Design and Sample
The Self-Help Intervention Program for HBP care (SHIP-HBP) was a community-based clinical trial designed to test the effects of a culturally tailored behavioral intervention in KAs residing in the Baltimore-Washington Metropolitan area. The SHIP-HBP combined a 6-week psycho-behavioral education with 12 months of BP self-monitoring and telephone counseling conducted by bilingual nurses. Details about the study design were reported elsewhere. 20
To be brief, the study employed a 2×2 factorial design with two key factors: a mode of education (i.e., in-class vs. mail) and intensity of telephone counseling (i.e., more intensive, bi-weekly counseling vs. less intensive, monthly counseling). Participants in both counseling groups used BP self-monitoring throughout the 12-month counseling period (). Study outcomes included systolic and diastolic BP levels and several psychosocial variables that were collected at: baseline, 3, 9, and 15 months (except for BP measurements made via self-monitoring at home). The positive BP changes in study outcomes at the 3-month follow-up were also reported. 20
In the present paper, we are reporting the study outcomes collected at 3 months through 15 months during which the telephone counseling was offered.
Eligibility criteria for the SHIP-HBP trial included: 40 to 64 years of age; systolic BP (SBP) ≥ 140 mm Hg and/or diastolic BP (DBP) ≥ 90 mm Hg on two separate occasions or being on antihypertensive medication; and self-identification as a KA. At baseline, 445 eligible KAs were recruited through ethnic churches, grocery stores, and ethnic newspaper advertisements. Over the course of the clinical trial, 85 participants dropped out, resulting in 359 KAs at the 15-month follow-up. There were no significant differences in sociodemographic characteristics between those who remained in the study and those who dropped out.
Study procedures were approved by the Institutional Review Board, and every participant provided written informed consent. During the baseline evaluation, trained bilingual nurses obtained three BP measurements at 1-min intervals, while demographic and psychosocial variables were measured via self-report. Participants were then assigned to either the in-class or the mail education groups to receive HBP-related education for a 6 week period. Upon completion of the 6-week education, participants were given a BP monitor (A&D UA-767) equipped with a tele-transmission system along with instructions. They were asked to measure their BP at home and transmit the BP data via phone over the following 6 weeks of the test period, during which it was made certain that all participants were able to measure and transmit their BP readings.
At 3 months (i.e., at the end of the 6-week test period), the KAs were randomly assigned to either the more intensive counseling (MC) or less intensive telephone counseling (LC) groups. The MI group received bi-weekly telephone counseling, while the LI group received less frequent, monthly counseling by a trained bilingual nurse for 12 months. Trained nurse counselors made the phone calls from a private office at a local community center. As a safety measure, nurses called the participants regardless of the counseling schedule if the transmitted BP readings were 160/100 mm Hg or higher and contacted a consulting physician on the study team. When the BP readings were 180/110 mm Hg or higher, nurses either called the patient immediately to ask him/her to visit a nearby emergency room or called his/her physician. Nurses followed these up with off-schedule phone calls and then with an additional call within a week for updates.
During the counseling period, participants were instructed to measure their BP three times both upon waking (AM reading) as well as before retiring to bed at night (PM reading). The participants were asked to perform both sets of triplicate measurements two or more times a week. The transmission device automatically saved up to 200 BP measurements. Nevertheless, in order to facilitate timely screening of abnormal BP readings and feedback, participants were asked to send their measurements via telephone at least once a week. Transmitted BP data was used to generate BP reports, which showed the time and frequency of both the BP measurements and BP readings. BP reports were made available to both the participant as well as the nurse counselor throughout the 12-month counseling period.
During the 12-month counseling period, BP was measured at home by participants using the A&D UA-767 (A&D Company, Ltd, Tokyo, Japan), which is a fully automatic device using the oscillometric method. Given varying numbers of BP measurements for each participant, it was determined that the bi-weekly means of SBP and DBP measurements be used to derive BP data for the current investigation (see details about compliance with home BP monitoring). In accordance with the guidelines laid out in the seventh report of the Joint National Committee on Prevention, Detection, and Treatment and Control of High Blood Pressure, BP control was defined as home-measured BP < 135/85 mm Hg (130/85 mm Hg for patients with diabetes).
Other assessments included sociodemographic characteristics, insurance status, and medical history measured at baseline through the study questionnaire. In addition, psychosocial outcome variables were measured at baseline, 3, 9, and 15 month follow-ups. Specifically, HBP knowledge was assessed using 12 items developed by the National HBP Education Program of the National Heart Lung Blood Institute as well as by using 14 items generated by the investigative team based on the literature review. The HBP knowledge instrument has previously been used in KAs. HBP beliefs were assessed using a 12-item questionnaire. The questionnaire asked participants to indicate whether they believed certain behavioral factors could help lower their BP (i.e., smoking, stress, weight, alcohol, salt, medication, coffee, diet, cholesterol, and exercise) and what the most important factors were in controlling their BPs. HBP self-efficacy was measured using a scale adapted from the HBP belief scale. The modified scale used 4-point Likert type items asking how confident the person was in managing HBP in 11 different areas such as reducing salt intake, taking HBP medicines, or eating fatty foods less frequently. Higher scores indicated higher levels of self-efficacy in the management of HBP. Depression was measured using the Kim Depression Scale for Korean Americans (KDSKA). The KDSKA contains 21 items assessing depressive symptoms in the following four categories: emotional, cognitive, behavioral, and somatic. Each item is presented on a 4-point Likert-type scale as a declarative sentence related to one symptom of depression from the previous one-week period. Higher scores represented more depressive symptoms. Finally, medication adherence was measured through the medication subscale from the Hill-Bone Compliance Scale. The scale consisted of nine items on a 4-point Likert-type scale that measured the self-reported degree of adherence to HBP medication regimen. Higher scores indicated lower adherence.
Only participants who self-reported at least one BP measurement were included in the statistical analysis (N=359). Two blood pressure outcomes (mean SBP and mean DBP measured bi-weekly) and five psycho-social outcomes (the mean scores of the HBP knowledge, HBP beliefs, HBP self-efficacy, KDSKA, and Hill-Bone Compliance Scale instruments measured at three follow-up assessments) were modeled using mixed-effects models to examine the rate of change over time, differences between intervention groups, and differences in the rates of change between intervention groups for each outcome. At each follow-up, only those subjects completing at least six of the nine items on the Hill-Bone Compliance Scale and taking BP medication were included in the adherence model (N=240). Each person could report up to 34 biweekly BP measurements during follow-up; in total, 5408 biweekly BP observations were collected from 352 participants. The average number of measurements per participant was 15.4.
Nonlinear rates of change were considered for the BP models; however, the more parsimonious linear model was deemed sufficient via visual examination of the model diagnostics. Mixed-effects models were implemented to account for the correlation induced by repeated measures (section 3.1 in Rabe-Hesketh and Skrondal, 2008; Gardiner et al., 2009). The models were adjusted for age, sex, marital status, education, work status, health insurance status, years of residence in the US, BP medication status, family history of high BP, and years of high BP. All modeling was done in R version 2.8.1 (R Development Core Team, 2008).