More than Conveying Information: Informed Consent as Speech Act

This article refers to:

Informed Consent: What Must Be Disclosed and What Must Be Understood?

In their target article, Millum and Bromwich (2021) situate their article against a backdrop of well-documented empirical research demonstrating that many participants have variable and often poor understanding of studies in which they enroll (Clapp and Fleisher 2018; Joffe et al. 2001; Koh et al. 2012; Montalvo and Larson 2014). Why do researchers continue with the “ritual” of informed consent when it is known that research participants clearly do not understand that to which they are consenting?

Millum and Bromwich’s target article argues that by unlinking the two elements of informed consent (i.e. disclosure by the researcher and understanding by the participant), participants can be considered as providing valid consent despite poor understanding. In their analysis, Millum and Bromwich state that the requirements for understanding can be viewed as parallel to the conditions for “the successful performance of the speech act of giving consent,” which often takes the form of a participant uttering the words “I consent to this research” or signing a form indicating the same.

However, though they invoke the concept of “speech act,” Millum and Bromwich do not discuss the crucial social contexts and procedural norms required for any speech act to take effect. Using the framework of speech acts, we ask, “what does informed consent do?” In doing so, we recognize the various other purposes of informed consent beyond and separate from the conveyance of information and argue that these additional purposes are an important justification for the ritual of informed consent.

The concept of “speech acts” was developed by philosopher of language J.L. Austin and was described in his William James Lectures delivered in 1955 titled “How to Do Things with Words” (Austin 1975). In these lectures, Austin provides several examples whereby uttering a sentence, something is done: some action is performed that changes something about the state of the world.

A woman dressed in a white gown standing before a man dressed in a formal suit and an ordained minister as well as gathered family and friends says “I do.” Her statement is not merely conveying information but formally changing the legal and social status of two people from single to married. A man, by writing in his will that “I give and bequeath my watch to my sister” is transferring ownership of his watch to his sister when he dies. A teacher saying “class is dismissed” is ending a pedagogical session and providing permission for students to get up from their seats. That is, with each of these statements, an individual conveys more than he or she (literally) says (Batten et al. 2018).

But saying these words or writing these words in any setting does not necessary result in the same effects. The same woman can say “I do” in response to any other question, or in the same sequence of conversation but the day before at a rehearsal for the ceremony, and they will not be the same speech act. The man can write “I give and bequeath my watch to my sister” in his personal diary or on a scrap piece of paper or as part of a work of fiction, and it will not have any effect. Certain accepted or conventional procedures must be in place and such procedures must be executed correctly and completely for the statements to have effect; there are many ways for this to fail.

In Millum and Bromwich’s article, they note that the conditions for a successful performance of the speech act of giving consent are that the participant understands (1) that she is giving consent; (2) how to exercise the right to give or refuse consent; and (3) to what she is consenting. According to their analysis, if the participant understands these three things, informed consent is valid. Given the examples provided by Austin, we may recognize that a participant can understand all three delineated statements and sign her name, but fail to provide a valid consent.

The context is critical to the efficacy of a speech act. Changing any aspect of a marriage negates the power of “I do”: if an ordained minister is not present, if there is no marriage license, if it is the wrong day, if it is spoken out of turn. Similarly, there are many possibilities for the invalidation of an informed consent.

The examples provided above focus our attention to the conditions of conventional procedures which must be met and allow us to ask: if those conditions are met what are the effects on the world? “I pronounce you…” creates a marriage. “I give and bequeath” transfers property. “Class is dismissed” terminates one activity and provides permission for disorder. While reading aloud a consent form may convey information, many other things are being done.

By looking at the effects of the speech act of the informed consent conversation, we find that informed consents are (a) a signal of respect for the research participant, to indicate and reinforce the participant’s autonomy and freedom to either sign or not sign an informed consent form and (b) an indication that the research project has and is going through the appropriate, institutionally approved and regulated process and therefore, that there are important protections in place.

Just as in a wedding, an ordained minister who stands before two people and declares “I now pronounce you husband and wife” does something with his words (blessing and actualizing the marriage), an informed consent only does something because other correct procedural steps have been taken. “I now pronounce you…” or “I do” indicate to those present that the couple has previously discussed and agreed to marriage, has obtained a marriage license, and that the relevant people and institutions have been involved to allow this marriage to be sanctioned. Similarly, the informed consent document indicates to the research participant that the researcher has followed necessary institutional and regulatory requirements (received IRB approval; obtained FDA approval for the trial if it is testing a drug, biologic or device; registered their trial; etc.).

Specifically, the procedures of informed consent signify to a research participant that appropriate protections are in place: an independent body (the IRB) has reviewed the research study and determined that it is ethically acceptable, that risks are minimized as much as possible, that risks of the trial are reasonable in relation to anticipated benefits, that there is fair subject selection, and that any adverse events will be reported for reevaluation of the risk-benefit ratio. The participant does not need to know what these required protections are or the details of who provides them.

The consent indicates that this research is institutionally sanctioned, whatever that may entail. Often in language there is a division of linguistic labor (Putnam 1973) which means that not everyone needs to have all knowledge and that non-experts trust that the experts are using language (or performing speech acts) in appropriate ways. This allows speakers to meaningfully talk about beech trees and elm trees even if they cannot tell the difference between them, since they can assume that there are experts who can. The research participant does not need to understand all that is required for approval of a protocol; they merely trust that the researcher (the expert) is authorized to proceed and that there are some sort of formal protections in place.

Millum and Bromwich find that the “very format [of written consent forms] communicates something false about the consent process”; in their legal-esque structure and small font, they are reminiscent of terms and conditions (which encourage the reader to skim) rather than documents which are meant to be read and understood. If we reframe informed consent as speech acts which indicate institutional blessing, the format of written consent forms actually successfully indicates that proper legal procedures have been undertaken and that the signing of these documents will perform an institutionally sanctioned autonomous agreement to participate.

Millum and Bromwich ultimately conclude that there is a distinction between what is required to be disclosed and understood for valid informed consent, which permits a “very thinly informed consent” to be acceptable. While we agree that it is acceptable that research participants do not fully understand, we arrive at this conclusion by recognizing informed consent as more than a process of information transfer.

As Millum and Bromwich ask, “why does it matter whether someone has given valid consent?” Many have questioned the goals of informed consent and whether or not it is achievable, recognizing that informed consent is aspirational and may serve other purposes (Burgess and O’Doherty 2019; Sreenivasan 2003; Wieten et al. 2019). By analyzing informed consents via the framework of speech acts, we find that consent does not primarily represent an information transfer, but rather a conveyance of respect toward research participants and a processual blessing by institutional and regulatory processes that provide at least some level of oversight and protection. Speech acts, like other theories and concepts in philosophy of language, can elucidate sources of confusion, specifically around the challenges of and misconceptions surrounding informed consent (Kim et al. 2013, 2015; Weinfurt 2018; Weinfurt et al. 2003). By understanding informed consent as performing a different utility than what is standardly perceived, we can develop more relevant metrics to measure whether informed consent has been achieved.