Approach 1 is the one most likely to look clean in a spreadsheet and least likely to survive review. If the chemical's IRIS assessment postdates the 2005 Guidelines for Carcinogen Risk Assessment (EPA/630/P-03/001F), an unadjusted adult slope factor applied to a child dose understates early-life risk for mutagenic carcinogens and fails EPA compliance for post-2005 assessments. The check is mechanical: open the IRIS record, read the assessment date, and refuse to publish a child risk number until that date is on the page.
Approach 2 is the winner for any assessment you must defend. It applies the Age-Dependent Adjustment Factor of 10x for exposure from birth to under 2 years and 3x for exposure from age 2 to under 16 years, weighted by the time spent in each band. The arithmetic is a duration-weighted sum, not a single blanket multiplier on the whole childhood window. If your exposure table has only one row labeled "child," you cannot yet run this approach; split the record first.
Approach 3 is permitted, not presumed. The Guidelines allow chemical-specific early-life data to replace the default factor when mode-of-action data justify it. That is a higher evidentiary bar than preferring a published study: you need a mode-of-action line of reasoning that explains why the default over- or under-protects for this chemical — the same act-on-weak-signals caution pattern the 1979 trichloroethylene review (discussed above) illustrates.
So the rule for this section is narrow and operational. Default to Approach 2 for post-2005 assessments, keep Approach 1 only as a labeled screening calculation that you do not present as compliant, and reserve Approach 3 for cases where the mode-of-action file is already assembled. The table's explicit winner is the ADAf-adjusted default, because it is the only one of the three that is both computable from a standard IRIS slope factor and defensible without an additional chemical-specific argument.

Costs and Numbers That Matter
For an exposure confined to birth through age 2, omitting the 10x Age-Dependent Adjustment Factor (ADAf) in the 2005 Guidelines for Carcinogen Risk Assessment (EPA/630/P-03/001F) leaves that band's risk contribution at one-tenth of the adjusted value. Because the factor is exactly 10, the error is a full order of magnitude, not a rounding or unit-conversion difference.
The factor is not a blanket multiplier for all childhood exposure. A child with equal annual dose from birth to age 16 receives a duration-weighted blend: 2 years at 10x and 14 years at 3x. The weighted average is (2 × 10 + 14 × 3) ÷ 16 = 3.875, so the correct blended factor is 3.875x, not 10x. Multiplying the entire childhood dose by 10 would overstate the age-2-to-under-16 band by more than a factor of three.
| Age band | Years | ADAf | Weighted contribution |
|---|---|---|---|
| Birth to age 2 | 2 | 10x | 20 |
| Age 2 to under 16 | 14 | 3x | 42 |
| Birth to under 16 | 16 | 3.875x | 62 |
The verification burden is small: one check of the chemical's IRIS page for the assessment date and whether an ADAf was applied. That single review is the difference between an estimate that carries the 10x birth-to-2 adjustment and one that silently drops it.
The check also prevents a second, subtler error: applying 10x to a whole 0–16 dose. The table shows why the two mistakes are not opposites in cost—one understates the infant band, the other overstates the older-child band—but both come from treating an age-specific factor as a childhood-wide factor.
Because the inputs are already in the IRIS record, the review adds no sampling, modeling, or laboratory cost; it is a documentation check against the 2005 Guidelines' age bands. It also establishes whether the assessment predates or postdates the 2005 Guidelines, which determines whether the factor is already embedded in the value.
The practical cost of skipping the check is therefore an unadjusted number that can understate the youngest exposure band by 10x while appearing complete. In a screening or priority-setting context, that is the largest arithmetic error available in this framework, and it is avoidable without new data or new measurements.

What the Evidence Does NOT Establish
That is a mode-of-action check, not permission to drop the adjustment silently. The practical rule is to record the basis for the mode-of-action determination, then state whether the default applies or whether a chemical-specific analysis supports a different treatment. If the record is thin, keep the default and flag the uncertainty. If the record is strong, cite the specific finding that justifies departing from it. Either way, the reasoning belongs in the workpaper next to the risk number.
IRIS values also need a label. EPA states that IRIS values represent the first two steps of the four-step risk assessment process described by the National Research Council and are not necessarily intended for use as regulatory levels. In plain terms, a slope factor or unit risk pulled from IRIS is a screening input. A risk estimate built from it is useful for priority ranking, comparing sites, and deciding what deserves closer review; it is not, by itself, a compliance determination. If a memo turns an IRIS-derived screening number into a pass/fail finding, the memo has crossed the line EPA's own caveat warns against.
The edge cases are where the adjustment rule is easiest to misapply. The table below separates situations where the ADAf framework is relevant from situations where it is not, based on the exposure window and the age of the assessment. It is the decision aid for this section only: it does not replace the mode-of-action check above, and it does not resolve which adjusted or unadjusted calculation is preferable in a full comparison.
| Situation | Does the 2005 ADAf rule need to be addressed? | What the analyst should document |
|---|---|---|
| Exposure occurs entirely after age 16 | No adjustment under the 10x/3x age bands; the ADAf rule is irrelevant to that exposure window | Show that the exposure period begins at or after age 16, so neither birth-to-under-2 nor age 2-to-under-16 exposure is being evaluated |
| Exposure includes birth to under 2 years | Yes, if the chemical's assessment and mode of action support applying the default | Identify the exposure portion in that band, the assessment basis, and the mode-of-action rationale for using or not using 10x |
| Exposure includes age 2 to under 16 years | Yes, if the chemical's assessment and mode of action support applying the default | Identify the exposure portion in that band and the rationale for using or not using 3x |
| IRIS assessment predates the 2005 Guidelines | Check the assessment date and basis before assuming the 2005 default structure is embedded in the value | State the assessment date, note that it predates EPA/630/P-03/001F, and explain whether the value already reflects any age-dependent treatment or whether the 2005 framework is being applied as a separate screening step |
| IRIS assessment postdates the 2005 Guidelines | Check the assessment's own treatment of early-life susceptibility before layering on a default | Confirm whether the assessment already incorporated the 2005 framework and whether the mode of action is genotoxic/mutagenic or non-mutagenic |
One more boundary: the Guidelines are dated 2005, and the citation to use is EPA/630/P-03/001F, Guidelines for Carcinogen Risk Assessment, National Center for Environmental Assessment, Office of Research and Development, Washington, DC. A pre-2005 IRIS assessment should not be treated as if it automatically contains the 2005 age-dependent framework. The check is simple: find the assessment date, find the mode-of-action discussion, and find whether early-life susceptibility was addressed. If any of those are missing, the risk number is not yet ready to be presented as adjusted.

Auditing a Child Risk Number
Start the audit with the inputs you actually have in hand: an IRIS oral slope factor in (mg/kg-day)⁻¹ and a child exposure history split into two age bands, 2 years at ages 0–<2 and 14 years at ages 2–<16. Before you multiply anything, check the assessment date on the IRIS record against the 2005 final Guidelines for Carcinogen Risk Assessment (EPA/630/P-03/001F, National Center for Environmental Assessment). If the chemical's evaluation predates the 2005 Guidelines, flag the record for separate treatment; if it postdates them, the age-dependent adjustment is on the table. As EPA's IRIS page states, IRIS values cover only the first two steps of the risk assessment process and are not necessarily intended as regulatory levels, so the date check belongs in your worksheet, not in a footnote.
Next, confirm the mode of action. The adjustment is keyed to mutagenicity, not to a general “child sensitivity” assumption. Pull the hazard narrative behind the slope factor and verify that the assessment characterizes the carcinogenic mode of action as mutagenic, consistent with the framework used for DNA-reactive agents in ICH M7 (R1) (2015) as discussed in the risk assessment literature on N-nitrosamines. If the mode of action is non-mutagenic or the record is silent, stop and document the basis for not applying the adjustment rather than silently carrying a default into the calculation.
Recompute the duration-weighted factor as derived in the Costs and Numbers section above: for this 2-year/14-year split the blended value is 3.875x, not 10x. Write the band-by-band arithmetic into the audit file so a reviewer can trace each term back to the exposure split.
Both directions of error are material, as the blended-factor arithmetic above shows: omitting the adjustment understates the blended child risk by roughly 3.9x, while applying 10x across all 16 years overshoots the duration-weighted result by roughly 2.6x. Neither mistake is cosmetic: one under-protects, the other distorts prioritization.
Use this copy-usable verification checklist before you sign the risk number: (1) IRIS assessment date confirmed against the November 2001 draft / 2005 final Guidelines; (2) mode of action confirmed as mutagenic in the assessment narrative; (3) exposure years entered by band, 2 and 14, with total 16; (4) duration-weighted factor recomputed as 62 / 16 = 3.875x, not 10x; (5) slope factor multiplied by the blended factor, with unadjusted and adjusted results both retained; (6) any decision to set the adjustment aside justified in writing with the specific record citation. If any line is blank, the number is not ready for review.

Decision Rules: Flag or Accept the Number
Once the assessment date and exposure window are known, the decision is mechanical: accept the IRIS number only if the age-dependent adjustment is already embedded or you have applied it yourself; otherwise flag it. This is the single accept-or-flag protocol that ties the earlier checks together, and it should be written into the workpaper before any slope factor or unit risk value is carried into a calculation.
Start with the date on the IRIS assessment. If the assessment postdates the 2005 Guidelines for Carcinogen Risk Assessment (EPA/630/P-03/001F) and the exposed population includes ages birth to under 2 years, require the 10x Age-Dependent Adjustment Factor (ADAf) in the calculation or reject the number for that use. The check is not whether the IRIS page mentions children in general; it is whether the value you are importing has already accounted for the early-life window in your exposure scenario. Because IRIS values are screening inputs rather than regulatory levels (as noted above), the burden stays with the assessor to confirm the intended use and the age applicability before accepting the value.
If the assessment predates 2005 and the IRIS page shows no ADAf, flag the value as unadjusted for early life and document that limitation. Do not silently use it for child exposure. The documentation should identify the chemical, the IRIS value used, the assessment date shown, the absence of an ADAf, and the exposed age groups in your scenario. That record matters because a pre-2005 value may be perfectly serviceable for an adult-only exposure case yet incomplete for a birth-to-under-2 case; the number itself does not tell you which situation you are in.
Apply the same rule to partial age windows. For a post-2005 assessment, exposure from birth to under 2 triggers the 10x check; exposure from age 2 to under 16 triggers the 3x check. If your scenario crosses both windows, keep the age bands separate rather than applying one factor to the whole exposure duration. If the exposure begins after age 16, the ADAf check does not supply a basis for an early-life adjustment, and the workpaper should say that the age window falls outside the adjustment bands described in the 2005 Guidelines.
The final accept-or-flag entry should be short and auditable: assessment date, exposure ages, ADAf status, mode-of-action status, and disposition. Accept only when the factor is applied or justified away in writing under the applicable rule. Flag when the assessment is pre-2005 with no ADAf, when a post-2005 early-life exposure lacks the required adjustment, or when a non-mutagenic mode of action is asserted without the mandated chemical-specific early-life evaluation. Anything less leaves a risk number in the file whose age basis has not been resolved.
What to do next
| Step | Action | Why it matters |
|---|---|---|
| 1 | On the IRIS chemical page, record whether the assessment predates or postdates the 2005 Guidelines before using its slope factor or unit risk. | This date determines whether the value is already adjusted or needs ADAf handling. |
| 2 | If the assessment is post-2005, apply ADAf 10x for ages 0–<2 and 3x for ages 2–<16 to the child's dose. | The 2005 Guidelines require these early-life adjustments. |
| 3 | If the assessment is pre-2005, flag the slope factor or unit risk as unadjusted and verify on the IRIS chemical page whether an ADAf was applied. | Pre-2005 values may not include the early-life factor. |
| 4 | Before reporting any risk number, apply the ADAf or explicitly justify not applying it. | A risk number must not omit the adjustment without justification. |
| 5 | Document the assessment date, age band, and ADAf status in the calculation notes. | This preserves the basis for the final risk estimate. |
| 6 | Re-check the IRIS chemical page and adjustment status before final use of the value. | Confirms the value's status has not changed. |
Frequently Asked Questions
When must I refuse to publish a child risk number using an unadjusted adult slope factor?
If the chemical's IRIS assessment postdates the 2005 Guidelines for Carcinogen Risk Assessment (EPA/630/P-03/001F), an unadjusted adult slope factor applied to a child dose understates early-life risk for mutagenic carcinogens and fails EPA compliance for post-2005 assessments.
What is the mechanical check before publishing a child risk number?
Open the IRIS record, read the assessment date, and refuse to publish a child risk number until that date is on the page.
What age bands does the Age-Dependent Adjustment Factor use?
It applies 10x for exposure from birth to under 2 years and 3x for exposure from age 2 to under 16 years.
How should the Age-Dependent Adjustment Factor be calculated across childhood?
The arithmetic is a duration-weighted sum, not a single blanket multiplier on the whole childhood window.
What must I do if my exposure table has only one row labeled "child"?
You cannot yet run the Age-Dependent Adjustment Factor approach; split the record first.
When can chemical-specific early-life data replace the default factor?
The Guidelines allow chemical-specific early-life data to replace the default factor when mode-of-action data justify it, requiring a mode-of-action line of reasoning that explains why the default over- or under-protects for this chemical.
Quick answers
| Why does an unadjusted adult slope factor applied to a child dose fail for a chemical whose IRIS assessment postdates the 2005 Guidelines for Carcinogen Risk Assessment (EPA/630/P-03/001F)? | It understates early-life risk for mutagenic carcinogens and fails EPA compliance for post-2005 assessments. |
| What is the mechanical check described for using an IRIS assessment in a child risk number? | Open the IRIS record, read the assessment date, and refuse to publish a child risk number until that date is on the page. |
| What Age-Dependent Adjustment Factor does Approach 2 apply for exposure from birth to under 2 years and from age 2 to under 16 years? | It applies 10x for exposure from birth to under 2 years and 3x for exposure from age 2 to under 16 years, weighted by the time spent in each band. |
| Why must an exposure table with only one row labeled "child" be split before running Approach 2? | Because the arithmetic is a duration-weighted sum, not a single blanket multiplier on the whole childhood window. |
| When may chemical-specific early-life data replace the default factor under Approach 3? | The Guidelines allow chemical-specific early-life data to replace the default factor when mode-of-action data justify it. |
Also worth reading: Why risk assessment is the most critical step in a successful financial audit: Why risk assessment is the · How to tailor risk assessment for complex financial audits: How to tailor risk assessment · Audit risk assessment: 5% anomaly threshold vs. materiality—expand testing: Audit risk assessment: 5% anomaly