Article Text
Abstract
This article introduces the notion of ‘routinisation’ into discussions of informed consent. It is argued that the routinisation of informed consent poses a threat to the protection of the personal autonomy of a patient through the negotiation of informed consent. On the basis of a large survey, we provide evidence of the routinisation of informed consent in various types of interaction on the internet; among these, the routinisation of consent to the exchange of health related information. We also provide evidence that the extent of the routinisation of informed consent is dependent on the character of the information exchanged, and we uncover a range of causes of routinisation. Finally, the article discusses possible ways of countering the problem of routinisation of informed consent.
- Autonomy
- Information Technology
- Informed Consent
Statistics from Altmetric.com
Introduction: informed consent and its limits
The moral value of obtaining informed consent from a patient before the initiation of a medical procedure involving the patient has been widely recognised and is today deeply embedded in current practice in the health sector.1 The value of informed consent is typically related to the role consent plays in protecting personal autonomy.2–7 Ideally, asking for consent endows the patient with the power to refuse or allow specific interventions in his or her life, or body and as such protects the patient's autonomy defined as the patient's freedom to form and act on his or her own reasons and purposes without the controlling influence of healthcare professionals or other external forces. However, the ability of informed consent to protect personal autonomy depends on the satisfaction of several conditions. In the literature on informed consent three conditions are generally recognised:2 ,8 ,9
-
That the patient receives adequate information about the medical procedure suggested and understands this information.
-
That the patient is competent to deliberate about the consequences of undergoing the suggested medical procedure and on this basis is able to decide whether or not to undergo the relevant medical procedure.
-
That the patient's consent is not coerced, manipulated or otherwise unduly influenced by healthcare personnel or others.
The clarification of each of these conditions has attracted extensive scholarly attention in recent years.10 For present purposes, it suffices that each of these conditions points to factors that may undermine, limit or threaten the ability of informed consent to protect the personal autonomy of the patient. If the patient is inadequately informed, incompetent or unduly influenced, then there is significant risk that the patient's choice to/not to consent does not originate in the patient's own plans and values, reasons and purposes.
There are, however, also some situations where it is generally accepted that the protection of personal autonomy through informed consent may be overridden or dispensed with. This holds for situations where:2 ,11 ,12
-
There are insurmountable practical difficulties in obtaining consent.
-
There are ethical considerations of greater weight that outweigh respect for personal autonomy.
This article studies situations exhibiting none of the above features that may undermine or override the protection of personal autonomy through the acquisition of informed consent. Thus, we shall start by assuming that all of the above conditions of the validity and relevance of informed consent are satisfied. We do, however, believe that even in such cases—even in the everyday practice of properly obtaining consent from normal patients—there is a threat to the protection of personal autonomy through informed consent: a threat that has received far too little attention in the literature. And that is the threat of ‘routinisation’. There is a discussion of ‘consent fatigue’ in the literature about bio-banking, that is, the phenomenon that people may stop giving consent if asked too many times.11 ,13 The phenomenon of routinisation is, however, the almost diametrically opposed phenomenon to consent fatigue since it involves an increased likelihood of consenting to repeated requests (see below).
Informed consent and routinisation
The basic claim of this article is that informed consent may lose its function as a protection of personal autonomy if it is routinised. Let us define routinisation the following way: the provision or refusal of informed consent has been routinised if, and only if, the provision or refusal happens as an unreflective, habitual act.14 If the provision or refusal of consent becomes an ‘act of routine’ in this sense then it will no longer protect personal autonomy. The protection of personal autonomy is essentially a matter of providing the patient with a ‘space’ to form and act on his or her plans and values, reasons and purposes. Reflection is crucial to the formation of plans and values, and reflection links plans and values, reasons and purposes to action through means–end reasoning. In asking for consent the patient is asked to reflect on and decide—on the basis of information about and understanding of the nature of a proposed medical intervention—if the intervention is in accordance with the patient's own plans and values, reasons and purposes. In this way, the asking for consent preserves a space for the reflection lying at the heart of exercising personal autonomy. However, as defined above, routinisation implies the suspension of reflection. The routinisation of informed consent means that the act of providing or denying consent has become routine: an unreflected, habitual act.
It is possible to distinguish two different types of routinisation:
-
Routinisation of materially identical decisions.
-
Routinisation of materially non-identical decisions.
The first type of routinisation occurs when the habit to consent or refuse only extends to one type of decision, for example, routine blood tests in a patient with a chronic illness. The second type occurs when the habit to consent or refuse is extended, without proper reflection to a range of materially non-identical decisions, for example, different blood tests as part of antenatal care.
In order to investigate whether routinisation actually occurs we have performed a survey in a Danish population of internet users. The specific research questions are:
-
Is the provision or refusal of informed consent in fact routinised?
-
To what extent is the routinisation of informed consent dependent on the nature of the content of the specific consent? (ie, is there evidence of routinisation across materially different contexts)?
-
What are the causes of the routinisation of informed consent?
The present study: routinisation in the use of ICT
Information and communication technology (ICT) provides an interesting ground for the study of routinisation for three reasons. First, because the users of ICT either implicitly or explicitly have to consent to the conditions of use of a given technology or service. Second, because the average user of ICT has to consent on numerous occasions. Hence it is relevant to study if the provision of consent may become a habit. Third, because the users of ICT are provided with or have access to extensive information on the conditions of use and the policies of protection of personal or private information. And therefore the users are in principle in a position to meaningfully reflect on the consequences of providing or refusing consent. In many cases, the agreement between the user and the provider of a specific ICT service is not labelled as ‘consent’ but rather as a ‘contract’. However, a routinisation of contractual agreement is in itself prima facie evidence of the routinisation of consent. A claim we shall return to below in the discussion.
The present study examines the extent to which users read the conditions of use and the policies of protection of personal and private information in various types of use of ICT. The cardinal assumption is that if the users regularly provide consent without reading such conditions and policies and/or the updates of these policies, then it implies, ceteris paribus the routinisation of their consent. Without reading the specific conditions of use and the specific policies of protection of personal information, a user cannot reasonably be said to have reflected on how the provision of consent in relation to the use of a specific ICT will affect his or her plans and values, reasons and purposes. And, the user cannot reasonably be said to have reflected on the consequences of providing consent in a way that ultimately amounts to protecting personal autonomy.
Our particular interest is in the routinisation of informed consent in the healthcare context, but the use of health related ICT services by the public is still relatively limited compared with the use of other services.15 We have therefore found it relevant to study the consent practices of users across a range of very commonly used ICT services including health related services. The different types or categories of use of ICT included in the present study are:
-
Installing of software and applications on computers
-
Webmailing (eg, Hotmail, Gmail, Yahoomail)
-
Netbanking
-
Online shopping (eg, Ebay, Amazon)
-
Exchanging information with public authorities online (eg, Skat.dk, the Danish tax authority website)
-
Social networking on the internet (eg, Facebook, Twitter, Youtube, Google+)
-
Using fitness and dieting websites (eg, Weightwatchers)
-
Using health information websites (eg, Netdoktor, the Danish equivalent to Netdoctor.co.uk).
The respondents are for each of the above categories asked if they read the conditions of use, updates of these conditions, the policies concerning protection of personal data and updates of these. In relation to fitness and health information websites, the respondents are also asked if they have a user-profile. This is important since establishing a user-profile requires the user to provide some health related information which is by legal definition sensitive information.
In order to clarify the causes of the possible routinisation of the handling of informed consent, those respondents who state that they are not reading the information are for each of the relevant categories asked to give their reason for not reading the conditions of use and policies concerning protection of personal data. The following reasons are provided as options:
-
I have had good experiences with the relevant provider
-
It was recommended by others
-
It is too often that one is required to read such conditions and policies
-
Such conditions and policies are all the same
-
There is little risk associated with sites such as these
-
The loss is insignificant if anything goes wrong
-
There is too much to read
-
The text is incomprehensible
-
It takes too long to read such conditions and policies
-
The conditions are not relevant for my use of the site
-
There is no good reason for not reading conditions and policies—I ought to do it.
Participants are asked to tick off one or more of these reasons, but are also given the possibility of stating alternative reasons. Participants are also asked to provide basic demographic information about age, gender and their level of ICT experience.
The questionnaire was piloted informally with a group of students before being finalised. It is delivered as a completely anonymous e-survey.
Respondents were recruited mainly through the social networking site Facebook. The research project was set up as an ‘event’, and individual users were then invited to participate in the event. A profile with key info on the research project was also set up—and with this profile as sender, individuals were invited to fill in the questionnaire. The survey is therefore not necessarily representative of any specific background population (see Discussion section below).
The statistical analysis was performed using the statistical package SPSS 16.16
Results
The survey received 1372 valid responses. Of these, 907 (66%) were from women and 465 from men (34%). The average age was 29.0 years (SD 11.7). The level of education can be seen in table 1. There were no significant differences in age and education between men and women.
Almost all respondent had used a computer for more than 5 years (96%), and the internet for more than 5 years (94%). In all, 25% used the internet more than 5 h daily, 69% 1–5 h and 6% less frequently. Men used the internet significantly more than women (p<0.000 Mann–Whitney test). The data concerning whether the respondents read the relevant information before using particular services can be seen in table 2.
The data concerning reasons for not reading the relevant information can be seen in table 3.
Discussion
The survey provides clear evidence of many users not reading the relevant conditions and policies and updates of these. The evidence in table 2 exhibits three patterns indicating the routinisation of consent.
First, there is a significant number of users who do not read the conditions and policies and updates of these across the various ICT services. More specifically, 60% of the users fail to read the conditions of use for at least two of the main categories of ICT services (Installing, Webmailing, Netbanking, Shopping, Public Authorities, Social Networking), and as many as 10% fail to read the conditions of use for all of the six main categories of ICT services. Overall, 59% of the users fail to read the policies for protection of personal information for at least two of the categories of ICT services, and as many as 14% fail to read the policies for protection of personal information for all of the six main categories of ICT services. This indicates the routinisation of materially non-identical decisions.
Second, since several of the categories cover ICT services which the users may be expected to have used several times or several different of (Installing of programs, Online shopping, Social Networking, Health Sites), it seems likely that a ‘no’ covers a failure to read conditions and policies on more than one occasion for a number of people. The fact that on average 37% read the conditions of use and 37.5% read the policies for the protection of personal information therefore must be assumed to cover a significant number of cases of use of ICT services where the relevant conditions and policies have not been read.
Third and finally, for all categories of ICT services there is a marked difference between the reading of conditions and policies the first time and the reading of updates. The minimal difference between the users who do not read conditions of use the first time and updates of these is 7.1% (Health Sites). The maximal difference is 36.5% (Online Shopping). The minimal difference between the users who do not read policies for protection of personal information the first time and updates of these is 4.7% (Health Sites). The maximal difference is 19.1% (Online Shopping). These figures clearly indicate the routinisation of materially identical decisions.
Table 3 showing the reasons for not reading the conditions and policies allows a further deepening of the notion of routinisation. So far we have assumed the failure to read conditions and policies to be indicative of routinisation. Although this may hold, table 3 allows further distinctions being made between levels of routinisation. Some of the reasons for not reading seem to imply the suspension of reflection to a greater extent than others. The reasons ‘Have to read too often’ and ‘All are the same’ thus both seem to indicate a general loss of interest in conditions and policies that leaves little room for reflection on specific conditions and policies. Between 72% (Installing) and 40% (Health Sites) of the respondents provide one or both of these reasons for not reading the conditions and policies.
Although ‘Good experience with provider’ and ‘Provider recommended’ are provided as reasons for not reading the conditions and policies—and hence imply that reflection on the specific conditions and policies have been suspended—they do not necessarily imply that reflection has been suspended altogether. The choice not to read may follow upon a reflection of the implications of using an ICT service based on past experience or the recommendations of others rather than a reading of the specific conditions and policies. The same argument could be applied to the reasons ‘Little risk with this type of site’ and ‘Loss insignificant’. They both imply suspension of reflection on the specific conditions and policies but are compatible with more general reflections on the implications of using a given ICT service. In order to distinguish this potentially less severe degree of routinisation it thus becomes of interest to determine the users who only provide reasons within this cluster of reasons. And, as an example, 24% of the users only provide one or more reasons within this cluster in relation to social network use.
It is worth noting that a significant number of users provide as reasons for not reading that there is ‘Too much to read’ and that it takes ‘Too much time to read’ (See table 3). It is not entirely clear how these reasons ties in with routinisation. They may reflect rationalisations of the already ‘routinised’ provision or refusal of consent, or they may reflect genuine problems of capacity or resources in the processing of information. Although routinisation may ensue in both cases, the former may cover more problematic cases of routinisation than the latter. Problems of capacity do not imply that the users do not value and understand the importance of providing or refusing consent or the lack of willingness to reflect on the consequences of this. Distinguishing between these cases could be done by taking into account the other reasons provided by these users (see above). Since in this article we are interested in the possible routinisation of informed consent more generally, and since the length of the conditions and policies of ICT services is specific to the field of ICT, we shall, however, leave such further enquiries aside.
Limits of the present study
The recruiting of participants among users of various websites and fora may increase the likelihood of over-representation of very experienced users and those particularly interested in the questions posed. A random sample of the background population would solve this problem but would also increase the likelihood of including people in the study with no or little experience. Since the study concerns the consent practices of users with some experience, we believe the current recruitment method is acceptable. The goal of the present study is to investigate whether routinisation occurs, not to provide any exact estimate of how common it is.
In the questionnaire the choice between reading and not reading is binary. The questionnaire does not incorporate the possibility that users may read the conditions and policies to different extent. One may suspect—not least because of this being the ‘socially acceptable behaviour’ (see below)—that being left with only these two options, the users only partially reading the conditions and policies will report that they do read these. The present study may consequently underestimate the level of routinisation to some degree.
On societal level, the reading of conditions and policies is arguably associated with the ‘responsible’ and ‘educated’ behaviour. If so, the users’ responses may represent idealisations of their behaviour in accordance with socially accepted or required behaviour of reading conditions and policies of the relevant kind.
In the mapping of the reasons provided by users we have focused entirely on their reasons for not reading the relevant conditions and policies. We do not know the reasons for reading the conditions. However, some possible reasons for reading the conditions and policies are not incompatible with the routinisation of consent. If, for example, the users read conditions because ‘they always do’ then this would clearly be good grounds for suspecting that reflection on the specific conditions and policies and their consequences for the individual user has been suspended. Again, this limitation on the present study implies that the level of routinisation may be underestimated.
In relation to the mapping of the reasons it may also be noted that the current study does little to clarify the relative weight of each of the reasons. Although each of the users may have several reasons for not reading the conditions and policies, it may be the case that for each user some of the reasons provided may play a greater role in the decision not to read the conditions and policies. This lack of distinction between the relative weight of the reasons may be of lesser importance in establishing the occurrence of routinisation, but leaves it more difficult to take action to prevent it.
Finally, it has to be noted that consent provided before the use of a given ICT service in most cases is explicit—the user is explicitly asked to ‘agree’ to certain condition and policies. In many cases, the updates of conditions and policies do not require an explicit act of consenting. Hence the lack of reading of updates may reflect unawareness of these updates rather than the routinisation of consent to updates. The routinisation of consent to updates may consequently be exaggerated in this study.
Summary and future research
In this article, we have introduced and defined the notion of routinisation of informed consent. We have argued that it may pose a hitherto overlooked threat to the ability of informed consent to protect personal autonomy. We have shown that there is evidence of routinisation of both materially identical and materially non-identical decisions in the consent behaviour of users of various kinds of ICT services and, more specifically, that there is evidence of routinisation in relation to fitness and dieting sites and health sites, and this applies to both people with and without a profile. Since routinisation may undermine personal autonomy and the validity of informed consent it is important that the factors leading to routinisation and their weight in the decision making of users are further explored in future research.
Footnotes
-
Contributors Both authors contributed equally to the design of the study, analysis and writing of the draft paper. Both authors are guarantors.
-
Funding None.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.