Introduction

Patients often read their medical records, and they may receive test reports, including radiology reports, in the mail. In general, these documents are written in medical jargon that is intended for documentation and for use by other medical professionals. Studies indicate that more than half of American adults have limited health literacy [11, 13, 17]. Even medically sophisticated patients may misinterpret some common medical jargon. Limited understanding and misinterpretation may affect health and patient satisfaction [13]. A Cochrane review showed that when providing patient-centered care, a model in which communication of health information is a key element, patients are more satisfied [4]. This underlines the pivotal role of patient information that is tailored to patients’ needs and understanding. Research on words used in provider–patient interactions show a strong affect on response to treatment and coping strategies [2, 3, 5–8]. To illustrate, the word “tear” refers to a signal abnormality or defect. In radiology reports, this term is commonly used to describe degenerative and often benign conditions of aging. A patient with no knowledge of the professional use of the word tear will likely interpret it colloquially and assume that they have damage in need of repair [16]. A study about the use of words by patients with chronic low back pain showed that the term “tear” was associated with poor perceived prognosis [14]. It is possible that this term might increase the chances that patients choose operative treatment. Two studies showed that incomprehensible and incomplete written information is associated with negative emotions and confusion [7, 10].

We sought to determine whether MRI reports that are rewritten to be understandable by a reader with an eighth-grade reading level, and whether the use neutral descriptive words and the most optimistic interpretations based on current best evidence will be better received by patients. Specifically, our primary study hypothesis was that the original and the reworded MRI reports would have comparable emotional valence. Valence is a term used by psychologists to refer to an object or event’s inherent attractiveness or aversiveness (ie, the quality that determines the emotional response). Our secondary hypotheses were that patient satisfaction, understanding, and perceived usefulness also would be comparable. Our third hypothesis was that equal numbers of patients would prefer the original or the reworded report.

Patients and Methods

Using an institutional review board-approved protocol, we enrolled patients visiting the orthopaedic hand and upper extremity outpatient office for this observational cross-sectional study. Pregnant patients and those who could not speak English were excluded, as mandated by the institutional review board. All patients were enrolled between May 2012 and July 2012.

We enrolled 102 patients, but two did not complete the questionnaire resulting in a study cohort of 100 patients. Of the 100 patients enrolled, 59 were women and 41 were men. Their mean (± SD) age was 51 ± 16 years (range, 21–80 years), the level of education was 15 ± 2.9 years (range, 4–24 years), and the majority of the patients worked full time (Table 1).

Table 1 Patient demographics*
Full size table

A list of all shoulder, elbow, wrist, and hand MR images ordered in 2011 was obtained from the radiology service. From this list, the first four MRI reports (one each shoulder, elbow, wrist, and hand) in chronologic order were selected with the following inclusion and exclusion criteria: (1) adult patient; (2) pain as the primary indication for the MRI; and (3) no malignancy presented or suspected. This resulted in reports about lateral epicondylitis (one), rotator cuff tendinopathy (two), trapeziometacarpal arthritis (three), and flexor carpi radialis tendinopathy (four) (Appendix 1 and Appendix 2).

The original reports were reworded to the recommended reading level for effective health education below the eighth-grade level [15]. This is easily checked using tools in Microsoft Word (Redmond, WA, USA). It is straightforward to explain medical things in simple, everyday language, although it is not something we are accustomed to doing. We used neutral descriptive words and the most optimistic interpretation based on current best evidence. For instance, words such as “tear” were replaced by more descriptive and accurate words such as hole, signal change, or defect. We also used analogies (eg, gray hair, bald spot) where appropriate. It was emphasized that the reports are simulated reports unrelated to their problem.

Each patient was presented with a report either on the shoulder, the elbow, the wrist, or the hand regardless of their reason for coming to the doctor. These were not reports on MR images that had been obtained on these patients nor were they related to their specific disorder. We were interested in the emotional content of the reports as judged by people not directly impacted by the reports. The original and reworded reports about the same condition were presented to the patient. We alternated the order of presentation to avoid bias. After reading each report, patients completed a questionnaire concerning their satisfaction with the presented report, usefulness of the presented content, the ability to understand the report, and their emotional valence using the Self Assessment Manikin (SAM) [3].

Patients’ satisfaction with each report was assessed using an 11-point Likert scale question: “On a scale between 0 meaning completely unsatisfied and 10 meaning completely satisfied, how would you feel to receive this report?” Two more questions were completed regarding the usefulness and understanding using 11-point Likert scale questions. Higher scores indicated greater usefulness and more understanding.

After each report, the SAM was obtained. The SAM is an affective rating system developed to study emotion [3]. It is a validated and reliable affective rating system tool for pleasure, arousal, and dominance (meaning control over the situation as opposed to feeling helpless) [1, 3]. Patients were requested to rate their emotional dimension pleasure from faces that range from happy (low score) to frowning (high score). Patients stated their arousal by selecting from an excited/worried (low score) face to a relaxed face (high score). The dominance dimension ranged from a small figure (no control; low score) to a large figure (full control; high score). For analyses, we scored the outcome of this questionnaire, ranging from 1 to 9 points. Additionally, after reading both reports, patients were asked which report they preferred.

According to a power analysis, 82 patients provide 80% power to detect a difference between MRI reports of 30% of a standard deviation in values for a specific type of emotional response with alpha set at 0.05. We enrolled 100 patients to have 25 patients for each report. We calculated means and SDs for continuous variables and frequencies for categorical variables. We used the nonparametric Mann-Whitney tests to compare the independent variables when not normally distributed. For the variables that were normally distributed, we used independent-samples t-tests.

We had one missing value for the question about satisfaction and one missing value for the SAM dimension pleasure. We imputed the mean to complete these values.

Results

Subjects felt more pleasant, calm, and in control when reading the reworded compared with the original reports. The pleasure score (mean ± SD) of all the original MRI reports was lower (4.6 ± 2.1) compared with the reworded reports (5.8 ± 2, p < 0.001). The arousal score (mean ± SD) of all the original reports was higher (4.1 ± 2.5) compared with the reworded reports (2.9 ± 1.8, p < 0.001). The dominance score (mean ± SD) of all the original reports was lower (4.4 ± 2.3) compared with the reworded reports (5.8 ± 2.2, p < 0.001) (Table 2).

Table 2 Bivariate analysis using the SAM questionnaire
Full size table

Analyzing the four different reports for differences by specific disease (acknowledging limited power), we found differences more for the tendinopathies than the arthropathy. The pleasure score was higher for the reworded rotator cuff tendinopathy reports compared with the original (p = 0.001). The arousal score was greater for the original lateral epicondylitis and wrist reports compared with the reworded reports (p = 0.01 and p = 0.04). The dominance score was greater for reworded rotator cuff tendinopathy report compared with the original (p < 0.001) (Table 3).

Table 3 Bivariate analysis using SAM questionnaire subreports
Full size table

Subjects found the reworded reports more satisfying, useful, and understandable than the original reports. The satisfaction score (mean ± SD) of all the original MRI reports was lower (5.1 ± 3.3) compared with the reworded reports (7.1 ± 2.8, p < 0.001). The usefulness score of all the original reports (mean ± SD) was lower (4.8 ± 3.5) compared with the reworded reports (7.1 ± 2.9, p < 0.001). The understanding score (mean ± SD) of all the original reports was lower (4.2 ± 2.3) compared with the reworded report (8.1 ± 2.6, p < 0.001) (Table 4).

Table 4 Bivariate analysis for satisfaction, usefulness, and understanding
Full size table

Of the four subreports, the satisfaction score was higher for the reworded rotator cuff tendinopathy report than the original (p < 0.001). The usefulness scores of the reworded tennis elbow, rotator cuff, and trapeziometacarpal arthrosis reports were higher than the original reports (p = 0.012, p < 0.001, and p = 0.024, respectively). The understanding scores were greater in all four reworded subreports (Table 5).

Table 5 Bivariate analysis of satisfaction, usefulness, and understanding subreports
Full size table

Seventy of 100 patients preferred the reworded MRI reports (Table 6). In each of four subreports, the reworded report was preferred by more patients (Table 7).

Table 6 Preferred report
Full size table
Table 7 Preferred subreports
Full size table

In bivariable analysis age affected emotional response, but otherwise demographic factors were not associated with emotional valence or report preference (Appendix 3 and Appendix 4).

Discussion

Diagnostic reports can be distressing, perhaps more so for patients with limited health literacy [7, 10, 13]. The aim of this study was to determine whether rewording of MRI reports results in better patient ratings of emotional valence, satisfaction, usefulness, and understanding. Furthermore, we wanted to find out whether the neutrally worded MRI reports rewritten at the eighth-grade reading level would be preferred by patients. Subjects felt more pleasant, calm, and in control on average after reading MRI reports reworded for readability and optimism. They also rated the reworded reports as more satisfying, useful, and understandable on average. Most patients preferred the reworded report over the original report.

These data should be interpreted in light of the fact that subjects were reading an MRI report that was unrelated to their problem and they may have a different emotional response if the report was their own. This was done for practical reasons and as pilot work to establish the merit of the concept given that a study of patients with each specific disease undergoing MRI would take much longer to complete. However, the subjects questioned were patients in a patient-provider setting and could envision receiving such a report. Other than the consistent approach of our team we did not control for the degree to which the report was modified. The reliability of rules for formatting reports or modifying existing reports could be the focus of future research. Measuring emotion is difficult, but the SAM questionnaire is a widely used and reliable method for quantifying affective influences of information on people [3]. Our subgroup analysis for different anatomic sites was secondary, underpowered, and hypothesis generating at best. The MRI reports used in this study were limited to orthopaedic conditions of the upper extremity and hand and did not encompass other regions and conditions of the human body. The reports did not include malignancies or dangerous conditions; therefore, we cannot generalize it to these conditions. We had only two uncompleted answers for which we imputed the mean. This is an accepted way to address missing values and two missing entries is a very low number [5].

These data suggest an advantage to including a summary of the report directed to the patient that is easy to understand and as optimistic as possible. This may encourage and reinforce optimal coping strategies, because distressing or confusing material may reinforce the natural human tendency to prepare for the worst, which psychologists call catastrophic thinking [8]. In addition, more satisfying patient information has been shown to reduce malpractice [6]. Understandable information allows the patient to be more involved with their medical care, which is associated with greater adherence to provider recommendations and better health outcomes [12].

The improved emotions and satisfaction of subjects reading reworded reports might translate to better overall satisfaction with care. Satisfaction with care can be at odds with good medical care and the example of opioid prescription was recently used to highlight this [9, 18]. Diagnostic reports with simple, accurate, dispassionate, and even optimistic wording can set up providers for more satisfying discussions of how to optimize health (ability to depend on one’s body) and help limit less satisfying conversations about why “we are not going to fix the tear found in my elbow/shoulder/wrist”.

Among the 30% (30 of 100) of subjects who preferred the original report, some commented that “the original report was more detailed” and some said the reworded report looked “too simple.” Others mentioned they preferred the original report because they want to unravel it with the help of the internet or their physician. We would argue that the technical information can be reported dispassionately and descriptively (eg, signal change, defect) more or less as it is now and that a less technical, more readable, and optimistic rewording of the interpretation portion of the report might be no less complete or less accurate. This combination might satisfy all patient interests. Any type of report will always benefit from discussion with a healthcare professional. Further research is needed to determine how best to word reports to provide correct and optimistic medical information with enough details yet still be understandable and without causing unnecessary distress.