Interventions to Improve Patient Comprehension in Informed Consent for Medical and Surgical Procedures: A Systematic Review

Abstract

Background

Patient understanding in clinical informed consent is often poor. Little is known about the effectiveness of interventions to improve comprehension or the extent to which such interventions address different elements of understanding in informed consent.

Purpose

To systematically review communication interventions to improve patient comprehension in informed consent for medical and surgical procedures.

Data Sources

A systematic literature search of English-language articles in MEDLINE (1949–2008) and EMBASE (1974–2008) was performed. In addition, a published bibliography of empirical research on informed consent and the reference lists of all eligible studies were reviewed.

Study Selection

Randomized controlled trials and controlled trials with non-random allocation were included if they compared comprehension in informed consent for a medical or surgical procedure. Only studies that used a quantitative, objective measure of understanding were included. All studies addressed informed consent for a needed or recommended procedure in actual patients.

Data Extraction

Reviewers independently extracted data using a standardized form. All results were compared, and disagreements were resolved by consensus.

Data Synthesis

Forty-four studies were eligible. Intervention categories included written information, audiovisual/multimedia, extended discussions, and test/feedback techniques. The majority of studies assessed patient understanding of procedural risks; other elements included benefits, alternatives, and general knowledge about the procedure. Only 6 of 44 studies assessed all 4 elements of understanding. Interventions were generally effective in improving patient comprehension, especially regarding risks and general knowledge.

Limitations

Many studies failed to include adequate description of the study population, and outcome measures varied widely.

Conclusions

A wide range of communication interventions improve comprehension in clinical informed consent. Decisions to enhance informed consent should consider the importance of different elements of understanding, beyond procedural risks, as well as feasibility and acceptability of the intervention to clinicians and patients. Conceptual clarity regarding the key elements of informed consent knowledge will help to focus improvements and standardize evaluations.

Informed consent is legally and ethically required before proceeding with invasive or high-risk clinical procedures.^1,2 Defined as the “process of communication between a patient and physician that results in the patient's authorization or agreement to undergo a specific medical intervention,”² valid informed consent requires patient understanding of the proposed intervention, including potential risks, benefits, and alternatives.³ Failure to obtain adequate informed consent compromises patient autonomy,^4,5 places patient safety at risk,^6–8 and legally may constitute negligence or battery.^4,9

Despite its critical importance to the provision of safe, high-quality, patient-centered health care, the process of informed consent in clinical practice is frequently inadequate,^3,10 and prior research has demonstrated that patient comprehension of the key elements of clinical informed consent is often poor.^11,12 Physicians receive little training in how to conduct informed consent discussions.¹³ Misunderstandings about consent requirements and goals,¹⁴ differing legal standards for informed consent disclosure,⁵ and the time pressures and competing demands of clinical medicine may also hinder the informed consent process. Many consent forms do not contain the key elements of informed consent¹⁵ or are written in a language too complex for many patients to understand.^16–18 Patients who do not speak English or have limited literacy are at increased risk for poor comprehension.^19,20

Although some argue that it is unrealistic to expect full patient understanding in informed consent,^21,22 organizations including the American Medical Association, the Agency for Healthcare Research and Quality (AHRQ), and the National Quality Forum have called for improvements in the informed consent process.^3,7,8 Interventions including revised consent forms, provision of additional written information, enhanced consent discussions, and the use of audiovisual technology have been developed and evaluated. However, to our knowledge, no study has systematically assessed interventions to improve informed consent for medical and surgical procedures. Previous reviews have addressed informed consent for research²³ or focused on older subjects²⁴ or psychiatric illness.²⁵ We conducted a systematic review to 1) identify and characterize the quality of trials of communication interventions to increase patient understanding in informed consent for medical and surgical procedures, 2) describe the key features of interventions and assess their relative effectiveness, 3) evaluate the elements of informed consent comprehension assessed, and, based on a synthesis of this literature, 4) provide recommendations for practice and future research.

Methods

Study Eligibility

To capture a broad range of studies, this review included both randomized controlled trials and controlled trials with nonrandom allocation published in English. Trials were included if they compared a standard informed consent process for a high-risk and/or invasive medical or surgical procedure with an enhanced process designed to improve comprehension. We included only studies that used a quantitative, objective measure to assess comprehension or recall of key elements of informed consent in both study arms and contained sufficient detail to characterize the instrument used. In addition, all studies addressed informed consent for a needed or recommended intervention in actual patients. Trials evaluating informed consent for research; trials evaluating a consent process for screening tests, educational programs, advanced directives, psychotherapy, prescription drugs, or cancer consultations; trials in which parents or surrogates obtained informed consent; trials of informed consent for contingency scenarios (i.e., a procedure that may be necessary in the future); and trials comparing different timing of informed consent were excluded.

Study Identification

We conducted a systematic literature search of MEDLINE (1949 to November 2008) and EMBASE (1974 to November 2008) for relevant articles. The search strategy was designed by a senior reference librarian with experience in systematic reviews. Search terms included informed consent, consent forms, comprehension, mental recall, patient education as topic, video recording, and tape recording. Web Appendix 1 describes the complete search strategy. In addition, we hand searched a published bibliography of empirical research on informed con-sent²⁶ and the reference lists of all identified relevant studies.

Study Selection

One reviewer (Y.S.) performed an initial screen by title and abstract to identify articles eligible for further review. All potentially relevant articles were retrieved in full text. A 2nd screen was then performed based on full-text review. Articles for which eligibility was not clear were reviewed by the study team (Y.S., A.F., R.S., D.S.), and inclusion/exclusion status was determined by consensus.

Data Extraction and Synthesis

We developed a standardized data abstraction form with input from all members of the study team (Web Appendix 2). Three reviewers (Y.S., R.S., A.F.) performed data abstraction, with 2 reviewers independently abstracting data from each study. Results were compared and agreed upon for all studies, with less than 5% of abstracted data requiring consensus.

Types of interventions and outcome measures

Studies were grouped by the type of communication intervention (written, audiovisual/multimedia, extended discussion, or test/feedback). In addition, we noted the elements of informed consent comprehension assessed (risks, benefits/indications, alternatives, and general knowledge), the timing of comprehension assessments relative to the informed consent process (immediately or delayed) and relative to the procedure (before, after, or both), and the differences in comprehension scores between study groups.

Study design

Studies were classified as randomized controlled trials if they were described as randomized using an appropriate method to generate the sequence of randomization (i.e., table of random numbers, computer generated). Studies described as randomized but with no description of the method used to generate randomization, or with a description of a nonstandard method of randomization (i.e., patients allocated alternately or according to hospital number), were classified as randomized controlled trials only if there was evidence of comparability of patient characteristics between groups. Studies without evidence of comparability between groups and studies with quasi-experimental designs (i.e., patients enrolled before implementation of the intervention v. similar patients enrolled after implementation of the intervention) were classified as nonrandomized controlled trials.

Quality of reporting

We initially calculated a Jadad score for each study but found this to be an inadequate quality measure for this group of studies because none were double blind.²⁷ We therefore created a structured 5-item rating tool based on domains recommended in an AHRQ evidence report²⁸ (Web Appendix 3). Specifically, we graded each eligible study (N=44) according to the adequacy of description of the study population, randomization, description of the informed consent intervention, outcome measurement, and results. Of note, this rating was based mainly on the quality of reporting and not the risk of bias. A subsample of 21 studies was graded by 2 investigators (Y.S. and A.F. or R.S.) to refine the quality criteria. All disagreements were resolved by consensus. Once we had ensured consistency in the quality rating system, the remaining studies were graded by 1 investigator (Y.S.).

Each of the 5 quality elements was rated as sufficient or insufficient. We then converted ratings for each element into a numerical value (sufficient=1 and insufficient=0) and created a composite quality rating score (range, 0–5) that gave each element equal weight. Articles with a quality rating score of 0 to 2 were considered poor quality, those with a score of 3 were considered fair quality, and those with a score of 4 to 5 were considered good quality. The quality rating score allowed us to observe whether assessing results from all studies and assessing results from only good-quality studies revealed similar patterns of benefit.

Results

Literature Search

The initial literature search identified 2083 citations: 1308 from MEDLINE and 775 from EMBASE (Figure 1). Of these, 1994 articles were excluded based on abstract or title, leaving 89 articles for full-text review. Hand searching of a published bibliography on informed consent²⁶ and the reference lists of all included studies identified an additional 33 studies, for a total of 122 articles for full-text review. Of these, 78 articles were excluded (15 insufficient detail or no objective measure of understanding, 14 not original research, 12 intervention for which informed consent not routinely sought, 12 no control group, 7 volunteers or contingency scenario, 6 informed consent for research, 5 compared different timing of informed consent, 3 parents consenting for minors, 2 not published in English, 1 duplicate article, 1 published abstract), leaving 44 studies that met our inclusion/exclusion criteria to be included in this systematic review.

Figure 1. Flowchart of systematic review.

Study Characteristics

Of the 44 studies, 23 addressed informed consent for medical procedures and 21 for surgical procedures (Tables 1–4). Medical procedures included anesthesia, radiographic imaging with intravenous contrast, chemotherapy, endoscopic procedures (colonoscopy, esophagogastroduodenoscopy, and flexible sigmoidoscopy), electroconvulsive therapy, and cardiac catheterization. Surgical procedures included cataract surgery; head and neck procedures; plastic surgery; orthopedic surgeries; intrathoracic, intraperitoneal, and arterial surgeries; and gynecologic procedures. More than half of the studies took place in the inpatient setting. Twenty trials were conducted in the United States; the remainder took place in the United Kingdom (12 trials), Australia (7), Canada (4), and New Zealand (1). The number of participants enrolled ranged from 18 participants²⁹ to 269 participants.³⁰ Overall, the amount of demographic information reported was poor. Studies also varied in the amount of demographic information they provided; 20 of the 44 studies did not specify the educational level of participants. Twenty-three of the 44 studies were randomized controlled trials.

Table 1. Results of Trials of Written Interventions.

Source	Procedure	Study Population		Intervention	Study Design	Outcome Measure				Results
		Age and Education	N			Knowledgea	Timing Relative to ICb	Timing Relative to Procedurec	Blindedd
Good-quality studies
Langdon and others (2002)41	Total hip arthroplasty	Mean age = 68 y; education NS	126	Illustrated sheet describing risks and general info using nonmedical terms in addition to structured verbal discussion	RCT	R, G	D	Before	X	Written information group had higher mean knowledge scores (48% v. 38%, P =0.0044)
Chan and others (2002)34	Thyroidectomy and parathyroidectomy	Mean age = 47 y; 63% with postsecondary education	125	Illustrated pamphlet describing risks in addition to standard verbal information from surgeon	RCTe	R	D	NS		Intervention group recalled more potential complications (50% v30%, P< 0.001)
Brown and others (2003)33	Otologic surgery	Mean age = 43 y; 48% with postsecondary education	50	Procedure-specific written handout describing risks in addition to standard verbal discussion	RCT	R	D	NS		No difference between handout and no handout groups in recall of complications (51% v. 56%, P=0.30)
Inglis and Farnill (1993)39	General anesthesia for surgery	Age range = 21-80 y; mean education = 10 y	40	More detailed information including numerical incidence of anesthesia risks v. routine information, both presented via oral and written script	RCTe	R	I	Before		Detailed information group more likely to correctly identify 2 out of 6 risks of anesthesia (P< 0.05)
Fair-quality studies
Armstrong and others (1997)30	Cosmetic surgery, elective hand surgery, and excision of minor skin tumors	Mean age = 44 y; 13% with A level or higher	269	Written sheet listing risks of procedure in addition to standard counseling with plastic surgeon	RCTe	R	D	After	X	Mean number risks recalled 1 wk after surgery higher in written group (score/7) (3.64 v. 2.95, P= 0.006)
Livesley and Rider (1993)43	Total hip and total knee replacement	Age NS; education NS	180	Booklet covering basic information about operation given 2 wk prior to surgery	NRCT	R, B, G	D	After		Booklet group more likely to give correct responses on 5 out of 9 knowledge questions (P = 0.001 to 0.05)
Courtney (1997)36	Anesthesia	Age NS; education NS	140	Brief anesthetic booklet (content not specified) sent in the mail	NRCT	R, G	I/D	Both		Booklet group had more correct responses to 2 out of 6 knowledge questions immediately after IC (P < 0.025 and P< 0.05); no difference between groups after surgery
Makdessian and others (2004)44	Rhinoplasty, face-lift surgery, or laser resurfacing	Mean age = 41 y; 45% university or higher	120	Pamphlet outlining risks of surgery in addition to standard oral consultation	RCT	R	D	NS		Pamphlet group recalled more risks of surgery (2.5 v. 1.5 of 5 risks; P< 0.001)
Villar and Hume (1988)48	Total hip replacement	Age NS; education NS	100	4-page information booklet about total hip replacement (content not specified) given on day of admission	NRCT	R, G	D	After		Booklet group more likely to give correct response on 1 out of 9 knowledge questions (P < 0.05)
Shurnas and Coughlin (2003)47	Forefoot surgery	Mean age = 51/53 y; 63% with postsecondary education	38	Patients were given a copy of the list of risks used to guide the IC discussion, which they read and signed	RCTe	R	D	After		No difference between groups in recall of risks 12 wk after surgery
Poor-quality studies
Lavelle-Jones and others (1993)12	Intrathoracic, intraperitoneal, and arterial surgeries	Median age = 61 y; education NS	265	Written operation information cards describing nature of operation, expected recovery, side effects, and risks in addition to standard discussion	NRCT	R, B, G	I/D	Both		Written information group had higher knowledge scores on day of discharge (4 v. 3 out of 6, P =0.015); however, no difference between groups immediately after IC or at postdischarge follow-up appointments
Clark and others (1991)35	Anesthesia	Age NS; education NS	233	Risk-specific anesthesia consent form in addition to standard preoperative discussion with anesthesia resident	NRCT	R	D	After		Intervention group recalled fewer risks of anesthesia (19% v. 33%; P<0.01)
Layton and Korsen (1994)42	Surgical removal of lower 3rd-molar teeth	Age NS; education NS	194	Written information sheet describing risks of surgery in addition to standard oral discussion	NRCT	R	D	After		Written information group had better recall of 2 out of 5 risks of the procedure (P<0.01 and P< 0.001)
Graham (2003)38	Breast radiotherapy	Mean age = 59 y; 36% college or equivalent	174	General written consent and toxicity-specific written consent compared with verbal consent only	NRCT	R	D	After		No clear association between consent type and recall of toxicities
Gerancher and others (2000)37	Epidural analgesia for labor and delivery	Age NS; education NS	113	Procedure-specific informed consent form reviewed and signed in addition to standard preanesthetic interview	NRCT	R, A, G	D	After	X	Intervention group had higher median knowledge scores 5–7 mo after preanesthetic discussion (90% v. 80%, P< 0.001)
Ashraff and others (2006)31	Elective hip replacement, knee replacement, knee arthroscopy, and shoulder surgery	Mean age = 66 y; education NS	110	Information leaflet describing risks and postoperative care sent 2 wk before the operation in addition to standard verbal information	RCT	R, G	D	Before		Leaflet group had higher median comprehension scores (40% v. 10%, P< 0.0001)
Askew and others (1990)32	Inguinal hernia repair, cholecystectomy, truncal vagotomy, and pyloroplasty	Age range = 23–82 y; education NS	100	Information sheet describing operation sent 2 wk before the operation in addition to standard IC	NRCT	B,G	D	After		Informed group more knowledgeable about details and indications for the operation (97% v. 69%, P< 0.001)
Winfield and others (1986)49	Excretory urography	Age range = 19-68 y; education NS	80	Procedure-specific IC form in addition to standard interview	NRCT	R, G	D	After		Group given IC form had higher knowledge score (73% v. 48%, P<0.01)
Jenaway (1993)40	Electroconvulsive therapy	Mean age = 57 y; education NS	57	Leaflet with basic information about electroconvulsive therapy	NRCT	G	D	Both		Leaflet group had greater knowledge (score/11) than control group before (3.1 v. 2.3; P= 0.03) and after (3.0 v. 1.9; P =0.002) treatment
Pesudovs and others (2006)45	Cataract surgery	Mean age = 70 y; education NS	50	Take-home copy of information sheet used as part of presurgical consult explaining risks/benefits of surgery	NRCT	R, B, A, G	I/D	Both		No difference between groups in knowledge at preop or postop assessment
Scanlan and others (2003)46	Cataract surgery	Mean age = 68 y; 55% grade 12 or higher	28	Take-home brochure with information about surgery in addition to standard IC discussion	NRCT	R, B, A, G	I/D	Both		Brochure group had less deterioration in recall from preop to postop assessment (75% to 64%, P =0.152) than controls (61% to 44%; P =0.009)

Note: NS = not specified; IC = informed consent; RCT = randomized controlled trial; NRCT = nonrandomized controlled trial.

^aIndicates elements of IC knowledge measured by the instrument (R = risks; B = benefits/indications; A = alternatives; G = general/other knowledge about procedure or not specified).

^bIndicates timing of the measurement of understanding relative to informed consent (I = immediately after informed consent; D = delayed; NS = not specified).

^cIndicates timing of the measurement of understanding relative to the medical/surgical procedure (before, after, both, or NS = not specified).

^dIndicates whether the measurement of understanding was described as blinded.

^eNo description of randomization method or described a nonstandard randomization method but with evidence of comparability between patient groups.

Table 4. Results of Trials of Test/Feedback Techniques.

Source	Procedure	Study Population		Intervention	Study Design	Outcome Measure				Results
		Age and Education	N			Knowledgea	Timing Relative to ICb	Timing Relative to Procedurec	Blindedd
Good-quality studies
Greening and others (1999)69	Electroconvulsive therapy	Mean age = 55 y; education NS	32	Patients told 10 basic items about ECT and asked to repeat back for total 3 trials or until all info recalled	RCT	G	NS/D	Both	X	Structured consent group recalled more items (score/10) when tested before ECT (mean 9.4 v. 6.4; P = 0.04); nonsignificant difference between groups after ECT (mean 6.9 v. 4.6; P= 0.11).
Fair-quality studies
White and others (1995)71	Percutaneous lung biopsy	Age range = 18–80 y; education NS	50	Patients required to repeat back 4 complications; repeated until patient could describe all 4	NRCT	R	D	After		Test/feedback group recalled more risks (score/4; mean 2.7 v. 1.6; P= 0.005)
Poor-quality studies
Wadey and Frank (1997)70	Anterior cruciate ligament reconstruction	Age NS; education NS	20	Patients required to repeat back risks and benefits described in surgical consultation and corrected until accurate	NRCT	R, B	D	NS		All patients (AT = 8) in the test/feedback group answered 3/3 questions correctly; in the standard group (N= 12), 4 patients answered 3/3 correctly, 7 patients answered 2/3 correctly, and 1 patient answered 0/3 correctly (P<0.03)

Note: ECT = electroconvulsive therapy; NS, not specified; IC = informed consent; RCT = randomized controlled trial; NRCT = nonrandomized controlled trial.

^aIndicates elements of IC knowledge measured by the instrument (R = risks; B = benefits/indications; A = alternatives; G = general/other knowledge about procedure or not specified).

^bIndicates timing of the measurement of understanding relative to informed consent (I = immediately after informed consent; D = delayed; NS = not specified).

^cIndicates timing of the measurement of understanding relative to the medical/surgical procedure (before, after, both, or NS = not specified).

^dIndicates whether the measurement of understanding was described as blinded.

Quality of Reporting

Overall, we rated 18 studies as good, 10 studies as fair, and 16 studies as poor quality (Tables 1–4). Common limitations included inadequate description of the study population, inadequate description of the method used to randomize or lack of randomization, inadequate description of the outcome measure, non-blinded outcome measures, and incomplete reporting of results.

Types of Interventions

Studies primarily tested 1 of 4 different types of informed consent interventions: 1) additional written information, 2) audiovisual/multimedia interventions, 3) extended informed consent discussions, and 4) test/feedback techniques.

Written interventions

Twenty-one studies evaluated written interventions (Table 1).^12,30–49 Written interventions included the use of a consent form with additional information specific to a single procedure^35,37,38,49 and the provision of additional written information ranging from a brief operation information card¹² to a 4-page information booklet.⁴⁸ The content of written interventions included additional information about the risks of the procedure in 16 of 21 studies^{12,30,31,33–35,37–39,41,42,44–47,49} and m ore general or unspecified information about the procedure in the remaining 5 studies.^{32,36,40,43,48} The majority of written interventions were provided in person at a preoperative visit or at the time of admission. Three studies evaluated information sheets mailed to patients prior to the procedure.^31,32,36 Only 9 of 21 studies included a copy of the written intervention they evaluated,^{12,32,34,37,39,45–47,49} and the reading level of the intervention was specified in only 2 studies.^39,45

The majority (16 of 21 studies) showed improvement in comprehension with the addition of written information. ^{12,30–32,34,36,37,39–44,46,48,49} However the degree to which comprehension improved varied between studies (Table 1) and was sometimes quite small. For example, in one study, the group receiving additional written information had higher scores on only 1 out of 9 knowledge questions asked.⁴⁸ Of the 16 studies in which written interventions were associated with improved comprehension, 3 studies were of good quality (Figure 2).^34,39,41 The remaining 5 studies found no difference in comprehension between groups^33,38,45,47 or worse recall of risks in the written intervention group.³⁵ Of the 5 studies in which written interventions did not improve comprehension, 1 study was of good quality (see Figure 2).³³

Figure 2. Benefit of interventions designed to improve comprehension in informed consent by intervention type and study quality.

Audiovisual or multimedia interventions

Fifteen studies evaluated audiovisual or multimedia interventions designed to improve understanding of informed consent (Table 2).^29,50–63 The majority of these studies examined the use of audiovisual materials in addition to standard informed consent procedures.^{29,50,54,55,57,58,62,63} Others evaluated audiovisual techniques in place of verbal or written information,^{52,53,56,59–61} and 1 study evaluated the use of a video in addition to or in place of an informed consent discussion.⁵¹ The content of audiovisual materials ranged widely, from a primary focus on risks of the procedure^52,56 to more comprehensive information about risks, benefits, alternatives, and general information about the procedure and recommended follow-up care.⁶⁰ Four studies evaluated interactive audiovisual programs that allowed patients to skip sections, request more information, self-pace, or test knowledge.^50,56,59,62 Noninteractive videos ranged in length from 5 min^53,58 to 16 min.⁵⁵

Table 2. Results of Trials of Audiovisual or Multimedia Interventions.

Source	Procedure	Study Population		Intervention	Study Design	Outcome Measure				Results
		Age and Education	N			Knowledgea	Timing Relative to ICb	Timing Relative to Procedurec	Blindedd
Good-quality studies
Agre and others (1994)51	Colonoscopy	Mean age = 61 y; 68% with postsecondary education	201	Video of physician describing risks, benefits, and alternatives to colonoscopy or video plus discussion with physician compared with discussion only	RCT	R, B, A	I	Before		Video and video plus discussion groups had greater knowledge scores (out of 13) than discussion-only group (mean video only 10.7, video plus discussion 11.0, discussion only 9.6; P<0.01)
Luck and others (1999)57	Colonoscopy	Age range = 20–88 y; 11% with postsecondary education	150	10-min video narrated by well-known actor about colonoscopy in addition to standard IC process	RCT	R, B, G	D	Before	X	Patients in video group had higher overall knowledge scores (mean 9.9 v. 8.3 out of 12; P= 0.0001)
Rossi and others (2005)61	Knee arthroscopy	Age range = 18-92 y; 61% with>12th-grade education	150	12-min educational video about knee surgery in place of standard consent discussion	RCT	R, B, G	I	Before		Mean comprehension was higher in the video group (79% v. 65%, P= 0.00001)
Shaw and others (2001)62	Colonoscopy	Mean age = 54 y; 58% college degree or greater	133	Interactive computer program at 8th-grade level about indications, complications, and outcomes of colonoscopy in addition to standard consent from endoscopist	RCT	R, B, A, G	I	Before	X	Interactive computer group had better comprehension overall (mean number correct responses 22 v. 19 out of 27;P<0.001)
Done and Lee (1998)54	Anesthesia	Mean age = 34 y; education NS	127	7-min video “About Your General Anesthesia” in addition to routine interview with anesthesiologist	RCT	R, G	I	Before		Video group was 6 times more likely to answer all knowledge questions correctly, after adjusting for preoperative state and anxiety (P<0.01)
Cowan and others (2007)53	CT scan with IV contrast	Mean age - 42 y; 31% with postsecondary education	107	5-min video of physician describing IV contrast in English or Spanish at 8 th-grade reading level in place of verbal information from emergency physician	RCT	R, B, A	I	Before		Video group had higher mean knowledge scores (difference of 20%; 95% CI for difference, 13%-28%)
Olver and others (2009)59	Chemotherapy	Mean age - 57 y; 27% with postsecondary education	101	Interactive multimedia CD-ROM about cancer and chemotherapy including interviews with doctors and patients in place of written information sheet	RCTe	R, B, G	D	After		No difference between CD-ROM and written groups in correct recall of number of drugs, treatment length, or treatment goal in intention-to-treat analysis
Astley and others (2008)52	Coronary angiography	Median age - 64 y; median 9-12 y education	99	Standardized information about risks presented verbally, in written form, or in a video form (including 4 types of memory cues)	RCT	R, B	I/D	Both		No difference between groups in recall of risks pre angiography, immediately postangiography, or at 30 d
Rossi and others (2004)60	Ankle fracture surgery	Age range = 17–62 y; 50% with postsecondary education	48	9-min videotape at 7th-grade level about risks, benefits, alternatives, and general information replaced verbal discussion	RCTe	R, B, A, G	I/D	Both		Video group scored higher on immediate (78% v. 55%, P= 0.0002) and follow-up (68% v. 54%, P=0.0139) knowledge questionnaires
Mason and others (2003)58	Tubal sterilization	Mean age = 34 y; 32% with A levels or higher	31	5-min video narrated by one of the study authors about sterilization in addition to standard consultation	RCT	R, A, G	NS	Before		Patients in video group scored higher on knowledge questionnaire (median 90% v. 58%;P<0.001)
Fair-quality studies
Zvara and others (1996)63	Anesthesia	Age NS; education NS	178	10-min video about anesthesia and surgery in addition to standard visit with anesthesiologist	NRCT	G	I	Before		Video group scored higher on 1 out of 6 knowledge questions (99% v. 86%; P= 0.03)
Hopper and others (1994)56	Venography, excretory urography, and CT scan with IV contrast	Mean age = 53 y; mean 13 y education	160	Interactive video consent of physician describing risks in place of written consent	RCT	R, B, G	I	Before		Video group scored higher on knowledge test (score/7; 5.2 v. 4.8; P<0.05)
Poor-quality studies
Ader and others (1992)50	Third molar extraction	Mean age - 24 y; mean education - 13 y	60	Interactive or noninteractive video about wisdom teeth, extractions, risks, and recovery in addition to standard consultation with the oral surgeon compared with standard consultation only	NRCT	R	I/D	Before		Both interactive and noninteractive video groups scored higher than standard consult-only group on a knowledge quiz (interactive video 73%, noninteractive video 85%, standard consult 40%; P< 0.0001)
Gaskey (l987)55	Anesthesia	Mean age - 33 y; education NS	40	16-min video about anesthesia in addition to standard preoperative visit with anesthesiologist	NRCT	G	NS	Before		No difference in mean knowledge score between groups
Westreich and others (1995)29	Electroconvulsive therapy	Median age - 63/65 y; mean education - 11 y	18	Informational video featuring an electroconvulsive therapy expert in addition to standard written consent	RCT	R, B, A, G	I	Before		No significant difference between groups in total knowledge

Note: NS = not specified; IC = informed consent; RCT = randomized controlled trial; NRCT = nonrandomized controlled trial; CT = computed tomography; IV = intravenous; CI = confidence interval.

^aIndicates elements of IC knowledge measured by the instrument (R = risks; B = benefits/indications; A = alternatives; G = general/other knowledge about procedure or not specified).

^bIndicates timing of the measurement of understanding relative to informed consent (I = immediately after informed consent; D = delayed; NS = not specified).

^cIndicates timing of the measurement of understanding relative to the medical/surgical procedure (before, after, both, or NS = not specified).

^dIndicates whether the measurement of understanding was described as blinded.

^eNo description of randomization method or described a nonstandard randomization method but with evidence of comparability between patient groups.

Of the 15 studies evaluating audiovisual interventions, 11 showed improved comprehension^{50,51,53,54,56–58,60–63} whereas 4 did not.^29,52,55,59 The degree of improved comprehension varied somewhat but was high overall (Table 2). For example, in one study, the video group was 6 times more likely to answer all knowledge questions correctly.⁵⁴ Eight of the 11 studies that showed improved comprehension were of good quality, and 2 of the 4 studies that showed no improvement in comprehension were of good quality (see Figure 2). Most of the studies that showed improved comprehension evaluated understanding of the elements of informed consent immediately after the informed consent intervention, although 2 good-quality studies tested understanding 1 wk⁵⁷ and 1 mo⁶⁰ after informed consent. Among the 4 studies that evaluated interactive audiovisual programs, 3 studies of mixed quality showed improved comprehension,^50,56,62 whereas 1 good-quality study found no difference in recall of information.⁵⁹

Extended informed consent discussion

Extended informed consent discussions were evaluated in 5 studies (Table 3).^64–68 The content of discussions varied from a sole focus on risks of the procedure⁶⁸ to a broader overview of risks, benefits, indications, alternatives, and general information about the procedure.⁶⁷ Two of the 5 studies showed improved comprehension (Figure 2).^64,66

Table 3. Results of Trials of Extended Informed Consent Discussions.

Source	Procedure	Study Population		Intervention	Study Design	Outcome Measure				Results
		Age and Education	N			Knowledgea	Timing Relative to ICb	Timing Relative to Procedurec	Blindedd
Good-quality studies
Lapid and others (2003)65	ECT	Age range = 29–86 y; 90% completed HS	40	Half-hour individual informational session with a psychiatrist reviewing answers to frequently asked questions about ECT in addition to standard pre-ECT education (discussion and videotape)	RCTe	R, B, A, G	I	Before	X	No difference between groups in understanding score (out of 6) pre- or post-IC (intervention group mean score pre-IC 5.5, post-IC 5.7; standard group mean score pre-IC 5.5, post-IC 5.8); both groups showed improved understanding after IC
Solomon and Schwegman- Melton(l987)66	Cardiac catheterization	Age range - 50-70 y; education NS	36	Half-hour individual structured teaching session with nurse reviewing risks, benefits, indications, and details of the procedure in addition to standard discussion with cardiology fellow	NRCT	R, B,G	I	Before		Structured teaching group had higher total knowledge score (out of 13; mean 11.5 v. 8.9; P<0.001)
Dodd and Mood (1981)64	Chemotherapy	Age range - 40-79 y; median education - 10.5 y	24	20-min 1–on-l visit from nurse in addition to standard IC procedure; patients in the information group received a review of information given during the IC process v. patients in the control group who received a review of other disease-related information	RCTe	R, B, G	D	NS		Information group recalled more names of drugs (78% v. 34%; P<0.05), side effects (52% v. 36%; P<0.01), lethal side effects (71% v. 21%; P<0.01), and indications for chemotherapy (86% v. 64% of indications correctly recognized, P<0.05) than control group
Fair-quality studies
Dawes and others (1992)67	Tonsillectomy or nasal surgery	Mean age = 35 y; 15% with A levels or higher	190	Group 1: informal interview; group 2: structured interview reviewing indications, risks, benefits and alternatives; group 3: structured interview and provision of written information sheet	NRCT	R, G	D	Before		No difference between groups in recall of operation details; structured interview groups recalled more complications but were also told more complications; informal interview group recalled greater percentage of the complications they were told (60% in group 1 v. 40% in group 2 and 43% in group 3; P<0.01)
Poor-quality studies
Hopper and others (1989)68	CT scan with IV contrast	Age NS; mean education = 13–14 y	100	Group 1: generic consent form, no verbal counseling; group 2: consent form with brief information about risks of contrast, no verbal counseling; group 3: consent form with detailed information about risks of contrast, no verbal counseling; group 4: detailed verbal counseling about risks from a physician, generic consent form	NRCT	R, B	D	After		Group 1 had lower knowledge scores (mean 38%); no difference between groups 2 (mean 68%), 3 (mean 63%), and 4 (mean 70%)

Note: CT = computed tomography; IV = intravenous; ECT = electroconvulsive therapy; NS = not specified; HS = high school; IC = informed consent; RCT = randomized controlled trial; NRCT = nonrandomized controlled trial.

^aIndicates elements of IC knowledge measured by the instrument (R = risks; B = benefits/indications; A = alternatives; G = general/other knowledge about procedure or not specified).

^bIndicates timing of the measurement of understanding relative to informed consent (I = immediately after informed consent; D = delayed; NS = not specified).

^cIndicates timing of the measurement of understanding relative to the medical/surgical procedure (before, after, both, or NS = not specified).

Among the 5 studies, 3 good-quality studies examined individual teaching sessions added to standard informed consent procedures.^64–66 Teaching sessions lasted 20 to 30 min and were conducted by a physician or nurse. In 2 of these good-quality studies,^64,66 teaching sessions resulted in improved overall comprehension (Table 3). In the 3rd study, the teaching intervention did not result in improved comprehension when added to a rigorous standard informed consent process, which already included both a discussion and an informational videotape.⁶⁵

The 2 additional studies that were of fair and poor quality compared extended informed consent discussions (with or without the provision of additional written information) to standard informed consent procedures.^67,68 These studies did not show significant gains in comprehension associated with more detailed verbal information.

Test/feedback techniques

Three small studies of variable quality compared test/feedback techniques with standard informed consent procedures (Table 4).^69–71 Test/feedback techniques involved asking patients to repeat back basic information they received as part of the informed consent discussion. Patients were given feedback until they were able to correctly verbalize specific elements of the informed consent discussion or until a maximum number of trials was reached. Test/feedback techniques were associated with improved understanding of informed consent in all 3 trials, 1 of which was of good quality (Figure 2).⁶⁹ In this good-quality study, comprehension was significantly higher in the intervention group when tested before the procedure, but this difference between groups was no longer significant when tested after the procedure (Table 4).

Outcome Measures

There was significant variation across studies with respect to the elements of informed consent comprehension assessed (Tables 1–4; Figure 3). The majority of trials (39 of 44 studies) assessed understanding of the risks of the procedure. Of the 39 studies that measured understanding of risks, the intervention improved comprehension in 28 studies (Figure 3). Other measured elements of comprehension included patient understanding of potential benefits/indications (20 of 44 studies; intervention improved comprehension in 14 studies), alternatives (10 of 44 studies; intervention improved comprehension in 7 studies), and general knowledge about the procedure (i.e., what organ is being operated on; 28 of 49 studies; intervention improved comprehension in 22 studies). The majority of studies (26 of 44) assessed only 1 or 2 elements of comprehension. Only 6 of 44 studies assessed all 4 elements of comprehension^{29,45,46,60,62,65}; of these, 3 studies found improved comprehension of all elements with the intervention.^46,60,62

Figure 3. Benefit of interventions designed to improve comprehension of informed consent by elements of informed consent comprehension assessed.

Instruments used to measure comprehension were not standardized across studies and ranged from brief written questionnaires to open-ended, in-person interviews. The timing of outcome assessments varied from immediately following informed consent to several months afterward. Although the majority of studies included an assessment of comprehension before the medical/surgical procedure, 13 studies assessed comprehension only after the procedure (Tables 1–4).^{30,32,35,37,38,42,43,47–49,59,68,71} The timing of outcome assessments with respect to the informed consent process or the medical/surgical procedure was not specified in 8 studies.^{33,34,44,55,58,64,69,70}

Discussion

This systematic review identified 44 studies that have evaluated a variety of interventions to improve patient comprehension in informed consent for medical and surgical procedures. Overall, we found fairly consistent evidence that additional written information, audiovisual/multimedia programs, extended discussions, and test/feedback techniques improve patient comprehension in informed consent, especially regarding risks and general knowledge about a procedure. Although the quality of reporting was mixed, separating out good-quality studies revealed a similar pattern of benefit. These results indicate that communication interventions are effective and support the need to improve the standard of care in clinical informed consent.

Our finding of overall benefit from interventions in the clinical setting differs from the results of a systematic review of interventions to improve understanding in informed consent for research, in which multimedia techniques and enhanced consent forms were found to have only limited success.²³ Although research with human subjects requires a formalized informed consent process subject to regulation and oversight,⁷² minimum standards governing informed consent for clinical care are less clear, and patient understanding of the information disclosed in informed consent for medical and surgical procedures is often poor.^3,12,73 Thus, in clinical practice, modest efforts may result in significant gains in patient understanding when compared with a standard, often inadequate, informed consent process.

These findings suggest several needed changes to the practice of informed consent for medical and surgical procedures. First, communication interventions should be promoted, as studies have demonstrated the effectiveness of a wide range of written, oral, and video techniques as a means to improve patient comprehension in clinical informed consent. Some argue that overwhelming evidence of inadequacies in the informed consent process signals a need to reconsider informed consent standards.^21,22 However, our findings of improved patient comprehension with informed consent interventions support continued efforts to improve the process of informed consent.

Second, in the absence of clear evidence demonstrating the superiority of one type of intervention over others, decisions regarding which type to adopt should consider the feasibility of the intervention in a specific clinical setting and the intervention's acceptability to clinicians and patients. The National Quality Forum, the American Medical Association, and AHRQ have focused on test/feedback techniques and the use of simplified written consent materials as promising interventions that may be straightforward to implement and most likely to improve patient-centered communication and patient safety.^3,7,8 Of note, written information must be provided in addition to a discussion, as the provision of written information alone does not constitute valid informed consent. In response, some institutions have revised consent forms to make them more understandable to patients, translated consent forms into languages prevalent in the community, and/or added a “teach-back” requirement to their informed consent process or the consent form itself. We were unable to demonstrate that these types of interventions were more effective than other approaches. However, our finding of improved patient comprehension in the majority of trials that evaluated additional written information as well as evidence of benefit in the small number of trials that evaluated test/feedback techniques supports the continued implementation and evaluation of these methods in clinical care settings.

Third, there does not appear to be consensus regarding what constitutes full informed consent. The American Medical Association describes informed consent as a communication process in which the patient's diagnosis; the nature, purpose, risks, and benefits of the proposed procedure; and the nature, risks, and benefits of alternatives to the proposed procedure, including the option of not receiving any treatment, should be discussed.² However, the content of interventions and the outcome measures of studies included in this review focused overwhelmingly on patient understanding of procedural risks. An undisclosed risk that occurs and causes harm may lead to a claim of malpractice.⁵ Thus, a focus on risks may indicate a primary concern with malpractice risk reduction and/or a misperception of informed consent as little more than a medical Miranda warning, with the sole purpose of advising patients of the procedural risks they may encounter.¹⁴ Only a small minority of studies included assessments of the risks, benefits/ indications, alternatives, and general knowledge about a procedure, suggesting that important aspects of patient understanding in informed consent are not routinely addressed and may not be well understood by clinicians and researchers. The doctrine of informed consent suffers further from competing legal standards regarding the amount of information that must be disclosed to patients. For example, some states have adopted a professional standard, meaning a physician must discuss what a reasonable physician would discuss under similar circumstances, whereas other states have adopted a patient-oriented standard, meaning a physician must discuss what a reasonable patient would find relevant in a similar situation.⁵ Conceptual clarity regarding the key elements of informed consent knowledge, and the alignment of legal and ethical requirements, will help to focus improvements to the informed consent process and standardize evaluations.

Fourth, particular attention should be paid to implementing interventions that are accessible to patients with limited literacy and/or limited English proficiency. These groups are at increased risk for poor comprehension^19,20 yet have not been the focus of the majority of studies evaluating understanding in clinical informed consent. Most studies included in this review did little to assess, much less address, the association of patient literacy or language with informed consent comprehension. The educational level of participants was not specified in half of the studies, reading level was specified for only 2 of the 21 studies evaluating written interventions, and only 1 study included non–English-speaking patients.⁵³ Future efforts to improve comprehension in informed consent should be tailored to patients at highest risk for misunderstandings.⁸

Although the majority (73%) of studies in this review found some benefit to interventions designed to improve patient understanding in informed consent, this finding was not universal, and a handful of good-quality studies found no improvement in comprehension with an enhanced informed consent process. Given the heterogeneity of included trials, possible explanations for differing results include differences in patient populations and clinical settings, the content and quality of the intervention, the quality of the “standard” informed consent process, assessments of comprehension (including the type of instrument used, timing, and elements of comprehension assessed), power limitations due to small sample sizes, or the quality of study methods.

Our findings of mixed benefit and the heterogeneity of trial quality suggest several avenues for further research. First, future intervention trials must include appropriate randomization techniques and blinded outcome assessments to ensure the validity of results. Second, outcome measures should consider more than simply patient understanding of the risks involved in a procedure, as other elements of comprehension, including benefits/indications, alternatives, and general knowledge about the procedure, are also important aspects of informed consent. Third, studies of informed consent should include patients with limited literacy and limited English proficiency and explore the differential effects of interventions in these subpopulations. Fourth, more research is needed to evaluate the comparative effectiveness of different types of interventions as well as the potential for combining methods (e.g., a video consent module incorporating a test/feedback technique). Fifth, future studies should also assess cost-effectiveness.

Our study has several limitations. First, we did not feel that it would be meaningful to make quantitative comparisons between studies because of variations in populations, settings, interventions, and outcome assessments. Consequently, we analyzed these data as a systematic review and not as a meta-analysis. As in all systematic reviews, we were limited in our ability to draw conclusions by the variable quality of included studies. We attempted to mitigate this effect by separating out the findings of good-quality studies (Figures 2 and 3). Second, although our search methods were systematic and drew on multiple sources, trials in this area are published in a wide range of journals and many were not identified in an initial search of MEDLINE. Thus, it is possible that we failed to identify studies that would meet our inclusion criteria. Third, we cannot exclude the possibility of publication bias, with negative studies remaining unpublished. However, we do not think that this is a serious limitation, as several of the smallest studies had negative findings. Fourth, although we attempted to give as much information as possible about the statistical significance of each study's findings in Tables 1 through 4, the degree of benefit was reported differently and varied widely between studies. Consequently, we were not able to summarize the statistical benefit associated with each intervention type or the degree to which statistically significant improvements in comprehension may have clinical or policy significance. Fifth, smaller studies may have been limited in their power to find a difference. However, negative results were not more common among studies with fewer participants. Sixth, given the heterogeneity and variable descriptions of communication interventions, we were unable to draw conclusions about the optimal content of each intervention type. Seventh, improved patient comprehension is only one important outcome of informed consent. Our review did not consider patient satisfaction, anxiety, trust, achievement of goals of care, or adherence to follow-up recommendations. Finally, we included only studies of immediately necessary procedures requiring documented informed consent. Considerable research on tools to improve patient understanding and involvement in medical decision making,^74–76 routine clinical care, and advance directives⁷⁷ did not inform this review.

In summary, studies conducted in a variety of settings across a range of clinical conditions have demonstrated the effectiveness of communication interventions in informed consent for medical and surgical procedures. We found evidence that additional written information, audiovisual/ multimedia interventions, extended discussions, and test/feedback techniques improve patient understanding, especially regarding risks and general knowledge. Wide variability in study quality and outcome measures suggests a need for consensus regarding the key elements of understanding in clinical informed consent to comprehensively address ethical and legal requirements. Research assessing the comparative effectiveness and feasibility of different types of interventions in diverse populations and settings is needed to support the further adoption of best practices in clinical informed consent.