全文

The informed consent process is both an ethical and a legal requirement before surgery. Appropriate and clear communication about the proposed surgery can assist the patient in becoming actively involved in decision making. Informed consent should include information about the nature and purpose of the treatment, its risks and consequences, and alternative courses of treatment 1 and 2. The need for adequate informed consent has been driven by several factors including increased expectations of surgical outcomes by both patients and surgeons and by medicolegal implications.

 

In Australia, legal cases such as Rogers v Whittaker (1992) found that “A doctor has a duty to warn a patient of a material risk inherent in the proposed treatment,” [3] thus raising the standards of information supplied to patients contemplating a surgical procedure. Of nearly 8000 Australian medicolegal cases resolved in the 7 years from 2002 to 2008, approximately 1 in 30 medical negligence claims and 1 in 9 conciliated complaints included allegations of problems with informed consent [4].

 

Several studies have examined use of videos and multimedia modules (MMs) in assisting with informed consent. Compared with standard consultation with or without written information, use of a video or MM led to improved patient understanding of the procedure they were to undergo 5, 6, 7, 8, 9, 10, 11, 12 and 13. To our knowledge, only 1 trial, by Mason et al [5] in 2003, has examined use of a video for a gynecologic procedure. We conducted a randomized controlled trial to compare the verbal consent process with or without the use of an MM for operative gynecologic laparoscopy. The hypothesis to be tested was that providing additional information in the form of an MM would improve patient knowledge without increasing anxiety levels.

 

Material and Methods

Ethical approval for the study was given by the Human Research and Ethics Committees of the Mercy Hospital for Women and Epworth Healthcare. Women who had been recommended to undergo operative laparoscopy for pelvic pain, were English speaking, and were able to give informed consent were approached for recruitment into the study. Women were recruited at 2 gynecology outpatient clinics: private rooms of one of us (P.M.) and the public endosurgery clinic of a major urban teaching hospital.

 

The authors developed a script for the MM education module for operative laparoscopy that included core information to be included in the module. A literature review was performed to look at the complications associated with operative laparoscopy. This information was used to determine an average risk for each complication and to present this in the MM. A 15-minute educational MM for operative laparoscopy was created, and comprised a mixture of voice, text, and 3-dimensional (3D) computer animation. Using a 3D graphics creation software program (Studio Max 4.0; Autodesk, Inc., San Rafael, CA), specific 3D animations were created that included all aspects of the core information. Four representative module screen shots are shown in Fig. 1.

 

 

The key points included in the MM were as follows: an explanation of the procedure, indication for the procedure, likely benefits of the procedure, possible risks of the procedure, and an explanation of the hospital process and discharge management.

 

Patients completed the routine surgical consent process with one of us (L.E. or P.M.). The physicians knew the information contained in the MM, and the process was performed in a standard, well-rehearsed manner with each patient. Key points of the surgical consent were documented on the hospital consent form and included risks of surgery (e.g., bowel injury, 1:1000). A decision was made not to use a script for the physician to read from because this breaks eye contact and reduces the clinician's ability to pick up on visual cues that patients may not understand the information being provided to them. If patients then agreed to be involved in a clinical trial, they were taken to another room and met by a research assistant. If patients consented to participate in the trial, they were randomized into the intervention or control group using allocations contained in sealed opaque envelopes that had been generated by a computer randomization program. The physician responsible for the routine verbal consent process was blinded to the allocation groups of the patients.

 

Demographic data collected included age, education level, previous surgery, and public or private patient status.

 

To assess anxiety levels, we used the 6-item version of the Spielberger 20-item State-Trait Anxiety Inventory (STAI). This is a validated short form and correlates well with the standard inventory but is simpler and quicker to use [14]. The STAI measures state anxiety (how one feels at the moment: “feel questions”) and trait anxiety (how one generally feels: “am questions”) (Fig. 2).

 

 

A 14-statement knowledge questionnaire (Fig. 3) was developed from the core information and evaluated through a number of revisions made by us. The final selection of 14 statements was achieved, with 7 of the 14 specifically related to possible complications of the proposed surgery. Each statement asked the patient to answer “True,” “False,” or “Unsure.” A correct answer was either True or False (Fig. 3), and participants were instructed to answer as honestly as possible without guessing. Patients who answered incorrectly or who chose the “Unsure” option received no score for that item. This was done to limit the number of guesses the patient made. Thus, for the purpose of analysis, the total score for each participant was a whole number, with a maximum score of 14.

 

 

A survey was also developed to assess patient satisfaction with the amount, method, and content of the information during the consent process. These questions were constructed using a 10-cm visual analog scale (VAS). In addition, patients in the intervention arm were asked if they felt more, less, or no difference in their anxiety level after watching the MM. Patients were also asked how they would prefer to receive informed consent in the future, i.e., from a physician only or from a physician plus the MM (Fig. 4).

 

 

After randomization, the women in the control group gave baseline demographic data, answered the knowledge questionnaire, and completed the STAI and the VAS survey. After randomization, the intervention group watched the 15-minute MM, then completed the knowledge questionnaire, STAI, and VAS survey regarding their perceptions.

 

Six weeks after original recruitment, all of the women were contacted again by the research assistant and completed the STAI and knowledge questionnaire for a second time. Inasmuch as waiting times for surgery differed between patients, it was decided that follow-up would occur at 6 weeks from recruitment for all patients. Patient perceptions were not remeasured because these related to the information delivery system, and it was believed that this would be less relevant to the participants at 6 weeks. All patient records were reviewed several months after completion of the trial to check for any surgical complications encountered.

 

Statistical Analysis

The number of patients needed was calculated on the basis of an analysis of covariance using baseline demographic variables. The calculation assumed a difference between groups of 3 points, with 3 SD, r2 = 0.15, α = 0.05, and ß = 80%. This gives a requirement of 17 patients randomized to each treatment group, which was surpassed in the present study.

 

Statistical analysis was performed using commercially available software (SAS/STAT version 9.2, SAS Institute, Inc., Cary, NC; and GenStat version 15.2, VSN International, Ltd., Hemel Hempstead, UK). Differences in the knowledge scores of the 2 treatment groups immediately after the intervention and 6 weeks later were estimated using analysis of covariance models. Further models were fitted to assess the sensitivity of the primary models to any imbalance between the intervention groups in the age and education distributions.

 

The State and Trait components of the STAI short form were analyzed using a linear mixed model, with the control vs intervention as a between-subject factor and initial vs 6-week results as a within-subject factor.

 

Comparison of the intervention groups insofar as questions about ease of understanding and appropriate amount of information, both measured on 10-cm VAS scales, was made using a bootstrapped comparison of means (10 000 samples with replacement). Comparison of the intervention groups insofar as consent approach preference was made using a permutation test (10 000 samples without replacement) [15].

 

Results

Recruitment took place at both sites, commencing in July 2012 and completed in December 2013. Forty-nine patients scheduled to undergo operative laparoscopy were asked to participate in the study (Fig. 5). Eight patients declined participation, most commonly because of concern about the time it would take. Forty-one patients consented to participate in the study and were randomized into 1 of 2 groups: 20 patients in the control arm, and 21 patients in the intervention arm. At 6 weeks, 5 patients were lost to follow-up, leaving 17 in the control arm and 19 in the intervention arm. Baseline demographic data for the study population are given in Table 1.

 

 

 

The means for the anxiety scores are given in Table 2. There was no significant interaction between the intervention effect and time (initial vs 6 weeks) for either State or Trait. The overall mean difference between control and intervention was small and not significant. For both variables (State and Trait anxiety scores), there was a small but significant overall decrease in anxiety scores over the 6 weeks.

 

 

At initial testing, the intervention MM group achieved a significantly higher mean (SE) score (11.3 [0.49]) for the knowledge questionnaire than did the control group (7.9 [0.50]) (p <.001) (Table 3). The difference between the treatment group mean was 3.4 (95% confidence interval [CI] 2.0–4.9). The control group achieved 56.4% correct answers and 43.6% incorrect answers (wrong answer, 16.1%; unsure, 27.5%). The intervention group achieved 80.9% correct answers and 19.1% incorrect answers (wrong answer, 14.3%; unsure, 4.8%).

 

 

Despite a robust randomization process, there seemed to be some baseline differences between the 2 groups insofar as age and education level (Table 1). Further analysis was performed on the knowledge scores to assess whether age and education may have an effect on outcomes. Adjusting for patient age, the mean estimates for initial knowledge scores were 7.7 and 11.5 for the control and intervention groups, respectively. Adjusting for patient age and education level, the mean estimates for Initial Knowledge score were 7.8 and 11.5 for the control and intervention groups respectively. Adjustments for age and education therefore make no meaningful difference in the estimates.

 

Six weeks after initial recruitment, the improved knowledge in the MM group was no longer apparent. The mean (SE) score for the intervention group was 8.4 (0.53), compared with 7.8 (0.50) in the control group (p = .44) (difference in mean scores 0.6; 95% CI for the difference, −0.5 to 2.2). At 6 weeks, the control group achieved 55.4% correct answers and 44.6% incorrect answers (wrong answer, 22.2%; unsure, 22.2%), and the intervention group achieved 60.5% correct answers and 59.5% incorrect answers (wrong answer, 27.7%; unsure, 12.8%).

 

There was also no statistically significant difference between the 2 groups in the change in score. After adjusting for initial knowledge, the mean (SE) changes in knowledge score were 1.50 (0.53) for the control group and 2.14 (0.49) for the intervention group. The point estimate of treatment group difference was 0.64, with an approximate 95% CI of −2.32 to 1.04. The difference in knowledge score between the 2 groups was statistically significant only at initial testing. At 6-week follow-up, there was no evidence to suggest that the mean knowledge scores in the treatment groups differed.

 

Patients in both groups were asked how well they understood the information and the outcome measured using a VAS. Both groups thought they understood the information well; in the MM group, the mean (SD) score was 8.5 (1.86), and in the control group was 8.2 (2.48). This difference was not statistically significant. (p = .64; t test and bootstrapped comparison of treatment group means, p = .65).

 

The results of the STAI showed no difference between the 2 groups in anxiety level at recruitment or at 6 weeks (Table 2). In addition to this validated method of scoring anxiety levels, the intervention group was asked whether they felt more anxious after watching the MM. Most patients (n = 17 [81.0%]) reported that viewing the MM either did not influence or reduced self-reported anxiety levels.

 

The intervention group was asked, once the consent process was completed, how they wished it had occurred. Sixteen patients (76.1%) answered physician plus MM; 1 patient (4.8%), MM alone; and 4 patients (19.6%), physician alone.

 

Patients were also asked how they would like to receive information if they were having an operation in the future. In the control arm, 8 patients (40.0%) stated they would prefer physician only and 12 (60.0%) stated physician plus an MM. In the intervention arm, 3 patients (14.3%) preferred physician only and 18 (86.0%) would prefer physician plus an MM. There was no statistically significant difference between treatment groups in terms of distribution of responses (permutation test, p = .07).

 

Discussion

Use of an MM was shown to improve patient knowledge about operative laparoscopy to treat pelvic pain at initial testing. The medical literature demonstrates that patients remember little of verbal information disclosed during the informed consent process and that their level of comprehension is often overestimated [16]. Finding ways to improve the informed consent process is therefore critical to aid patient autonomy in decision making and to assist the physician in building a healthy relationship with the patient and in authorizing appropriate treatment plans.

 

Similar studies have demonstrated that knowledge scores deteriorate over time. Cornoiu et al [6] compared an MM with verbal consent alone or pamphlet only arms for patients scheduled to undergo knee arthroscopic surgery. They found that patients in all 3 groups exhibited deterioration of knowledge over time. The MM group still performed better than the other 2 groups at follow-up, and this difference remained statistically significant. In the present study, the MM group seemed to demonstrate greater deterioration in knowledge than did the intervention group. This occurred primarily with regard to questions that asked a specific number (e.g., rates of bowel injury, death, and gas embolism) (Fig. 3). Patients in the MM group were more likely to remember the answers to these specific questions, but only at initial testing. This may be because the module presented the risk pictorially as well as verbally and was therefore easier to remember at the time of patient education. Furthermore, in the MM group the information was reinforced because they received it twice, verbally from the physician and from the MM. However, by 6 weeks, the intervention group had lost this advantage.

 

Clinicians need to be aware of this deterioration in knowledge. Patients give consent to proceed with a surgical procedure at a specific time, and it is the surgeon's responsibility to ensure that the patient is in a position to make as informed a decision as possible. If the surgeon obtains consent from a patient many weeks before their operation, important surgical points should be reiterated. The patient may remember little on the day of surgery because of the demonstrated degradation in information retention over time. In addition, the surgeon should document the key points of the informed consent process for medicolegal reasons because patients are unlikely to remember them.

 

When designing the MM, we believed that outlining significant complications was an essential component of the core information. Seven of 14 questions on the knowledge questionnaire related specifically to complications. Of 41 patients in the study group, 2 experienced complications, a recognized bowel injury and a minor postoperative infection (cystitis). Both women were in the intervention arm of the trial and had therefore seen the MM. The bowel injury was in a 42-year-old woman with severe adhesions. During a difficult pararectal dissection, the rectum was inadvertently entered. The defect was oversewn laparoscopically with 3× interrupted 2/0 polydioxanone sutures, and the patient made a full recovery. A meta-analysis by Brosens et al [17] found that diagnostic and minor operative laparoscopy are associated with an .08% risk of bowel injury, and in major operative laparoscopy the risk increases to 0.33%. On review of the knowledge questionnaire, the patient involved answered the question about bowel injury correctly at initial testing and at 6 weeks (Fig. 3) and made the following comment to the research team: “Having watched the video helped me deal with the complication as I knew that it could happen and I knew that it could be dealt with if recognized appropriately. I didn't need as much explanation after my operation.”

 

There is some disparity in the literature insofar as patient anxiety levels with use of videos. Some authors have found that use of further information in the form of videos or MM reduces anxiety, and others have found no change in anxiety levels. We found that use of the MM in addition to verbal consent did not improve anxiety scores; however, anxiety levels were not higher in patients exposed to the MM. We did note that mean anxiety levels for all the study participants decreased during the 6 weeks, which is not surprising because most patients will find a surgical discussion with their physician to be stress provoking.

 

The lack of any significant effect of video or MM use on anxiety is in keeping with the findings of other researchers. Mason et al [5] used a video to improve informed consent for women undergoing laparoscopic sterilization. They found no difference in mean anxiety scores in either intervention group (verbal only or verbal plus video). Cornoiu et al [6] found that anxiety scores were similar in all intervention arms. Conversely, Freeman-Wang et al [18] and Thomas et al [19] reported that the use of videos reduced anxiety. However, both of these trials included patients with a diagnosis of cancer or a precancerous lesion, in whom baseline anxiety levels were understandably high and any information given was likely to alleviate stress.

 

Many physicians have the perception that giving patients too much information about their surgery, in particular the risks of surgery, may increase distress and anxiety. The present study and the medical literature do not support this view. In a review of the literature we were unable to find any study in which the use of audiovisual information (e.g., video or MM) led to increased anxiety levels. There is evidence that patient anxiety levels are either not affected or will decrease with additional information 5, 6 and 19.

 

A potential limitation of the present study is that the surgeons providing verbal consent were aware that the patient might be part of the trial. Although the surgeons did not know whether the patients would agree to be recruited or to which group they might be randomized, they may have attempted to provide more thorough information during the informed consent process. If this is correct, it might have improved the scores achieved by the control group. However, the intervention MM group still demonstrated significantly greater knowledge scores.

 

It could be argued that a limitation of the present study is reporting bias. Although our knowledge questionnaire has not yet been validated, its format was based on similar reported studies 5 and 6. At a minimum, the questionnaire has undergone a process to ensure that it has both face and content validity. There is no criterion standard for testing knowledge in patients undergoing operative laparoscopy to treat pelvic pain; thus we are unable to assess concurrent validity for our tool. The information surgeons give their patients is likely to be extremely varied, with emphasis placed on the surgeon’s own experiences and, in particular, complications they have encountered. The strength of our knowledge questionnaire is that it was carefully constructed by several specialist gynecologists using evidence-based medical literature of laparoscopic complications, and its design encouraged patients not to guess the correct answer, because of the “Unsure” option with each statement (Fig. 3). Future work examining the best information delivery system could focus on validating tools to test patient knowledge. Although such a tool does not currently exist, the present study demonstrates the strengths of an MM tested using a well-designed questionnaire.

 

The rationale for informed consent is to improve patient autonomy, i.e., to enable the patient, armed with the necessary information, to be able to make his or her own decision about the need for surgery [20]. Clearly, there can be discrepancies in what the individual patient may want to hear about the proposed surgery and what the surgeon believes the patient should know. Use of a standardized MM can assist the surgeon in giving the patient consistent general information about the operation, and the verbal consent process enables the surgeon to outline the risks and benefits of the operation specific to that individual. It is likely that surgeons often fall short of the perfect informed consent process. Braddock et al [16] demonstrated that only 9% of 2535 clinical decisions made by patients after consultation with primary care physicians and surgeons met the criteria for completely informed decision making. The addition of a well-designed MM can be a useful tool to aid the patient's decision making.

 

It was beyond the scope of the present study to examine whether improving the informed consent process will reduce litigation against physicians. Careful documentation and good communication are essential to good medical practice and provide clear evidence of patient-centered care [21]. The risk of litigation often depends on patient dissatisfaction due to lack of communication or rapport with the physician [22]. Therefore, use of an MM should not replace the physician-patient relationship. However, it can be a useful adjunct, as demonstrated in our trial, to improve patient knowledge about the recommended procedure and increase understanding about potential complications and their management.

 

In conclusion, use of an MM improves patient understanding about their operation in the short term, does not increase patient anxiety levels, and was found by patients to be useful and their preferred way of receiving consent. An MM is a novel method of aiding the informed consent process. It should not be a substitute for physician-patient discourse but a useful extra source of information for the patient. An MM can facilitate understanding of crucial facts, ensure that all patients receive the same information, and aid patients in becoming active in their own medical care, thus facilitating patient autonomy.

 

Acknowledgements

We thank Mr. Sean McGuigan, biostatistician with Epworth HealthCare, and Associate Professor Graham Hepworth, Statistical Consulting Centre at the University of Melbourne, for expert assistance with statistical analysis.