Study Design
CONSORT guideline was used for conducting this trial. This randomized, controlled clinical trial with two parallel arms recruited 35 physicians and 240 hypertensive patients presenting to primary health care centers in Mashhad, Iran, from September 2013 to August 2014. In Iran, healthcare centers are a public sector to provide primary health care services. This primary health care services are provided free of charge in public facility and run by health care professionals such as physicians, family health experts, nutritionists, midwives, and psychologists.
Participants
In this study, physicians were eligible for inclusion in the study if they (a) did not attend in the communication skills training, (b) participated as member in the Ministry of Health (MOH) sites, and (c) have experience to provide health care services to at least five patients with blood pressure. Physicians’ ages ranged from 26 to 51 (32.03 ± 9.12); they were residents and they provide primary care services, diagnosis, and treatment for patients in the healthcare centers.
This study involved hypertensive patients in interaction with the study of physicians. Patients were eligible for this work if they (a) have uncontrolled blood pressure; (b) carry a diagnosis of hypertension on at least five previous clinic visits; (c) used at least one antihypertensive medication in the last three months, and (d) were age 18 or older. They were excluded if patients (a) did not give informed consent; (b) had suffered mental illnesses and disability.
Sampling and randomization
Sampling began upon the approval of the ethics committee of Mashhad University of Medical Sciences (IR.MUMS.REC.1392.125) and the registration of the study at the Iranian Registry of Clinical Trials (IRCT20160710028863N24). Data collection and recruitment are conducted based on standard guidelines that are described in detail elsewhere [20].
In this study, random allocation of healthcare centers was conducted prior to individual recruitment (Figure 1). Six healthcare centers were randomly selected from a list of centers. Healthcare centers were randomly assigned (as the unit of randomization) to control with usual care (n = 3 health care centers) and intervention (n = 3) groups. All physicians in 6 healthcare centers were screened, of whom 35 physicians meet the inclusion criteria and they were invited to participate in the study. Then, they completed a Health Literacy Assessment Questions (HLAQs) to assess their communication skills before and after the intervention. The ratio of physicians to patients in his study was 1:7 in each of the clinics. Physicians in the intervention group received educational training, while the control group did not receive educational training.
In this study, 940 patients’ medical records were randomly screened, of whom 242 met the inclusion criteria and 2 participants were lost from follow up. Finally, 240 hypertensive patients were included in the data analysis. These patients were invited to attend in this survey and asked to submit an informed consent form. Eligible patients who had completed the pretest questionnaires were randomly assigned to the intervention (n= 119) and control (n=121) groups (Figure 1) using a computerized table of random numbers and block randomization with block sizes of 4 and 6. The randomization was performed by a member of the study team who was blind to patient assignment and sampling. During the study period, health team members and physicians were aware of allocated groups. Primary and secondary outcomes were assessed by a person not involved in the sampling and group allocation.
After the physicians' training was conducted, all enrolled patients returned to the health care centers (both control and intervention) for follow-up office visits and asked them to complete a survey. A socio-demographic questionnaire, HLAQs, patient medication adherence, and provider counseling were completed by the patients.
The sample size used for this study was estimated based on power analysis in which a power of 0.8 with a confidence interval of 95% and an alpha of 0.05 for all statistical analyses.
Data collection tools
The primary outcome in the 6th month included the proportion of patients with controlled BP based on a reduction in systolic and diastolic BP. The secondary outcomes were promoting patients’ health literacy skills, self-efficacy and medication adherence in hypertensive patients.
In this study, we used guidelines for cross-cultural adaptation (CCA) to translate all questionnaires into Persian [20] (Table S1). The structure intent and content of questionnaires and their relevant items were maintained in the Persian version, which leads to maintaining the original scoring system in the adapted questionnaire. In this study, we used the content and face validity assessment methods to measure the validity of the relevant questionnaires (Table S1). The reliability of the HLAQs, self-efficacy, and adult primary care were confirmed using a test-retest on 30 people.
The content validity index (CVI) for the HLAQs, self-efficacy, and adult primary care questionnaire scales were 87%, 86.2%, 89% and content validity ratio (CVR) for these questionnaires were 88%, 82% and 91% respectively. The Cronbach’s alpha was estimated as 0.91 for HLAQs, 0.92 for self-efficacy, and 0.93 for adult primary care questionnaire. Likewise, the Intra-class Correlation Coefficient (95% Confidence Interval) was estimated as 0.97 (0.79 to 0.99) for HLAQs, 0.93 (0.82 to 0.95) for self-efficacy, and 0.95 (0.87 to 0.98). A demographic questionnaire, health literacy assessment questions (HLAQs), Chew’s screening questions (CSQ), adult primary care questionnaire, and chronic disease self-efficacy questionnaire were used in this study. The demographic questionnaire, CSQ, adult primary care questionnaire, and chronic disease self-efficacy questionnaire were completed by patients in both groups before (baseline) and 1 and 6 months after completion of the intervention (follow up) through a self-report. To assess providers–patients' communication skills, the HLAQs was also completed by physicians in both groups before and 1 and 6 months after completion of the intervention.
In this work, uncontrolled BP was considered as an “average systolic BP exceeded 140 mm/Hg or diastolic BP above 90 mm/Hg; or BP above 130/80 for patients with diabetes or renal insufficiency” [20]. Three BP measures were conducted in each medical checkup according to the BP guideline, using an appropriately sized cuff and a calibrated automated BP monitor (Snoqualmie, WA, USA, BPTru device) after five minutes of rest, in the sitting position, with no less than 1 min between measurements. The average of the second and third measures was used [21]. Measurements were completed for eligible patients and physicians in both the control and intervention groups at baseline (before educational training), and 1 and 6 months after completion of the educational training (follow up).
Physician communication skills
Health Literacy Assessment Questions (HLAQs) were used to examine physician’s communication skills at baseline and follow-up [22, 23]. This is a reliable and valid instrument in health care concepts to assess provider-patient communication skills. This tool includes three sections (spoken, written, and communication skills, patient-provider collaboration and support of patient). This scale contained 33 items with a five -point scale ranging from 0 (poor) to 4 (excellent) and higher scores suggesting physicians have sufficient skills to improve patients’ communication needs, quality of care (e.g., respect and friendliness) and health‐literacy limitations, and successfully address them [23].
Patient health literacy
The Chew’s Screening Questions (CSQ) was used to examine patients’ HL level. This scale has been developed based on the rapid estimate of a test of functional HL in adults (TOFHLA) and adult literacy in medicine (REALM) [24, 25]. This questionnaire includes 4 questions related to a problem learning about medical conditions, confident to fill medical forms and help read hospital material. Responses for all items range from 1 (never) to 5 (always), which combined into a single patient satisfaction composite.
Medication adherence
We used the Adult Primary Care Questionnaire to examine the effect of physicians’ communication training on improving medication adherence through 29 items [26]. Responses for the 2 first items were bi-optional (1 = no, 2= yes), while other items were 4-points Likert Scale ranging from 1 (never) to 5 (always). These items asked the patients to assess whether or not physicians guide them to hypertension-related issues in improving medical care and patient concerns about medications. [23, 27].
Patients’ self-efficacy
The Chronic Disease Self-efficacy Questionnaire was implemented to examine patients’ confident related to their regularity tasks or certain activities [28]. This tool included 33 items to examine the following scales: patients’ information about the disease, exercise, communicate with the physician, manage the disease, obtain help from the community, social activities, depression items and manage/control symptoms and shortness of breath. All these items are scored based on five points- Likert from 1 (not at all confident) to 5 (totally confident) [28].
Intervention
Eligible physicians in the intervention group were invited to receive training, aiming to promote their communication skills with hypertensive patients. The intervention in this study was conducted based on 3 sessions of Focus –Group Discussion (FGD) and 2 workshops (10 hours per session).
The FGD sessions, which involved physicians in the intervention group (17 physicians), was conducted by a health educator and cardiologist. In these sessions, physicians expressed their views and experiences of HL improvement in health care clinics with focusing on social problems and patient’s communication needs. In the FGD session, physicians also discussed the existing challenges used to create a successful therapeutic physician-patient relationship. A third session, they concluded that poor clinician support system, patient HL skills and physician communication abilities (medical writing skill, verbal ability, and collaboration) are main barriers to control BP because the physicians do not regularly use HL informed skills in communicating with patients and their families (Table 1).
Two full-day workshops were conducted by a cardiologist in two consecutive weeks to improve communication skills between physicians and their patients. These workshops were designed based on the Health Literacy in Practice (HLP) model and standard treatment algorithms for hypertension care and management based on the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure Guidelines (JNC 7). HLP is a facilitated method to increase HL skills, medication adherence and patients’ self-efficacy among patients with chronic disease. Based on this method, physicians must first understand their patients’ HL limitations and communication needs to effectively communicate with them. Likewise, this method emphasizes on patients’ understanding of the self-management abilities and patients’ concerns (side effect and medications) in order to teach them correct methods to take medications [22, 23].
In the intervention group, the physician’s training was conducted using a self-assessment checklist and training package. The main target of these workshops was enhancing the physicians’ counseling skills that affect the behavior change in hypertensive patients. All physicians were trained to: help the patient in overcoming barriers to hypertension treatment, improve engagement with patients, advice the patient about specific health behavior, detect the patient’s sources in changing this behavior, and improve patients’ HL skills to take higher responsibility for their own self-care behaviors (Table 1).
Data analyses
In the present study, the bivariate analyses (t-test, chi-square, ANOVA) were used to compare clinical and socio-demographic characteristics between experimental and control groups. The linear regression testing the effects of selected covariates (physition-patients’ communication skill) on the dependent variable (pations’ skills, medication adherence, self-efficacy, diastolic blood pressure (DBP) and systolic blood pressure (SBP) scores). The random-effects least squares regression model was conducted to cluster patients within the physician. This model included the main effects of study arm assignment (control vs. intervention), time period (baseline vs. follow-up), and their interaction. Statistical Package for Social Sciences software (SPSS 16, Chicago, Illinois) and R version 3.0.2 were used to produce accurate estimates.