5.2 Responses to Charge Questions
The third topic on the agenda had only the following charge question: "ATSDR has reviewed studies that employed analytical, observational, and questionnaire techniques for identifying children who exhibit soil-pica behavior. What are the strengths and weaknesses of these methods? Are there other methods available to identify people with soil-pica behavior? Considering the pros and cons of each method, which method do you think ATSDR should use to identify people with soil-pica behavior?"
When addressing these issues, the panelists referred to their earlier responses (Section 3.3) on the utility of the different methods for characterizing the prevalence of soil-pica behavior. In summary, because analytical, observational, and questionnaire approaches all have their limitations, the panelists recommended that future studies use multiple approaches to identify soil-pica children. The panelists added that ATSDR needs to validate the findings of these approaches with epidemiologic studies that measure biomarkers of exposure to give the greatest degree of confidence in the results.
The panelists' other comments on means for identifying soil-pica children focused on how to conduct effective surveys. Following are their comments.
Behaviors to consider in survey questions. Several panelists recommended that ATSDR review the existing literature on soil ingestion questionnaires before developing their own surveys (SD, NF, BL). For instance, one panelist indicated that his research on children living in Midvale, Utah, has found no association between the frequency of hand washing, as reported by parents, and blood lead levels (BL). Another panelist added that his research found soil ingestion rates to be essentially uncorrelated with several behaviors, such as thumb sucking, hand washing, and use of pacifiers (SD). Both panelists thought that ATSDR could select survey questions judiciously by reviewing the results of these and other studies on soil ingestion.
Illustrating the utility of this suggestion, the panelists debated the need for including questions about hand washing on soil ingestion surveys. One panelist recommended that surveys address this topic, but two others noted that their own research found hand washing to be uncorrelated with other metrics of soil ingestion (i.e., analytical data from a tracer study and blood lead levels) (SD, BL). Given this precedent, another panelist questioned the utility of including detailed questions about personal hygiene on a soil ingestion survey, particularly since questions are directed at parents whose responses might be unreliable (NF).
Specific suggestions for survey questions. The panelists offered several suggestions for specific survey questions that can help ATSDR identify soil-pica children. One panelist, for instance, suspected that ATSDR can adequately characterize parents' perceptions of soil ingestion among their children by asking just five carefully crafted questions (NF). She recommended using staged questions that start by focusing on general information (e.g., "does your child eat soil or dirt?") and end by addressing more specific information (e.g., "is this a weekly event?"). Another panelist recommended that surveys use questions and response options that are unambiguous (BL), (e.g., "how often have you observed your child put soil or dirt in his or her mouth?", with response options of "never," "once a month," "once a week," "once a day," or "several times a day.") He thought such descriptive response options are needed to derive semi-quantitative accounts of soil-pica. Another panelist agreed with the approach of using specific, unambiguous questions (SD). He explained that his research has found responses to direct and specific questions about a child eating dirt and soil to be associated with higher levels of soil ingestion as determined analytically, while responses to indirect questions were essentially unrelated to analytic measures of soil ingestion.
Another panelist provided two more suggestions on developing specific questions (NF). First, to avoid recall bias, she recommended that ATSDR ask parents to remember recent events, as opposed to events that occurred months or years earlier (e.g., "how often have you observed your child put soil or dirt in his or her mouth in the last week?" rather than "have you ever seen your child put soil or dirt in his or her mouth?") Second, this panelist cautioned that using surveys to derive information on soil ingestion rates would be challenging. She thought asking parents about behaviors that lead to dramatically different ingestion rates (e.g., eating with fingers as compared to putting handfuls of soil into their mouth) might help identify the children who likely ingest the greatest quantities of soils, but she stressed that surveys are not the best approach for getting quantitative data on soil ingestion rates.
Other considerations for surveys. The panelists listed several other suggestions for ATSDR to consider when developing soil ingestion surveys. First, one panelist noted that face-to-face interviews conducted by properly trained individuals typically generate the most detailed information, though she acknowledged that this approach is very time consuming and expensive (NF). This panelist added that ATSDR should draft soil ingestion surveys to assess the frequency of specific behaviors, and not to determine simply if these behaviors occur. Other recommendations included noting on surveys that the questions pertain to typical children's behavior (BL), ensuring that surveys are not too long (BL), and having trusted and respected individuals (e.g., community leaders or pediatricians) administer surveys (DV).