emergency department visits estimated for the noninstitutionalized population for that period. This estimate was twice as high as would be expected from the NHIS-derived estimate presented earlier.
Unlike NAMCS and the NHAMCS outpatient file, females (550 per 100,000) had a slightly higher rate of visits than males (520 per 100,000), although this difference was far narrower than in the NHIS estimates. The age distribution differed from the NHIS, NAMCS, and NHAMCS outpatient subset, with those 65 years of age and over having the lowest rate of all three age groups. Once again, examination of narrower age strata, especially for those younger than 18 years of age, was beyond the scope of this analysis. As was the case with the outpatient NHAMCS file, the visit rate for whites (520 per 100,000) is lower than that for blacks (830 per 100,000). Rates of poisoning-related visits ranged from a low of 480 per 100,000 persons in the Midwest to a high of 560 per 100,000 persons in the West, a regional pattern that, once again, varied in comparison to each of the other datasets. Of these emergency department visits, 18 percent resulted in a subsequent admission to a hospital, and an additional 8 percent were transferred to another facility, while 39 percent were referred to another physician or clinic.
In 2001, questions were added to the survey concerning whether the patient was seen in the emergency department in the past 72 hours and whether the visit was initial or follow-up. Of those with nonmissing data, 97 percent had not been seen in the emergency department in the past 72 hours, and 95 percent were an initial visit (data not shown in Table 3-6).
The most common ICD-9 codes were “toxic effects of other substances (venom; bites of venomous snakes, lizards, and spiders; tick paralysis)” (989.5), “poisoning by unspecified drug or medicinal substance” (977.9), and “suicide and self-inflicted poisoning by tranquilizers” (E950.3). A much higher percentage of emergency department poisoning visits were associated with a reason or symptom of poisoning (40 percent) than from either the NAMCS or NHAMCS outpatient files. Nonetheless, as was the case with NAMCS and the outpatient NHAMCS file, the increase in observations and estimated visits that would be obtained by adding in records with a patient reason or symptom of poisoning that lacked a concomitant ICD-9 code remains negligible: an additional 6 percent to the estimate (1,514,000 as opposed to 1,428,000 cases).
The percentage of total emergency department visits estimated to be associated with poisoning in this analysis is slightly higher here than one published by McCaig in 1996 using 1993–1996 NHAMCS emergency department files (1.1 percent of all visits in that study compared with 1.5 percent here) (McCaig and Burt, 1999). This difference is likely due to differences in the definition of poisoning used in the earlier study, which was limited to visits with a poisoning-related E-code rather than utilizing