Writing Effective Queries

Effective queries are clear, specific, neutral, and non-leading. A query should describe the issue and request verification or clarification without telling the site what answer to provide. Leading queries are problematic because they may influence the site to change data without proper source review. The purpose of a query is to obtain accurate data or documented explanation, not to force a preferred value. A poor query might say, "Please change temperature to 37.0 because 370 is wrong." This is leading and assumes the correct value. A better query would say, "Temperature is recorded as 370 degrees Celsius, which is outside the expected range. Please verify against the source document and correct or clarify." This query identifies the problem, explains why it is being queried, and asks the site to verify. Queries should include enough context for the site to respond efficiently. The participant ID, visit, form, variable, and issue should be clear. If the issue involves two fields, both should be mentioned. For example, "The discharge date is recorded as 2026-05-02, but the admission date is recorded as 2026-05-10. Please verify both dates against the source documents and correct or clarify." The query does not assume which date is wrong.

Queries should be concise. Long queries may be difficult to interpret. They should avoid blame, ambiguous language, and unnecessary technical jargon. They should also avoid asking multiple unrelated questions in one query. If two independent issues exist, separate queries may be clearer. The data manager should prioritize queries. Critical queries affecting safety, eligibility, primary outcomes, or serious adverse events should be addressed quickly. Minor spelling issues or non-critical optional fields may have lower priority. Excessive low-value queries can overwhelm sites and distract from critical quality issues. Table 5.6: Poor and Improved Query Wording