Patient Comprehension of Emergency Department Care and Instructions: Are Patients Aware of When They Do Not Understand?

Introduction 

In the emergency department (ED), the effective exchange of information between patients and health care providers is critically important to patient care. Yet, the chaotic nature of the environment and transient interactions pose significant challenges to communication.1, 2, 3

Previous research suggests that patients often have difficulty understanding discharge instructions provided in the ED. Frequently, written materials exceed patients' literacy levels, which may contribute to problems with comprehension.4, 5 Direct assessments of patient and caretaker comprehension after ED discharge have demonstrated difficulties with recalling diagnoses and discharge instructions.6, 7, 8, 9 These deficits have been shown to exist immediately after an ED visit and thus are not merely a function of people forgetting information over time.

Deficiencies in patient comprehension represent communication failures. Identifying and addressing these problems are essential steps to improving patient care. During the past 20 years, most of the research on patient-provider communication has been conducted in primary care settings and has focused on patient-physician interactions.10 Numerous studies have shown that patients desire information about their medical care and identify communication as a critical element of their interactions with health care providers.10, 11, 12, 13, 14 In the ED, communication is recognized as a key factor in patient satisfaction, and problems with communication have been found to be a leading cause of patient complaints.15, 16, 17, 18, 19, 20, 21

So why do communication failures occur? The answers are complex and multifaceted. They reflect countless factors on the part of the patient, physician, health care team, and the environment. Unfortunately, there are many aspects of the ED setting and our interactions with patients that are difficult to control, and ED providers cannot hope to ensure perfect communication with every patient. We can, however, strive to minimize communication failures by characterizing them and determining why they occur and how we may intervene to reduce or prevent them. Patient comprehension serves as a meaningful measure of what the patient takes away from their visit and thereby provides a valuable tool for communication research.22, 23, 24

Previous research in the ED setting has focused predominantly on patient understanding of their diagnosis and discharge instructions. Assessments of patient understanding of their ED care (tests and treatments) have not been included. Because insight into patient comprehension difficulties in this area may provide additional valuable information about the communication process, it is important to include this domain in measures of patient comprehension.

The objective of this study, therefore, was to assess patients' comprehension across several distinct domains of the ED visit and also to determine patients' awareness of deficiencies in their comprehension. The authors are not aware of any previous research that has specifically explored patients' insight into their comprehension deficits. This research is an important step in directing future efforts to identify and intervene with patients who demonstrate comprehension deficits.

Materials and Methods 

Methods of Measurement 

Enrolled subjects participated in an audiotaped interview that assessed their subjective understanding of the information provided by their medical team, their recall of their ED care and discharge instructions, their satisfaction with their care, and their behavioral intentions on discharge. Participants were allowed to refer to their discharge instructions during the interview. Subjects were compensated with a gift certificate ($5).

If it was not possible to conduct a face-to-face interview with the participant, the research assistant attempted to schedule a telephone interview. All telephone interviews were completed within 24 hours of discharge to minimize recall bias.

The interview guide was developed by all of the authors, who met as a group biweekly throughout study development and data collection. In addition, input was obtained from survey and qualitative research experts at our institutions. The interview guide was extensively tested with colleagues and friends, and their feedback was used to make changes to the wording of questions. Additional changes were incorporated after an initial phase of data collection, according to a review of preliminary results, as well as feedback from the research assistants.

During the interview, participants rated their perceived comprehension on a 5-point scale (poor to excellent) for each of 4 domains: (1) diagnosis and cause, (2) ED care (tests and treatments), (3) post-ED care (prescriptions, ancillary measures, and follow-up), and (4) return instructions. After 19 patients had been enrolled, an additional question assessing perceived difficulty of understanding was added in an attempt to provide a more sensitive indicator. Figure 1 provides the wording of both question types.

View Large Image Download to PowerPoint

Figure 1. Patient interview questions to assess perceived comprehension.

All interviews were transcribed in full. The accuracy of transcription was assessed by review of each transcript by an author (K.G.E.) immediately after transcription. If there were missing words or any areas of concern, then the complete audiotape was reviewed and appropriate changes were made to the document. Patient (or caretaker) comprehension was determined for each domain of the ED visit according to the concordance between direct patient recall and ED chart review. A 5-category concordance coding scale (no concordance, minimal concordance, partial concordance, near concordance, complete concordance) and specific guidelines were established by the consensus of all the authors after reviewing a subset of cases. The patient interview guide, as well as the concordance coding scale and guidelines, are provided as Appendix E1 and E2 (available online at http://www.annemergmed.com). Two authors (coders) rated each case independently. For the majority of cases, 1 physician coder and 1 nonphysician coder were assigned to each case. In a minority of cases, 2 physicians were assigned. We required that a physician code each case to help interpret whether or not the patient's description was consistent with the medical record. Discrepancies between the 2 scores for each case and domain were discussed by the 2 authors during a debriefing session, before each provided a final score. Coders did not have to agree on their final scores.

Results 

During the study period, 366 patients were approached and 175 agreed to participate (48%). Of the 175 patients who agreed to participate, 29 did not complete their telephone interview (25 were not able to be reached and 4 refused at the telephone contact). Thus, 146 participants were enrolled in the study. The 195 patients (52%) who refused to participate were asked several demographic questions, their chief complaint, and the reason for their refusal. A total of 358 patients were missed during our study period.

Of the 146 patients who were enrolled, 141 completed the interview (97%). For one patient, our audiotape recorder malfunctioned and the interview was lost. The remaining 4 interviews were incomplete because of interruptions (ie, patient's cab arrived, telephone conversation disconnected, patient no longer desired to continue). One patient who failed the Mini-Cog actually completed the interview, but this case was subsequently dropped. Thus, our total sample population is 140. Caretaker interviews were conducted for 2 cases. Telephone interviews were conducted for 36 cases (26%).

Demographic data for our sample and refusal populations are provided in Table 1. Education data were obtained only for the sample population. No differences were noted in the demographic data or overall satisfaction scores for the sample and refusal populations.

Table 1. Demographic data for sample and refusal populations.

	Demographic data	Sample (n=140)	Refusals (n=195)
	Age (y, SD)	39(15);range19–83	43(19);range18–93
	Sex (% female)	82(59)	107(55)
	Race (%)
	Black	27(19)	49(25)
	White	95(68)	131(67)
	Other	12(9)	10(5)
	Missing	6(4)	5(3)
	Education (%)
	Less than high school	14(10)
	High school graduate	35(25)
	College	64(46)
	Graduate school	26(19)
	Overall satisfaction (1–5 scale)	4.1	4.2

Initial comprehension ratings given by the 2 independent reviewers were in agreement 63% (353/560) of the time, with a range of 56% to 70% agreement across the 4 domains of the ED visit. Approximately 16% (34/207) of the disagreements reflected a lack of medical knowledge on the part of the nonphysician reviewer. Only 18% (38/207) of the disagreements reflected differences of greater than 1 category between the 2 reviewers. Four percent (8/207) of the disagreements were due to one reviewer giving the domain a score of NA (not able to score). The remainder of the disagreements was due to coder error, indicating that one coder had missed information or failed to follow a scoring guideline. None of the disagreements was due to differences in interpretation of the coding guidelines, and, after debriefing, there was 99% agreement in the final ratings given by the 2 reviewers.

Table 2 provides illustrative examples of what each comprehension rating represents, with data provided from patient interviews and corresponding ED chart review. We provide examples for each of the 4 domains of the ED visit, with ratings ranging from no concordance to near concordance.

Table 2. Examples of comprehension scores.

	Examples	Patient Interview	ED Chart Review	Omitted or Discordant Information	Rating Given
	Diagnosis and cause	No diagnosis, “They don't know.”	Pelvic inflammatory disease	Not aware of diagnosis documented by MD	No concordance
	ED care	Tests: blood tests with unknown results (patient states results not available for 3 days); culture test; Treatments: “shot to ease my stomach to stop the vomiting …that's all”	Tests: CBC, Chem7, LFTs, Amy/Lip, Ca/Mg/Phos—all NL; UA (5 WBC, 5 RBC, 3+ protein) with culture and sensitivity sent; Treatments: IVF 500 mL, Reglan IV, prednisone PO, ciprofloxacin PO	Tests: not aware of NL results for blood tests or abnormal UA results; Treatments: omits IVF, prednisone, ciprofloxacin	Minimal concordance
	Post-ED care	Meds: Vicodin Ancillary measures: none Follow-up: 4-5 days at clinic	Meds: Motrin, Vicodin if pain still uncontrolled Ancillary measures: drink plenty of fluids Follow-up: 5-7 days at clinic	Meds: omits Motrin Ancillary measures:omits fluids	Partial concordance
	Return instructions	“…vomiting or shortness of breath or diarrhea or something like that”	Return for fever, vomiting, chest pain, shortness of breath, worsening abdominal pain or other concerns	Omits fever, chest pain	Near concordance

CBC, complete blood cell count; Chem7, chemistry panel; LFT, liver-function tests; Amy/Lip, amylase/lipase; WBC, white blood cells; RBC, red blood cells; NL, Normal; IVF, intravenous fluid; UA, urinalysis.

Table 3 shows the distribution of comprehension ratings for all patients, across all 4 domains of the ED visit. Each patient contributed up to 4 comprehension ratings, 1 for each of the 4 domains of interest: diagnosis and cause, ED care, post-ED care, and return instructions. Our results indicate that the majority of domains were rated as complete concordance (62%, or 341 domains). However, for 25% (136 domains) of the scored domains, patients' comprehension was rated as near concordance, and for 14% (76 domains) their comprehension was partial concordance or worse (ie, minimal or no concordance).

Table 3. Distribution of comprehension scores.

	Distribution	Total Number of Scored Domains⁎ (%)	95% CI
	Complete concordance	341(62)	57–65
	Near concordance	136(25)	21–28
	Partial concordance (or worse)	76(14)	11–17

⁎ Seven domains were scored as not able to code and so are not included.

Although many patients demonstrated excellent comprehension (complete concordance) in 1 or more domains, only a minority of patients demonstrated such comprehension in all 4 domains. As shown in Figure 2, only 22% of patients had complete concordance in all 4 domains, meaning that 78% of patients demonstrated deficient comprehension in at least 1 domain. For 51% (71/140) of patients, we documented deficiencies in 2 or more domains. Five patients (4%) demonstrated deficient comprehension in all 4 domains. For those participants with a comprehension deficit in only 1 domain, nearly a third (32%) of the time, concordance was rated as partial or worse for the deficient domain. For those cases with 2 or more domains deficient, the majority (61%, or 43/71) demonstrated at least 1 domain of partial concordance or worse.

View Large Image Download to PowerPoint

Figure 2. Comprehension deficiencies. A deficient domain is one in which concordance was rated as less than complete concordance (ie, near, minimal, partial, or no concordance).

Table 4 shows the distribution of comprehension deficiencies across the 4 domains of the ED visit. Patients had the greatest difficulty with post-ED care, and 34% of comprehension deficiencies occurred in this domain. Patients seemed to have the least difficulty recalling their diagnosis, with only 15% of deficiencies occurring in this domain.

Table 4. Comprehension deficiencies by domain of ED visit.

	Comprehension	Total Number of Deficient Domains⁎ (%)	95% CI
	Diagnosis and cause	32(15)	11–21
	ED care	61(29)	23–35
	Post-ED care	73(34)	28–41
	Return instructions	46(22)	16–28

⁎ A deficient domain is one in which concordance was rated as less than complete concordance (ie, near concordance, minimal concordance, partial or no concordance).

A sensitivity analysis comparing interviews conducted immediately after discharge to those performed by telephone revealed no significant differences in the distribution of comprehension (concordance) scores or the proportion of participants with comprehension deficiencies. For telephone interviews, comprehension deficiencies occurred with the greatest frequency in the domain of ED care (53%), whereas post-ED care was a similar proportion (39%) and diagnosis and cause was a smaller proportion (11%).

Table 5 presents the results for subjective understanding for all patients (or caretakers) and shows how these results correspond with comprehension ratings. In the first row, the table shows that study participants perceived comprehension deficits or difficulty with understanding in 21% of all scored domains. The next 2 rows divide the scored domains into 2 groups: those with complete concordance and those with less than complete concordance. The table shows that patients perceived comprehension deficits or difficulty in a similar proportion of these subgroups (22% and 20%, respectively).

Table 5. Awareness of comprehension deficiencies.

	Awareness	Total Number	Number With Perceived Deficit/Difficulty (%)
	Total scored domains	553⁎(140 Patients×4 domains)	118(21;95%CI18–25)
	Domains with complete concordance	341	76(22;95%CI18–27)
	Domains with less than complete concordance	212	42(20;95%CI15–26)

⁎ Seven domains were scored as not able to code and so are not included.

Overall, our findings indicate that the majority of patients with comprehension deficits failed to perceive them. Patients perceived difficulty with comprehension for only 1 in every 5 domains (20%) that were found to have less than complete concordance. This trend was consistent across all 4 domains of the ED visit, with less than 30% of patients indicating comprehension difficulties for domains with deficiencies. Although participants receiving higher comprehension ratings (near concordance) were less likely to perceive comprehension difficulties than those with the lowest scores (minimal or no concordance), the majority of all deficiencies were not perceived by patients (or caretakers) even for the lowest ratings.

Twenty-two percent of patients (or caretakers) perceived comprehension difficulties for domains given scores of complete concordance (Table 5). Among those patients with perfect scores (ie, complete concordance for all 4 domains), 39% (12/31) perceived difficulty with comprehension in at least 1 domain.

Limitations 

Our study has several important limitations. First, it was conducted at 2 teaching hospitals in southeastern Michigan, limiting its generalizability. Second, our comprehension scores may have been subject to variation across cases because the same 2 authors did not score each case. We tried to minimize this concern and standardize the scoring process by establishing a detailed set of guidelines for scoring. In addition, each case was scored independently by 2 authors and discrepancies were discussed before final scoring. Third, concordance coding (to obtain comprehension scores) was limited in some cases by the number and detail of follow-up questions asked by the research assistant during the interview. To minimize subjective judgment, we established clear guidelines on how to handle these situations. In extreme cases in which significant information was missing because of inadequate follow-up questions, a score of NA (not able to code) was given. In those cases in which the patient (or caretaker) provided incomplete answers, a score was provided according to what was available from the transcript and patients were not scored lower for information that was lacking because of inadequate follow-up questioning. This approach would, if anything, overestimate comprehension scores. Fourth, the use of telephone interviews introduced the possibility of decreased recall compared to those interviewed at ED discharge because of the delay in the interview. Although sensitivity analysis did not reveal any significant differences between these groups, the increased frequency of comprehension deficiencies for ED care in telephone interviews likely reflects rapid deterioration in recall for this information with time. Fifth, our questions about patient perception of comprehension and difficulty with understanding were likely subject to bias. Some patients (or caretakers) might have felt uncomfortable or embarrassed about reporting comprehension difficulties even if they perceived them. Because the intention of our study was to consider whether or not one can identify patients with comprehension deficits by asking them directly, this is an acceptable limitation.

Discussion 

In this sample of ED patients from 2 health systems, many patients have poor comprehension of multiple aspects of their ED care and discharge instructions. Seventy-eight percent of the patients demonstrated a comprehension deficiency in at least 1 domain of their ED visit. These deficits were most common for the category of post-ED care, raising significant concerns about patients' ability to adhere to discharge instructions and recommendations after leaving the ED. Moreover, our study suggests that we cannot simply ask patients to identify their comprehension deficiencies because the majority did not report difficulties in areas in which deficits were objectively demonstrated.

Despite increased concern during the past decade about the importance of communication in the ED, the high rates of deficiencies in patient (or caretaker) comprehension we found are similar to those found in studies conducted in the 1990s by Logan et al8 and Spandorfer et al.9 In the former study, 153 adult patients were interviewed after ED discharge and 63% were found to demonstrate incorrect answers in at least 1 part of the patient's discharge instructions (diagnosis, home care, and follow-up plan). In the second study, Spandorfer et al9 found that for each of 3 categories of the ED visit (diagnosis, medications, and follow-up instructions), approximately 40% of 217 adult patients demonstrated deficient understanding of discharge instructions. It is likely that the frequency of comprehension deficits was somewhat higher in our study largely because our study design considered an additional category for ED care and also included a scoring system for partial but incomplete comprehension.

Our study builds on previous findings by assessing patient understanding across the entire ED visit, rather than just considering recall of the diagnosis and discharge instructions. It remains unclear which of the identified deficits are most clinically relevant. Future work will need to explore how the domain of the deficit and the score correspond with outcomes, such as adherence, recidivism, and morbidity.

Although we found comprehension deficits across all domains of the ED visit, it is striking that the highest frequency arose in the category of post-ED care (medications, ancillary measures, and follow-up). A similar trend has been found in previous studies that have demonstrated that patients have greater difficulty recalling information about their medications and home care than their diagnoses.6, 7, 8 A provocative example from our study is one patient who was diagnosed with pelvic inflammatory disease and did not understand that she was prescribed an antibiotic.

Although we are not aware of studies demonstrating a direct relationship between patient comprehension, adherence, and outcomes, there are some early data to support this. In a pediatric ED study by Grover et al,7 parents' ability to recall their child's follow-up appointment time(s) at discharge was the only variable found to be associated with compliance with these appointments. Such findings emphasize the importance of identifying why patients demonstrate the greatest comprehension deficits for post-ED care instructions. It is possible that some of the difficulty stems from the fact that much of the post-ED care information is provided at the end of the visit, when the patient is anxious to leave and may feel less inclined to ask questions. In addition, it is likely that in some instances these instructions are never discussed but simply written on discharge paperwork that the patient is unaware of, unable to read, or chooses to ignore.

Our study takes an important step in exploring comprehension difficulties in the ED by assessing patients' awareness of identified deficits. It is of great concern that only about 20% of patients reported comprehension difficulties for domains with deficiencies. Overall, it is important to keep in mind that the results for our perceived comprehension assessment are subjective. In such settings, patients' answers are heavily influenced by experience, as well as personal attitudes and standards for what they need and want to understand. This bias makes it inappropriate to place value on relative scores and limits us to using this tool simply as a means for identifying those who would be likely to identify comprehension difficulties if asked. Our findings would strongly suggest that, if we rely on such an approach, we will miss the majority of patients with comprehension deficits.

These findings are generally consistent with results of work by Crane6 in 1997. In this study of 314 patients, comprehension rates were found to be 59% (average across all parts of the discharge instructions), whereas greater than 90% of patients reported that they had been provided adequate explanations of their instructions after an ED visit.

Overall, our study findings suggest that we need better strategies for identifying patients who are having difficulty understanding their care and instructions in the ED. One strategy might be to do what we did in this study: test patients' recall immediately after discharge, through strategies such as asking patients to repeat information in their own words. Recent research in primary care settings indicates that although physicians rarely assess patients' recall or comprehension of new concepts during outpatient visits, diabetic patients of physicians who consistently make these assessments with strategies such as “closing the loop” have significantly better glycemic control than patients of physicians who do not.25 This technique may improve communication by making physicians aware of comprehension deficits and thereby facilitating individualized patient education.25 This study found that physicians' assessments of patient recall did not prolong visits, perhaps because physicians immediately identified areas of poor comprehension and focused discussion.

Efforts to anticipate, identify, and address communication failures are critical to improving patient care. As part of our work, we plan to explore physicians' ability to predict patients' deficits, as well as specific visit and patient characteristics that are associated with poor comprehension. In addition, the content and organization of discharge instructions should be considered as a possible means of improving comprehension. Instructions may help to improve understanding if they clearly emphasize all domains of the visit (diagnosis, ED care, post-ED care, and return instructions).

View Large Image Download to PowerPoint

Part 1.

View Large Image Download to PowerPoint

Part 2.

View Large Image Download to PowerPoint

Part 3.

View Large Image Download to PowerPoint

Part 4.

View Large Image Download to PowerPoint

Part 5.

View Large Image Download to PowerPoint

Part 6.

View Large Image Download to PowerPoint

Part 7.

View Large Image Download to PowerPoint

Part 8.

View Large Image Download to PowerPoint

Part 9.

View Large Image Download to PowerPoint

Part 10.

View Large Image Download to PowerPoint

Part 11.

View Large Image Download to PowerPoint

Part 12.

View Large Image Download to PowerPoint

Part 13.

View Large Image Download to PowerPoint

Part 14.

View Large Image Download to PowerPoint

Part 15.