Can poison control data be used for pharmaceutical poisoning surveillance?

Abstract

Objective: To determine the association between the frequencies of pharmaceutical exposures reported to a poison control center (PCC) and those seen in the emergency department (ED).

Design: A statewide population-based retrospective comparison of frequencies of ED pharmaceutical poisonings with frequencies of pharmaceutical exposures reported to a regional PCC. ED poisonings, identified by International Classification of Diseases, Version 9 (ICD-9) codes, were grouped into substance categories. Using a reproducible algorithm facilitated by probabilistic linkage, codes from the PCC classification system were mapped into the same categories. A readily identifiable subset of PCC calls was selected for comparison.

Measurements: Correlations between frequencies of quarterly exposures by substance categories were calculated using Pearson correlation coefficients and partial correlation coefficients with adjustment for seasonality.

Results: PCC reported exposures correlated with ED poisonings in nine of 10 categories. Partial correlation coefficients (r(p)) indicated strong associations (r(p)>0.8) for three substance categories that underwent large changes in their incidences (opiates, benzodiazepines, and muscle relaxants). Six substance categories were moderately correlated (r(p)>0.6). One category, salicylates, showed no association. Limitations Imperfect overlap between ICD-9 and PCC codes may have led to miscategorization. Substances without changes in exposure frequency have inadequate variability to detect association using this method.

Conclusion: PCC data are able to effectively identify trends in poisonings seen in EDs and may be useful as part of a pharmaceutical poisoning surveillance system. The authors developed an algorithm-driven technique for mapping American Association of Poison Control Centers codes to ICD-9 codes and identified a useful subset of poison control exposures for analysis.

Figure 1

American Association of Poison Control Centers (AAPCC) codes, mapped into one of 92 International Classification of Diseases, Version 9 (ICD-9) based substance categories using a stepwise algorithm as shown above. The 10 most common categories of exposures in the emergency department (ED; table 1) were used for the analysis and are labeled ‘Common ED Substance Category’; the remainder are labeled ‘Other Substance Category.’

Figure 2

Scatter plots of poison control exposures versus emergency department exposures by quarter. NSAIDs, non-steroid anti-inflammatory drugs.