It's not just what you say, it's also how you say it: Opening the ‘black box’ of informed consent appointments in randomised controlled trials

Social Science & Medicine 68 (2009) 2018–2028

Contents lists available at ScienceDirect

Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed

It’s not just what you say, it’s also how you say it: Opening the ‘black box’ of informed consent appointments in randomised controlled trials Julia Wade a, *, Jenny L. Donovan a, J. Athene Lane a, David E. Neal b, Freddie C. Hamdy c a

Department of Social Medicine, University of Bristol, 39 Whatley Road, Clifton, Bristol BS8 2PS, United Kingdom University of Cambridge Department of Oncology, Box 279 (S4), Addenbrooke‘s Hospital, Hills Road, Cambridge CB2 0QQ, United Kingdom c Nuffield Department of Surgery, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, United Kingdom b

a r t i c l e i n f o

a b s t r a c t

Article history: Available online 11 April 2009

Randomised controlled trials (RCTs) represent the gold standard methodology for determining effectiveness of healthcare interventions. Poor recruitment to RCTs can threaten external validity and waste resources. An inherent tension exists between safeguarding informed decision-making by participants and maximising numbers enrolled. This study investigated what occurs during informed consent appointments in an ongoing multi-centre RCT in the UK. Objectives were to investigate: 1] how study staff presented study information to participants; 2] what evidence emerged as to how well-informed participants were when proceeding to randomisation or treatment selection; and 3] what aspects of the communication process may facilitate improvements in providing evidence of informed consent. Qualitative analysis of a purposive sample of 23 recruitment appointments from three study centres and involving several recruitment staff applied techniques of thematic, content and conversation analysis (CA). Thematic analysis and CA revealed variation in appointment content and structure. Appointments were mostly recruiter-led or participant-led, and this structure was associated with what evidence emerged as to how participants understood information provided and whether they were in equipoise. Participant-led appointments provided this evidence more consistently. Detailed CA identified communication techniques which, when employed by recruiters, provided evidence as to how participants understood the choices before them. Strategic use of open questions, pauses and ceding the floor in the interaction facilitated detailed and systematic exploration of each participant’s concerns and position regarding equipoise. We conclude that the current focus on content to be provided to achieve informed consent should be broadened to encompass consideration of how information is best conveyed to potential participants. A model of tailored information provision using the communication techniques identified and centred on eliciting and addressing participants’ concerns is proposed. Use of these techniques is necessary to make potential participants’ understanding of key issues and their position regarding equipoise explicit in order to facilitate truly informed consent. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: UK Informed consent Recruitment Randomised controlled trial (RCT) Conversation analysis Shared decision-making Prostate cancer

Introduction The randomised controlled trial (RCT) is accepted as the gold standard methodology for evaluating effectiveness of healthcare interventions. However, low accrual threatens the power of RCTs, external validity of findings, and may necessitate additional investment of research resources (Britton et al., 1998). Maximising recruitment is therefore crucial. Barriers to recruitment may be clinician-related (perceived lack of resources, time constraints, loss of professional autonomy and concern about impact on doctor– patient relationships) or patient-related (difficulties with informed

* Corresponding author. Tel.: þ44 0117 9287362. E-mail address: [email protected] (J. Wade). 0277-9536/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.socscimed.2009.02.023

consent, uncertainty, or preferences for particular treatments, Mills et al., 2006; Ross et al., 1999). RCTs can only be undertaken where there is ‘equipoise’: that is, there is no evidence to show that a person would be advantaged or disadvantaged by being allocated to a particular treatment (Chard & Lilford, 1998; Freedman, 1987). Ethically potential recruits must be given sufficient information to make an informed decision about participation (ICH, 1996; World Medical Association, 2004). In practice, they should be fully informed, in equipoise and accept randomisation to determine treatment (Bower, King, Nazareth, Lampe, & Sibbald, 2005; Mills et al., 2003). This raises questions as to what level of detail of information is required for a person to give informed consent, and when and how it should be provided (Boulton & Parker, 2007). Current practices leave room for improvement: in one RCT 51% of participants believed the doctor

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

had chosen their treatment and only 23% knew they had been randomised (Hietanen, Aro, Holli, & Absetz, 2000). Participants frequently fail to understand the rationale for RCTs (Featherstone & Donovan, 1998, 2002; Robinson et al., 2004). A systematic review, attempting to identify optimal methods of obtaining informed consent showed that providing more information led to greater understanding of the nature of RCTs and rights to withdraw or choose treatment but lower consent rates (Edwards, Lilford, are Thornton, & Hewison, 1998). While content and presentation of written participant information sheets (PIS) are highly standardised (Grossman, Piantadosi, & Cohavey, 1994; NRES, 2005), the content and quantity of spoken information provided in informed consent appointments are not monitored and their effects on participant understanding are unknown (Brown, Butow, Butt, Moore, & Tattersall, 2004). Providing an accurate PIS is not enough to ensure comprehension of key issues (Dixon-Woods et al., 2007). Potential participants need time to discuss and understand concepts of randomisation and equipoise (Featherstone & Donovan, 1998; Jenkins, Fallowfield, Shouhami, & Sawtell, 1999; Mills et al., 2003). A small number of studies have audio- or video-taped informed consent appointments. These found recruiters were poor at initiating discussions of the participant’s perspective (Brown, Butow, Ellis, Boyle, & Tattersall, 2004; Tomamichel et al., 1995), checking participant comprehension of information (Brown, Butow, Ellis, et al., 2004; Jenkins et al., 1999; Tomamichel et al., 1995) and explaining key concepts such as randomisation and equipoise (Albrecht, Blanchard, Ruckdeschel, Coovert, & Strongbow, 1999; Brown, Butow, Ellis, et al., 2004; Donovan et al., 2002; Jenkins et al., 1999). Brown, Butow, Butt, et al. (2004) proposed a typology to evaluate the content and quality of information given by oncologists seeking informed consent for clinical trials. They proposed a) strategies to promote collaborative decision-making, b) a specific sequence of topics for discussion, c) wording to convey key trial concepts and d) communication techniques to avoid subtle coercion and render conflicts of interest explicit. Applying this typology in evaluating informed consent consultations in clinical trials for cancer treatment, Brown, Butow, Ellis, et al. (2004) found considerable variation in practice. Many oncologists paid lip service to shared decision-making, by merely offering the option of delaying treatment decision. Key content was often omitted: the rationale for randomisation was covered in less than half of consultations. Moreover, as the authors acknowledge, while this methodology identifies whether the recruiter has raised an issue, it does not capture how participants interpret the information or whether the invitation to express views is taken up. A growing body of literature demonstrates an association between communication behaviour and patient outcomes in treatment decision-making (Arora, 2003; Ong, de Haes, Hoos, & Lammes, 1995). Shared decision-making (Charles, Gafni, & Whelan, 1999) is advocated in life-threatening diseases such as cancer (Gatellari, Butow, & Tattersall, 2001; Ong et al.,1995). Yet current approaches to informed consent conform more closely to the informed model of decision-making than the shared model (Charles et al., 1999). The emphasis is on what information must be provided in one direction, from health professional to patient, with the patient deliberating on treatment options and reaching a decision. Shared decision-making (two-way exchange of information between professional and patient, joint deliberation and joint decision-making, Charles et al., 1999) has been advocated for informed consent discussions (Brown, Butow, Butt, et al., 2004), but there is little evidence of it taking place in practice (Brown, Butow, Ellis, et al., 2004). This study aimed to open the ‘‘black box’’ of what goes on during informed consent appointments in a large ongoing multi-centre RCT, the ProtecT study (Prostate testing for cancer and Treatment,

2019

investigating effectiveness and cost effectiveness of three treatments for Prostate cancer, Donovan et al., 2003). The objectives were to investigate: 1] how study staff presented study information to participants; 2] evidence that emerged as to how well-informed participants were when proceeding to randomisation or treatment selection; 3] aspects of the communication process that may facilitate improvements in evidence of informed consent. It was intended that findings would be used to develop advice to improve informed consent appointments to RCTs more generally. Methods Main study design The ProtecT (Prostate testing for cancer and Treatment) study involved a programme of prostate specific antigen (PSA) testing amongst men in the community, inviting those with localised prostate cancer to be randomised to one of three treatments (Donovan et al., 2003). Multi-centre research ethics committee approval was obtained. Men aged 50–69 years in GP practices in nine UK centres were invited to an appointment and given detailed written and spoken information about the implications of having a PSA test, uncertainties about treatments and the need for a randomised trial of treatment. Those consenting to testing who had abnormal results were offered further diagnostic tests including biopsy. Those diagnosed with localised prostate cancer attended an appointment with a urologist to discuss the diagnosis and be given a detailed written PIS and introductory spoken information about the treatment trial comparing radical surgery (RS), radical conformal radiotherapy (RT) and active monitoring (AM, regular PSA tests and treatment as required). Participants attended a longer informed consent appointment (‘information appointment’) scheduled a week later, to allow an opportunity to absorb the information given. The information appointment was with a trained research nurse whose aim was to assist the man to reach an informed decision about whether to participate in the RCT and consent to randomisation, or choose treatment outside the RCT. Design of present study The study reported here is a qualitative study embedded within the main RCT. Qualitative research was integrated into both feasibility (Donovan et al., 2002) and main studies: information appointments were routinely tape-recorded to investigate differences in recruitment rates between centres and help train new staff (Donovan et al., 2008). This qualitative study used audio-recordings of information appointments to investigate interaction between recruiters and potential participants and was primarily conducted by JW (appointed to the ProtecT team in 2005 so not previously involved in ProtecT research). Participants Purposive sampling ensured the selection of a wide range of audio-recordings of appointments: from three study centres, over an extended time-period, including appointments conducted by several recruitment staff with different approaches to obtaining informed consent, and among participants with a range of sociodemographic characteristics. Variations in randomisation rates (the percentage of those eligible consenting to participate in the RCT) occurred between centres and over time (Donovan et al., 2008) and facilitated sampling of appointments for this study. Examples of each of three outcomes defined a priori to reflect potentially varying levels of informed consent were included:

2020

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

a) Participant consented to randomisation and accepted treatment allocation. Ideally this outcome reflects high levels of informed consent but may conceal poor understanding or passive acceptance of the trial. b) Participant consented to randomisation but rejected treatment allocation. This outcome, where the allocation was rejected immediately, was assumed to indicate a lack of equipoise reflecting difficulties with the informed consent process. Ideally participants should be in equipoise when randomised therefore unlikely to reject the allocation immediately. c) Participant refused randomisation and chose treatment. Ideally this outcome reflects high levels of informed decision-making but may conceal poor levels of understanding. Procedure Audio-recordings were fully transcribed with all personal data anonymised to ensure confidentiality. Thematic, content and conversation analyses were used in combination. Thematic analysis involved the inductive identification of codes from the data (Krippendorff, 2004; Silverman, 2001) using constant comparison techniques pioneered in grounded theory (Glaser & Strauss, 1967; Miles & Huberman, 1994). Analysis was undertaken in several ways: on each transcript separately, then with groups of transcripts analysed together by outcome, by centre and early or late in the time-period, until saturation was reached. As themes emerged, the researcher returned to the raw data looking for negative or deviant cases to test emergent themes and ensure these were well grounded in the data. Content analysis involved identification and definition of codes before searching for these in the data (Krippendorff, 2004; Neuendorf, 2002) and was applied specifically to identify evidence of participant views on treatments and evidence of equipoise at the point of deciding whether to be randomised. Conversation analysis (CA) rests on the principle that all utterances are social actions connected in sequences. As sequences of actions have patterns, what one participant says and does is shaped by what has previously been said and will in turn shape what will come (Maynard & Heritage, 2005; Silverman, 2000). Sections of the transcript for detailed scrutiny using CA were identified during thematic analysis and transcribed in more detail (see Appendix B). CA enabled exploration of overall structural organisation as well as the sequence, design and timing of conversational turns (Drew & Heritage, 1992) including overlapping talk and the organisation of turn-taking (Schegloff, 2000). Participant contributions were analysed in the context of the conversation with the recruiter with the aim of finding evidence about how information was received and interpreted, the success or failure of the information exchange, and to identify patterns most likely to facilitate and provide evidence of participant understanding (Heritage & Robinson, 2006). Data analysis proved to be a lengthy and highly complex process as the focus moved back and forth between thematic analysis, content analysis and CA as findings emerged, with the methods of each contributing to the findings separately and in combination. To check coding reliability, four other qualitative researchers independently evaluated twelve appointments. Codes were crosschecked with JW and areas of disagreement resolved with reference to the raw data. Researchers were blinded by the trial coordinator (JAL) as to which study centre was involved and where the appointment fell within the period studied. Sampling and analysis continued concurrently so that data saturation determined sample size. JW completed the analysis and produced detailed descriptive accounts, which were discussed in detail with JD to produce the findings.

Findings Appointments (n ¼ 23) were sampled from three study centres over a three-year period (November 2000–January 2004), involved 12 recruitment staff, and included a range of potential recruitment outcomes and participants across the age and social spectrum (Table 1). Most participants had the commonest clinical cancer stage, and were accompanied during appointments (Table 1). Thematic analysis of appointment content and CA of appointment structure revealed variations in the presentation and content of information provided by recruitment staff. Content analysis then identified evidence of participant views and equipoise when deciding whether to be randomised. Application of CA to analyse turn-design and turn-taking subsequently revealed patterns of communication that either facilitated or impeded evidence of participant views becoming available. Findings are detailed in turn below.

How study staff presented study information to participants Recruitment staff were given a checklist to ensure they covered essential study information: diagnosis, advantages and disadvantages of treatments, determining treatment outside the trial, the need for an RCT, the purpose of randomisation, the right to refuse participation or take time to consider, and inviting a decision as to how to proceed. As senior nurses, they delivered this information using their clinical experience, with considerable variation in their approach. CA of structural organisation revealed that the format of most appointments (n ¼ 20) fell into one of two categories: recruiter-led or participant-led. Three appointments combined elements of both (Table 2). Recruiter-led appointments Recruiter-led appointments were most common (n ¼ 12). In these, recruiters controlled and dominated the agenda (Appendix A, Box 1), providing information in checklist sequence. Participants were invited to comment or ask questions as each topic was concluded. These appointments followed proposed models of good practice by delivering requisite information in sequence and systematically attempting to elicit participant views (Brown, Butow, Butt, et al., 2004; ICH, 1996). Table 1 Characteristics of sample of appointments. Category

Number of appointments (N ¼ 23)

Centre Centre A Centre B Centre C

5 11 7

Outcome Randomised & accepted allocation Randomised & rejected allocation Chose treatment

12 4 7

Age of participants (years) 50–59 60–69

6 17

Social class Managerial & professional Other

13 10

Clinical stage T1 T2

16 7

Appointment attendance Alone Accompanied

6 17

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

2021

Table 2 Structural format, content analysis and outcome regarding randomisation.

Participant

Structure RL ¼ recruiter-led PL ¼ participant-led

Views on treatments expressed

In equipoise?

Randomised þyes/no ¼ took time to consider before decision

Decision þaccepted/þrejected ¼ took time to consider before decision

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

RL RL RL Mixed RL RL Mixed RL RL RL RL RL RL RL PL PL PL PL Mixed PL PL PL PL

No Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

? U* ? U   U ? U U U* U  ?      U* U* U* U

Yes Yes Yes þYes þNo No Yes Yes Yes Yes Yes Yes No Yes No No No No Yes Yes þYes Yes Yes

Rejected Accepted Accepted Accepted – – Accepted Accepted Accepted þAccepted Accepted Accepted – þAccepted – – – – Rejected þRejected Accepted þ Rejected Accepted

U denotes evidence that participant was in equipoise;  denotes evidence that participant was not in equipoise; ? denotes insufficient evidence to make a judgement; U* denotes participant who initially expressed a preference but during the appointment has revised their thinking. These participants expressed uncertainty as to whether they were in equipoise, but indicated they would consider all three treatments if allocated.

Participant-led appointments Eight appointments followed a ‘‘question and answer’’ format: the participant effectively controlled the agenda, putting questions to the recruiter in the order they chose, persisting until satisfied with the information provided (Appendix A, Box 2). Checklist content and sequence were not necessarily followed and some appointments deviated markedly from proposed models of good practice (Brown, Butow, Butt, et al., 2004) both in content and order of presentation. However, thematic analysis revealed more evidence of participants asking questions and expressing concerns about treatments or study processes (Appendix A, Box 2, lines 73– 75, 78–80, 89–91) than in recruiter-led appointments. Evidence of participant views and equipoise Findings of the CA of structural organisation, content analysis and eventual outcome as regards decisions to be randomised are summarised in Table 2. Content analysis revealed that in some appointments (P1, P3, P8), views of participants about treatments were largely absent prior to the decision about randomisation, despite invitations to voice views. CA of these appointments showed that unmarked responses (e.g. ‘‘mm’’, ‘‘yeah’’, ‘‘right’’ Appendix A, Box 1) predominated. Unmarked responses may function to encourage a speaker to continue their turn or may indicate implicit rejection of the relevance of information given (Heritage & Sefi, 1992). These appointments provided insufficient evidence to determine whether or not participants fully understood the implications of treatment options. In these appointments plus P14, there was insufficient evidence to judge whether the participant was in equipoise. All these appointments were recruiter-led. By contrast, all participant-led appointments provided evidence about participant views of treatments and their position regarding equipoise. Appointment structure was therefore associated with the extent to which participant views about treatments and their position regarding equipoise were voiced. Recruiter-led appointments provided supposedly essential information systematically, but did not necessarily

enable participant views to emerge. Participant-led appointments, by contrast, provided information in a tailored way: some ‘‘essential’’ topics were omitted, but participant views were always expressed. Participant-led appointments therefore consistently enabled the recruiter (and researcher) to reach judgements about participant levels of understanding and position regarding equipoise. Communication strategies to elicit patient views Participant views emerged in some recruiter-led appointments, indicating that structural factors alone did not account for all differences (Table 2). CA of turn-design and turn-taking identified communication strategies that were associated with successful elicitation of participant views about treatment and equipoise regardless of structure and these strategies are illustrated below. Question design, pauses and overlapping speech The use of open questions elicited views and beliefs rapidly, allowing recruiters (R) to identify which issues participants (P) wished to address (lines 1 & 8):

Fragment 1 1

R So .hhh how did you get on with your information sheet?(0.5)

2

P Alright I ju:st feel tha:t you know on balance um (0.8) the surgery is

3

maybe (0.5) the quickest a:nd well I know that there’s there’s

4

drawbacks to it [but I think it’s maybe (1.0)

5

R

6

P the surest thing to (.) to do something about it, a wee bit concerned

7

[Mm hm.

about the radiotherapy, that was one [I thought]

8

R

9

P Oh it just seems a long time to me to be going on and (.)

10

[Tell me(.)] why

[talking about five days a week for some time (P22 randomised to RT, takes time to consider then rejects allocation)

2022

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

Open questions were highly effective, but not always sufficient to elicit concerns. At times recruiters needed to combine open questions with deliberate pauses. In recruiter-led appointments, recruiters presented large amounts of detailed information about a treatment, before inviting questions. Some participants needed time to formulate questions, particularly when recruiters spoke at length. The following extract follows the recruiter’s exposition of surgery lasting 3 min 49 s. During this time the participant contributed 55 turns to the conversation, the vast majority of them (50) consisting of minimal turns (e.g. ‘yeah’, ‘mm’, ‘right’), with only five major turns (e.g. ‘About the size of a walnut or something is it?’).

Another successful technique for eliciting views was for the recruiter to cede the floor rapidly where overlapping speech occurred. In Fragment 4, the recruiter was explaining randomisation, in response to participant questions. The participant attempted to take the floor (line 4): the recruiter ceded at once (line 3 ‘yeah’). This, combined with the use of continuers (lines 6, 8, 10, 13, 16, 18), enabled him to raise concerns about impotence.

Fragment 4 Fragment 2 1.

1.

R so u::m d’you wanna ask me anything about the

2. 3.

P

4.

R

5.

surgery?

3.

No not really [uh:::m

4.

[It’s usually (.) five to seven days in hospital Dxx if everything goes according to plan .hhh (P3, randomised to RP and accepts).

The invitation for questions (lines 1–2) was phrased as a closed question, inviting a yes or no response (contrast with ‘What are your thoughts on surgery?’). The participant’s response ‘no,’ (line 3) was qualified (‘not really’) and followed by ‘uh:::m’, functioning to hold his turn while he reflected. The recruiter, however, failed to react to these signals, beginning her next turn in overlap with him (line 4). She acknowledged the overlap by leaving a micro pause (line 4) for him to take the floor (Schegloff, 2000) but when he failed to do so, she continued her description of surgery, again dominating with minimal contributions from the participant. By contrast the recruiter in Fragment 3 left an exceptionally long pause (line 2, over 4 s) after asking why this participant rejected RT. This recruiter combined pauses with the use of continuers (e.g. ‘mm’, ‘right’ lines 4, 11, 14, 16), successfully encouraging this participant to explain his thinking.

R compare A treatment with treatment B (0.8) then

2.

what we do is what we call randomised um st(0.8) studies .hhh[which P

5.

(.) yeah

[and we haven’t(.) yes (.) I mean er (1.4) I mean I I’m fifty nine

6.

R Yeah=

7.

P =er (1.4) er it’s caught at an early stage I [still have

8.

R

9.

P an active active- a- active sex life

10.

R Yeah

11.

P

12.

[Yeah

.hhh so (.) uh radical surgery (1.4) might not be that appropriate (0.5) I’m not trying to talk myself out of [That

13.

R

14.

P

15.

[Right= =if that if that’s the best thing that’s the best thing that’s the >[way it goes<

16.

R

17.

P

on the other hand if if I was eighty years old (.)

18.

R

Mmm hmm

19.

P

and this cancer was really advanced

20.

R

Yeah

21.

P

and I didn’t have a sex life (0.8) th- the options would (1.2) or or the

22.

[Yeah

choices might be more clear cut (P4, randomised to AM after seeking further opinion from GP & accepted allocation)

Fragment 3 1.

R

Yes I was going to say, was it just the thought of side effects that put you off? (4.1)

2. 3.

P

4.

R

=Mm=

5.

P

=I I felt that it was a big leap (.) in (1.3) such an early stage (0.8)

6.

R

right although you’ve got the advantage of knowing that the tumour

that is there is sort of basically killed by the radio[therapy.

7. 8.

Partly but=

P

9.

[Yeah what I meant was u:h wwith it being so small (.) the the (1.1) mo:re (.) important of the two (0.8) uh

10.

such as operation [or (.) radiotherapy

11.

R

12.

P

seemed to be (0.6) taking a sledgehammer to crack a nut at this .hh

14.

R

mm hm

15.

P

and there’s no movement (.)

16.

R

right

17.

P

if it does if the PSA does start to show then (1.1) yes I’d think about it

particular moment because (1.0) it’s small (.)

13.

18.

[Mm.

then (P15, not randomised, chose monitoring)

Where recruiters failed to cede the floor, expression of concerns could be prevented or delayed (see Fragment 2 above). Fragment 5 shows multiple attempts by a participant to articulate concerns before eventual success. This exchange began with the participant‘s request to be randomised. He also voiced reservations about surgery (lines 5, 6 and 8) and denied fear was an issue and was attempting an explanation (line 8) when the recruiter increased the volume of her voice, signalling her desire to take the floor (Schegloff, 2000, line 9). She and the participant’s wife reassured him fear was normal (despite his denial of fear) and that he would receive optimum care (lines 9–29). He successfully signalled his wish to take the floor and explained his anxieties related to heavy lifting duties at work (lines 30–42). Three more times he attempted to articulate fear of incontinence (Fragment 6, lines 94, 134–136 and 171–172), succeeding on the final attempt. Twice his meaning was misinterpreted and inappropriate reassurance given (e.g. about the cancer being localised, lines 107–109) before he finally articulated his concerns (lines 171–172). Had this recruiter and his wife ceded the floor more readily, allowing him space to articulate his concerns, much time would have been saved and the risk of never addressing this man’s concerns would have been avoided.

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

Fragment 5

Fragment 6

1. P

93.

I’ll go for it right (0.8) [and then I’ll

2. Wife 3.

[with

4. R

[Yeah there’s no

5. P 6. 7. R

[No it’s not that because (.) I’m a bit wary of (.) the knife you know you know= =Right=

8. P

it’s not that I’m frightened it’s (.) just because=

9. R

=THAT’S [very normal that’s very normal

10. P 11. R 12. Wife

but I’m worried ab (.) about you know the possibilities of (.)

95.

Wife It coming back that’s what he’s [thinking of

96.

R

97.

P

back (0.8) it’s it’:s I’m hoping everything’s going to be

99.

[nor gonna be working

101. P

(.)

103. P

that

one of the bigges t centres [in the country the bes t place to be [S- s- see me (0.8) me (0.9) me job when you’re saying you know like I do a lot of h- heavy work [you know [Mmm

33. P

and I do and I use me body an awful lot like for =

34. R

=yeah

35. P

heavy (.) [lifting you know

36. R

[Yeah

(continued discussion, decision to be randomised; allocated RP) 132. N

so how do you feel about it though (1.2)

133. P Well (2.6) it’s it’s (.) of the three (.) it’s the one I didn’t want (.) but (.) as 134.

you say or the way you put it like it it’s mu- the (.) I’m (1.0) I’m not (0.9)

135.

really frightened of the operation (0.7) I’m just hoping that everything

136.

(0.6) [ you know will be alright

137. R 138.

[Oh yeah (0.6) ‘cos at the end of the day if I’d come in here said you know this morning (.) well you need to have the operation (0.5)

139. Wife you’d have accepted it (continued discussion about surgery)

37. P

[and er

38. R

[oh right

39. P

I can see me having problems in that field you know [like

40. R

[Mmm

104. R the thing is that nobody would give you those guarantees

[well I would be frightened

32.

41. P

[Yeah

102. R obviously but [I mean

30. P 31.

[hey? No not so much coming

98.

(reassurance about hospital being centre of excellence) 29. R

[Yeah [but there’s

100. R [NORMAL when we get in [there

[yeah yeah I mean [the thi:: ng is

P I know when you do it (.) the cancer’s going to go (.) yeah (.) that I do

94.

[Just to see what it comes up

2023

165. Wife You’ve got to think positive 166. R [yeah

you know like I will do a lot of damage down below because I I just

42.

don’t think I just attack yeah=

43. R

=No so maybe you’ll need a bit more guidance on how long to stay off

You will [be looked

167. W

[You’ve got to think [positive

168. P 169.

[ I know I’m just worried case [once the

170. Wife [I know (P11 randomised to RP and accepts)

171. P

operation’s over and everything something might you know like

172.

bladder might not work as well or (this might not work as well)

173. Wife But they can do things [to put that right 174. R

Systematic eliciting and addressing concerns Eliciting participants’ beliefs and treatment preferences was not necessarily easy, even when recruiters used the above techniques. Some men found difficulty expressing concerns about sensitive issues, e.g. incontinence or impotence. In Fragment 7 the recruiter answered the participant‘s concern as he initially articulated it (lines 1–2) with general comments about the advantages of surgery (lines 5–9). The man’s emphatic repetition that he saw no advantage to surgery (line 15) indicated this had not addressed the issue. He then revealed his concern was incontinence (lines 18–19). Once explicit, the recruiter could address this. The process of communication in appointments where concerns were explored in depth was therefore cyclical: repeat questioning and probing enabled participants to articulate concerns. Recruiters then provided specifically tailored information and elicited further concerns until, ideally, all had been addressed. Some men expressed treatment preferences and a reluctance to be randomised as appointments began, and became more convinced of these as discussion continued. However, where recruiters elicited preferences or concerns systematically and addressed them specifically, several came to appreciate the uncertainty regarding which treatment was best and to accept randomisation. Others developed new preferences. Fragment 8 illustrates how a participant articulated this revision process aloud. This participant disliked the idea of regular blood tests (lines 6–7) and assumed these were only required under AM (lines 24, 27–31) but was informed

[ That’s right yeah yeah

175. Wife with things like that if you [ did 176. R

[ Yeah

177. Wife have a lot of bother with your bladder they could help you (P11 randomised to RP and accepts)

they were required for all treatments (lines 8, 10–12). His comment ‘Mmm that’s very interesting’ (line 36), a marked acknowledgement in CA, indicated acquisition of new information. His next comment revealed he now viewed AM as an option but reservations about radiotherapy persisted (lines 38–39), so discussion shifted to eliciting and addressing these concerns, providing further evidence of him revising his views (Fragment 9, lines 270–271). This man chose to be randomised and was assigned RT. After time to reflect, he rejected the allocation and chose AM which he had initially rejected. The exchanges in Fragments 8 and 9 suggest that information provided during the appointment influenced his final choice. Where recruiters failed to elicit and address concerns as systematically as this, there was little evidence to judge a man’s level of knowledge or understanding. When one participant expressed a strong aversion to surgery (Fragment 10, lines 3 and 11), the recruiter made no attempt to explore his reasons but continued her description of treatments (lines 12–13). This recruiter described the treatments, the rationale for the trial and purpose of randomisation in detail and only then, at the point of

2024

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

Fragment 7

Fragment 8

1

the one thing that I really can’t see an upside of (.) is the

1

um >is the surgery<

2

R

Mmm

3

4

P

.hhh u::m (.)

4

5

R

>I mean I suppose< that one of the plus sides can be that (.) some

P

2 3

Wife and would you go with the other one with (0.5) monitoring if it come up P

(0.8) I might (.) but I’m (0.7) I keep thinking that’s the easy way out and

5

R2

[well

P

[you have to keep coming back and getting blood tests don’t you (0.5)

[Well you would anyway [irrespective,

um (.)

6

people will believe that because the (.) prostate gland has been

6

7

removed (.)

7

[(I’m with you)

8

R1 P

8

P

9

R

[is that the idea?

[as far as they’re concerned the cancer is in the pot

9

10 P

Yeah (0.7)

10 R1

11 R

but there is a (0.6) a risk that some of it has been left [behind and er

12 P 13 14 R

[(3 syllables) heh

[yeah. we still follow everybody up (0.8) irrespective of what they have (0.8)

11

u:m so even if it was active monitoring (.) radiotherapy (.) surgery (0.7)

12

um (0.6) men still have to come to clinic (0.6)

heh [no

(explanation about frequency of blood test)

[Mmm

18 P

15 P

No I I I honestly can’t see an upside (.)

16 R

[Yeah

17 P

[for me in in that (0.7) either of the other two u:m (1.3) I’m I’m I’m wide

18

open about .hhh (.) the thing that particularly does concern me (0.8) u:m

19

is becoming a dribbling incontinent (P5 not randomised, chose brachytherapy outside study)

=and that’s just for a blood test [is it?

19 R1

[um that’s just for the blood test and (.)

20

and occasionally the doctor might do a (.) you know rectal examination

21

but not every time (.) just maybe once a year (0.8) u:m (0.9) so:=

22 P

=That’s the PSA (.) THAT’s with the u:h (1.4)

23 R1

>We do the PSA blood test<

24 P

Sorry that’s for the (1.2) moni[toring?

25 R1 26

making a decision regarding randomisation, elicited the participant’s preference for AM. Rather than explore the reasons, the recruiter simply accepted it (Fragment 11). Structural analysis showed this recruiter presented essential trial information and invited questions from the participant. Detailed analysis revealed a failure to elicit participant views other than in these two instances. This participant’s selection of AM may have been an informed decision, but there was little evidence available to substantiate this. Ascertaining the level of equipoise Randomisation only becomes a rational and ethical option for an individual when they are well informed and in a position of ‘equipoise’ (Chard & Lilford, 1998). As men absorbed information and revised their views during appointments, it was often difficult to judge whether they were sufficiently close to equipoise to consider randomisation, or whether they retained clear treatment preferences (e.g. Fragments 8 and 9). This could be made explicit by recruiters asking whether a participant would be prepared to accept each of the three treatments, or one that he had expressed concerns about (e.g. Fragment 9 line 344–346). This strategy helped expose remaining reservations for discussion (e.g. Fragment 8 lines 38–39). Recruiters also used this technique very effectively to emphasise that randomisation was not recommended if people were not in equipoise (Fragment 12). This strategy was notably lacking from appointments where participants were randomised then immediately rejected the allocation (P1, P19). Its use might have exposed their lack of equipoise and led to further discussion and clarification before proceeding either to randomisation or treatment choice.

Discussion This study investigated the information exchanged during RCT informed consent appointments to examine how study staff presented information to participants, explore evidence that emerged

[>Monitoring. But even even with< with the surgery or the radiotherapy the (.) [we do the same.

27 P

[They’ll do blood tests?=

28 R1

=We’d still do the blood tests and we’d still do a rectal examinations (.)

29 P

all that follow-up even after you’ve had the

30 R1

Yes (0.5) [yes.

31 P 32 R1 33 P 34 R1

[Prostate removed? Yeah [wouldn’t have thought there’d be anything there] [Doctors can can coz coz the doctor feels to see whether,

35 36 P

[(it is still healthy). [(if something’s there) (4.1) Mm that’s very interesting (1.8)

37 R

But that would probably be once [a year

38 P 39

[I’m STILL TORN between the two ((laughs)). Definitely have to (rule out the radiotherapy).

(P22)

as to how well-informed participants were when proceeding to randomisation or selecting treatment, and identify aspects of the communication process that facilitated this evidence becoming available. Previous research has suggested that highly structured consent processes are more effective in ensuring essential information is conveyed (Brown, Butow, Ellis, et al., 2004; Dunn & Jeste, 2001). Other research, however, has indicated that people make sense of information in complex and unexpected ways (DixonWoods et al., 2007; Featherstone & Donovan, 2002). This study found that recruiter-led appointments were more likely than participant-led appointments to conform to research guidelines by providing required information systematically (ICH, 1996), but evidence as to how participants understood this information was lacking from several recruiter-led appointments. Even when recruiters employed strategies conventionally identified as promoting shared decision-making (e.g. inviting questions, Brown, Butow, Butt, et al., 2004), participant contributions were limited. In these cases, there was insufficient evidence for the recruiter or researcher to judge the participant’s level of understanding or position regarding equipoise. By contrast, participant-led

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

2025

Fragment 9

Fragment 11

116. P

Well what does it involve then (.) what is (.) I

308 R

So Bxxxx what what would you like me to do now

117.

mean (0.6) I mean you’ve got some treatment before you

309 P

I would like to (0.5) carry on monitoring

118.

get the radiotherapy=

310 R

(1.0) ok

119. R1 =What happens with the radiotherapy is you um (.) .hhh

311 P

(0.8) based on (.) all the facts that you’ve given me (5 syllables)

120.

[the gentleman

312 R

mm (.) okey doke .hhh you do understand that there is (.) a possible risk

[see this is (.) you get something [you take a pill to .

313

of it [being ((2 syllables))

121. P 122. R1 123.

314 P

[You do you get (.) you get hormone treatments

316

(recruiter responds to 5 further queries about radiotherapy ) 270. P

(1.8) mm (.) that’s very good, that is (2.4) it’s started me

271.

thinking .hhh o::h dear (further discussion of radiotherapy and surgery)

336 R1

court.

338 P

Thank you [((laughs)).

339 R1

317

certainly in the next (1.2) year=

318 P

=coming up for monitoring and that doesn’t bother me

319 R

Mm hm. Good. (P13 refused randomisation and chose monitoring)

[((laughs)) (1.0) I say we’re only here to give you the information [you know.

341 P

[I know (3.2) You (.) go through the

342

randomisation if you want but as I say I mean it may be helpful with

343

your statistics but I may not (2.0) follow what’s (.) [come up

344 R 345

[Yeah yeah (.) but as I say hopefully (.) in the majority of men (0.8) you won’t need other treatment is is you know

so what shall we do then ((laughs)) (2.1) the ball’s in your

337

340

[yeah I understand there are [risks

315 R

[Would you consider if it did come back just as if it came back as radiotherapy (.) would you

346

consider it?=

347 P

=I’d think about it, but I (.) I’d think about it (P22 randomised to RT and rejects after taking time to consider)

appointments might not cover all ‘essential’ information or follow a routine structure. However, they consistently provided evidence that participants were able to articulate concerns, giving the recruiter and researcher insight into participant levels of understanding and enabling a judgement as to whether they were in equipoise. Further analysis using CA revealed variations in recruiters’ communication strategies that enabled or impeded participants in expressing their views. Recruiters who systematically used open questions and pauses and readily ceded the floor were more successful in enabling participants to articulate concerns than those who failed to utilise these techniques, regardless of whether appointments were recruiter-led or participant-led. The degree to

Fragment 10 1

R

2

Wife (0.5) I dunno if he wants to have the surgery

So do you want to ask me anything about the surgery?

3

P

4

R

(.) the answer to that is no ((laughs))=

5

P

=straight away

which recruiters were systematic in applying these strategies determined both the extent to which concerns were voiced and the presence of evidence in the interaction that concerns had been addressed. In particular, the degree to which recruiters successfully explored and exposed participant beliefs about treatments was crucial in providing evidence about whether participants were in equipoise. These findings suggest the focus on what content must be provided in informed consent appointments should broaden: how recruiters approach this task is equally important. The most effective appointments for providing evidence that participants were well-informed were those where recruiters applied the strategies identified above to explore participants’ concerns systematically and establish clearly whether participants were in equipoise. Least effective appointments were those in which recruiters reproduced information provided in the PIS and invited questions. Concepts of equipoise and randomisation, implications of treatments, right to choose treatment or withdraw remain the bedrock of informed consent discussions and should be included in the PIS (Brown, Butow, Butt, et al., 2004; ICH, 1996). However, these appointments present a unique opportunity for the twoway exchange of information. If exploration of an issue demonstrates it is well understood or irrelevant to an individual’s decision-making, time may be better spent on more relevant issues. Recruiters should use the informed consent appointment to explore individual views about trial or treatment processes systematically, and establish participant positions regarding equipoise. This can be done within a recruiter-led or participantled framework; the very nature of the strategies employed may favour participant-led approaches. There will always be a tension in obtaining informed consent between the objective of maximising recruitment and ethical concerns to ensure that participation or non-participation is fully informed (Boulton & Parker, 2007). These findings highlight the pivotal relationship between whether a recruiter has clarified the presence or absence of equipoise and the decision made regarding randomisation. Randomisation is only ethical where evidence of equipoise emerges (World Medical Association, 2004). If a recruiter

6

R

But keep an open mind

7

P

Aye well I have

8

R

[Keep an open mind

Fragment 12

9

P

[uhuh uhuh

1. 2. 3. 4. 5.

10 R Yeah 11 P 12 R 13

yeah well at this point in time (0.5) [no (.) no way [right (.) aggressive treatment is radiotherapy

ok (.) ok (0.8) the other (P13)

R

The important thing for you to think is would you feel (.) you wouldn’t feel happy but would you be (0.6) comfortable to um think about all three of them (0.9) If that’s the case, then randomisation’s probably for you but if you think there’s no way I could consider one of those then it isn’t (P10, randomised to RP and accepts)

2026

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

fails to elicit views or concerns effectively, there will be insufficient evidence of equipoise. Application of strategies identified here may help protect participants against coercion: the systematic use of open questions and pauses to encourage participants to voice views, concerns and preferences makes leading questions less likely. Audio-recording of appointments enables practice to be monitored and training of recruiters targeted (De Salis, Tomlin, Toerien, & Donovan, 2008; Donovan et al., 2008). When the strategies identified are applied systematically, numbers opting to be randomised and accepting the allocation can reach high levels (Donovan et al., 2008). Some participants in this RCT used discussions with the recruiter to confirm treatment preferences. For others, systematic eliciting and addressing of concerns regarding treatments or trial processes resulted in a perspective shift from treatment preference to consent to randomisation and acceptance of allocation. Variation in participant preferences for information is well documented (Degner & Sloan, 1992; Gatellari et al., 2001). The proposed approach would enable discussion to be tailored to the needs of the individual in a truly patient-centred way (GMC, 2008) incorporating principles of shared decision-making (Charles et al., 1999). Findings have implications for the training of professionals who recruit to RCTs. Strategies that proved most effective in eliciting participant views were relatively simple (open questions, long pauses, readily ceding the floor) and some have been identified elsewhere as highly effective in eliciting information from patients in doctor–patient consultations (Heritage & Robinson, 2006; Schwabe, Howell, & Reuber, 2007). Some, such as long pauses, do not occur naturally in communication, but training should ensure that recruiters are comfortable applying such techniques. A major strength of this study was the innovative application of CA in addition to thematic and content analyses, allowing assessment of the contributions of all parties, interacting in real time (Maynard & Heritage, 2005). This enabled a more patient-centred analysis of the information exchange (Ong et al., 1995), with participants equal parties in communication. Previous investigations have focussed only on conduct of recruiters (Albrecht et al., 1999; Brown, Butow, Ellis, et al., 2004; Jenkins et al., 1999; Tomamichel et al., 1995). By examining more and less successful exchanges, it was possible to identify patterns that were facilitating or inhibiting information exchange. Evaluation of future interventions applying these techniques to improve informed consent appointments would benefit from application of CA as a method of analysis. The study has some limitations. Use of audio- rather than videorecordings resulted in the exclusion of some non-verbal information. Data were limited to audio-recordings of appointments, without triangulation with observations or interviews to verify findings with the parties. Additional data might have shed light on motives for particular patterns of communication, and levels of insight about the impact of communication practices, but use of appointments meant that the focus of the research was on the presentation of information and reaction of RCT participants in a ‘real-life’ situation, without later interpretation. Data available for analysis reflected what would be available during appointments to recruiters seeking to satisfy themselves that participants were well informed and in equipoise before proceeding to randomisation. The study was limited to investigating recruitment in one RCT, but the plausibility of the findings suggests applicability more generally, particularly in view of the very different options offered in the ProtecT study, ranging from no active intervention to invasive surgery. Although this study focussed on how study staff presented information to participants, evidence emerged that participant strategies could override recruiter strategies. Further research is

needed to investigate the impact of participant communication on information exchange.

Conclusion Ensuring high levels of informed consent in RCTs requires considerable commitment by recruiters. The current focus on what information should be provided to participants must be broadened to include consideration of how information is best conveyed and how it is interpreted. A model of tailored information provision using the positive communication strategies identified to elicit and address participants’ concerns and clearly establish their position regarding equipoise is proposed. When these techniques were systematically employed, more evidence emerged of participants’ understanding of key issues, with the result that evidence was available to judge whether consent for randomisation was truly informed.

Acknowledgements The ProtecT study is funded by the UK NIHR Health Technology Assessment Programme (projects 96/20/06, 96/20/99). The authors are particularly grateful to three anonymous reviewers, whose detailed comments clarified our thinking considerably. The authors would also like to acknowledge the tremendous contribution of all members of the ProtecT study research group, and especially the following: Elizabeth Salter, David Gillatt, Phillip Powell, Prasad Bollina, Sue Bonnington, Debbie Cooper, Andrew Doble, Alan Doherty, Emma Elliott, Pippa Herbert, Peter Holding, Joanne Howson, Mandy Jones, Roger Kockelbergh, Howard Kynaston, Teresa Lennon, Norma Lyons, Hilary Moody, Stephen Prescott, Pauline Thompson. Merran Toerien, Zelda Tomlin, Isabel de Salis (supported by the Quartet research programme funded by the MRC Health Services Research Collaboration) and Nicky Mills (supported by an MRC Fellowship) contributed to the qualitative research. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Department of Health.

Appendix A. Examples of appointment structures 1. Recruiter-led appointment 1. R Okay erm so like I say as far as we’re aware it’s an early localised 2. prostate cancer so the chances of this progressing are quite small but [4 lines explanation of localised prostate cancer and introduction to treatments] 7. P Right 8. R that’s the three treatments that I just quickly wanted to to run through [6 lines] 9. P Yeah 10. R Um the worry is that you know obviously um there’s other aspects 11. of your health as you’re getting older [3 lines] 15. P No 16. R So um a- a- as I say it can be a lot of years before some men even 17. experience any problem from it um but then um it would be a case of 18. sometimes we may have missed the boat as far as an actual treatment is 19. concerned. 20. P Right 21. R but there’s still facilities for treating any symptoms that develop . [5 lines in which surgery is introduced] 27. P Yeah 28. R Um and it is a very vascular organ so the majority of men do need a 29. blood transfusion and it involves a seven to ten day stay in hospital [12 lines on surgery]

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028 (continued )

2027

(continued )

1. Recruiter-led appointment

2. Participant-led appointment

42. P Mmm 43. R But there are treatments that are available in lots of cases um for the 44. impotence 45. P Right 46. R Uhm (pause) the radiotherapy [5 lines on radiotherapy] 52. P Yeah 53. R Going every day Monday to Friday for the treatment [10 lines on radiotherapy, including side effect of tiredness] 63. P Right 64. R that’s something else which tends to be cumulative as as things go on, not 65. everyone, not everyone, but there is, you know, there is some risk um and 66. um obviously the prostate gland is still there so there is a risk that the prostate 67. cancer can return 68. P Mmm 69. R Um so basically they’ve all got their pros and cons 70. P Yeah 71. R um we have no evidence to show that any one is any better than 72. another. [4 lines] 77. P Right 78. R We’ll be looking at any aspects of their health which have been affected 79. from any treatment they’ve received and hopefully in years to come we’ll be 80. able to say that, you know, for men with early localised 81. P This is the best one 82. R This is the best way to deal with it 83. P Right okay 84. R Um so that’s that. Do you have any questions for me, I mean I’ve talked 85. long enough now I think. 86. P OK, just picking up on the um on the radiotherapy one question about 87. radiotherapy. (P5 not randomised, chose brachytherapy outside the study)

[16 lines explanation about MRI] 55. P How do you know whether it is a high-grade disease or not? Or you 56. don’t know at this stage? 57. R Well yes we get that from the pathology results from the biopsy [3 lines explanation] 58. P Sorry to be jumping around all over the place 59. R No no, these are questions you need to ask. [continues explanation of disease grade and stage for 11 lines] 71. P Made notes all over the place 72. R That’s fine. (2 secs) It’s a lot of information you will find to take on board. 73. P Um I must say at this stage I don’t fancy being just monitoring. It um, I 74. would rather something was done with it. And I know you want me to take 75. whatever the computer throws up but I don’t have to do that do I? 76. R No, you certainly don’t, no. Can I just ask you what what um what 77. puts you off the active monitoring? 78. P Well if I have got cancer I would rather get rid of it. Um um 79. because if you leave it it ultimately becomes the tiger, yeah. And this for me is 80. not a rehearsal you have only one go you know. 81. R There’s a balance. It is a balance really of weighing up the um yes 82. the radical treatments do have the potential you know of of cure um but do 83. you actually need such aggressive treatment? And likewise with the active 84. monitoring um um it is a case of should you really have the radical treatment 85. or should you be monitored. Um and as far as we can tell all three 86. treatments are equal because we don’t know the answers to these questions. 87. Um and it’s that’s basically you have got to look at the advantages and the 88. disadvantages of um of each of the treatments 89. P My preferred treatment would be the radiotherapy and if I can go into 90. that a little further, I have got a friend that I play golf with and he had 91. treatment, radiotherapy he started off by getting jabs in his stomach [continues] (P7 randomised to AM and accepted allocation)

Appendix B. CA transcription conventions 2. Participant-led appointment 1. R (referring to patient information sheet) right so how did you get on? 2. P Well I went through it and um I’ve got quite a few questions I would like 3. to ask. 4. R Ok 5. P Um I don’t know how to start with it. They said ok you have early 6. prostate cancer which appears to be confined to the prostate gland. How do 7. you know this? 8. R Well what we think or what we can tell from your results from the blood 9. test is that it’s a localised so it is within the prostate. People um have 10. prostate cancer that’s outwith the prostate tend to have a higher PSA a higher 11. blood test [11 lines further explanation] 23. P So what is my current PSA level can you tell me 24. R 3.3 25. P 3.3 so that is not too bad is it? 26. R It is slightly above the norm, in actual fact until about 18 months ago um 27. at your age they would have considered that a biopsy would not have been 28. done, that would have been on the normal threshold but now it’s just a lower 29. threshold 30. P So what is the norm now? 31. R 0 to 3 is the normal threshold for your age range and then for older men 32. it’s 0 to 5 [11 lines explanation about PSA levels] 44. P Well (pause) let’s see, if the PSA level rises the doctor will do further 45. tests to see if the cancer is spreading. What further tests will he do? A biopsy? 46. R No he would do a scan, a bone and an MRI scan which is a scan of the 47. pelvic area and the prostate [3 lines description of MRI scans] 51. P What does MRI stand for 52. R It stands for magnetic resonance imaging 53. P Is that ultra sound. 54. R No it’s not ultrasound, it’s magnetic imaging

References Albrecht, T. L., Blanchard, C., Ruckdeschel, J. C., Coovert, M., & Strongbow, R. (1999). Strategic physician communication & oncology clinical trials. Journal of Clinical Oncology, 17(10), 3324–3332. Arora, N. K. (2003). Interacting with cancer patients: the significance of physicians’ communication behaviour. Social Science & Medicine, 57, 791–806.

[] ¼ (0.4)

:::

Square brackets Equals sign Time in round brackets Period in round brackets Colons

.

Period

,

Comma

?

Question mark

here HERE

Underlining Capitals Degree signs ‘More than’/‘less than’ symbols

(.)





>here< Hem hm Lau(h)gh () (here) .hhh

‘h’ in brackets within a word Empty round brackets Word(s) in round brackets

Overlapping talk No space between turns Intervals within or between talk (measured in tenths of a second) Discernable silence, too short to measure Extension of preceding sound (the more colons, the greater the extension) Closing intonation (not necessarily the end of a sentence) Continuing intonation (not necessarily between clauses of sentences) Rising intonation (not necessarily a question) Emphasis Loud, relative to surrounding talk Soft, relative to surrounding talk Speeded up, relative to surrounding talk Laughter voiced separately from talk (may also be represented by Hah or Heh) Laughter interpolated within talk Transcriber could not hear what was said Possible hearing of what was said In breath Atkinson & Heritage, 1984, pp. ix–xvi

Atkinson, J. M., & Heritage, J. (1984). Structures of Social Action: Studies in Conversation Analysis. London: Cambridge University Press. Boulton, M., & Parker, M. (2007). Informed consent in a changing environment. Social Science & Medicine, 65, 2187–2198. Bower, P., King, M., Nazareth, I., Lampe, F., & Sibbald, B. (2005). Patient preferences in randomised controlled trials: conceptual framework and implications for research. Social Science & Medicine, 61, 685–695.

2028

J. Wade et al. / Social Science & Medicine 68 (2009) 2018–2028

Britton, A., McKee, M., Black, N., McPherson, K., Sanderson, C., & Bain, C. (1998). Choosing between randomised and non-randomised studies: a systematic review. Health Technology Assessment, 2(13). Brown, R. F., Butow, P. N., Butt, D. G., Moore, A. R., & Tattersall, M. H. N. (2004). Developing ethical strategies to assist oncologists in seeking informed consent to cancer clinical trials. Social Science & Medicine, 58, 379–390. Brown, R. F., Butow, P. N., Ellis, P., Boyle, F., & Tattersall, M. H. N. (2004). Seeking informed consent to cancer clinical trials: describing current practice. Social Science & Medicine, 58, 2445–2457. Chard, J. A., & Lilford, R. J. (1998). The use of equipoise in clinical trials. Social Science & Medicine, 47(7), 891–898. Charles, C., Gafni, A., & Whelan, T. (1999). Decision-making in the physician–patient encounter: revisiting the shared treatment decision-making model. Social Science & Medicine, 49, 651–661. Degner, L. F., & Sloan, J. A. (1992). Decision making during serious illness: what role do patients really want to play? Journal of Clinical Epidemiology, 45, 941– 950. De Salis, I., Tomlin, Z., Toerien, M., & Donovan, J. (2008). Qualitative research to improve RCT recruitment: issues arising in establishing research collaborations. Contemporary Clinical Trials, 29, 663–670. Dixon-Woods, M., Ashcroft, R. E., Jackson, C. J., Tobin, M. D., Kivits, J., Burton, P. R., et al. (2007). Beyond ‘‘misunderstanding’’: written information and decisions about taking part in a genetic epidemiology study. Social Science & Medicine, 65, 2212–2222. Donovan, J., Hamdy, F., Neal, D., Peters, T., Oliver, S., Brindle, L., et al. (2003). Prostate testing for cancer and treatment (ProtecT) feasibility study. Health Technology Assessment, 7(14). Donovan, J., Lane, A., Peters, T., Brindle, L., Salter, E., Gillatt, D., et al. (2008). Development of a complex intervention improved randomisation and informed consent in a randomized controlled trial. Journal of Clinical Epidemiology. Donovan, J., Mills, N., Smith, M., Brindle, L., Jacoby, A., Peters, T., et al. (2002). Quality improvement report: improving design and conduct of randomised trials by embedding them in qualitative research: ProtecT (prostate testing for cancer and treatment) study. BMJ, 325, 766–770. Drew, P., & Heritage, J. (1992). Analyzing talk at work: an introduction. In P. Drew, & J. Heritage (Eds.), Talk at work: Interaction in institutional settings (pp. 3–65). Cambridge: Cambridge University Press. Dunn, L. B., & Jeste, D. V. (2001). Enhancing informed consent for research and treatment. Neuropsychopharmacology, 24(6), 595–607. Edwards, S. J. L., Lilford, R. J., Thornton, J., & Hewison, J. (1998). Informed consent for clinical trials: in search of the ‘best’ method. Social Science & Medicine, 47(11), 1825–1840. Featherstone, K., & Donovan, J. L. (2002). Why don’t they just tell me it straight, why allocate it? The struggle to make sense of participating in a randomised controlled trial. Social Science & Medicine, 55, 709–719. Featherstone, K., & Donovan, J. L. (1998). Random allocation or allocation at random? Patients’ perspectives of participation in a randomised controlled trial. BMJ, 317, 1177–1180. Freedman, B. (1987). Equipoise and ethics of clinical research. New England Journal of Medicine, 317(3), 141–145. Gatellari, M., Butow, O. N., & Tattersall, M. H. N. (2001). Sharing decisions in cancer care. Social Science & Medicine, 52, 1865–1878. General Medical Council. (2008). Consent: Patients and doctors making decisions together. London: GMC. Glaser, B., & Strauss, A. (1967). The discovery of grounded theory: Strategies for qualitative research. New York: Aldine.

Grossman, S. A., Piantadosi, S., & Cohavey, C. (1994). Are informed consent forms that describe clinical oncology research protocols readable by most patients and their families? Journal of Clinical Oncology, 12, 2211–2215. Heritage, J., & Robinson, J. D. (2006). The structure of patients’ presenting concerns: physicians’ opening questions. Health Communication, 19(2), 89–102. Heritage, J., & Sefi, S. (1992). Dilemmas of advice: aspects of the delivery and reception of advice in interactions between health visitors and first-time mothers. In P. Drew, & J. Heritage (Eds.), Talk at work: Interaction in institutional settings (pp. 359–417). Cambridge: CUP. Hietanen, P., Aro, A. R., Holli, K., & Absetz, P. (2000). Information and communication in the context of a clinical trial. European Journal of Cancer, 36, 2096–2104. ICH harmonised tripartite guideline for good clinical practice for research E6 (R1). (1996). . Jenkins, V. A., Fallowfield, L. J., Shouhami, A., & Sawtell, M. (1999). How do doctors explain randomised clinical trials to their patients? European Journal of Cancer, 35(8), 1187–1193. Krippendorff, K. (2004). Content analysis: An introduction to its methodology (2nd ed.). Thousand Oaks, CA: Sage Publications. Maynard, D. W., & Heritage, J. (2005). Conversation analysis, doctor–patient interaction and medical communication. Medical Education, 39, 428–435. Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook. London: Sage Publications. Mills, E. J., Seely, D., Rachlis, B., Griffith, L., Wu, P., Wilson, et al.. (2006). Barriers to participation in clinical trials of cancer: a meta-analysis & systematic review of patient-reported factors. Lancet Oncology, 7, 141–148. Mills, N., Donovan, J. L., Smith, M., Jacoby, A., Neal, D. E., & Hamdy, F. C. (2003). Patients’ perceptions of equipoise are crucial to trial participation: a qualitative study of men in the ProtecT study. Controlled Clinical Trials, 24, 272–282. Neuendorf, K. A. (2002). The content analysis guidebook. Thousand Oaks, CA: Sage Publications. NRES. (2005). Information sheets and consent forms, guidance for researchers and reviewers. Version 3.2. National Research Ethics Service, National Patient Safety Agency. Accessed 22.10.07. Ong, L. M. L., de Haes, J. C. J. M., Hoos, A. M., & Lammes, F. B. (1995). Doctor–patient communication: a review of the literature. Social Science & Medicine, 40, 903–918. Robinson, E., Kerr, C., Stevens, A., Lilford, R., Braunholtz, D., & Edwards, S. (2004). Lay conceptions of the ethical and scientific justifications for random allocation in clinical trials. Social Science & Medicine, 58, 811–824. Ross, S., Grant, A., Counsell, C., Gillespie, W., Russell, I., & Prescott, R. (1999). Barriers to participation in randomised controlled trials: a systematic review. Journal of Clinical Epidemiology, 52(12), 1143–1156. Schegloff, E. A. (2000). Overlapping talk and the organization of turn-taking for conversation. Language in Society, 29(1), 1–63. Schwabe, M., Howell, S. J., & Reuber, M. (2007). Differential diagnosis of seizure disorders: a conversation analytic approach. Social Science & Medicine, 65, 712–724. Silverman, D. (2000). Doing qualitative research: A practical handbook. London: Sage Publications. Silverman, D. (2001). Interpreting qualitative data: Method for analyzing talk, text and interaction. London: Sage Publications. Tomamichel, M., Sessa, C., Herzig, S., de Jong, J., Pagani, O., Willems, Y., et al. (1995). Informed consent for phase 1 studies: evaluation of the quantity and quality of information provided to patients. Annals of Oncology, 6, 363–369. World Medical Association. (2004). Declaration of Helsinki, ethical principles for medical research involving human subjects. Tokyo: World Medical Association General Assembly.