INTRODUCTION

According to gold standards of high-quality, modern medical care, patients should be informed about and involved in their care.1,2,3 This patient right is mandated by law in many countries.4,5 Patient recall and comprehension of medical information are prerequisites for providing informed consent, making informed treatment decisions, lifestyle, and self-management, and adhering to treatment recommendations.6,7 However, medical information is often complex and place great demands on both physicians’ information giving and knowledge translation skills and the patients’ capacity to understand, remember, and ultimately act on the information received.8 The process of informing patients involves a dynamic interplay between physicians’ skills in presenting information in a clear, relevant, and actionable way, and patients’ health literacy skills.3 This review focuses on physicians’ information exchange practices and associated patient-related behavioral outcomes.

Physicians have a moral and professional obligation to provide high-quality information to patients and secure their comprehension.4,9 Although physicians often assume that their explanations and instructions are easy to understand, they are often misunderstood by their patients.10,11 Patients commonly forget or misunderstand 40–80% of the information provided by physicians.11,12,13,14 The personal and societal costs of ineffective information giving are high: non-adherence to treatments15, medical errors16, longer hospital stays, frequent re-admissions17, patient complaints and litigations18, poor patient health19,20, and healthcare costs.21

Effective information giving requires a complex interaction of content, form, and use: speakers formulate what (the content) and choose how to say something to achieve their goal. Most of the medical literature has focused on the content of the information. Evidence that informational content on its own promotes patient outcomes is poor and, if present, most studies have focused on visual or written information in addition to the medical talk.21,22,23 Little attention has been given to how the medical information is provided by physicians during a consultation. Addressing this knowledge gap, we performed an initial scoping review, where we identified a range of strategies for effective information giving for different purposes, i.e., to support patient comprehension, persuade patients, build a relationship, or report facts objectively.24 Whether using communication strategies for providing medical information improves patient-related outcomes remains, to the best of our knowledge, unknown.25 There is some evidence for an association between general physician communication skills and patient outcomes, but these systematic reviews and/or meta-analyses report on generic communication interventions and show equivocal results.26,27,28,29,30 Without evidence for how the information is provided by physicians in the medical talk, it is difficult to identify what features of information-giving are associated with better outcomes and how best to design training programs to optimize the effectiveness of the information exchange.

In this systematic review, we describe the reported effects of physicians’ information-giving strategies on patient-related outcomes, as well as the features of these strategies and of the randomized controlled trials (RCTs) testing them.

METHODS

Protocol

This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) guidelines.31 The review protocol is registered in PROSPERO (ID: CRD42019115791).

Eligibility Criteria

Relevant RCTs were identified through a previously reported scoping review of physicians’ information-giving strategies in the dialog with patients.24 RCT studies were eligible for this study if they had a low risk of selection bias assessed with the Risk of Bias 2 (RoB2) tool and if they tested the effect of specified information-giving strategies used by physicians in dialog with patients/analog patients across any medical setting on patient-related outcomes. Studies based on a mix of physicians and other healthcare professionals were included.

Eligible interventions included consultations in which defined strategies for communicating medical information to patients were tested. Eligible comparisons comprised any type of controls.

Outcomes were patient-related, broadly categorized into patient information recall and behavioral outcomes. We excluded outcomes that were assessed in only one eligible study (i.e., satisfaction, quality of life, anxiety, stress, patient’s perceived physician competence). We also excluded trust outcomes because of the scientific debate about the conceptual, methodological, and empirical fragility of trust in the medical relationship32, especially in relation to physicians’ information giving.33

Search Strategy and Data Sources

We searched the databases MEDLINE, Embase, PsycINFO (Ovid), and Cochrane Central Register of Controlled Trials from inception to 24 April 2020 without restrictions. We developed the search strategy with an expert medical librarian (HS). Initial search terms were gathered from a set of key articles, then using an iterative process to develop the final search strategy based on relevant key terms and subject headings (Appendix Table 4). We also screened the reference lists of included or relevant articles to retrieve additional references.

Study Selection

Screening for inclusion in the initial scoping review was conducted independently by five pairs of reviewers. Conflicts were solved by discussion with a third reviewer. Screening for RCTs to be included in this study was performed by three reviewers (J. M., H. C. L., L. K. J.) based on unequivocal low risk of selection bias from RoB2 assessment, reported previously.24

Data Extraction

Data extraction was performed by pairs of researchers. Data on the reported effects of the included interventions on patient-related outcomes were extracted using a predefined document. When different data on the same outcome were reported, we selected the information with greater reliability in terms of type of measure. Authors were contacted to retrieve missing or incomplete data.

Specific data were selected and extracted to describe the studies, Table 1. Details about the information-giving strategies were also extracted from studies, reported in Table 2. Unique information-giving strategies were considered to be the modified minimal units of actions concerning information provision. These were extracted word by word (“Specific message/strategy” in Table 2), organized into strategy types, and strategy types were classified into main categories based on underlying mechanisms of functioning (Table 2, the categorization process is reported elsewhere).24

Table 1 Characteristics of Information Provision Interventions Assessing Patient Information Recall and Behavioral Outcomes
Full size table
Table 2 Information-Giving Intervention, Strategy(ies), Strategy Type(s), and Strategy Category(ies) Targeted by Each Study
Full size table

Data Analysis

As expected, and described in the study protocol, due to the high heterogeneity of studies, interventions, and outcomes (confirmed also by statistical analyses: χ2 = 98.62, p < 0.001; I2 = 92% for studies including a behavioral outcome; χ2 = 11.33, p = 0.25; I2 =21% for studies including information recall as outcome), a meta-analysis with pooled quantitative summary estimates was deemed inappropriate. Therefore, all effects reported, study by study, were qualitatively synthetized, descriptively summarized without summary estimates in tables, and visualized through forest plots obtained with Review Manager version 5.4.1.

RESULTS

Overview of Studies

We initially screened 9423 abstracts and 175 full-text articles of which 39 were included in the initial scoping review.24 Of these, 17 studies were RCTs with low risk of selection bias and eligible for the systematic review (Fig. 1). Two articles tested two different interventions with different participants each38,39, and one of these used one common control group.39

Figure 1
figure 1

PRISMA Flow chart.

Full size image

Trials were published between 1994 and 2020, and included 8256 patients or analog patients. The average age of patients in the 17 studies was 48 years (SD = 17.13). Effects of studies are reported separately for the two main outcomes: information recall and behavioral outcomes.

Information Recall Studies

The effect of information-giving interventions on recall or knowledge outcomes was investigated in eight of the 17 articles, reporting 10 different interventions. Information recall was most frequently assessed as a combination of free recall and recognition. All recall measures were self-reported and assessed immediately after the intervention. One study assessed self-reported parental knowledge/understanding of the child’s condition before and after the intervention.41

Characteristics of Studies Assessing Information Recall

The number of patients involved in these studies ranged from 3040 to 25337 (Table 1). Explaining clinical issues (7 of 10 studies) was the most frequent information provision task. In eight of the 10 interventions, participants were analog patients. Former patients were used in three experimental interventions reported in two articles.37,38 Six interventions, reported in four articles, were tested in a fictional experimental setting using video-vignettes36,37,38,39, while the others were conducted in real life.

All the 10 interventions tested unique groups of strategies, six studies tested cognitive aid strategies, and four relationship-oriented strategies (Table 2). Nine out of 10 studies also tested one single strategy. Overall, the most frequently tested strategies were emotional responsiveness during information-giving dialogs36,38,39, information structuring34,38,40, and use of visual demonstrations during oral information giving.34,35,41

Effects of Interventions on Information Recall

Seven out of 10 interventions showed a positive effect on information recall, with two studies reporting significant changes34,35 and two interventions included in one study reporting significant changes in recognition but not in free recall39 (Table 3 and Fig. 2). Ackermann et al.34 evaluated the effect of structuring (e.g., akin to a book where high-level information is presented as “title and chapter headings” to low-level information as the text) versus non-structuring of the information given at discharge on the amount of information freely recalled by students (acting as analog patients) with different levels of prior medical knowledge. All 234 participants assigned to the structured discharge consultation significantly increased the number of items recalled (17% increase of recall performance) compared to those receiving non-structured information. The effect was particularly pronounced among those with the least prior medical knowledge (42% increase of recall performance). Bennet et al.35 tested a visual method utilizing diagrams to illustrate key points included in the informed consent form on a small sample of 32 patients compared to two control conditions, usual care and “teach-the-teacher” condition where patients are asked to repeat the key points of the informed consent. They found that this visualization strategy significantly increased the number of items recognized by patients compared to the usual care condition, without increasing the average time needed. They did not detect differences between the visualization strategy and the teach-the-teacher strategy, but the latter required more time. Visser et al.39 recently compared the effects of oncologists’ emotion-oriented speech and emotion-oriented silence during extensive information-giving sequences on free recall and recognition, compared to giving limited space for emotional disclosure. Both these strategies enhanced recognition but not free recall, with no apparent influence on patient emotional stress level.

Table 3 Summary of Results for Intervention and Control Conditions for Each Study
Full size table
Figure 2
figure 2

Forest plot of recall or knowledge after information provision interventions.

Full size image

None of the other interventions reported a statistically significant improvement in information recall. Notably, the most recent study by Lehmann et al. with one of the largest samples in this group of articles (N= 148)38 tested two of the strategies found to have a positive significant effect in other studies (emotional-responsiveness and information structuring) but reported no improvements in recall. They found that trust may play a conflicting role in recall because enhanced trust decreased recall, and that patients’ personal characteristics (age, gender, education, health literacy) confounded recall outcomes.

Behavioral Outcomes Studies

Behavioral outcome measures were included in nine of the 17 studies. Behavioral outcomes were assessed objectively in seven and by self-report measures in two studies48,51 and included alcohol consumption42, weight loss43, blood lipid levels46, smoking cessation47, and treatment-related changes like medication side effects48, use of a treatment49, or likelihood of undergoing a treatment51, participation in screening44, and written durable power of attorney.45

Characteristics of Studies Assessing Behavioral Outcomes

The number of patients involved in the nine studies assessing behavioral outcomes ranged from 6647 to 305346 (Table 1). Most studies (8 of 9) were conducted in real-life settings, and one used video-vignettes.51 The most frequent clinical task performed was improving health-related behaviors (6 of 9).

In general, each intervention tested multiple information provision strategies (Table 2). The most frequently included strategies were persuasive (5 of 9) and cognitive aid strategies (4 of 9). Two interventions included cognitive aid strategies combined with objectivity-oriented49 or relationship-oriented strategies51; two interventions included persuasive strategies combined with objectivity-oriented strategies.44,45

Effect of Interventions on Behavioral Outcomes

Eight of the nine studies reported significant improvements in behavioral outcomes (Table 3 and Fig. 3). All the interventions that included a strategy aimed at persuading patients and influencing their thinking and behavior, by being directive43,45,49, providing argumentations44,48, or negatively framing the message, reported positive significant improvements on patients’ behaviors. Aveyard et al.43, Grimaldo et al.45, and Mazza et al.49 all tested the effect of a direct recommendation from the doctor to engage in extra-visit activities and of planning a follow-up. These strategies were provided alone43 or in combination with other information-giving strategies45,49, and led patients to reduce their weight43, to write a durable power of attorney after 12 months45, or to use contraceptives after 2 months49. Both Lamb et al.48 and Boguradzka et al.44 reported a significant positive impact of providing medical information with full disclosure of benefits and disadvantages (in the case of Boguradzka et al.44 together with structured information) on patients’ experienced side effects and participation in screening, respectively. Kim et al.47 tested another persuasive strategy for framing an information message: stressing losses and framing the message negatively. The inclusion of three aversive sentences on consequences of smoking led 66.7% of patients to quit smoking after 6 months compared to the 30.3% in usual care.

Figure 3
figure 3

Forest plot of behavioral outcomes after information provision interventions.

Full size image

Both Ockene et al.42 and Saha and Beach51 tested patient-centered communication strategies in the direction of enhancing patient cognitive processing and understanding of the information (e.g., open-ended questions, lay language, demonstrating with visuals and/or gestures). Saha and Beach51 also included information provision strategies supporting the physician-patient relationship like expressing empathy, being responsive to the patients’ concerns, and personalizing the information to make it relevant to the patient’s context. Both studies demonstrated that these interventions improved patients’ health-related behaviors, in the direction of reducing alcohol usage42 or increasing the analog patients’ likelihood of undergoing bypass surgery.

The only study that did not report a clear, positive, and significant improvement in patients’ behavioral outcomes (in this case, blood lipid levels, coronary risk, and the frequency of reaching lipid targets) tested the role of repeatedly discussing information on patient’s risk for future cardiovascular events showed in a graphical format with a computer printout.46 The findings reported were at the border of significance and Grover et al. discussed that choices in the study may have underestimated the intervention arm.

Discussion

To the best of our knowledge, this is the first systematic review investigating the effects of information-giving strategies on patient outcomes across different types of medical settings exclusively including RCTs with low risk of bias. This review of 17 RCTs involving 8256 patients provides strong indications that using deliberate communication strategies when providing information can be more effective in improving patient outcomes than not using deliberate strategies. This main finding enriches results from previous systematic reviews showing how physician communication in general26,28 and written or visual information outside the medical consultation can improve patient outcomes.22,23,52,53 It sheds light on the particular importance of oral information giving, which is routinely used by physicians in their daily practice and do not require additional resources. If oral information provision is deliberately enhanced by specific strategies to frame the information, this may be a powerful tool for improving important cognitive and behavioral outcomes of patients, as well as many other related outcomes.54

Four out of 10 studies testing physicians’ information giving strategies on patient information recall reported a positive significant effect. These studies were quite similar in terms of tested strategies, clinical setting, and study design. All but one study testing physicians’ information giving strategies on patients’ behavioral outcomes reported positive significant effects, also on objectively measured outcomes such as weight or blood lipid levels changes. These studies were very heterogeneous including a wide range of strategies and behavioral outcomes’ types, and the findings may have different interpretations and implications.

One possible interpretation involves the extremely different nature of the two considered outcomes and related communication goals (facilitating understanding, changing behaviors), reflected in the use of distinct information framing strategies. In particular, information provision interventions with information recall as outcome mostly tested cognitive aid strategies (like information structuring) in fictitious settings, with the purpose of explaining clinical issues. Some of these studies also tested relationship-oriented strategies (like emotional responsiveness): studies testing relationship-oriented strategies were those showing lowest effects on information recall. One of these studies pointed out the intervening effect of relationship- and trust-related variables on recall38, which can potentially explain the reduced impact of this group of studies on recall. The relationship between trust and information recall needs further investigations and may represent a challenge in clinical practice33, potentially suggesting a need for physicians to emphasize the importance that patients question their information giving, particularly if patients seem to defer to their authority. On the other hand, for the goal of changing patients’ beliefs or behaviors, persuasive strategies generally yielded strong effects. This supports suggestions provided in a JAMA viewpoint on the essential function of persuasion in medical communication.55 Information messages aimed at encouraging patients to engage in certain health behaviors may particularly benefit from deliberate embedding within a persuasion frame. While in this systematic review we focus on explicit persuasive information strategies so that patients become engaged in certain beliefs or behaviors, naturalistic studies have also showed that persuasive attempts can be used in subtle, implicit ways by physicians.56 Combined, these results call for a discussion about appropriate and deliberate use of persuasion in physician information giving.

Information recall trials were mostly conducted in fictitious settings and tested unique and consistent strategies, while behavior outcome trials were mainly conducted in real-life settings and tested multiple types of strategies. Real-life studies may introduce more variation in the intervening variables and participants. This may produce greater effects on patient outcomes as patients may find the intervention more relevant to them and/or rely more on the physician’s advice compared to individuals participating as in the shoes of patients or in fictitious scenarios. Previous research has indicated that analog patients are as reliable as actual patients to evaluate physicians’ communication behaviors57,58, but this depends on their engagement and by how the scenarios are designed.59 Overall, this may indicate the need for a stepwise approach: (1) map behaviors that deserve specific testing, (2) ascertain their potential efficacy in experimental settings, and (3) when variables and mechanisms in play are ascertained, determine how the tested strategy function in real-life settings to produce the desired changes.

Finally, even if the findings reflect information provided for two different communication goals and therefore the strategies used differed substantially, they all have in common the element of information shared by the physician, paired with specific strategies. This may provide some insights about the complex interplay among physicians’ information giving, patient information recall, and patient behaviors. The reported effects on behavioral outcomes may be explained by mechanisms that go beyond the information exchange and involve patients’ perceptions, knowledge, beliefs, attitudes, and intentions to change.60,61 Alternatively, patients may prioritize key items of information to remember, those perceived to enable and motivate certain behaviors. Future studies should explore what information patients prioritize as most important to remember, and also what is the minimum number of recalled items necessary to enable engagement in desired behaviors (e.g., participate in screening programs, lifestyle behavior change). Recently, the “learning by doing” pedagogic approach has been stressed, which considers behaviors as facilitators of learning experiences.62 Future studies are needed to understand the dynamics between cognitive and behavioral learning processes as a result of different combinations of information contents and strategies, including possible intervening elements such as patient attitudes, perceptions, beliefs, and knowledge.

Strengths and Limitations

There are several limitations to this study. First, the extent of positive findings in the included studies may be related to publication biases. Second, findings may need to be interpreted separately for studies assessing recall and behavioral outcomes, even if the two outcomes have physicians’ information giving as common denominator. Third, we were able to identify only 17 relevant studies to include with rather small samples, despite all being low-risk RCTs.

Strengths include the rigorous, comprehensive search completed in 2020, and resulted in an overview of a largely unexplored key clinical skill. Findings reflect a strict selection of high-quality articles based on rigorous screening and quality assessment procedures. The study provides a valuable knowledge base for future studies and practical indications for physicians for successfully conveying information to their patients.

Conclusions

Providing medical information using specific framing strategies appears to improve patient information recall or health-related behaviors. The study offers insights about specific strategies that physicians can deliberately use to frame medical information to reach defined communication goals and improve patient outcomes. Future studies should test the identified strategies with larger samples, in real-life settings to test cognitive aid strategies for securing patient recall, disentangle the complex interplay between different types of strategies concurrently used to deliver similar messages, and teaching courses on information sharing including framing strategies. Finally, future studies should also investigate on the other part of the puzzle, namely to investigate patients’ strategies to make sure physicians understand the information they provide.63