The Association of Health Literacy with High-Quality Home Blood Pressure Monitoring for Hypertensive Patients in Outpatient Settings

Worldwide hypertension (HT) guidelines recommend use of home blood pressure monitoring (HBPM) in patients with persistent suboptimal blood pressure (BP) readings. It is not clear how patients with limited health literacy could perform HBPM to assist BP control. This study aimed at finding the association between HBPM and patients from lower socioeconomic classes, particularly on the effect of health literacy or educational level. Three electronic databases (MEDLINE, EMBASE, and PubMed) were searched for primary studies with keywords including educational level, health literacy, numeracy, home blood pressure monitoring, accuracy, and quality. The PRISMA guideline was followed. The quality of the literature was assessed by the Cochrane tool and modified Newcastle-Ottawa Scale. Nineteen interventional studies and 29 cross-sectional studies were included. Different populations used different cutoffs to report patients' educational level, whereas health literacy was rarely measured. Three studies used psychometric validated tools to assess health literacy. The quality of HBPM could be assessed by the completion of the procedures' checklist or the number of HBPM readings recorded. The association between subjects' health literacy or educational level and the quality of HBPM was variable. The interventional studies showed that increasing professional-patient contact time could improve patients' knowledge, efficacy, and quality of HBPM. Conclusion. Patients' educational level and literacy were not the limiting factors to acquire high-quality HBPM. High-quality HBPM could be achieved by the structured educational intervention. The quality and amount of evidence on this topic are limited. Therefore, further studies are warranted.


Background
Among hypertensive patients, 10% to 50% of their office blood pressure (BP) readings are higher than the home blood pressure readings [1]. In patients presented with uncontrolled hypertension in our daily practice, home blood pressure monitoring (HBPM) (also known as self-blood pressure monitoring (SBPM)) is an essential monitoring option especially for patients with a suspected white coat effect or masked hypertension. It has become an important recommendation in most international hypertension management guidelines [2][3][4]. HBPM was also shown to improve office BP readings, increase BP control rate, and enhance the quality of life at low patient risk [5]. e beneficial effect could be reassured when HBPM is delivered with other forms of interventions, such as patient education or drug titration [5,6]. erefore, health care professionals routinely recommend HBPM to patients with high office BP.
Compared with standard ambulatory BP measurement, HBPM had mean sensitivity of 85.7% (78.0% to 91.0%) and specificity of 62.4% (48.0% to 75.0%) in diagnosing hypertension [7]. e relatively large range of sensitivity and specificity highlighted various factors affecting the accuracy of HBPM. e HBPM readings may be inaccurate due to patients' inappropriate operation, withholding undesirable readings, or inaccurate automated devices. Parker et al. indicated that there was an end-digit preference for zero numbers and specific-value preference for readings just below the alert threshold among patients in self-reporting their BP [8]. In addition, as there were large varieties of home blood pressure monitors available in the market, their accuracy could be questionable. An assessment done by Ringrose et al. revealed that most home BP devices were not accurate to within 5 mmHg [9].
As a result, health care professionals could prescribe inappropriate management according to inaccurate HBPM records.
From a patient's perspective, performing HBPM is not always an easy task. Patients with lower self-efficacy, lower educational levels, or lower health literacy may need special interventions to assist home monitoring. Fletcher et al. illustrated patients' and health care professionals' concerns of HBPM in a qualitative review, as HBPM involved interpretation, attribution, and action [10]. e quality of HBPM could be highly operator-dependent. Any inaccurate readings or wrong interpretation may lead to patient anxiety, overdiagnosis, or overtreatment due to falsely high home BP readings. On the other hand, falsely low home BP may lead to false reassurance, underdiagnosis, or poor drug compliance.
Different elements are required to perform high-quality HBPM as described in clinical guidelines. ey include access to accurate BP monitors, skills, and knowledge to perform HBPM, motivation to perform HBPM regularly, and accurate reflection of HBPM readings to their health care providers. Patients may not have the hardware, skill, and knowledge to implement successful HBPM. ey need health care providers' instruction and feedback to practice HBPM independently. eir skills and BP records should be reviewed regularly in order to ensure their compliance with HBPM protocol, such as measurement preparation, procedure, and how to record BP readings. In a busy primary care practice, time constraints may preclude physicians from taking time to educate HBPM and review patients' home BP records.
Conventionally, many studies assessed patients' educational level as part of the sociodemographic background instead of assessing patients' health literacy specifically. "Health Literacy" (HL) is the patient's ability to read, interpret, and respond to the information during health care activities. It was defined by the American Medical Association in 1999 as "the constellation of skills, including the ability to perform basic reading and numeral tasks required to function in the healthcare environment" [11]. Underprivileged patients, such as those from lower socioeconomic class, those with lower educational levels, or those with limited health literacy or numeracy, were found to have a poorer outcome in overall noncommunicable diseases [12]. In addition, patients with inadequate health literacy were more likely to have poorer disease knowledge, poorer selfefficacy, and misconception in cardiovascular disease [13]. ey may also encounter greater barriers in performing accurate HBPM. Few studies have focused on whether the underprivileged patients were able to perform HBPM as good as middle or above socioeconomic class patients. Given the large and increasing global disparities of BP control in hypertensive patients from the low-income population, there is a clinical urge to formulate suitable interventions which could help patients achieve desirable BP targets [14]. Most of the existing review papers focused on the BP outcomes of the global hypertensive population after different HBPM interventions [5][6][7]10]. So far, there is a limited understanding of how the socially disadvantaged population could successfully perform high-quality HBPM that could subsequently improve their BP control. is study aimed to find out the association between patient health literacy (including educational level and other related socioeconomic factors) and HBPM, which may or may not lead to improvement of BP control. e finding will be particularly useful to the low-income hypertensive populations.

Methods
We performed a systematic review using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guideline to investigate the association between HBPM and patients' health literacy or educational level [15].

Selection of Studies.
We included all original research articles evaluating adult practice or attitude towards HBPM or SBPM, which include knowledge, skills, and practice towards HBPM or SBPM. e articles should contain an association of subjects' ability to read, understand, and follow instructions, such as educational level, health literacy, and numeracy with either HBPM or overall BP control. e articles could include processes of HBPM or SBPM, practices such as the prevalence of HBPM or SBPM, and skills or knowledge of HBPM or SBPM.
ere was a preferable analysis of the association between study subjects' ability and quality of HBPM or SBPM.
We excluded studies with neither analysis of HBPM or SBPM practice nor subjects' educational level or health literacy.

Search Strategy.
We performed a web-based search of the MEDLINE, EMBASE, and PubMed databases. We also screened the reference list of all relevant studies (snowball search). Studies published in English from 1910 to present were included.
We defined two main search concepts ("self-blood pressure monitoring" and "appropriateness of self-care activities") and combined the search by "AND." We used the MeSH term "blood pressure monitoring, ambulatory" or the term "self-blood pressure monitoring" or "self-blood pressure measurement," "home blood pressure monitoring," "self-measurement," and "blood pressure." We then used the term "health literacy," "mathematics," "numeracy," "educational level," "educational status" or "health knowledge, attitude, practice." We limited searching by studies for adults (age > 18) e whole syntax is shown in Table 1. e date of the last search was on 27 November 2019.

Selection of Publications.
We went through a two-step selection process. We first read the titles and abstracts. Studies meeting all inclusion criteria above were identified as potentially appropriate. We then analyzed the full texts of the selected articles according to the inclusion and exclusion criteria. Reasons for exclusion were documented.
Two independent review authors (SF and MD) did the whole selection process. Disagreements between us were resolved by consensus. A review author (either CW or BC) was consulted if disagreement persisted.
e selection of the studies is based on the following criteria: (1) e article is an original study, which includes a detailed study method for the assessment of the risk of bias. (2) e quantitative studies included assess the association between health literacy or numeracy or educational status of patients and their practice or knowledge on SBPM or HBPM or evaluate the interventions to enhance HBPM practice by enhancing the efficacy of HT patients. e studies centered on adult patients with an established diagnosis of hypertension. e studies that focused on diagnostic tests, screening of hypertension, and hypertension in pregnancy were excluded.

Data Extraction.
We extracted bibliographic data (author, publication year, title, and journal), study design, setting, country, inclusion and exclusion criteria, subject recruitment, study population characteristic (age and gender), and date and duration of the study. We registered the tools used for assessing the outcome measurement and if there is an association between the subjects' health literacy and the appropriateness of HBPM. We retrieved all outcome categories. Finally, we extracted the HBPM-related interventions and patient outcome particularly for subjects with relatively low health literacy or educational level.

Quality Assessment.
Critical appraisal was independently recorded by reviewers to allow comparison. Risk of bias was assessed by considering relevant domains to interventional studies, including participant selection, measurements of variables, and controlling for confounding, in line with the Cochrane Collaboration's Grading of Recommendations Assessment, Development, and Evaluation (GRADE) tool for assessing the risk of bias [17,18]. Each domain was rated with "high," "low," or "unclear" according to the risk of bias, with free text explanations. In order to maximize relevance to nonrandomized studies, the Newcastle-Ottawa Scale (NOS) for cross-sectional studies was used [19]. Two authors (SF and MD) assessed the individual study by three domains which are selection (maximum 5 stars), comparability (maximum 2 stars), and outcome (maximum 3 stars) resulting in total NOS grade. e summation of the 3 domains' number of stars resulted in the total NOS score. Very good studies scored 9-10 stars, good studies scored 7-8 stars, satisfactory studies scored 5-6 stars, and unsatisfactory studies scored 0-4 stars.

Data Synthesis.
We separately collected the cross-sectional studies and the interventional studies data for narrative synthesis. Different assessments of educational level and/or health literacy, HBPM device, technique, and quality of HBPM were recorded. Figure 1 shows the systematic search and selection of relevant studies adopting the PRISMA guideline 2009 [16]. 195 studies were identified from MEDLINE, EMBASE, and PubMed. Bibliographies of primary studies and review articles meeting the inclusion criteria were searched manually to identify 15 further eligible studies. 182 unique studies in total were included for the screening of abstracts. After reviewing the abstracts, 77 studies were excluded because the studies did not assess HBPM or self-BP monitoring, or the research subjects were not hypertensive patients, nor was there any association between HBPM and patients' educational status or health literacy. 105 studies were included for full-text assessment of eligibility. Finally, 48 studies (19 randomized controlled trials and 29 cross-sectional studies) were included in the data synthesis. e results of cross-sectional studies are shown in Table 2. e included studies were performed in North America, Europe, and Asia from 2003 to 2019. Most of the participants were patients with hypertension. Two of the studies surveyed pharmacists and primary care providers such as nurses and physicians about their clinical practice of HBPM. Study sites included a populational survey, recruited in community organizations, primary care clinics, or outpatient clinics in the hospital. Most studies demonstrated a positive relationship between subjects' educational level or health literacy is associated with owning BP monitors at home, performing it regularly or recording the measurements accurately. Five out of twenty-nine studies reported a negative association between patient educational level or other social factors and practice of HBPM. e BP outcomes of patients were included in 5 studies: 2 studies showed a positive association of HBPM and BP control, while 3 studies did not demonstrate any better BP control. e quality assessment by Newcastle-Ottawa Score for cross-sectional studies found the studies ranged from very good (grade 9/10) to unsatisfactory (grade 2/10). Most unsatisfactory studies got low sampling scores. Table 3 shows the results of the 19 interventional studies. Most studies were performed in North America, while 2 of International Journal of Hypertension (1) Home monitoring/or home care/ (2) Reading/or health education/or health literacy/ (3) Patient compliance/ (4) Education program/or interdisciplinary education/or education/or health education/or patient education (5) Blood pressure monitoring, ambulatory (6) Self-care/ (7) Educational status/ (8) Patient education as topic/ (9) "Reproducibility of results"/ (10) Essential hypertension/or White coat hypertension/or masked hypertension/or MEDLINE / (11) 1 or 5 or 6 (12) 2 or 4 or 7 or 8 (13) 3 or 9 (14) 10 and 11 and 12 and 13 (15) Limit 14 to (English language and (MEDLINE or "PubMed not MEDLINE")) (1) Blood pressure monitoring/ (2) Home monitoring/or home care/ (3) Self-monitoring/ (4) Reading/or health education/or health literacy/5. Health disparity/6. Patient compliance/ (7) Reproducibility/ (8) Essential hypertension/ (9) White coat hypertension/ (10) 1 or 2 or 3 (11) 8 or 9 (12) Education program/or interdisciplinary education/or health education/or patient education/ (13) 4 or 5 or 12 (14)      International Journal of Hypertension   them were performed in Europe and one of them was performed in Hong Kong. ere were different modes of HBPM interventions, such as providing home BP monitors, patient education, and training intervention, record and feedback system to HBPM measurements, training, and updating knowledge to health care providers.

HBPM in the Included Studies.
ere were various types of home BP monitors involved in selected studies. ey included automatic electronic branchial devices, electronic semi-automated branchial devices, manual mercurial sphygmomanometer, and electronic wrist devices. e possession of home BP monitors was related to higher educational level and/or income status [20,25,41,44], while the frequency of HBPM as clinician recommended was not necessarily related to the educational level. Ragot et al. found that 90% of HBPM users did not receive information about HBPM use [37]. Other studies demonstrated that the HBPM quality might not be related to the educational level. When patients were instructed to use HBPM by health care providers, there was consistent reporting of regular HBPM use [22,28,36,46]. ere was no psychometrically validated tool to assess the quality of HBPM. e quality of HBPM was assessed by different tools defined by authors in different studies. Some used the number of successfully documented or transmitted BP readings over the number of expected BP readings as high-quality HBPM. Flacco [33].

Educational Status.
Most studies assessed the subjects' educational status. e educational level was usually selfreported as part of personal characteristics. e assessment method could be highly heterogeneous. Most studies categorized educational attainment into different levels of schools: primary schools, middle schools, high schools, and colleges, but the cutoff level and the number of categories highly varied. ree studies included "illiterate" as one of the educational status categories [25,43,45]. Other studies also included years of education for data synthesis. If the number of years was defined as binary categories, their cutoff years could vary from 5 years to 12 years.

Health Literacy.
HL was not commonly assessed in studies of HBPM. Only 3 studies used 4 different validated health literacy (or numeracy) scales for assessment. Kim et al. used the High Blood Pressure Health Literacy Scale [57]. ey did not categorize the subjects as high or low HL. ey measured the change in HL before and after the intervention. Shi et al. used the Chinese Health Literacy Scale for Hypertension [40]. More than half of their study subjects (55.3%) had low health literacy. Rao et al. used the Rapid Estimate of Adult Literacy in Medicine-Short Form (REALM-SF) and the 3-item numeracy measure [38]. Less than one-third of the subjects (31%) had low numeracy. ese three studies found a strong association between health literacy or numeracy with educational status.

Other Assessments of Patients' Ability.
Apart from educational status and health literacy, six studies quantified the subjects' ability by different knowledge scores. e scoring items included knowledge about hypertension, hypertension complications, hypertension comorbidity, and HBPM [20,27,31,32,37,57].

Association between HBPM and Educational Status or
Health Literacy. In some studies, subjects with higher educational levels were not found to use more HBPM [21,28,29,31,35,64]. However, in one study, a larger proportion of subjects with higher educational levels used HBPM [25]. Subjects who believed HBPM could help BP control performed more regular HBPM [21,36,41]. Some studies showed that patients with higher educational levels, higher HL, or higher numeracy could perform higher quality of HBPM, such as better compliance with HBPM procedure and more complete or accurate HBPM record [26,38,45,47]. Flacco et al. did not find such an association. Instead, they found that patients performed higher quality of HBPM if they received HBPM instructions from pharmacists or doctors than if they received them from nonprofessionals [28].

Interventions to Improve HBPM for Patients with Different HLs.
e interventional studies described different complex interventions that targeted patients with uncontrolled hypertension. Some of them targeted socially disadvantaged subjects, i.e., Korean American older patients and African Americans [57,62]. Kim

Discussion
In this review, the educational status might or might not be associated with HBPM practice, quality, and compliance. e finding was particularly significant for patients with lower socioeconomic status. It was known that patients with lower educational status and lower income had a higher risk of hypertension and more nonadherence to antihypertensive treatment, subsequently leading to poorer clinical outcomes [67]. Health care providers could consider HBPM as an intervention which could improve drug adherence and BP control. Important elements included coaching patients on proper selection of HBPM devices and correct HBPM techniques (e.g., accurate recording of home BP readings). It could be done by providing home devices that store multiple BP readings, or uploading readings to smartphones or computers, or transmitting them directly to electronic health records.
Although various hypertension guidelines recommended the use of HBPM in diagnosis, the overall possession and compliance to HBPM were suboptimal.
We also identified that after structured training, socially disadvantaged patients could have significant improvement in HL, self-efficacy, and BP control. We suggested a structured intervention for identifying patients with low HL and offering training of hypertension self-care including HBPM, with the effect of improving BP control and increasing patients' HL. In view of the fact that less than 40% of the hypertensive patients had optimal BP control in different populations, the promotion of high-quality HBPM in patients with uncontrolled hypertension should be a clinical priority.
ere was no structural or validated tool for assessing the quality of HBPM. e quality of HBPM depends on a validated BP device, competence of patients to perform HBPM on their own with a correct method and frequency, a record of accurate HBPM reading, and sharing of that record with health care professionals. Studies in this review modified recommendations of HBPM procedures from various international guidelines. Some studies also adopted the teletransmission of BP readings via the electronic system. We, therefore, suggest future research to develop a patientfriendly protocol to assess high-quality HBPM.
We also found that only a small proportion of studies focused on the assessment of HL and the outcome of hypertension control. e most common assessment is the educational status, which may be unrelated to patients' performance or compliance with antihypertensive treatment.
e EHS Guideline proposed that the first step to tackle patients with poorly controlled chronic illness should be patient-centered care: to identify patients' barriers to better control the disease [1]. For instance, health care professionals should be well-equipped with communication techniques with low HL patients. Rajah et al. summarized that healthcare professionals should use everyday language and teach-back method and provide patients with reading materials and aids. However, the most commonly reported barrier regarding patient-centered care is time constraints.

Strength of the Study.
is is the first study focusing on HBPM and its association with patients' education level, including health literacy, numeracy, and other socially related factors. Most studies were performed in community or outpatient settings, where primary care observation or intervention could be applied.

4.2.
Limitations of the Study. Heterogeneous assessment of HBPM or SBPM, the prevalence of HBPM, and the educational status of patients are limitations of this study. Validated assessment of health literacy is sparse.

Conclusion
Patients' educational or health literacy levels were not limiting factors to acquire skills and knowledge of HBPM. High-quality HBPM could be achieved by structured educational interventions. Complex interventions involving patient education, providing valid home BP monitors, and facilitating patient-clinician communication may improve BP control. ose interventions should be tailor-made to subjects with low educational levels, which could be equally effective in improving the overall BP control.

BP:
Blood pressure C: Comparability CVD: Cardiovascular disease DBP: Diastolic blood pressure EHS: European hypertension society FU: Follow-up HBPM: Home blood pressure monitoring HL: Health literacy N: No NA: Not available/applicable NOS: Newcastle-Ottawa Scale O: Outcome OR: Odds ratio RCT: Randomized controlled trial S: selection SBP: Systolic blood pressure SBPM: Self-blood pressure monitoring Y: Yes.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
SF, MD, CW, and BC participated in the study design and data synthesis. SF and MD performed the search, and SF wrote the first draft of the manuscript. MD, CW, and BC commented on this draft and performed critical revisions. All authors have read and approved the manuscript.