Health literacy assessment instruments: literature review

Bibliographic searches were performed in the electronic databases of PubMed, Web of Science and BVS – LILACS. DeCS (Health Science Descriptors) and MeSH (Medical Subject Headings) descriptors were used, combined by Boolean operators.

Selection criteria

Articles published between 2009 and 2014 in Portuguese, English or Spanish, that used general health literacy assessment instruments or instruments that allow this type of analysis. Articles that did not approach health literacy in the title or abstract were excluded, as well as articles that did not cite health literacy assessment instruments and outcomes, repeated articles, reviews, case studies and series of cases. Two reviewers analyzed and selected the articles.

Results

A variety of health literacy assessment instruments was identified. The most frequently used instruments were the Test of Functional Health Literacy in Adults and the Rapid Estimate of Adult Literacy. The design of most studies was cross-sectional observational, with greater scientific output in the United States and Europe. Categorization of health literacy into levels and presentation of the respective frequency distributions prevailed. Speech, Language and Hearing Sciences studies that include health literacy assessment are scarce.

Conclusion

Researchers have utilized a variety of instruments and methods to assess health literacy, with predominance of clinical screening tests focusing on functional health literacy, administered in the setting of outpatient clinics and hospitals.

Health literacy; Health evaluation; Methodology; Review; Speech, language and hearing sciences

Authors, year	Country	Design	Casuistics	Instruments*	Results
Sleath et al., 2014⁽¹¹⁾	United States of America	Multi-center, cross-sectional, observational.	228 English-speaking adults with glaucoma in self-administration of medications.	REALM	14.5% of patients obtained a score of 0-60 in REALM and 83.3% obtained a score of 61-66. There was no association between HL and adherence to medication.
Izard et al., 2014⁽¹²⁾	United States of America	Cross-sectional observational.	50 patients with prostate cancer (aged ≥ 21 years) and 50 health professionals.	¹⁾ REALM-SF ²⁾ Graphical Literacy Scale	Mean score in REALM-SF: 6.8 (maximum: 7, SD: 1.0), equivalent to high HL.
van der Vaart et al., 2014⁽¹³⁾	The Netherlands	Pretest/ posttest.	360 patients with rheumatoid arthritis aged 20-86 years.	¹⁾ MFCCHL– Ishikawa (Version in Dutch) ²⁾ 5-item version of PEPPI-5	Mean score in HL: 38.6 (scale from 14 (low HL) to 56 (high HL)) (SD: 7.2). Significant association between portal use and HL: Higher HL = higher frequency of access to the portal.
Tsai et al., 2014⁽¹⁴⁾	Taiwan	Cross-sectional observational.	347 women living in the regions of great Taipei or Taoyuan (north of Taiwan).	¹⁾ THLS-modified version ²⁾ HPLP Chinese	Mean score in HL: 2.60 (maximum: 4, SD: 0.55) = moderate HL. 33.2% of the participants presented inadequate HL and 66.8%, adequate HL. There was an association between HL and health-promoting behaviors.
Wei et al., 2014⁽¹⁵⁾	Taiwan	Cross-sectional observational.	752 Taiwanese adults (372 from Taipei and 380 from Hualien).	MHLS	Mean score in MHLS: 43.2 (maximum: 50, SD: 6.6). Adequate HL: 68.6%, marginal HL: 25.3%, and inadequate HL: 6.1%.
Smedberg et al., 2014⁽¹⁶⁾	15 European countries	Multi-center, cross-sectional, observational.	8,344 pregnant women or women with children younger than one year.	SBSQ	HL among non-smokers and smokers: low: 5.0% and 7.4%, medium: 39.7% and 47.4%, high: 55.3% and 45.1% (p-value <0.001). Women with low HL were more likely to continue smoking during pregnancy.
Lupattelli et al., 2014⁽¹⁷⁾	Countries in Europe, Oceania, and the Americas	Multi-center, cross-sectional, observational.	4,999 pregnant women.	SBSQ	High HL: 54.5%, medium: 40.3% and low: 5.2%. HL was significantly associated with non-adherence to medication, risk perception and beliefs about the medication.
Nesbitt et al., 2014⁽¹⁸⁾	United States of America	Cross-sectional observational.	612 patients with cardiac insufficiency, age ≥ 18 years, rural area.	STOFHLA	Mean score in HL: 25.5 (maximum: 36, SD: 8.8). HL had no influence on health-related quality of life.
Hirsh et al., 2014⁽¹⁹⁾	United States of America	Randomized clinical trial.	300 patients with rheumatoid arthritis, age ≥ 18 years.	STOFHLA	28% of patients with inadequate or marginal HL. Mean score in HL: 77.6 (maximum: 100; SD: 23.5%). Limited HL is a predictor of the presence of significant discrepancies between verbal and written assessments.
Riebl et al., 2013⁽²⁰⁾	United States of America	Cross-sectional, observational - pilot study	60 adults aged ≥ 21 years.	NVS	Mean score in HL: 5.2 (ranges from 0-6: 4-6 = adequate HL). Individuals with lower level of schooling presented lower HL scores.
Inoue et al., 2013⁽²¹⁾	Japan	Cross-sectional observational.	269 patients with type 2 diabetes, age ≥ 20 years and < 75 years.	¹⁾ MFCCHL– Ishikawa ²⁾ Question with a Likert scale.	Functional HL: 3.36 (SD: 0.58); communicative: 2.65 (SD: 0.68); critical: 2.28 (SD: 0.59). HL and patient-physician communication were associated with the understanding and management of diabetes.
Jovic´-Vranes et al., 2013⁽²²⁾	Serbia	Cross-sectional observational (instrument validation).	120 patients, age ≥ 18 years.	¹⁾ TOFHLA ²⁾STOFHLA (Serbian versions)	Adequate HL: 59.0% and 82.9%, limited HL: 41.0% and 17.1% (respectively in TOFHLA and STOFHLA). Sex, age, level of schooling, self-perception of health and presence of chronic disease were associated with HL.
Kiser et al., 2012⁽²³⁾	United States of America	Randomized clinical trial.	99 adult patients with chronic obstructive pulmonary disease.	STOFHLA	Low HL: 36%. Patients with low or high HL benefited from the intervention of self-management of an inhalation technique.
Anger et al., 2012⁽²⁴⁾	United States of America	Cross-sectional observational - pilot study	36 women with pelvic floor disorder, aged between 42 and 94 years.	TOFHLA	Mean score in HL: 93 (maximum 100). HL score decreased as age increased. Patients had low recall of the diagnosis and poor understanding of their pelvic floor condition, despite the fact that 97.2% of them presented high HL.
Coughlan et al., 2012⁽²⁵⁾	Ireland	Cross-sectional observational - pilot study	199 patients, age ≥ 18 years, attending an anticoagulation clinic.	REALM	Median of the HL score: 64 (maximum: 66). Low HL: 19.1%, adequate HL: 80.9%. There was an association between HL and level of schooling.
Apolinário et al., 2012⁽²⁶⁾	Brazil	Cross-sectional observational (instrument validation).	226 elderly patients, age ≥ 60 years, both sexes.	SAHLPA (full and short versions: SAHLPA-50 and SAHLPA-18)	Score mean: 37.7 (SD: 9.0). Inadequate HL: 66% (SAHLPA-50). The developed instrument presents good validity and consistence.
Eriksson-Backa et al., 2012⁽²⁷⁾	Finland	Cross-sectional observational.	281 elderly patients aged 65-79 years, both sexes.	Questionnaire adapted from eHEALS	Greater vulnerability to the obtention/use of health information: low level of schooling and self-perception of health, disinterest and low proactivity in the search for health information.
Convery et al., 2011⁽²⁸⁾	Australia	Cross-sectional observational.	80 adults with hearing loss aged 45-90 years, both sexes.	STOFHLA	Mean score in HL: 34 (adequate HL), (maximum: 36; SD: 4.61). Higher HL levels were associated with increased likelihood of completing the task in an independent and successful way.
van der Vaart et al., 2011⁽²⁹⁾	The Netherlands	Cross-sectional observational.	227 patients of a rheumatology clinic aged <70 years.	MFCCHL– Ishikawa (Version in Dutch)	Mean score in the functional HL: 3.4 (SD: 0.55), communicative: 2.8 (SD: 0.64) and critical: 2.0 (SD: 0.67).
Myaskovsky et al., 2011⁽³⁰⁾	United States of America	Multi-center, cross-sectional, observational.	275 subjects with spinal cord injury, wheelchair users for at least one year, age ≥16 years.	¹⁾ REALM – Revised. ²⁾ 9-item Healthcare System Distrust Scale.	Mean score in HL: 7.37 (maximum: 8, SD: 1.55). Afro-Americans obtained a lower HL compared to whites.
Hahn et al., 2011⁽³¹⁾	United States of America	Cross-sectional observational (instrument validation).	608 English speakers, primary care patients, age ≥21 years.	Health LiTT	The proportions of correct answers in the Health LiTT ranged from 30% to 95%. The test provides reliable and valid scores. Lower scores were associated with lower income and level of schooling, absence of previous use of computers and self-report of poor reading skills.
Rawson et al., 2010⁽³²⁾	United States of America	Cross-sectional observational (instrument validation).	155 patients referred to outpatient cardiology.	¹⁾ REALM ²⁾ METER	Mean score: METER: 36.1 (SD: 5.0; maximum: 40); REALM: 62.2 (SD: 6.5; maximum: 66). METER measures HL in a fast and practical way.
Fink et al., 2010⁽³³⁾	United States of America	Multi-center randomized clinical trial.	502 patients with indication for elective surgeries.	REALM	HL mean score: 62.3 in the group with the repeat back technique and 62.5 in the group without the technique.
Fink et al., 2010⁽³⁴⁾	United States of America	Multi-center randomized clinical trial.	502 patients with indication for elective surgeries.	REALM	Patients with limited HL: 16% with score <60 and 2% with score <40 in REALM. HL was not significantly associated with understanding.
Sox et al., 2010⁽³⁵⁾	United States of America	Cross-sectional observational.	15 parents/guardians of students undergoing diagnostic analysis for ADHD	TOFHLA	HL score ranged from 79 to 100. The inclusion of user support items in the developed application was based on HL or on experience of technology.
Ishikawa et al., 2009⁽³⁶⁾	Japan	Cross-sectional observational.	134 patients with type 2 diabetes being regularly monitored at an outpatient clinic.	¹⁾ MFCCHL– Ishikawa ²⁾ PPPM - adapted ³⁾ IPCDP - adapted	Functional HL: 3.4, communicative: 2.5 and critical: 2.0 (maximum = 4). HL is related to patient-physician communication and to patient’s perception of the process.
Carthery-Goulart et al., 2009⁽³⁷⁾	Brazil	Cross-sectional observational (instrument validation).	312 healthy patients, age ≥ 18 years, both sexes.	STOFHLA (in Portuguese)	32.4% of patients with inadequate or borderline performance in the instrument. Among elderly patients, 51.6%. Association between S-TOFHLA results and level of schooling. The instrument is adequate to the Brazilian population.

Instrument	Characteristics	Versions / Languages
Test of Functional Health Literacy in Adults (TOFHLA) ^{(22,24,38,39,40)}	Assessment object: Functional health literacy. Assessment method: Self-administered, with previous oral instructions. For reading comprehension items, sentences with absent words are presented. Subjects select one out of four answer options to fill in the gaps. Texts are extracted from real materials from the health context, such as medical forms and instructions for health examinations. For mathematical literacy items, materials related to medical prescriptions are presented and the patient is required to answer the questions orally. Assessment items: TOFHLA: 50 reading comprehension items and 17 numerical skills items. STOFHLA: 36 reading comprehension items and four numerical skills items. Administration time: 22 minutes for TOFHLA and 12 for STOFHLA. Classification of outcomes: For TOFHLA, the score ranges from 0 to 100 (0 to 50 for reading comprehension, the same for mathematical literacy). The results establish levels of functional health literacy. Inadequate (0-59 points), marginal (60-74 points) and adequate (75-100 points). For STOFHLA, the score also ranges from 0 to 100: 0 to 28 in the mathematical literacy items (each item is worth 7 points) and 0 to 72 in the reading comprehension items (each item is worth 2 points). Levels of functional health literacy: Inadequate (0-53 points), marginal (54-66 points) and adequate (67-100 points).	1.TOFHLA: Original version in English. 2. TOFHLA-Spanish: Version in Spanish. 3. STOFHLA (Short TOFHLA): Short version. 4. STOFHLA- Spanish: Short version in Spanish. 5. TOFHLA and STOFHLA in Serbian⁽²²⁾. 6. STOFHLA in Portuguese⁽³⁷⁾.
Rapid Estimate of Adult Literacy in Medicine (REALM) ^(26,41,42)	Assessment object: Functional health literacy. Assessment method: The patient reads out loud a list of 66 medical terms organized in increasing order of difficulty. Each word pronounced correctly is worth 1 point. Assessment items: 66 terms in the area of health. Administration time: 2 to 3 minutes. Classification of outcomes: The score ranges from 0 to 66 and identifies health literacy as inadequate (0-44 correct answers), limited (45-60 correct answers) and adequate (61-66). REALM attributes 1 out of 4 estimated literacy degrees: 1) 3^rd series and below (score from 0 to 18), 2) 4^th to 6^th series (19 to 44 points), 3) 7^th to 8^th series (45 to 60 points) and 4) 9^th series and over (61 to 66 points).	1. REALM short form⁽⁴³⁾. 2. REALM – Revised: Version with 8 items⁽⁴¹⁾. 3.SAHLSA -50 and 18: Versions adapted to Spanish, including comprehension assessment⁽²⁶⁾. 4. SAHLPA 50 and 18: Version of SAHLSA adapted to Portuguese.
Measure of functional, communicative and critical health literacy (MFCCHL– Ishikawa) ⁽⁷⁾	Assessment object: Functional, communicative and critical health literacy, developed for patients with type 2 diabetes. The objective is to assess the ability to extract, understand and use health-related information. Assessment method: Self-administered questionnaire of health literacy scales. If the patient requests, he can be helped by an assistant. Assessment items: 3 sub-scales with 14 answer items: 5 items for functional HL, 5 for communicative HL, and 4 for critical HL. Administration time: Not specified. Classification of outcomes: Each item is assessed according to a 4-point Likert scale, ranging from 1 (never) to 4 (frequently). The score of the items of each scale is added and divided by the number of items in that scale. Therefore, the score in each sub-scale ranges from 1 to 4. The scores are inverted for the functional HL sub-scale so that higher scores indicate higher health literacy.	1. (MFCCHL– Ishikawa): Version in Dutch.
Set of Brief Screening Questions (SBSQ) ⁽⁴⁴⁾	Assessment object: Health literacy (without specification of HL level). The test enables to identify patients with inadequate HL. Assessment method: Interview. Assessment items: Three screening questions with answers according to a 5-point Likert scale. Administration time: There is no maximum time stipulated for filling in the questionnaire. Classification of outcomes: HL classified as inadequate, marginal or adequate. The test’s authors report that the optimal cutpoint depends on the test’s accuracy and prevalence of inadequate HL, as well as on tests’ costs and classification of false positives. Other studies^(12,13) in which answers in a 4-point Likert scale were used established a score ranging from 0 to 12 and classified HL as low (0 to 5 points), medium (6 to 9 points) and high (10 to 12 points).	1. SBSQ with 3 questions and answers in a 4-point Likert scale. Classification of HL in low, medium and high^(16,17). - Remark: Studies have administered the test in up to 18 countries and translated it to their respective languages¹⁷.

Academia Brasileira de Audiologia Rua Itapeva, 202, conjunto 61, CEP 01332-000, Tel.: (11) 3253-8711, Fax: (11) 3253-8473 - São Paulo - SP - Brazil
E-mail: revista@audiologiabrasil.org.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] Corresponding author:Suzana Raquel Lopes Marques. E-mail: suzilps@yahoo.com.br