English-Language Competency Using Written Tests of Health Literacy: Limitations

English-Language Competency Using Written Tests of Health Literacy Limitations

The major limitation of this study was the inability to directly assess a person’s verbal English. The use of a written test instead of an oral language test probably overestimated the number of patients who speak English, because verbal comprehension precedes written abilities. Although neither language competency test had been validated in the ED, the REALM and the STOFHLA were chosen because of the brevity and ease of administration. Davis and others had found that the REALM takes 1-5 minutes to perform by personnel with minimal training and could be used to estimate patient literacy for use in primary care patient education and medical research. In another study, the same authors found that the REALM provides a good estimate of a patient’s reading ability, with concurrent validity when measured against standardized reading tests. STOFHLA was found to be a valid and reliable indicator of a patient’s ability to read health-related information. We did not use a comparative group in this study, limiting its applicability. The study was also limited by the testing time and number of questions. However, we felt it was best to perform validated tests, whatever their length, for the study.

A sixth-grade reading level was used as the cutoff for English-language competency based on studies of minimal competency in English speakers for understanding healthcare-related communication. Powers found that 40% of the ED patients could not read at the eighth-grade level, and 20% were considered functionally illiterate. This is unfortunate, since ED patient-directed materials were found to range from eighth-to-13th grade on the Fry index. In a review of the readability of consent forms, Mader and Player found that the mean readability level for informed consent was 10th grade in the 88 informed consents that were examined. Discharge instructions were found by Spandorfer to be written at the 1 lth-grade reading level, although most patients were found to have a mean reading ability of sixth grade.

We used Hispanic surnames as the means to identify Hispanic patients. However, it would be more exact to ask all patients about their language background. We did not differentiate the patient’s country of origin, a fact that may have influenced their abilities. The study was also limited by the inner-city sample used in the study. The study was limited by the convenience enrollment method, rather than the use of a consecutive sample. Those participants who did not complete the test limited the study. It is uncertain whether this study has broader applicability to other populations or subsets of Hispanic patients. The study used the physicians’ and nurses’ interpretation of the patients’ ability to speak English, and no training in the process was performed; no formal tests were offered the physicians and nurses to make this determination.

We were not able to determine if some patients refused to take the test because of its length or concern about giving proper information about themselves. There was no attempt to coerce any patient who stated they did not speak English or refUsed to participate. The study did not attempt to confirm the legal status of any of the participants or question rationale for nonen-rollment. There was no change in care or delay in care for those that participated in the study.