admissions to the hospital, except obstetric and newborn patients. During that time, there were 20,441 eligible patients who stayed overnight and 7124 short-stay (1-day) or day-surgery patients admitted to the hospital. Medical devices used by these patients included 20,948 urinary catheters, 3530 central venous catheters, 8197 arterial catheters, and 5687 ventilators. During the study period, 80 device-related adverse events were reported by hospital personnel using the electronic medical record. A total of 7059 flags were triggered, of which 552 (7.8%) were considered hazards or adverse events. When the medical records of patients were examined, 95% of the reports were confirmed. The telemetry checklist revealed 593 false alarms and 4 missed ectopic events that were not otherwise recorded by the surveillance methods. None of these problems caused harm to the patient. In the postdischarge telephone interview, 8% of patients reported having a problem with medical devices; 7% of these were confirmed. ICD-9-code-based surveillance triggered 1122 flags that most frequently involved orthopedic, vascular, and dialysis devices, as well as urologic catheters (25%, 19%, 15%, and 9%, respectively) ( Table 1.) Over one fourth (26%) of the problems detected by ICD-9-code-based surveillance involved short-stay patients whose problems were mostly with dialysis (47%) and orthopedic devices (33%). Seventy-two percent of the surveillance detections were confirmed by a medical record review. No one detection method identified a majority of the total number of hazards or adverse events. Overall, the rate of detection of adverse events causing patient harm was 1.6 per 1000 hospitalized patients (95% confidence interval [CI], 0.9-2.5) for incidence reports, 27.7 per 1000 patients (95% CI, 24.9-30.7) for computer flags, and 64.6 per 1000 patients (95% CI, 10.4-69.1) for ICD-9 codes. The overall rate of detection was 83.7 per 1000 patients (95% CI, 78.8-88.6).
|
|---|