On a scale of 0(least likely) to 10 (most likely), how likely is it that this problem formulation presents a workable problem that can be managed based on best available information to date?

Answer Explanations

• 8

  Expert 1

  I believe this is a workable problem formulation, although risk management may have to be adapted to the lack of toxicity data on most components of the mixture.

• 7

  Expert 10

  The draft text is workable, but there can be a significant number of unknown PFAS in some drinking waters. Humans will also be exposed to precursors to perfluoroalkyl acids and it's uncertain how to include these substances in risk assessments. Some consideration (e.g. a degree of precaution) should be given to these issues in setting the threshold values in drinking water.

  The number of PFAS selected for subgrouping will be limited by data availability on exposure and toxicology.

• 3

  Expert 8

  As stated in 2.1 the problem formulation statement needs refinement. The UCMR3 provides information on the levels of 6 PFAS (PFOA, PFOS, PFNA, PFHxS, PFHpA and PFBS) in the public water supply. The criteria of using pchem properties to identify other PFAS present in drinking water is an insufficient criteria by itself, and would need consideration of use, pathway to drinking water, and actual measurement in drinking water.

• 5

  Expert 9

  This problem formulation presents a workable problem provided that those charged with its evaluation consider alternative or broader approaches to the problem of protecting human health from PFAS in drinking water. Therefore, the likelihood of the problem formulation presenting a workable problem isn't the problem formulation itself, but the perceptions, flexibility, and adaptability of those addressing the problem formulation/managing the problem. For example, the problem formulation is heavily weighted toward toxicological data, exposure data, and physical-chemical properties of PFAS themselves. For many PFAS, these data are limited or non-existent. Thus, the problem may be limited to only those PFAS for which "sufficient data" exist. The reality is that many PFAS lack specific data and thus new approaches to assessing their potential risks may need to be developed and/or considered.

• 2

  Expert 11

  Given the issues identified above, it is not clear to me what is being proposed in the statement and how workable the statement would be. As a result, I ranked the workability of the statement as a 2. In addition, since I am not fully familiar with the data available PFAS , I am not in a position to answer this question on the basis of technical feasibility.

• 2

  Expert 7

  Needs much more precision.
  Experience has shown that laws and regulations include CAS numbers for clarification. However, CAS numbers should be carefully applied - and not being used when they cause ambiguity: PFOS has many CAS numbers:
  a) perfluorooctanesulfonic acid (CAS number 1763-23-1)
  b) pPerfluorooctane sulfonic acid, potassium salt (CAS number 2795-39-3
  c) perfluorooctane sulfonic acid, ammonium salt (CAS number 29081-56-9
  d) perfluorooctane sulfonic acid, lithium salt 29457-72-5
  e) perfluorooctanesulphonate 307-35-7
  f) perfluorooctane sulfonate 45298-90-6
  g) Perfluorooctane sulfonic acid, diethanolamine salt 70225-14-8
  etc. These are only linear PFOS compounds, which a chemical laboratory determines typically as a salt; thus, no differentiation can be made but the list implies that each of the compounds needs to be controlled.

• 8

  Expert 4

  I think the scope (PFAS in drinking water / setting MCLs) is limited enough that there is a high likelihood it can be completed. However note that “setting an MCL” is a risk management task that would be based, in part, on the risk assessment.

• 7

  Expert 2

  For PFAS such as PFOS, PFOA, PFHxS and PFNA, there is adequate information to perform risk assessment. For other PFAS such as long chain carboxylates, poly/perfluoroethers, data are limited. Furthermore, how PFAS precursors, which may not be measurable in an environmental matrix, but upon exposure biotransform to a persistent end stage PFAS metabolite in human bodies, are considered in risk assessment? In fact, majority of PFAS in commerce are precursors and without adequate information about their usage and biotransformation lot of uncertainty will be introduced. My suggestion would be to start making subgroups of PFAAS first. Those subgroups should have some commonality such similar structure or similar mode of toxicity etc. Available evidence suggests that each subgroup may have different mode and potency in terms of toxicity. Once those subgroups are identified, we can tackle the risk assessment one by one.

• 7

  Expert 5

  PF can be tailored to represent/frame the issue as specifically or as loosely as desired. Each extreme has positives and negatives. It seems that the most likely impediment to success may be the ability to assign individual PFAS chemicals to groups, and the availability of enough toxicological information to assign potency values to wither the index chemical for the group, or "reliable-enough" data (e.g., in vitro concentration-response data) for other components to either reaffirm adequacy of grouping or to evaluate potency relative to the index chemical. While this response centers on grouping, the issue of grouping and potency may be the two most important issues to resolve.

• 5

  Expert 6

  I am ambivalent about whether the problem, as formulated, can be "managed based on best available information to date" hence my 5 rating. Given the extent of the gaps in knowledge of toxicological effects of PFAS, relative potencies and differences in environmental fate/handling, there is strong likelihood that attempts to construct a robust risk assessment framework for multiple PFAS exposures will fall short of objectives. However, the approach outlined in the draft Problem Formulation statement seems to offer the best opportunity to construct a risk assessment framework that could at least partially address the problem and provide a way forward.

• 5

  Expert 3

  The Problem Formulation above presents a workable problem in that the scope of exposure is limited to drinking water. The bigger question is whether the assessment will provide the information needed to achieve the goal of scientifically supported risk management options under SDWA.

  The Conceptual Model for the assessment (as noted above) needs to be developed. For example, "general population" needs to be defined; presumably this will include all ages with exclusions based on bioavailability. For example the Conceptual Model can exclude in utero exposure for those PFAS that can't cross the placenta. Statements could be made regarding early life post-natal exposure for breast milk exposure vs. water exposure; this model could include water used in formula if that is congruent with the purpose of the assessment. Rather than including water used in cooking and /or bathing, the Conceptual Model could be restricted solely to drinking water consumed orally as water. Statements in the Conceptual Model can be made about potential exclusion of inhalation exposure of PFAS from drinking water based on considerations of volatility.

  The Problem Formulation states that resources will include current data on occurrence in drinking water SOURCES. MCLs are established for finished drinking water as it leaves the processing plant. The Conceptual Model needs to describe the assumptions or data that will be used in estimating PFAS exposure at the tap (or more generally as water enters the distribution system).

  The Conceptual Model needs to describe the health endpoints to be assessed (adverse outcomes that one is trying to reduce as a consequence of the risk management). This leads to descriptions of the populations at risk and refines the scope of the risk assessment. For example, the analyst asks whether developmental effects observed or expected? If they are not, the population can be restricted to post-developmental ages.

  These are not meant to be annoying questions or points. Rather the intent is to defines more clearly if the risk assessment is designed to provide information needed for scientific support in choice of risk management options.