ORIGINAL ARTICLE
Evaluating Informed Consent for Colonoscopy Rohan M. Gett, MBBS(Hon), BSc(MED), FRACS, MS,*w Anuk R. Cooray, MBBS(Hon),* Darren Gold, MSc, FRACS, FRCS(Gen), FRCS(Eng), MBBS,*w and Mark Danta, BMed, MD, FRACP*
Background: This study aimed to collect and analyze data on patient knowledge of colonoscopy and their preferences regarding the provision of information about the procedure. Specifically, how much detail patients know about different aspects of the procedure and through which methods they best understand risk are evaluated and demographic correlations identified. Materials and Methods: The study sample consisted of colonoscopy patients from 2 colorectal surgeons and a gastroenterologist at St Vincent’s Public Hospital, Sydney for the period August 1 to November 1, 2010. A voluntary questionnaire was performed in the waiting room before colonoscopy. The questionnaire collected data on patient demographics; patient-perceived knowledge of the procedure; and understanding and preferences of various communication formats. Results: Measures of patient-perceived knowledge about colonoscopy were significantly lower than those that would be preferred by patients (P = 0.002). Those with higher levels of education preferred communication of colonoscopy-related information via a leaflet form, whereas those with lower levels preferred verbal information from a doctor or nurse (P = 0.049). The most preferred format for explaining the risk of perforation was the pie graph, followed by both the 1000-person pictograph and absolute risk ratios. Conclusions: Patients received suboptimal levels of information about colonoscopy compared with their preferences. Key areas for improvement include providing more understandable information about the risks of colonoscopy. A combination of written information, diagrams and graphs, and then a discussion of this information to check the understanding is likely to be most effective. Further research into the communication of risk, with larger groups of patients, is likely to help clinicians in gaining fully informed consent in all patients. Key Words: colonoscopy, informed consent, patient knowledge, communicating risks
(Surg Laparosc Endosc Percutan Tech 2014;24:345–352)
I
n Australia, the number of colonoscopies continues to grow at a rate of approximately 8% a year.1 This growth in colonoscopy increases the pressure on doctors to diagnose more pathology and to provide adequate informed consent regarding the complications of colonoscopy. The Received for publication June 12, 2013; accepted January 26, 2014. From the *St Vincent’s Clinical School, University of New South Wales, NSW; and wDepartment of Surgery, St Vincent’s Hospital, Melbourne, Vic., Australia. This study received Human Research Ethics Committee Executive approval at St Vincent’s Hospital (File No. 10/052). The authors declare no conflicts of interest. Reprints: Rohan M. Gett, MBBS(Hon), BSc(MED), FRACS, MS, Suite 607 St Vincent’s Clinic, 438 Victoria St, Darlinghurst, NSW 2010, Australia (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
risk of perforation and hemorrhage in colonoscopy is low (0.02% to 0.2% and 0.05% to 1.0%,2,3 respectively), but requires disclosure and discussion because of the high rate of associated morbidity. An Australian study demonstrated that the majority of patients was either satisfied or very satisfied with the current informed consent process for colonoscopy,4 however, there is substantial variability in the information provided to patients5,6 and many colonoscopists do not provide adequate information to patients.7–10 A Western Australian audit,5 for consent practices in colonoscopy, identified that 12% of interns (doctors in their first postgraduate year) did not always mention perforation or hemorrhage and >50% of interns overestimated colonoscopic failure and perforation rates. These inadequacies in the consent process are potentially widespread and may make clinicians susceptible to litigation.5 Determining how much information is required for patients to give informed consent is difficult and there is a paucity of literature to assist in this process. Studies show that the information required to provide adequate consent is highly variable.6 Many patients, and indeed clinicians, judge the importance of information by the perceived severity of the complication rather than the frequency. Other patients prefer information about all or even none of the complications, irrespective of risk and severity. Colonoscopy patients, in particular, were less likely to want “no information” about complications than gastroscopy patients.6 One study found that professionals tended to underestimate health consumers’ desire for information, with about 40% of consumers wanting more information than they were given.11 This study aimed to compare the amount of information that patients receive compared with the amount they expect. Unfortunately, the effective communication of health risk information to patients is not frequently studied and thus, there is limited scientific evidence on how best to represent quantitative risk data.12,13 Most of the research focuses on the effect of different formats on accuracy and comprehension. Some studies have demonstrated that visual formats of risk, compared with textual or written presentations, assist in the perception, understanding, and interpretation of quantitative information.14,15 In particular, pictographs (eg, 1000 person dot/icon diagrams) have been associated with a more accurate overall impression of risk and numerical knowledge of that risk compared with text formats16 and graphical formats, including bar and pie graphs.17–19 Very little research examines the formats most preferred by patients in explaining the risks of colonoscopy. This study investigated which of following formats was most appropriate for explaining risks in colonoscopy with particular regard to colonoscopic perforation: absolute risk ratios, relative risk ratios, pie graphs, 1000-person pictographs, and logarithmic risk scales.
Volume 24, Number 4, August 2014
www.surgical-laparoscopy.com |
345
Surg Laparosc Endosc Percutan Tech
Gett et al
MATERIALS AND METHODS Participants The study sample consisted of colonoscopy patients from 2 colorectal surgeons and a gastroenterologist at St Vincent’s Public Hospital, Sydney, Australia for the period of August 1, 2010 to November 1, 2010. Forty-five participants were mailed information about the study by their gastroenterologist or colorectal surgeon. Information about the procedure itself was delivered in the usual way to participants through written information and preprocedural visits in clinics or on the wards by a senior consultant surgeon/gastroenterologist or registrar. The questionnaire was performed in the waiting room before colonoscopy and participation was voluntary. Remuneration was not provided and all data were deidentified when recorded. Written consent from patients participating in the study was obtained before performing the questionnaire.
Questionnaire This questionnaire was designed to evaluate patient knowledge of the procedure before the endoscopy and the patients’ preferences for the methods of explaining risk. The questionnaire consisted of 3 sections: demographics, medical history, and understanding colonoscopy. Demographic information was collected to assist in correlating patient knowledge and preferences with age, culture, and education level. Within the medical history section, indications, number of previous colonoscopies, and any comorbidities that could affect patients’ understanding of colonoscopy were recorded. The third section of the questionnaire evaluated participants’ understanding of colonoscopy. Participants were asked how much they knew about different aspects of the procedure and then how much detail they would prefer to know. Patients also indicated a preferred mode for risk communication and gave a preference to 1 of 5 distinct formats that explained the risk of perforation. The ease of understanding of each format was also recorded. A final question measured how patients perceived the risk of perforation after their procedure. The 5 risk communication formats chosen included 2 text and 3 graphical formats. The first 2 formats compared the risk of perforation to “everyday” risks using absolute and relative risk ratios (Figs. 1, 2). The pie graph (Fig. 3) and 1000-person pictograph (Fig. 4) used surface area to demonstrate the risk of perforation during colonoscopy when compared with the lifetime risk of colorectal cancer. The pictograph was created based on studies reporting patient preference for separately highlighted blocks (over individually highlighted dots)20 and the apparent faster, more accurate perception of horizontal pictographs.21 The final risk scale format (Fig. 5) used a logarithmic scale, useful in very rare risks (< 1 in 1000),22 to compare colon perforation and colorectal cancer risk.
Volume 24, Number 4, August 2014
Knowledge variables were measured by the amount of information a participant thought they knew about each aspect of the procedure. Total patient knowledge about colonoscopy was calculated as the average of all the various knowledge subvariables or aspects of the procedure measured, including: purpose of colonoscopy, procedural details, benefits, risks or complications, effectiveness, limitations or disadvantages, alternative investigations, sedation and anesthetics, preparation, and recovery. For preferred level of knowledge, participants were asked to indicate the amount of information they desired about the procedure, risks, benefits, and limitations. Participants were also asked to select one preferred mode of communication from: (1) An interactive computer program (this involved the patient viewing a computer-based presentation in the waiting room, completing a test of their understanding and receiving feedback). (2) Information leaflet (here, written information is handed to the patient in paper form and can be taken home afterwards). (3) Verbal information from a doctor or nurse. (4) No communication. (5) Other. The 5 risk communication formats were evaluated by participants for ease of understanding with patients asked for their overall preference. Finally, the participants were asked to indicate their subsequent perception of the risk of perforation.
Statistical Analysis Statistical analyses were performed using SPSS statistical software (version 17.0, SPSS Inc.). The ordinal 5-point scales for knowledge and understanding were assumed to be equally spaced and parametric so they could be converted to interval variables. This allowed for independent samples t tests when comparing means of binary categories and paired samples t tests when compared with variables with the same scale.
RESULTS Study Population A total of 45 patients were scheduled for colonoscopy during the study period. Of these, 9 did not complete the procedure and 4 chose not to participate. The remaining sample (n = 32) comprised of 17 men and 15 women, ranging in age from 21 to 73 (M = 48.3 y, SD = 14.8 y). Some participants chose not to complete all of the demographic or communication items and this was entered as missing data. The demographics of the sample are described in Table 1. Age data clustered around the 50 years point (kurtosis = 0.86). There was a good distribution of participants among across sex, country of birth, education level, and previous colonoscopy experience.
Patient Knowledge
FIGURE 1. Colonoscopic perforation risk compared to colorectal cancer and other adverse events.
346 | www.surgical-laparoscopy.com
Few participants (n = 2, 6.3%) wanted little or no information about the procedure, the rest (n = 30, 93.7%) preferred to know “somewhat” to every detail of information about colonoscopy, its risks, and benefits. When compared with participants’ preferred level of knowledge, information regarding: how colonoscopy is performed (P = 0.012); risks or complications (0.009); limitations or disadvantages (0.000); alternatives (0.000); sedation and r
2014 Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Evaluating Informed Consent for Colonoscopy
FIGURE 2. Colonoscopic perforation risk compared to colorectal cancer, assault and unexpected death.
anesthetics (0.003); and what happens after the procedure (0.011), were all significantly lower than preferred levels (paired samples t tests, 2-tailed significance) (Fig. 6). Furthermore, preferred overall knowledge was significantly higher (P = 0.002) than total knowledge about the procedure by + 0.66 ± 0.20 (mean difference ± SE) levels. Demographic correlations were rare; however, previous experience of colonoscopy correlated with significantly higher knowledge of what happens after the procedure (independent samples t test, 2-tailed P = 0.008, mean difference ± SE = + 1.25 ± 0.44 levels), but not with total knowledge level (independent samples t test, 2-tailed P = 0.213, mean difference ± SE = + 0.36 ± 0.28). Australian indigenous ethnicity significantly correlated with lower knowledge levels in several measures, and age accounted for 16.3% of the variability in knowledge of preparation (Table 2). These last 2 relationships may be affected by the small sample size of indigenous Australians (n = 2, 6.3%). There were no statistically significant relationships between total knowledge level and: sex, being born overseas, having a language other than English mainly spoken at home, level of education, age, and whether the colonoscopist was a colorectal surgeon or gastroenterologist (Table 3).
Preferred Communication Methods The most preferred mode of communication of information regarding colonoscopy was by information leaflet (n = 20, 62.5%), followed by verbal information from a doctor or nurse (n = 9, 28.1%). The preferred mode of communication had a statistically significant relationship with the level of education of participants (Fischer exact 2-tailed, P = 0.049). Those with higher levels of education preferred leaflet information, whereas those with
increasingly lower levels preferred verbal information from a doctor or nurse more (Fig. 7). When compared with absolute risk ratios (P = 0.008), relative risk ratios (0.000), and the risk scale (0.000), the pie graph was significantly easier to understand for the risk of perforation (paired samples t tests, 2-tailed significance) (Fig. 8). Similarly, the 1000-person pictograph was significantly easier to understand than relative risk ratios (P = 0.011) and the risk scale (P = 0.000). The most preferred format for explaining the risk of perforation was the pie graph (n = 17, 54.8%), followed by absolute risk ratios (n = 6, 19.4%) and the 1000-person pictograph (n = 6, 19.4%). There was a trend in which those with lower levels of education preferred the pie graph, whereas those with higher levels preferred the absolute risk ratios and risk palette, but this was not statistically significant (Fig. 9). The majority of participants perceived the risk of perforation as a “very low risk” (n = 18, 60%), followed by “almost no risk” ( = 8, 26.7%), and “low risk” (n = 4, 13.3%).
DISCUSSION In this study the majority of patients were aged between 40 and 60 years, reflecting the role of colonoscopy in screening and detection of colorectal cancer in this segment of the population. Demographic correlations were difficult to identify due to the small sample size of the study. Consent was obtained by consultant level staff rather than by a range of health workers from interns to attending physicians. Despite the seniority of the clinicians, gaps in patients’ knowledge and understanding still existed. Total knowledge levels about colonoscopy were lower than that preferred by patients (P = 0.002) but were consistent with other research suggesting inadequate information is given to patients.7–10 Information about the process of
FIGURE 3. Pie graphs of colorectal cancer risk and colonoscopic perforation risk. r
2014 Lippincott Williams & Wilkins
www.surgical-laparoscopy.com |
347
Gett et al
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
FIGURE 4. Thousand person pictograph of colon cancer risk.
colonoscopy, risks of the procedure, limitations, alternative options, sedation techniques, and recovery were identified as significant gaps in patient knowledge. This is of concern to clinicians given that specific “material risks” such as perforation and hemorrhage should be disclosed to avoid breaching duty of care requirements and to reduce medical negligence claims with approximately 40% of malpractice cases involving endoscopy relating to issues with consent.23,24 The marginal P value for the significant linear correlation between age and knowledge of preparation for colonoscopy suggests an error in inference due to multiple comparisons. A post hoc analysis of covariance using Tukey’s range test could perhaps better approximate the significance of this correlation. In contrast, the relationship between previous colonoscopy experience and increased knowledge about events after the procedure is far stronger and more logical (mean difference ± SE = + 1.25 ± 0.44 knowledge levels). One implication of this would involve explaining the outcomes for colorectal cancer screening
FIGURE 5. Relative risk of adverse events.
348 | www.surgical-laparoscopy.com
before the colonoscopy to improve patients’ knowledge. It is surprising that mean total knowledge level increased only 0.36 levels with previous experience of colonoscopy and did not reach statistical significance. This could be due to the small sample of the study, poor retention of colonoscopyrelated knowledge, or that patients are not increasing their understanding of the procedure despite undergoing endoscopy repeatedly. While conducting the interviews, 6 participants whose preferred mode of communication was via an information leaflet or verbal information from a doctor or nurse, TABLE 1. Sample Demographics
n = 32 Age (mean ± SD) (y) Sex M/F [n (%)] Country of birth [n (%)] Australia UK and Ireland Europe (including former USSR) NZ, Pacific Islands, PNG South East Asia Indian subcontinent and other Asia Other N/A Level of education [n (%)] rSecondary school qualification Certificate/trade Associate/bachelor’s degree Graduate degree N/A Previous colonoscopy [n (%)] Yes No N/A
r
48.3 ± 14.8 17 (53.1)/15 (46.9) 15 5 5 2 1 3 0 1
(46.9) (15.6) (15.6) (6.3) (3.1) (9.4) (0.0) (3.1)
6 8 12 4 2
(18.8) (25.0) (37.2) (12.5) (6.3)
17 (53.1) 13 (40.6) 2 (6.3)
2014 Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Evaluating Informed Consent for Colonoscopy
FIGURE 6. Patient knowledge of colonoscopy consent and procedure (means and standard deviations).
commented that they preferred to have both these forms together. This is consistent with a randomized Swiss25 trial that demonstrated combined written and oral information
led to more favorable quality assessments of the endoscopy information compared with oral information alone. It is advisable for doctors to provide procedure-specific leaflets
TABLE 2. Significant Correlations With Knowledge Level
Nature/Measure of Independent Variable (s) Categorical with 2 independent groups
Interval paired groups
Independent Variable(s) Previous colonoscopy experience Aboriginal or Torres Strait Islander Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge
Interval independent group
Age
Knowledge Variable(s) (Dependent) Knowledge of after the procedure
Tests
Independent samples t test Total level of knowledge (many Independent subvariables also significant) samples t test Knowledge of how Paired colonoscopy is performed samples t test Knowledge of risks or Paired complications samples t test Knowledge of limitations or Paired disadvantages samples t test Knowledge of alternatives Paired samples t test Knowledge of sedation Paired and anesthetics samples t test Knowledge of after the Paired procedure samples t test Total level of knowledge Paired of colonoscopy samples t test Knowledge of preparation Correlation Linear regression
Significance (2-tailed) 0.008 0.000 0.012 0.009 0.000 0.000 0.003 0.011 0.002 0.03 0.03
Scores Mean difference ± SE = + 1.25 ± 0.44 95% CIs, + 2.15, + 0.35 Mean difference ± SE = 1.61 ± 0.52 95% CIs, + 0.56, + 2.67 Mean difference ± SE = + 0.61 ± 0.23 95% CIs, + 0.15, + 1.08 Mean difference ± SE = + 0.742 ± 0.27 95% CIs, + 0.20, + 1.29 Mean difference ± SE = + 1.24 ± 0.25 95% CIs, + 0.73, + 1.75 Mean difference ± SE = + 1.98 ± 0.28 95% CIs, + 1.42, + 2.55 Mean difference ± SE = + 0.87 ± 0.27 95% CIs, + 0.33, + 1.42 Mean difference ± SE = + 0.74 ± 0.27 95% CIs, + 0.19, + 1.30 Mean difference ± SE = + 0.66 ± 0.20 95% CIs, + 0.26, + 1.06 Pearson correlation = 0.403 B score = 9.915; R2 = 0.163
CI indicates confidence interval.
r
2014 Lippincott Williams & Wilkins
www.surgical-laparoscopy.com |
349
Surg Laparosc Endosc Percutan Tech
Gett et al
TABLE 3. Nonsignificant Correlations With Total Knowledge Level
Variables Categorical with 2 independent groups Sex Born overseas English second language Proceduralist Previous experience of colonoscopy Ordinal with 2 or more independent groups Level of education Interval independent group Age Preferred level of knowledge
Tests
Independent samples t test Independent samples t test Independent samples t test Independent samples t test Independent samples t test
Significance (2-tailed)
0.923 0.664 0.095 0.900 0.213
One-way ANOVA
0.163
Correlation Linear regression Correlation
0.831 0.831 0.553
Linear regression
0.553
offering information about common and serious complications and then tailor their verbal communication to patient-specific risks or information. The positive correlation between lower education level and preference for verbal communication may be enough to encourage both flexibility with verbal information about colonoscopy and a greater sensitivity to the fact that patient preferences in the consent process can be affected by demographics. This pattern was also reflected in a Canadian cancer study demonstrating that younger patients and those with higher levels of education were more likely to prefer increased control over decision making.26
Volume 24, Number 4, August 2014
Both the pie graph and 1000-person pictographs received statistically significant preferences in terms of ease of understanding and overall popularity. Despite various studies advocating the more accurate overall impression and numerical knowledge of pictographs, the pie graph was preferred in this study to the 1000-person pictograph (n = 17, 54.8% and n = 6, 19.4%, respectively). This was especially the case at lower educational levels (Fig. 9). Schwartz et al27 recently found that less than half the patients in their study could name a single risk of colonoscopy and surprisingly, most had received information about colonoscopy from other sources rather than their doctor. Cleary, there is improvement required so that clinicians communicate the risks of colonoscopy to patients in a format that can be understood and recalled. In our study, the pie graph was preferred by patients with lower levels of education even though the overall preferred communication style was verbal. An opportunity therefore exists to use the pie graph prominently within colonoscopy information leaflets and then to complete informed consent with further tailored discussion with the patient. More research on patient preferences, perceptions, understanding, and interpretation is needed to justify the use of pictographs over other graphical formats such as bar and pie graphs. It is important that the preferences of researchers do not bias this process, as they will likely be consistent with those of a higher educational level and validate formats less accessible to those with lower levels of education. Given that health care risks are usually uncertain and complex, Paling22 suggests that doctors should be flexible in choosing from a “toolbox” of visual aids including depictions of absolute risks, risk scales, and pictographs. Finally, the highly clustered perception of the risk of perforation (n = 32, 100% participants felt perforation was “almost no risk,” “very low risk,” or “low risk”) should emphasize a widespread ability of patients to understand risk with the appropriate use of communication modes and formats. It is evident that colonic perforation, although rare, is a “material risk” for patients.
Limitations of Study and Future Directions An important limitation of this study is the small sample size. Even though some trends have been calculated as statistically insignificant, some inferences can be made in applying a normal distribution more relevant to a larger
FIGURE 7. Preferred mode of communication by education level.
350 | www.surgical-laparoscopy.com
FIGURE 8. Ease of understanding by explanation of risk. r
2014 Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Evaluating Informed Consent for Colonoscopy
after viewing videotape information may help clinicians in the future. The patients from our study still preferred the informative leaflet but many now feel that digital sources of information will become preferred in the near future. Thus, more information can be embedded in websites or provided digitally to improve informed consent. For Agre and colleagues, even though the demonstrated benefit was slim, there was nevertheless improvement. Perhaps colonoscopy patients in the future might have information leaflets, be asked to watch digital presentations, and then finally their understanding would be checked by a discussion with the doctor. Overall, while printed materials, diagrams, and audiovisual materials can be useful adjuncts to patient’s decision making, they are aides and not substitutes for doctor-patient discussion. In these discussions, the patients should be given enough time to consider information and the colonoscopist should solicit and answer questions.29
CONCLUSIONS
FIGURE 9. Preferred mode of risk explanation by education level.
sample size. Potential confounders in this study include the use of a single public hospital and only 3 different colonoscopists, although analysis of key demographics helped evaluate and control this confounder. Another consideration for the study is that patients in the waiting room for colonoscopy have undergone bowel preparation and can be disorientated or even distressed. It is also possible that the order of risk communication formats could make them successively easier to understand. This is unlikely, however, considering the low understanding score for the final risk scale format. In addition, the colonoscopists may have adjusted their consent process for the study itself. Nonetheless, important gaps in patient knowledge were still established. Finally, some doctors might only complete their whole consent process, immediately before sedation, in the colonoscopy procedure room itself. This practice could reduce patient knowledge as measured by the questionnaire. The similarly low levels of patient knowledge despite previous experience of colonoscopy, however, suggest that this effect is minimal. Future studies should determine those aspects of informed consent considered to be absolutely essential both to the patient and the busy clinician. With regard to colonoscopy, Schwartz and colleagues’ summary of fully informed consent under 3 headings: benefits, risks, and alternatives is useful. The main elements of risk being: perforation, bleeding, anesthetic risks, bowel preparation risks, and infection. Although this study has not specifically addressed the benefits or alternatives to colonoscopy, the discussion of risk could be improved with the use of diagrams such as the pie graph and then anticipating that some patients, particularly the less educated, may also require an informative, detailed yet simple discussion from the doctor.27 Agre et al’s28 research, which demonstrated some improvement in the understanding of consent-related issues r
2014 Lippincott Williams & Wilkins
In this small study, the patients received less information about colonoscopy then they preferred. Key areas for improvement include providing more understandable information about the risks of colonoscopy. Varying levels of education within a group of patients make it important for clinicians to tailor their explanation of risk to the patients. As education levels decrease and possibly, literacy is reduced; then more verbal information is often required rather than relying on the patient’s correct understanding of text, graphs, and diagrams. The pie graph is 1 effective way to communicate risk. It is the clinician’s role at times, to summarize the information and to present the facts in a simple way. However, clinicians cannot assume which form of communication suits all patients and therefore, a combination of written information, diagrams, and graphs and then a discussion of this information to check understanding is likely to be most effective. Further research into the communication of risk, with larger groups of patients, is likely to help clinicians in gaining fully informed consent in all patients. REFERENCES 1. Bourke J. Making every colonoscopy count: Ensuring quality in endoscopy. J Gastroenterol Hepatol. 2009;24:S43–S50. 2. Rathgaber SW, Wick TM. Colonoscopy completion and complication rates in a community gastroenterology practice. Gastrointest Endosc. 2006;64:556–562. 3. Viiala CH, Zimmerman M, Cullen DJE, et al. Complication rates of colonoscopy in an Australian teaching hospital environment. Intern Med J. 2003;33:355–359. 4. Segarajasingam DS, Pawlik J, Forbes GM. Informed consent in direct access colonoscopy. J Gastroenterol Hepatol. 2007; 22:2081–2085. 5. Seow CH, Leber JM, Ee Hc, et al. Survey of consent practices for inpatient colonoscopy and endoscopic retrograde cholangiopancreatography at a tertiary referral center. J Gastroenterol Hepatol. 2006;21:1340–1345. 6. Brooks AJ, Hurlstone DP, Fotheringham J, et al. Information required to provide informed consent for endoscopy: an observational study of patients’ expectations. Endoscopy. 2005; 37:1136–1139. 7. Macsween HM, Canadian Association of Gastroenterology. Informed Consent: Guidelines for Gastrointestinal Endoscopy. Can J Gastroenterol. 1997;11:533–534. 8. Newton J, Hawes R, Jamidar P, et al. Survey of informed consent for endoscopic retrograde cholangiopancreatography. Dig Dis Sci. 1994;39:1714–1718.
www.surgical-laparoscopy.com |
351
Gett et al
Surg Laparosc Endosc Percutan Tech
9. O’Sullivan S, Crippen C, Ponich T. Are patients informed when they consent to ERCP? Can J Gastroenterol. 2002; 16:154–158. 10. Triantafyllou K, Stanciu C, Kruse A, et al. Informed consent for gastrointestinal endoscopy: a 2002 ESGE survey. Dig Dis. 2002;20:280–283. 11. Strull WM, Lo B, Charles G. Do patients want to participate in medical decision making? J Am Med Assoc. 1984;252:2990–2994. 12. Edwards A, Thomas R, Williams R, et al. Presenting risk information to people with diabetes: evaluating effects and preferences for different formats by a web-based randomised controlled trial. Patient Educ Couns. 2006;63(3 SPEC. ISS.):336–349. 13. Waters EA, Weinstein ND, Colditz GA, et al. Formats for improving risk communication in medical tradeoff decisions. J Health Commun. 2006;11:167–182. 14. Cleveland WS, McGill R. Graphical perception: theory, experimentation, and application to the development of graphical methods. J Am Stat Assoc. 1984;79:531–554. 15. Lipkus IM, Hollands JG. The visual communication of risk. J Natl Cancer Inst Monogr. 1999;25:149–163. 16. Tait AR, Voepel-Lewis T, Zikmund-Fisher BJ., et al. The effect of format on parents’ understanding of the risks and benefits of clinical research: a comparison between text, tables, and graphics. J Health Commun. 2010;15:487–501. 17. Hawley ST, Zikmund-Fisher B, Ubel P, et al. The impact of the format of graphical presentation on health-related knowledge and treatment choices. Patient Educ Couns. 2008;73:448–455. 18. Burkell J. What are the chances? Evaluating risk and benefit information in consumer health materials. J Med Libr Assoc. 2004;92:200–208. 19. Fagerlin A, Wang C, Ubel PA. Reducing the influence of anecdotal reasoning on people’s health care decisions: is a
352 | www.surgical-laparoscopy.com
20. 21.
22. 23. 24. 25.
26. 27. 28. 29.
Volume 24, Number 4, August 2014
picture worth a thousand statistics? Med Decis Making. 2005; 25:398–405. Barratt A, Trevena LT, Davey H. Use of decision aids to support informed choices about screening. BMJ. 2004;329: 507–510. Price M, Cameron R, Butow P. Communicating risk information: the influence of graphical display format on quantitative information perception—accuracy, comprehension and preferences. Patient Educ Couns. 2007;69:121–128. Paling J. Strategies to help patients understand risks. BMJ. 2003;327:745–748. Levine EG, Brandt LJ, Plumeri P. Informed consent: a survey of physician outcomes and practices. Gastrointest Endosc. 1995;41:448–452. Gerstenberger PD, Plumeri PA. Malpractice claims in gastrointestinal endoscopy—analysis of an insurance industry database. Gastrointest Endosc. 1993;39:132–138. Felley C, Perneger TV, Goulet I, et al. Combined written and oral information prior to gastrointestinal endoscopy compared with oral information alone: a randomized trial. BMC Gastroenterol. 2008;8:22. Degner LF, Sloan JA. Decision making during serious illness: what role do patients really want to play? J Clin Epidemiol. 1992;45:941–950. Schwartz PH, Edenberg E, Barrett PR, et al. Patient understanding of benefits, risks, and alternatives to screening colonoscopy. Fam Med. 2013;45:83–89. Agre P, Kurtz RC, Krauss BJ. A randomized trial using videotape to present consent information for colonoscopy. Gastrointest Endosc. 1994;40:271–276. Zuckerman MJ, Shen B, Harrison ME, et al. Informed consent for GI endoscopy. Gastrointest Endosc. 2007;66:213–218.
r
2014 Lippincott Williams & Wilkins
Evaluating Informed Consent for Colonoscopy Rohan M. Gett, MBBS(Hon), BSc(MED), FRACS, MS,*w Anuk R. Cooray, MBBS(Hon),* Darren Gold, MSc, FRACS, FRCS(Gen), FRCS(Eng), MBBS,*w and Mark Danta, BMed, MD, FRACP*
Background: This study aimed to collect and analyze data on patient knowledge of colonoscopy and their preferences regarding the provision of information about the procedure. Specifically, how much detail patients know about different aspects of the procedure and through which methods they best understand risk are evaluated and demographic correlations identified. Materials and Methods: The study sample consisted of colonoscopy patients from 2 colorectal surgeons and a gastroenterologist at St Vincent’s Public Hospital, Sydney for the period August 1 to November 1, 2010. A voluntary questionnaire was performed in the waiting room before colonoscopy. The questionnaire collected data on patient demographics; patient-perceived knowledge of the procedure; and understanding and preferences of various communication formats. Results: Measures of patient-perceived knowledge about colonoscopy were significantly lower than those that would be preferred by patients (P = 0.002). Those with higher levels of education preferred communication of colonoscopy-related information via a leaflet form, whereas those with lower levels preferred verbal information from a doctor or nurse (P = 0.049). The most preferred format for explaining the risk of perforation was the pie graph, followed by both the 1000-person pictograph and absolute risk ratios. Conclusions: Patients received suboptimal levels of information about colonoscopy compared with their preferences. Key areas for improvement include providing more understandable information about the risks of colonoscopy. A combination of written information, diagrams and graphs, and then a discussion of this information to check the understanding is likely to be most effective. Further research into the communication of risk, with larger groups of patients, is likely to help clinicians in gaining fully informed consent in all patients. Key Words: colonoscopy, informed consent, patient knowledge, communicating risks
(Surg Laparosc Endosc Percutan Tech 2014;24:345–352)
I
n Australia, the number of colonoscopies continues to grow at a rate of approximately 8% a year.1 This growth in colonoscopy increases the pressure on doctors to diagnose more pathology and to provide adequate informed consent regarding the complications of colonoscopy. The Received for publication June 12, 2013; accepted January 26, 2014. From the *St Vincent’s Clinical School, University of New South Wales, NSW; and wDepartment of Surgery, St Vincent’s Hospital, Melbourne, Vic., Australia. This study received Human Research Ethics Committee Executive approval at St Vincent’s Hospital (File No. 10/052). The authors declare no conflicts of interest. Reprints: Rohan M. Gett, MBBS(Hon), BSc(MED), FRACS, MS, Suite 607 St Vincent’s Clinic, 438 Victoria St, Darlinghurst, NSW 2010, Australia (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
risk of perforation and hemorrhage in colonoscopy is low (0.02% to 0.2% and 0.05% to 1.0%,2,3 respectively), but requires disclosure and discussion because of the high rate of associated morbidity. An Australian study demonstrated that the majority of patients was either satisfied or very satisfied with the current informed consent process for colonoscopy,4 however, there is substantial variability in the information provided to patients5,6 and many colonoscopists do not provide adequate information to patients.7–10 A Western Australian audit,5 for consent practices in colonoscopy, identified that 12% of interns (doctors in their first postgraduate year) did not always mention perforation or hemorrhage and >50% of interns overestimated colonoscopic failure and perforation rates. These inadequacies in the consent process are potentially widespread and may make clinicians susceptible to litigation.5 Determining how much information is required for patients to give informed consent is difficult and there is a paucity of literature to assist in this process. Studies show that the information required to provide adequate consent is highly variable.6 Many patients, and indeed clinicians, judge the importance of information by the perceived severity of the complication rather than the frequency. Other patients prefer information about all or even none of the complications, irrespective of risk and severity. Colonoscopy patients, in particular, were less likely to want “no information” about complications than gastroscopy patients.6 One study found that professionals tended to underestimate health consumers’ desire for information, with about 40% of consumers wanting more information than they were given.11 This study aimed to compare the amount of information that patients receive compared with the amount they expect. Unfortunately, the effective communication of health risk information to patients is not frequently studied and thus, there is limited scientific evidence on how best to represent quantitative risk data.12,13 Most of the research focuses on the effect of different formats on accuracy and comprehension. Some studies have demonstrated that visual formats of risk, compared with textual or written presentations, assist in the perception, understanding, and interpretation of quantitative information.14,15 In particular, pictographs (eg, 1000 person dot/icon diagrams) have been associated with a more accurate overall impression of risk and numerical knowledge of that risk compared with text formats16 and graphical formats, including bar and pie graphs.17–19 Very little research examines the formats most preferred by patients in explaining the risks of colonoscopy. This study investigated which of following formats was most appropriate for explaining risks in colonoscopy with particular regard to colonoscopic perforation: absolute risk ratios, relative risk ratios, pie graphs, 1000-person pictographs, and logarithmic risk scales.
Volume 24, Number 4, August 2014
www.surgical-laparoscopy.com |
345
Surg Laparosc Endosc Percutan Tech
Gett et al
MATERIALS AND METHODS Participants The study sample consisted of colonoscopy patients from 2 colorectal surgeons and a gastroenterologist at St Vincent’s Public Hospital, Sydney, Australia for the period of August 1, 2010 to November 1, 2010. Forty-five participants were mailed information about the study by their gastroenterologist or colorectal surgeon. Information about the procedure itself was delivered in the usual way to participants through written information and preprocedural visits in clinics or on the wards by a senior consultant surgeon/gastroenterologist or registrar. The questionnaire was performed in the waiting room before colonoscopy and participation was voluntary. Remuneration was not provided and all data were deidentified when recorded. Written consent from patients participating in the study was obtained before performing the questionnaire.
Questionnaire This questionnaire was designed to evaluate patient knowledge of the procedure before the endoscopy and the patients’ preferences for the methods of explaining risk. The questionnaire consisted of 3 sections: demographics, medical history, and understanding colonoscopy. Demographic information was collected to assist in correlating patient knowledge and preferences with age, culture, and education level. Within the medical history section, indications, number of previous colonoscopies, and any comorbidities that could affect patients’ understanding of colonoscopy were recorded. The third section of the questionnaire evaluated participants’ understanding of colonoscopy. Participants were asked how much they knew about different aspects of the procedure and then how much detail they would prefer to know. Patients also indicated a preferred mode for risk communication and gave a preference to 1 of 5 distinct formats that explained the risk of perforation. The ease of understanding of each format was also recorded. A final question measured how patients perceived the risk of perforation after their procedure. The 5 risk communication formats chosen included 2 text and 3 graphical formats. The first 2 formats compared the risk of perforation to “everyday” risks using absolute and relative risk ratios (Figs. 1, 2). The pie graph (Fig. 3) and 1000-person pictograph (Fig. 4) used surface area to demonstrate the risk of perforation during colonoscopy when compared with the lifetime risk of colorectal cancer. The pictograph was created based on studies reporting patient preference for separately highlighted blocks (over individually highlighted dots)20 and the apparent faster, more accurate perception of horizontal pictographs.21 The final risk scale format (Fig. 5) used a logarithmic scale, useful in very rare risks (< 1 in 1000),22 to compare colon perforation and colorectal cancer risk.
Volume 24, Number 4, August 2014
Knowledge variables were measured by the amount of information a participant thought they knew about each aspect of the procedure. Total patient knowledge about colonoscopy was calculated as the average of all the various knowledge subvariables or aspects of the procedure measured, including: purpose of colonoscopy, procedural details, benefits, risks or complications, effectiveness, limitations or disadvantages, alternative investigations, sedation and anesthetics, preparation, and recovery. For preferred level of knowledge, participants were asked to indicate the amount of information they desired about the procedure, risks, benefits, and limitations. Participants were also asked to select one preferred mode of communication from: (1) An interactive computer program (this involved the patient viewing a computer-based presentation in the waiting room, completing a test of their understanding and receiving feedback). (2) Information leaflet (here, written information is handed to the patient in paper form and can be taken home afterwards). (3) Verbal information from a doctor or nurse. (4) No communication. (5) Other. The 5 risk communication formats were evaluated by participants for ease of understanding with patients asked for their overall preference. Finally, the participants were asked to indicate their subsequent perception of the risk of perforation.
Statistical Analysis Statistical analyses were performed using SPSS statistical software (version 17.0, SPSS Inc.). The ordinal 5-point scales for knowledge and understanding were assumed to be equally spaced and parametric so they could be converted to interval variables. This allowed for independent samples t tests when comparing means of binary categories and paired samples t tests when compared with variables with the same scale.
RESULTS Study Population A total of 45 patients were scheduled for colonoscopy during the study period. Of these, 9 did not complete the procedure and 4 chose not to participate. The remaining sample (n = 32) comprised of 17 men and 15 women, ranging in age from 21 to 73 (M = 48.3 y, SD = 14.8 y). Some participants chose not to complete all of the demographic or communication items and this was entered as missing data. The demographics of the sample are described in Table 1. Age data clustered around the 50 years point (kurtosis = 0.86). There was a good distribution of participants among across sex, country of birth, education level, and previous colonoscopy experience.
Patient Knowledge
FIGURE 1. Colonoscopic perforation risk compared to colorectal cancer and other adverse events.
346 | www.surgical-laparoscopy.com
Few participants (n = 2, 6.3%) wanted little or no information about the procedure, the rest (n = 30, 93.7%) preferred to know “somewhat” to every detail of information about colonoscopy, its risks, and benefits. When compared with participants’ preferred level of knowledge, information regarding: how colonoscopy is performed (P = 0.012); risks or complications (0.009); limitations or disadvantages (0.000); alternatives (0.000); sedation and r
2014 Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Evaluating Informed Consent for Colonoscopy
FIGURE 2. Colonoscopic perforation risk compared to colorectal cancer, assault and unexpected death.
anesthetics (0.003); and what happens after the procedure (0.011), were all significantly lower than preferred levels (paired samples t tests, 2-tailed significance) (Fig. 6). Furthermore, preferred overall knowledge was significantly higher (P = 0.002) than total knowledge about the procedure by + 0.66 ± 0.20 (mean difference ± SE) levels. Demographic correlations were rare; however, previous experience of colonoscopy correlated with significantly higher knowledge of what happens after the procedure (independent samples t test, 2-tailed P = 0.008, mean difference ± SE = + 1.25 ± 0.44 levels), but not with total knowledge level (independent samples t test, 2-tailed P = 0.213, mean difference ± SE = + 0.36 ± 0.28). Australian indigenous ethnicity significantly correlated with lower knowledge levels in several measures, and age accounted for 16.3% of the variability in knowledge of preparation (Table 2). These last 2 relationships may be affected by the small sample size of indigenous Australians (n = 2, 6.3%). There were no statistically significant relationships between total knowledge level and: sex, being born overseas, having a language other than English mainly spoken at home, level of education, age, and whether the colonoscopist was a colorectal surgeon or gastroenterologist (Table 3).
Preferred Communication Methods The most preferred mode of communication of information regarding colonoscopy was by information leaflet (n = 20, 62.5%), followed by verbal information from a doctor or nurse (n = 9, 28.1%). The preferred mode of communication had a statistically significant relationship with the level of education of participants (Fischer exact 2-tailed, P = 0.049). Those with higher levels of education preferred leaflet information, whereas those with
increasingly lower levels preferred verbal information from a doctor or nurse more (Fig. 7). When compared with absolute risk ratios (P = 0.008), relative risk ratios (0.000), and the risk scale (0.000), the pie graph was significantly easier to understand for the risk of perforation (paired samples t tests, 2-tailed significance) (Fig. 8). Similarly, the 1000-person pictograph was significantly easier to understand than relative risk ratios (P = 0.011) and the risk scale (P = 0.000). The most preferred format for explaining the risk of perforation was the pie graph (n = 17, 54.8%), followed by absolute risk ratios (n = 6, 19.4%) and the 1000-person pictograph (n = 6, 19.4%). There was a trend in which those with lower levels of education preferred the pie graph, whereas those with higher levels preferred the absolute risk ratios and risk palette, but this was not statistically significant (Fig. 9). The majority of participants perceived the risk of perforation as a “very low risk” (n = 18, 60%), followed by “almost no risk” ( = 8, 26.7%), and “low risk” (n = 4, 13.3%).
DISCUSSION In this study the majority of patients were aged between 40 and 60 years, reflecting the role of colonoscopy in screening and detection of colorectal cancer in this segment of the population. Demographic correlations were difficult to identify due to the small sample size of the study. Consent was obtained by consultant level staff rather than by a range of health workers from interns to attending physicians. Despite the seniority of the clinicians, gaps in patients’ knowledge and understanding still existed. Total knowledge levels about colonoscopy were lower than that preferred by patients (P = 0.002) but were consistent with other research suggesting inadequate information is given to patients.7–10 Information about the process of
FIGURE 3. Pie graphs of colorectal cancer risk and colonoscopic perforation risk. r
2014 Lippincott Williams & Wilkins
www.surgical-laparoscopy.com |
347
Gett et al
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
FIGURE 4. Thousand person pictograph of colon cancer risk.
colonoscopy, risks of the procedure, limitations, alternative options, sedation techniques, and recovery were identified as significant gaps in patient knowledge. This is of concern to clinicians given that specific “material risks” such as perforation and hemorrhage should be disclosed to avoid breaching duty of care requirements and to reduce medical negligence claims with approximately 40% of malpractice cases involving endoscopy relating to issues with consent.23,24 The marginal P value for the significant linear correlation between age and knowledge of preparation for colonoscopy suggests an error in inference due to multiple comparisons. A post hoc analysis of covariance using Tukey’s range test could perhaps better approximate the significance of this correlation. In contrast, the relationship between previous colonoscopy experience and increased knowledge about events after the procedure is far stronger and more logical (mean difference ± SE = + 1.25 ± 0.44 knowledge levels). One implication of this would involve explaining the outcomes for colorectal cancer screening
FIGURE 5. Relative risk of adverse events.
348 | www.surgical-laparoscopy.com
before the colonoscopy to improve patients’ knowledge. It is surprising that mean total knowledge level increased only 0.36 levels with previous experience of colonoscopy and did not reach statistical significance. This could be due to the small sample of the study, poor retention of colonoscopyrelated knowledge, or that patients are not increasing their understanding of the procedure despite undergoing endoscopy repeatedly. While conducting the interviews, 6 participants whose preferred mode of communication was via an information leaflet or verbal information from a doctor or nurse, TABLE 1. Sample Demographics
n = 32 Age (mean ± SD) (y) Sex M/F [n (%)] Country of birth [n (%)] Australia UK and Ireland Europe (including former USSR) NZ, Pacific Islands, PNG South East Asia Indian subcontinent and other Asia Other N/A Level of education [n (%)] rSecondary school qualification Certificate/trade Associate/bachelor’s degree Graduate degree N/A Previous colonoscopy [n (%)] Yes No N/A
r
48.3 ± 14.8 17 (53.1)/15 (46.9) 15 5 5 2 1 3 0 1
(46.9) (15.6) (15.6) (6.3) (3.1) (9.4) (0.0) (3.1)
6 8 12 4 2
(18.8) (25.0) (37.2) (12.5) (6.3)
17 (53.1) 13 (40.6) 2 (6.3)
2014 Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Evaluating Informed Consent for Colonoscopy
FIGURE 6. Patient knowledge of colonoscopy consent and procedure (means and standard deviations).
commented that they preferred to have both these forms together. This is consistent with a randomized Swiss25 trial that demonstrated combined written and oral information
led to more favorable quality assessments of the endoscopy information compared with oral information alone. It is advisable for doctors to provide procedure-specific leaflets
TABLE 2. Significant Correlations With Knowledge Level
Nature/Measure of Independent Variable (s) Categorical with 2 independent groups
Interval paired groups
Independent Variable(s) Previous colonoscopy experience Aboriginal or Torres Strait Islander Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge Preferred level of knowledge
Interval independent group
Age
Knowledge Variable(s) (Dependent) Knowledge of after the procedure
Tests
Independent samples t test Total level of knowledge (many Independent subvariables also significant) samples t test Knowledge of how Paired colonoscopy is performed samples t test Knowledge of risks or Paired complications samples t test Knowledge of limitations or Paired disadvantages samples t test Knowledge of alternatives Paired samples t test Knowledge of sedation Paired and anesthetics samples t test Knowledge of after the Paired procedure samples t test Total level of knowledge Paired of colonoscopy samples t test Knowledge of preparation Correlation Linear regression
Significance (2-tailed) 0.008 0.000 0.012 0.009 0.000 0.000 0.003 0.011 0.002 0.03 0.03
Scores Mean difference ± SE = + 1.25 ± 0.44 95% CIs, + 2.15, + 0.35 Mean difference ± SE = 1.61 ± 0.52 95% CIs, + 0.56, + 2.67 Mean difference ± SE = + 0.61 ± 0.23 95% CIs, + 0.15, + 1.08 Mean difference ± SE = + 0.742 ± 0.27 95% CIs, + 0.20, + 1.29 Mean difference ± SE = + 1.24 ± 0.25 95% CIs, + 0.73, + 1.75 Mean difference ± SE = + 1.98 ± 0.28 95% CIs, + 1.42, + 2.55 Mean difference ± SE = + 0.87 ± 0.27 95% CIs, + 0.33, + 1.42 Mean difference ± SE = + 0.74 ± 0.27 95% CIs, + 0.19, + 1.30 Mean difference ± SE = + 0.66 ± 0.20 95% CIs, + 0.26, + 1.06 Pearson correlation = 0.403 B score = 9.915; R2 = 0.163
CI indicates confidence interval.
r
2014 Lippincott Williams & Wilkins
www.surgical-laparoscopy.com |
349
Surg Laparosc Endosc Percutan Tech
Gett et al
TABLE 3. Nonsignificant Correlations With Total Knowledge Level
Variables Categorical with 2 independent groups Sex Born overseas English second language Proceduralist Previous experience of colonoscopy Ordinal with 2 or more independent groups Level of education Interval independent group Age Preferred level of knowledge
Tests
Independent samples t test Independent samples t test Independent samples t test Independent samples t test Independent samples t test
Significance (2-tailed)
0.923 0.664 0.095 0.900 0.213
One-way ANOVA
0.163
Correlation Linear regression Correlation
0.831 0.831 0.553
Linear regression
0.553
offering information about common and serious complications and then tailor their verbal communication to patient-specific risks or information. The positive correlation between lower education level and preference for verbal communication may be enough to encourage both flexibility with verbal information about colonoscopy and a greater sensitivity to the fact that patient preferences in the consent process can be affected by demographics. This pattern was also reflected in a Canadian cancer study demonstrating that younger patients and those with higher levels of education were more likely to prefer increased control over decision making.26
Volume 24, Number 4, August 2014
Both the pie graph and 1000-person pictographs received statistically significant preferences in terms of ease of understanding and overall popularity. Despite various studies advocating the more accurate overall impression and numerical knowledge of pictographs, the pie graph was preferred in this study to the 1000-person pictograph (n = 17, 54.8% and n = 6, 19.4%, respectively). This was especially the case at lower educational levels (Fig. 9). Schwartz et al27 recently found that less than half the patients in their study could name a single risk of colonoscopy and surprisingly, most had received information about colonoscopy from other sources rather than their doctor. Cleary, there is improvement required so that clinicians communicate the risks of colonoscopy to patients in a format that can be understood and recalled. In our study, the pie graph was preferred by patients with lower levels of education even though the overall preferred communication style was verbal. An opportunity therefore exists to use the pie graph prominently within colonoscopy information leaflets and then to complete informed consent with further tailored discussion with the patient. More research on patient preferences, perceptions, understanding, and interpretation is needed to justify the use of pictographs over other graphical formats such as bar and pie graphs. It is important that the preferences of researchers do not bias this process, as they will likely be consistent with those of a higher educational level and validate formats less accessible to those with lower levels of education. Given that health care risks are usually uncertain and complex, Paling22 suggests that doctors should be flexible in choosing from a “toolbox” of visual aids including depictions of absolute risks, risk scales, and pictographs. Finally, the highly clustered perception of the risk of perforation (n = 32, 100% participants felt perforation was “almost no risk,” “very low risk,” or “low risk”) should emphasize a widespread ability of patients to understand risk with the appropriate use of communication modes and formats. It is evident that colonic perforation, although rare, is a “material risk” for patients.
Limitations of Study and Future Directions An important limitation of this study is the small sample size. Even though some trends have been calculated as statistically insignificant, some inferences can be made in applying a normal distribution more relevant to a larger
FIGURE 7. Preferred mode of communication by education level.
350 | www.surgical-laparoscopy.com
FIGURE 8. Ease of understanding by explanation of risk. r
2014 Lippincott Williams & Wilkins
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Evaluating Informed Consent for Colonoscopy
after viewing videotape information may help clinicians in the future. The patients from our study still preferred the informative leaflet but many now feel that digital sources of information will become preferred in the near future. Thus, more information can be embedded in websites or provided digitally to improve informed consent. For Agre and colleagues, even though the demonstrated benefit was slim, there was nevertheless improvement. Perhaps colonoscopy patients in the future might have information leaflets, be asked to watch digital presentations, and then finally their understanding would be checked by a discussion with the doctor. Overall, while printed materials, diagrams, and audiovisual materials can be useful adjuncts to patient’s decision making, they are aides and not substitutes for doctor-patient discussion. In these discussions, the patients should be given enough time to consider information and the colonoscopist should solicit and answer questions.29
CONCLUSIONS
FIGURE 9. Preferred mode of risk explanation by education level.
sample size. Potential confounders in this study include the use of a single public hospital and only 3 different colonoscopists, although analysis of key demographics helped evaluate and control this confounder. Another consideration for the study is that patients in the waiting room for colonoscopy have undergone bowel preparation and can be disorientated or even distressed. It is also possible that the order of risk communication formats could make them successively easier to understand. This is unlikely, however, considering the low understanding score for the final risk scale format. In addition, the colonoscopists may have adjusted their consent process for the study itself. Nonetheless, important gaps in patient knowledge were still established. Finally, some doctors might only complete their whole consent process, immediately before sedation, in the colonoscopy procedure room itself. This practice could reduce patient knowledge as measured by the questionnaire. The similarly low levels of patient knowledge despite previous experience of colonoscopy, however, suggest that this effect is minimal. Future studies should determine those aspects of informed consent considered to be absolutely essential both to the patient and the busy clinician. With regard to colonoscopy, Schwartz and colleagues’ summary of fully informed consent under 3 headings: benefits, risks, and alternatives is useful. The main elements of risk being: perforation, bleeding, anesthetic risks, bowel preparation risks, and infection. Although this study has not specifically addressed the benefits or alternatives to colonoscopy, the discussion of risk could be improved with the use of diagrams such as the pie graph and then anticipating that some patients, particularly the less educated, may also require an informative, detailed yet simple discussion from the doctor.27 Agre et al’s28 research, which demonstrated some improvement in the understanding of consent-related issues r
2014 Lippincott Williams & Wilkins
In this small study, the patients received less information about colonoscopy then they preferred. Key areas for improvement include providing more understandable information about the risks of colonoscopy. Varying levels of education within a group of patients make it important for clinicians to tailor their explanation of risk to the patients. As education levels decrease and possibly, literacy is reduced; then more verbal information is often required rather than relying on the patient’s correct understanding of text, graphs, and diagrams. The pie graph is 1 effective way to communicate risk. It is the clinician’s role at times, to summarize the information and to present the facts in a simple way. However, clinicians cannot assume which form of communication suits all patients and therefore, a combination of written information, diagrams, and graphs and then a discussion of this information to check understanding is likely to be most effective. Further research into the communication of risk, with larger groups of patients, is likely to help clinicians in gaining fully informed consent in all patients. REFERENCES 1. Bourke J. Making every colonoscopy count: Ensuring quality in endoscopy. J Gastroenterol Hepatol. 2009;24:S43–S50. 2. Rathgaber SW, Wick TM. Colonoscopy completion and complication rates in a community gastroenterology practice. Gastrointest Endosc. 2006;64:556–562. 3. Viiala CH, Zimmerman M, Cullen DJE, et al. Complication rates of colonoscopy in an Australian teaching hospital environment. Intern Med J. 2003;33:355–359. 4. Segarajasingam DS, Pawlik J, Forbes GM. Informed consent in direct access colonoscopy. J Gastroenterol Hepatol. 2007; 22:2081–2085. 5. Seow CH, Leber JM, Ee Hc, et al. Survey of consent practices for inpatient colonoscopy and endoscopic retrograde cholangiopancreatography at a tertiary referral center. J Gastroenterol Hepatol. 2006;21:1340–1345. 6. Brooks AJ, Hurlstone DP, Fotheringham J, et al. Information required to provide informed consent for endoscopy: an observational study of patients’ expectations. Endoscopy. 2005; 37:1136–1139. 7. Macsween HM, Canadian Association of Gastroenterology. Informed Consent: Guidelines for Gastrointestinal Endoscopy. Can J Gastroenterol. 1997;11:533–534. 8. Newton J, Hawes R, Jamidar P, et al. Survey of informed consent for endoscopic retrograde cholangiopancreatography. Dig Dis Sci. 1994;39:1714–1718.
www.surgical-laparoscopy.com |
351
Gett et al
Surg Laparosc Endosc Percutan Tech
9. O’Sullivan S, Crippen C, Ponich T. Are patients informed when they consent to ERCP? Can J Gastroenterol. 2002; 16:154–158. 10. Triantafyllou K, Stanciu C, Kruse A, et al. Informed consent for gastrointestinal endoscopy: a 2002 ESGE survey. Dig Dis. 2002;20:280–283. 11. Strull WM, Lo B, Charles G. Do patients want to participate in medical decision making? J Am Med Assoc. 1984;252:2990–2994. 12. Edwards A, Thomas R, Williams R, et al. Presenting risk information to people with diabetes: evaluating effects and preferences for different formats by a web-based randomised controlled trial. Patient Educ Couns. 2006;63(3 SPEC. ISS.):336–349. 13. Waters EA, Weinstein ND, Colditz GA, et al. Formats for improving risk communication in medical tradeoff decisions. J Health Commun. 2006;11:167–182. 14. Cleveland WS, McGill R. Graphical perception: theory, experimentation, and application to the development of graphical methods. J Am Stat Assoc. 1984;79:531–554. 15. Lipkus IM, Hollands JG. The visual communication of risk. J Natl Cancer Inst Monogr. 1999;25:149–163. 16. Tait AR, Voepel-Lewis T, Zikmund-Fisher BJ., et al. The effect of format on parents’ understanding of the risks and benefits of clinical research: a comparison between text, tables, and graphics. J Health Commun. 2010;15:487–501. 17. Hawley ST, Zikmund-Fisher B, Ubel P, et al. The impact of the format of graphical presentation on health-related knowledge and treatment choices. Patient Educ Couns. 2008;73:448–455. 18. Burkell J. What are the chances? Evaluating risk and benefit information in consumer health materials. J Med Libr Assoc. 2004;92:200–208. 19. Fagerlin A, Wang C, Ubel PA. Reducing the influence of anecdotal reasoning on people’s health care decisions: is a
352 | www.surgical-laparoscopy.com
20. 21.
22. 23. 24. 25.
26. 27. 28. 29.
Volume 24, Number 4, August 2014
picture worth a thousand statistics? Med Decis Making. 2005; 25:398–405. Barratt A, Trevena LT, Davey H. Use of decision aids to support informed choices about screening. BMJ. 2004;329: 507–510. Price M, Cameron R, Butow P. Communicating risk information: the influence of graphical display format on quantitative information perception—accuracy, comprehension and preferences. Patient Educ Couns. 2007;69:121–128. Paling J. Strategies to help patients understand risks. BMJ. 2003;327:745–748. Levine EG, Brandt LJ, Plumeri P. Informed consent: a survey of physician outcomes and practices. Gastrointest Endosc. 1995;41:448–452. Gerstenberger PD, Plumeri PA. Malpractice claims in gastrointestinal endoscopy—analysis of an insurance industry database. Gastrointest Endosc. 1993;39:132–138. Felley C, Perneger TV, Goulet I, et al. Combined written and oral information prior to gastrointestinal endoscopy compared with oral information alone: a randomized trial. BMC Gastroenterol. 2008;8:22. Degner LF, Sloan JA. Decision making during serious illness: what role do patients really want to play? J Clin Epidemiol. 1992;45:941–950. Schwartz PH, Edenberg E, Barrett PR, et al. Patient understanding of benefits, risks, and alternatives to screening colonoscopy. Fam Med. 2013;45:83–89. Agre P, Kurtz RC, Krauss BJ. A randomized trial using videotape to present consent information for colonoscopy. Gastrointest Endosc. 1994;40:271–276. Zuckerman MJ, Shen B, Harrison ME, et al. Informed consent for GI endoscopy. Gastrointest Endosc. 2007;66:213–218.
r
2014 Lippincott Williams & Wilkins
