Informed consent (IC) lies at the crux of patient-centred care in medicine. It affirms patient autonomy and should convey vital information, including the nature, risks, benefits and alternatives of a procedure.1 Despite its acknowledged importance and purpose, IC in endoscopy is inconsistently performed.2–5 Even when performed, it does not always fulfil its purpose with patients having seemingly limited understanding.6,7 Patient dissatisfaction and complaints may arise as a result.8–10

Traditionally, IC involves a verbal discussion prior to a procedure. Differences in the process arise from inherent variations by different health professionals discussing the procedure and patient variability of comprehension based on educational, ethnic, cultural and socioeconomic factors.7,11 Inconsistency is amplified by variation and uncertainty of what should be discussed and in particular, risks of the procedure. Endoscopy guidelines state IC of risk should be “procedure-, circumstance- and patient-specific, even if the likelihood is very small.”1

Different interventions have been trialled to improve the IC process. While extended discussions appear effective at improving patient understanding for research consent,12 both time and staff constraints in endoscopy may limit feasibility and practicality of this. Increasingly, multimedia education tools such as video are acknowledged as useful aids for this purpose.13 Endoscopy-specific studies have demonstrated that patients may score better in knowledge-based tests with video IC for endoscopy.3,14–16 In addition, there is no increase in anxiety and improved satisfaction, consistent with an understanding that patients prefer full disclosure.14,15,17–21

Use of video-assisted IC has not been widely adopted in gastrointestinal endoscopy. The absence of population-specific evidence in light of known patient variability, in conjunction with lack of evidence of feasibility or acceptability to both patient and clinician may explain this. Clinical equipoise remains as to the best method to obtain IC in terms of patient satisfaction, preference and provision of information.

We hypothesise that a video-assisted IC on the day of the procedure may be more effective and preferred by patients and staff as a means to educate and consent than traditional verbal IC.

Materials and methods

The study was performed at Tauranga Hospital, a secondary centre with 349 inpatient beds, servicing the Western Bay of Plenty in New Zealand including a catchment of over 230,000 people. The majority of endoscopy cases are triaged directly to endoscopy from community referrals.

Participants

All consecutive patients aged 18 years or older scheduled for an outpatient gastroscopy, colonoscopy or both between 1 June 2017 and 1 December 2017 were considered for enrolment in the trial. Patients were excluded if they were having a procedure other than colonoscopy or gastroscopy such as Endoscopic Retrograde Cholangiopancreatography (ERCP) or a planned therapeutic intervention such as dilatation or variceal banding. Further exclusions included patients who would ordinarily require a third party to sign consent (due to age, intellectual disability, lack of capacity, those visually or hearing impaired) and patients with a language barrier such that an interpreter was required.

Study design

Consistent with standard practice in our department, all patients were sent an information booklet containing details of the procedure at least one week prior to their appointment. Upon arrival to the unit and before randomisation, eligible patients were given a written participation information sheet explaining the purpose and requirements of the study. If agreeable to participation, patients were then randomly assigned to either the ‘verbal’ or ‘video’ group with allocation concealed in opaque envelopes, along with the patient questionnaire. Envelopes were pre-made, shuffled and then distributed to participants by administration staff not involved in the remainder of the trial.

The control group received the standard practice of verbal, nurse-led consent. All nurses participating had successfully completed competency training, including direct observational practice evaluation in taking IC.

Intervention and video design

Two videos were created, professionally filmed and narrated by a nurse investigator (TF). The videos were 3.5 minutes long and designed to cover all aspects of legal IC. Subtitles were included to highlight individual risks as they were mentioned. Validity of the videotape script content was established by submission to the Gastroenterology Department and Hospital Media Staff.

The intervention was the demonstration of the video, played in the patient bed space and displayed on an A5-sized tablet with the use of earphones in order to minimise contamination bias with neighbouring patients. Following the video, patients had the opportunity for further discussion and to ask questions. Patients were then asked to sign a standard hospital consent form and fill out the questionnaire.

Outcomes

The primary outcomes measured were the recollection of procedural risks (sum of all correct answers for risk recall items) and patient experience. Following the procedure and prior to discharge, patients were asked to record if they felt their experience was “similar, better or worse” than the video or verbal explanation that was provided to them. Secondary outcomes included breakdown of individual risk recall items, patient perception about understanding and taking of IC, risks, alternatives and anxiety.

Outcome assessment

The questionnaire consisted of 22 items (see supplementary file), designed and adapted from earlier validated studies where available.5,22 The majority of the questionnaire was formatted using tick boxes, intended to limit ambiguity. It was divided into three segments: baseline demographic variables, knowledge and preference/expectation.

To assess knowledge, an open format asking about risks of the procedure was adapted.23 Participants were asked to list the risks that they could recall. In addition, we asked at the end of IC whether patients felt “more anxious/less anxious/no different”.

Post-procedure, after the sedation had worn off, participants were given back the study questionnaire to complete the remaining two questions regarding the experience of the procedure as compared to the explanation received during their IC.

At the end of the study period, a separate questionnaire was given to participating nurses to gather feedback on the effect of use of the video on time and patient flow within the department.

Statistical analysis

We aimed to detect an effect size in the sense of Cohen (proportion of the baseline pooled standard deviation) of 0.5 at 80% power for a False Discovery Rate-corrected significance level of 2.5%.24 To do so, we required 78 participants per arm; assuming 20% missing values, 98 participants per arm were required to meet these criteria. We therefore aimed to recruit 100 participants per arm.

The primary analysis was carried out in an intention-to-treat analysis set. Linear regression was used to analyse the risk recall score. Logistic and proportional odds regression were used to analyse binary and multinomial outcomes respectively. Rare categories (<5%) in multinomial outcomes were collapsed with an adjacent category. The models were assessed for adjustment by baseline demographic variables during a blind review, absent of any knowledge regarding allocation.

Missing data was multiply imputed in 15 replications of the data set. Subgroup analyses were carried out in prespecified categories (prior procedure, level of education, type of procedure) using interactions between subgroup indicator and group allocation. The nominal significance level was set at 5% against two-sided alternatives, and multiplicity controlled using False Discovery Rate.

Results

A total of 280 potential participants were assessed for the study, with 80 excluded, leaving 100 patients randomised to each of the verbal and video groups (Figure 1). The majority of patients were in the 50–75 year-old age group (Table 1). Considerably more patients in verbal arm attended university (44%) with more in the video arm having been educated to a high-school level (56%). Ninety-two percent of patients in the video group indicated they received the pre-endoscopy information booklet, read it and found it useful. This is compared to 98% in the verbal group, with all except one patient reporting it was useful.

Figure 1: Study design and exclusions.

Table 1: Baseline patient characteristics.

‍

There was no significant difference in means between the total number of correctly identified risks recalled in the video compared to the verbal group: difference of means 0.11 (95% CI (-0.19, 0.41) p=0.46). Similarly, there was no significant difference between the groups in the reported patient experience: experience better compared to experience similar or worse OR 0.86 (95% CI (0.45, 1.63) p=0.55). A total of 192/196 patients from both groups reported that the procedure was as expected or better than what was explained to them during the IC process (Table 2).

Table 2: Procedural experience relative to explanation during IC.

‍

There was a signal towards improved recall of bleeding (Table 3) as a potential risk in the video as compared to the verbal arm, but it did not reach statistical significance (p=0.059). There is mild evidence for an effect on this particular recall risk, with significance noted in two of the sensitivity analyses (see Appendix). Following bleeding and perforation, there was a marked drop-off in recollection in both groups of the remaining five risks that were discussed, with only one person in the video group recalling missed pathology as a risk (Table 3).

Table 3: Recollection of individual risks.

‍

Nearly all patients in both arms of the study (>96%) felt following the IC process that they had a good understanding, knew what to expect and why the procedure was taking place. Very few had questions remaining at the conclusion of their IC (Table 4).

Table 4: Patient perception.

‍

Patients verbally consented reported to be less anxious about the procedure (Odds Ratio Estimate 0.38: 95% CI (0.18–0.82) p=0.014). This result was supported by the sensitivity analyses (see Appendix).

Following the study, all nurses who took part completed the survey to measure feasibility and acceptability.

12/17 (71%) preferred the video over verbal IC. 8/17 (47%) thought it saved time, 4/17 (24%) thought it took more time and 5/17 (29%) thought it made no appreciable difference.

Discussion

This study revealed that video-assisted consent made no significant difference to the IC process in terms of patient recollection compared to the usual verbal IC performed by an experienced nurse. A new and unanticipated finding of the trial was a striking inability to recall risk, despite utilising a combination of techniques shown to be of benefit in other studies, including written information sent to the patients one week prior to examination.25 This result may represent an overestimation of true information retention as the questionnaire was provided to patients immediately following their IC. Further subgroup analyses did not demonstrate differences based on education, ethnic, or endoscopic experience. Insofar as we are aware, this is the first paper that specifically assesses and reports patient recollection of specific risks in the endoscopy setting. Clinicians need to be aware of this poor recollection and may need to consider modifying how to prioritise information delivery during the IC process. It should also serve to highlight to other specialities, with procedures or operations more complex or riskier than endoscopy, that special care and adequate time must be taken to allow patients to comprehend and question during the IC process.

Despite the lack of recollection of procedural risks, nearly every patient in both groups reported that they understood the procedure, felt like they knew what to expect, why it was being performed and only very few had questions remaining at the conclusion of the process. Following the procedure, they felt their experience was similar or better to what was discussed during the IC. This result is very encouraging and reassuring. One supposition to explain these incongruous findings is that perhaps one cannot equate recall with understanding or comprehension, which may be more complex. It has been described that “people often make reasonable decisions but cannot later recall the premises that supported the reasoning or the process that led to the conclusion. Nevertheless, they might well have understood it at the time.”26 As medical professionals, we rely heavily on objective, categorical and quantitative data to guide clinical practice and the available studies on IC reflect this. To our knowledge, this is the only study in endoscopy IC to report on patients’ perception of this process regarding both their feeling of understanding and fulfilment of expectations. Patient satisfaction is a key factor in reducing misunderstanding, disputes and complaints, with legal cases frequently due to failures in communication as opposed to failures in treatment.27 Perhaps patient satisfaction, infrequently measured in studies,28 as well as objective recall could be integrated into future outcomes or measures of success in trials of IC.

Regarding the amount of information provided during IC, other endoscopy video IC studies measured recollection of procedural risks based on only three items. With the aforementioned ambiguous guidance as to what should be discussed, it is naturally of concern to the authors that in our study, only 1/157 responses could recall missed pathology. The risk of this has been found to be 2–6% for cancer29,30 with 43% of metachronous colorectal cancers attributed to missed lesions on prior colonoscopy.31 However, the reported 0.06% risk of perforation,32 or 0.1–0.3% risk of post-endoscopy infection33 is often emphasised and better recalled. The poor recollection of risks, perhaps augmented by a stressful pre-operative setting, may inform clinicians, endoscopy units and future studies of IC to modify how they prioritise information delivery during the IC process.

Lastly, acceptability and feasibility to staff is vital for any new intervention. The majority of nurses preferred using the video, and nearly half thought it saved time. This may be an underestimation, confounded by the learning curve with additional time taken to run the study. Limitations of this study include the lack of blinding of the nursing staff to patient allocation group, which was not feasible in our setting. Despite a clear and user-friendly questionnaire there were substantial missing observations arising from incomplete survey forms. After randomisation there were considerable differences in baseline demographics of the two groups. Although these differences may in principle have led to chance bias, the blind review and adjustment for demographic covariates allays this risk. Decreased anxiety in the verbal arm, perhaps a reflection of the inherent reassurance with human interaction is difficult to confirm, as a formal State-Trait Anxiety Inventory (STAI) questionnaire34 was not utilised in favour of a single simplified and non-validated question “more anxious/less anxious/no different”.

In summary, this study showed no difference between video-assisted or verbal IC in terms of recollection of procedural risks, which were poorly recalled, or patient fulfilment of expectation, experience or perception of understanding, which was universally high. Further action to prioritise information delivery during IC is required. Future studies in this field should include patient-centred outcomes as a measure of success.

Appendix

Supplementary files

1. Sensitivity analyses: Complete data (no imputation).

‍

2. Sensitivity analyses: Unadjusted data (imputed data only).

‍

3. Sub-group analysis based on patient baseline characteristics.

Previous scope

Risk recall score (difference of mean)

P value

‍

4. Subgroup analysis of patient experience with “experience better” compared to reference category of “experience similar or worse” for video vs verbal groups.

‍

Consent questionnaire

‍