Resident Recognition of Low Literacy as a Risk Factor in Hospital Readmission

Abstract

BACKGROUND

Low literacy is associated with poor self-management of disease and increased hospitalization, yet few studies have explored the extent to which physicians consider literacy in their patient care.

OBJECTIVE

To examine trainee recognition of low literacy as a potential factor in patient adherence and hospital readmission.

DESIGN AND PARTICIPANTS

Randomized study of 98 Internal Medicine residents and medical students. Trainees reviewed a case history and completed a questionnaire pertaining to a fictional patient's hospital readmission. Case version A contained clues to suggest limited patient literacy skills, while version B did not. Responses were reviewed for mention of low literacy and educational strategies recommended for low-literate patients.

RESULTS

Few trainees raised the possibility of low patient literacy, even when provided clues (25% in Group A vs 4% in Group B, P=.003). Furthermore, while most trainees listed patient education as an important means of preventing another readmission, only 16% suggested using a strategy recommended for low-literate adults.

CONCLUSION

Few trainees recognized low literacy as a potential factor in patient nonadherence and hospital readmission, and few recommended low-literate educational strategies. Medical residents and students may benefit from additional training in the recognition and counseling of low-literate patients.

Low patient literacy is associated with poor disease-related knowledge and self-management, worse self-reported health status, poor treatment adherence, a 30% to 50% increased odds of hospitalization, and higher annual health care costs.^1–11 Physicians tend to overestimate their patients' literacy level, and despite abundant evidence, may not consider limited literacy skills as a factor in patient care.^12–14 Experts recommend educating low-literate patients through the use of simple language, having patients repeat back instructions to confirm understanding, and use of assistance devices, such as pillboxes, graphic medication schedules, pictographs, and other multimedia aids.^14–16

In this study, we used a fictional case history to examine the factors that medical students and residents considered responsible for the nonadherence and readmission of a patient with congestive heart failure, as well as the strategies they proposed to encourage patient adherence. We hypothesized that trainees would not recognize low literacy as a risk factor for nonadherence and hospital readmission unless prompted that the patient may have trouble reading and understanding health-related information.

METHODS

Setting and Participants

The study took place at a large, university-affiliated Internal Medicine training program whose primary training site is an inner-city hospital. All medical students and residents attending a noon educational conference on hospital readmission were eligible to participate. The conference was held at all 4 teaching hospitals associated with the program.

Study Design

Conference attendees were asked to review a case history describing a retired mechanic being readmitted for an exacerbation of congestive heart failure after not taking his medications appropriately and missing a follow-up appointment. Participants were unaware that 2 versions of the case existed. The scenario in version A suggested that the patient relied upon his daughter to read his pill bottles and appointment slips. version B contained no such text (see Appendix A). Prior to distribution, the cases were stacked in random order, as per a computer-generated sequence. Randomization was stratified by teaching hospital. The cases and an accompanying survey were then distributed sequentially to participants at the time of enrollment. Participants who received version A of the case history were considered randomized to Group A, and those who received version B were considered randomized to Group B. Participants were told that the study concerned hospital readmission, with no mention of low literacy. The study design and materials were approved by the Emory University Institutional Review Board, and all participants provided consent.

The 1-page survey contained primarily open-ended questions (see Appendix B). Trainees listed the factors that they believed contributed to the patient's nonadherence and hospital readmission and the measures they would take to prevent another hospitalization. They also provided demographic data including gender and year in training. The study, which took approximately 20 minutes to complete, was followed by an educational conference on preventing hospital readmission, with emphasis on recognition of low literacy.

Data Analysis

Survey responses were reviewed for specific mention of low patient literacy, including “poor reading skills,”“low literacy,”“illiteracy,” or similarly clear reference to low literacy. A secondary analysis identified surveys in which lack of patient education was considered a contributing factor, even if literacy was not specifically mentioned. We also reviewed surveys for suggested use of a recommended low-literate technique (e.g., simple language, confirmation of patient understanding, “teach-back” technique) or assistance device (e.g., pillbox, graphic medication schedule).^14–16 One author (C.P.) coded all surveys using prespecified criteria, and the other (S.K.) reviewed the scoring. Any differences were resolved by discussion. Descriptive statistics and χ² tests were performed using SAS 6.12 software. α was set at 0.05.

RESULTS

Ninety-eight trainees completed surveys (response rate 95%), including 11 medical students, 43 interns (postgraduate year (PGY)-1), and 43 residents (PGY-2 and PGY-3). One survey was returned without response to the year in training question. There were similar percentages of male (54%) and female (40%) responders, with 4 nonresponders to the question on gender (Table 1).

Table 1.

Description of Participants and Summary of Responses for Subjects Prompted (Group A) Versus Not Prompted (Group B) to Low Patient Literacy

	Group A (N=48)	Group B (N=50)
Gender, no. (%) male†	22 (46)	32 (64)
Year in training†
Residents, no. (%)	18 (38)	25 (50)
Interns, no. (%)	22 (46)	21 (42)
Medical students, no. (%)	7 (15)	4 (8)
Considered hospitalization preventable, no. (%)‡	43 (100)	44 (100)
Estimation of complexity of medication regimen, mean‡§	4.9	4.8
Estimation of adherence, mean‡¶	3.5	3.4
Mentioned literacy as a potential factor in readmission, no. (%)*	12 (25)	2 (4)
Suggested educational strategy recommended for low-literate adults, no. (%)**	13 (27)	3 (6)

^†Percentages relative to total subjects in each group. Gender not reported by 4 participants (3 in Group A, 1 in Group B). Year in training not reported by 1 participant in Group A.

^‡Percentages relative to number of valid responses for each item. Preventability of hospitalization not provided by 11 participants (5 in Group A, 6 in Group B). Regimen complexity and patient adherence not assessed by 1 participant in Group B and 1 participant in Group A, respectively.

^§On a Likert scale of 1 “not at all complex” to 7 “very complex.”

^¶On a Likert scale of 1 “not at all compliant” to 7 “very compliant.”

^*P=.003

^**P=.005

Overall, only 14% of trainees suggested low patient literacy as a factor in readmission or nonadherence. Among those prompted by version A of the case history, 12 (25%) mentioned that the patient may have poor literacy, versus 2 (4%) in Group B (P=.003) (Table 1). Residents were more likely than interns to comment on low patient literacy (10, or 23% vs 2, or 5%, P=.013).

Most trainees (71%) mentioned educating the patient as a means of preventing hospital readmission, primarily by suggesting written instructions. Only 16% of trainees specifically suggested using an educational strategy recommended for low-literate adults. Group A was more likely to do so (13, or 27% vs 3, or 6% in Group B, P=.005) (Table 1). Only 7 of the 14 trainees (50%) who mentioned low literacy as a possible contributing factor suggested an educational technique recommended for low-literate patients. These responses did not vary by year in training.

The secondary analysis revealed similar results. A larger number of trainees, 37 (38%), recognized that lack of patient education may have contributed to the nonadherence and hospital readmission. Trainees more frequently mentioned patient education as a contributor when prompted by the case history (25, or 52% in Group A vs 12, or 24% in Group B, P=.004). Again, residents were more likely than interns to comment on this issue (26, or 60% vs 7, or 16%, P<.001). Only 11 of the 37 trainees (30%) who mentioned patient education as a possible contributing factor recommended an educational technique recommended for low-literate patients.

DISCUSSION

When reviewing a common clinical scenario of nonadherence and hospital readmission, few residents and interns mentioned the patient's literacy skills as a potential contributing factor. Providing hints that the patient might have limited literacy skills did improve recognition of low literacy, but rates of recognizing and appropriately dealing with literacy issues remained low. Even when residents recognized low literacy, half still did not recommend an appropriate means of patient education, recommending only written instructions.

Residents were significantly more likely than interns to recognize and consider low literacy as a factor in adherence, probably as a result of additional clinical experience with low-literate patients. However, they were not more likely to suggest a recommended educational technique.

Our results are congruent with other reports that physicians overestimate the reading ability of their patients and infrequently use recommended educational strategies. Bass showed that residents perceived only 10% of their patients to have literacy problems when actually greater than 1/3 did.¹² Similarly, Schillinger et al.¹⁴ demonstrated that only 20% of physicians confirm patient understanding by having the patient repeat back instructions.

Despite the recommendations of 2 national expert panels that health literacy training be added to medical curricula, few such training programs are described in the medical, public health, and education literature.^17,18 While our study suggests that some awareness of literacy develops naturally during residency, the need for health literacy curricula remains high. Such education should promote awareness of how to recognize low-literate patients, as well as strategies which may enhance understanding and adherence in this population.

There were limitations to our study. First, the investigators were not blinded to the version of the case history, A versus B, when reviewing survey responses. However, trainees had to make clear reference to low literacy to be considered having recognized low literacy as a contributing factor. Second, the prompts about low literacy in the case history may have been too subtle, or not reflective of the clues that can be gathered during a real-life patient encounter. However, the significant difference between Groups A and B suggests that the prompt was effective. Further, in patient care, clues about the patient's reading ability may be subtle, as patients are commonly ashamed of their difficulty.¹⁹ Third, residents may also use recommended techniques, such as using simple language, in real-life encounters without considering them specific strategies to prevent readmission. Fourth, the study was limited to a single institution. It is possible that recognition of low literacy would be even less common at other institutions, as the present group of trainees has extensive clinical exposure to a low-literate population, as well as faculty with expertise in health literacy and physician-patient communication. Another consideration is that, despite consensus among experts that certain strategies are appropriate for educating low-literate patients, there is limited published evidence confirming the effectiveness of these techniques.^14–16

This investigation raises several questions for further research. What techniques do residents use to teach their low-literate patients? How can these skills be implemented in medical school and residency training? What impact will use of these skills have on patient outcomes and hospital readmission rates? We have begun addressing some of these questions at our institution through a health literacy training program for Internal Medicine residents.

In conclusion, our study suggests that medical students and residents do not routinely recognize low literacy as a potential factor in patient nonadherence and hospital readmission. Even when they consider literacy, trainees still commonly recommend written instructions without supplementing these instructions with strategies recommended for educating low-literate patients. These findings indicate a need for further training in the recognition of poor reading skills, the concept of health literacy, and the use of recommended educational strategies.