Short messaging service (SMS) messages for improving

Using Short Messaging Service (SMS) text messages as an intervention is one way to apply modern technology to chronic disease management. SMSs are very commonly used, low cost and easy-to-use even for older individuals. An SMS support system is also feasible and simple enough to be used wide-scale in non-research settings, which makes it a rather attractive subject of non-research in hypertension management.

Earlier research on SMSs in clinical use in the early 2000’s and 2010’s suggested that SMSs support may lead to behavioral changes and improve BP control, but published evidence was quite mixed and limited (126-129). After our study was conducted, two large reviews by Palmer et al. in 2018 and Islam et al. 2019 showed 2 – 4 mmHg reductions in SBP in a 6 - 12 month follow-up, but the evidence about long term (12 months or more) treatment adherence and BP control is still quite limited and based on very few studies. So far, there is no published evidence about the effectiveness of SMS treatment support combined with the use of a checklist. In Finland, there are no previous studies published in peer-reviewed scientific journals on SMS effectiveness in BP reduction.

As for adverse effects of SMS interventions, the only known potential adverse events may include road traffic accidents, but the studies reported in this review have not reported any adverse effects (130).

2.8.2 Checklists in improving treatment adherence and BP control

Checklists are commonly used cognition aids in non-medical settings, especially in high intensity fields of work, such as the aviation industry and the military. The aim of the use of a checklist is to reduce human errors often under highly complex and

stressful conditions. The core principle is simple: the checklist guides users through accurate task completion by collecting all the relevant criteria of

consideration concerning a specific process and presenting them as a go-through list. Thus, the use of a checklist serves as a decision and memory aid and allows for more predictable outcomes of system operation and more consistent quality of work despite individual and collective cognitive weaknesses. (131,132)

There is some evidence suggesting that checklists could improve the quality of care in medical settings, too. Checklists were first introduced in medical settings in 2006 and have since been proven to improve the safety and quality of care in surgery, anesthesia and intensive care unit settings in most, but not all studies (133). Furthermore, checklists could also have a positive effect on adherence to guidelines, process outcomes of care and treatment compliance, at least in in-patient care (134-136). A very interesting new area of research is the use of electronic checklists, showing positive reports from several clinical areas (132).

However, there is insufficient evidence of the use of checklists in the treatment of chronic diseases and no evidence concerning the initiation of antihypertensive medication in outpatient care. It is also unclear if using a checklist can improve BP control in long-term follow-up and in a primary care setting. Furthermore, no evidence exists about the effectiveness of using a checklist combined with an SMS treatment support.

Checklists might, in theory, also contribute to adverse events, such as delays in treatment or burden-related errors by health care professionals. However, there is no evidence of any edverse events that checklists might actually have contributed to.

2.9 HEALTH-RELATED QUALITY OF LIFE IN HYPERTENSION

Hypertensive patients are mostly asymptomatic unless their BP is extremely high.

From that viewpoint, the negative association between hypertension and health-related quality of life (HRQoL) is somewhat surprising. However, the correlation between awareness of hypertension and decreased HRQoL was reported as early as the 1970s in a landmark study by Haynes et al. and has been confirmed since then in several studies and systematic reviews (20,137,138). The magnitude of HRQoL impairement associated with hypertension has usually been quite small but clear (20,138). Earlier research has also suggested that HRQoL deterioration is associated more with the awareness of being sick (“labeling”) - not with the elevated BP or the side effects of the treatment per se (18,139).

It is important to note, though, that the negative effect of hypertension on HRQoL does not seem to be inevitable and that in some studies, the HRQoL of hypertensive patients has even improved over time (19,140) What explains this contradictoy evidence and whether it is possible to systematically avoid the deterioration of HRQoL in hypertension remain unclear. Earlier research has indicated that practical differences in hypertension treatment may affect patients’

HRQoL (140-143). Another possible explanation is the study design. The majority of earlier studies on this field have been cross-sectional studies on patients with existing hypertension, but changes in HRQoL can only be examined in longitudinal studies and the likelihood of detecting these changes is best with newly diagnosed hypertension. Yet another explaining factor may be the choice of HRQoL

instrument. Most commonly, HRQoL has been measured by means of a profile-based instrument, the Short-Form Health Survey (SF-36)(144). The studies that have used the preference-based EQ-5D instrument (i.e. the score reflects how good or bad a certain health state is according to the preferences of the general population), have reported EQ-5D scores of hypertensive patients as quite close to the general population average and hypertension typically presents a very small decrease in HRQoL, compared to other major chronic conditions (145,146). We need more real-life evidence from longitudinal studies to better understand the factors affecting the HRQoL of hypertensive patients. Also, too little is known about HRQoL with newly diagnosed hypertensive patients in primary care settings, especially with the EQ-5D instrument. Furthermore, the current emphasis on patient-centered care places great significance on investigating HRQoL as a very potential patient-centered outcome measure for hypertensive patients (21).

3 AIMS OF THE STUDY

The purpose of this thesis was to test the effectiveness of a checklist for the initiation of antihypertensive medication and text message support in improving the quality of hypertension treatment compared with usual care in adult

hypertensive primary care patients in real world health care settings. The specific aims of the substudies were:

1. To investigate whether using a checklist in the initiation of new antihypertensive medication would improve hypertensive patients’

perceived treatment-related information, motivation and behavioral skills, focusing especially on setting and knowing the adequate BP target. (Study I).

2. To assess whether personalized text message support, together with a checklist, would improve systolic BP control as compared with usual care during the first year of new antihypertensive therapy (Study II).

3. To investigate the use of antihypertensive medication and hypertension-related use of health care services during the first treatment year (Study II).

4. To assess hypertensive patients’ experiences and opinions about checklist use and text message support at 12 months follow-up (Study II).

5. To investigate cardiovascular risk factors and perceived HRQoL of newly diagnosed hypertensive patients during the first year of antihypertensive medication (Study III).

6. To examine what factors might be associated with changes in HRQoL during the first year of antihypertensive medication (Study III).

4 SUBJECTS AND METHODS

4.1 METHODS

The Check and Support Study (ClinicalTrials.gov reference NCT02377960) was a cluster-randomized controlled trial carried out in a real-life primary care setting in Finland. The study was conducted in accordance with the principles of the

Declaration of Helsinki and the ethical standards of the institutional review board of the Hospital District of Northern Savo (reference 63/2014) and written informed consent was obtained from all the study patients. Study reporting is in line with the Consolidated Standards of Reporting Trial (CONSORT) 2010 guidelines.

4.1.1 Setting and study design

The study was conducted between January 27^th, 2015 and March 6^th, 2018 in eight primary care study centers in Central Finland. The study centers included six public sector health centers, one private sector occupational care center and one public sector health center with occupational health care services. Together, the study centers provide primary health care services for a population of approximately 200,000 persons in both urban and rural areas. At the beginning of the study, all participating physicians in both study arms received basic information about the study and a short lesson on then-current Finnish hypertension practice guidelines 2014. The study centers were then grouped into comparable pairs and

randomized by principal investigator to function either as an intervention (n = 4) or control (n = 4) center using a two-block randomization list. The pairing was done to match the following attributes: Location (urban–rural), center size (small–large) and selection of services (occupational health care service or not). Due to real-life study settings and nature of interventions, the study was unblinded. For the last analysis (Study III), both study groups were combined and the study design was converted into an observational study.

4.1.2 Subjects

Study patients were recruited by treating physicians when initiating a new

antihypertensive medication in routine practice. In participating study centers, the physicians asked their patients to take part in the study if they met the inclusion criteria: (1) aged between 30 and 75 years, (2) about to start antihypertensive

medication for the first time, (3) a clinical diagnosis of hypertension (diagnosis was made by treating physician), (4) possession of a mobile phone, (4) ability to read SMS messages, (5) ability to take care of their personal medication, (6) ability to perform home BP measurements and (7) an agreement on using electric drug prescriptions (standard care in Finland). The exclusion criteria were: (1)

unwillingness to give informed consent or to take part in the study, (2) pregnancy, (3) a malignant disease that was determined by the treating physician to have an impact on life expectancy, (4) having or suspected of having depression or psychosis (based on clinical evaluation, previous EMR notes and a valid screening questionnaire for depression), (5) atrial flutter or atrial fibrillation (based on clinical evaluation and previous EMR notes) , (6) SBP > 200 mmHg, (7) DBP > 120 mmHg, (8) fast onset or worsening of hypertension, (9) hypokalemia (K < 3.3 mmol/l) or (10) kidney disease, defined as an eGFR < 45 ml/min/1.73m2, or proteinuria (albumin-creatinine ratio > 30 mg/mmol, 24-h protein excretion > 500 mg/day, night urine albumin > 200 μg/min, or proteinuria in urine dipstick test).

4.1.3 Study measures, outcomes and data collection 4.1.3.1 Baseline measures

4.1.3.1.1 Clinical measurements

Each study patient’s baseline office BP, height, weight and waist circumference were measured by the treating physician. Office BP was measured three times from the left arm with a Microlife WatchBP Home (A or N type) automatic oscillometric monitor after five minutes of rest in the sitting position (147). The default cuff was a wide-range (arm circumference 22–42 cm) semi-rigid conical cuff. However, large (arm circumference > 42 cm) and small cuffs (arm

circumference < 22 cm) were also available if needed. The same BP monitor that was used for office BP measurement was then loaned to the study patients to be used for home BP measurements. Both written and oral instruction for adequate home BP measurements were given to all study patients. Office BP was defined as the mean of three office measurements. Home BP was defined as the mean of all home measurements over a seven-day period (three measurements twice daily, between 6 - 9 a.m. and 6 – 9 p.m.). Waist circumference was measured at the midpoint between the lower border of the patients’ rib cage and the iliac crest.

BMI (kg/m²) was calculated by dividing the patient’s weight (kg) by the square of his/her height (m). An electrocardiogram (ECG) was taken and the following laboratory tests were performed: fasting plasma cholesterol, fasting plasma

glucose, plasma creatinine, plasma potassium, eGFR (The Chronic Kidney Disease Epidemiology Collaboration CKD-EPI equation (148)) and proteinuria (measured with the albumin excretion rate using one of the three alternative methods: nightly urine, diurnal urinary protein excretion, or spot urine albumin-creatinine ratio). In assessing whether the treating physician had set an adequate BP target, the EMR notes and checklist were both taken into account.

4.1.3.1.2 Questionnaires

All the patients filled in a questionnaire on basic demographics, smoking habits (Heaviness of Smoking Index) and alcohol use with alcohol consumption questions from the alcohol use disorders identification test (AUDIT-C) (149,150).

The three AUDIT-C questions each had scores ranging from 0 to 4 points and were summed for a possible score of 0 to 12. University- or college-level education was considered higher education.

Exercise habits were assessed using the Frequency-Intensity-Time (FIT) questionnaire, which has one question each on the frequency, efficiency, and duration of exercise, with scores from 1 to 5, 1 to 5 and 1 to 4 points, respectively (151). The scores were then multiplied and the possible sum score (FIT-index, range 1–100) was calculated, with < 36 indicating low, 37–63 moderate and 64 or more high physical activity.

HRQoL was assessed using the EuroQol five-dimension (EQ-5D, 3L)

questionnaire, which covers five dimensions of health: mobility, self-care, usual activities, pain/discomfort and anxiety/depression (152). In the EQ-5D, patients give each dimension a score of one to three, according to three levels of condition severity: 1 = “no problems”, 2 = “moderate problems” or 3 = “severe problems”. The given scores were then converted into a single EQ-index by applying scores from the British valuation set, with a scale from −0.59 to 1.00, with a value of 1

indicating the best possible health state and functional ability (152,153). EQ-5D questionnaire also includes a standard vertical 20-cm visual analogue scale (EQ-5D VAS) for recording perceived current overall health state, with a score of 0

indicating “the worst imaginable” and a score of 100 “the best imaginable” health state.

In addition, each study patients’ 10-year risk for a first fatal CVD event was estimated based on sex, age, SBP, smoking and total cholesterol level and using the Systematic COronary Risk Evaluation system (SCORE). The SCORE risk level estimation was categorized into low (< 1%), moderate (≥1 to < 5%), high (5–9%) or

very high risk (≥10%). In the SCORE evaluation, patients with documented CVD, chronic kidney disease (stages 3–5), diabetes or very high levels of individual risk factors are automatically considered to be at a high or very high risk of CVD, without a formal risk estimation. Hypertension-mediated organ damage may also increase CVD risk to a higher level. (3). A questionnaire also included questions on each of the three elements of the informationmotivationbehavioral skills (IMB) -model (154).

4.1.3.2 Follow-up measures at 12 months 4.1.3.2.1 Clinical measurements

At the follow-up visit, all the baseline clinical measurements were repeated. In addition, EMR notes were analyzed to assess medication information and hypertension-related use of health care services.

4.1.3.2.2 Questionnaires

All the baseline questionnaires were repeated at the 12-month follow-up visit. As part of the questionnaire at 12 months, patients in the intervention group were also asked to give feedback about the usefulness of SMS support (‘How useful did you find SMSs for supporting your treatment?’) and significance of the checklist (‘In addition to SMSs, you filled in a checklist with your treating physician to support your treatment. How important did you find it?’) with an 11-point numerical rating scale (0 = not useful or important at all, 10 = very useful or important). In addition, participants were asked to evaluate their willingness to receive hypertension treatment-related SMSs in the future (‘Would you like to receive SMS support for your treatment in the future?’; Yes or No).

4.1.3.3 Outcomes and data collection 4.1.3.3.1 Primary outcomes

The primary outcome of the study was the proportion of study patients achieving the SBP target at the 12-month follow-up. Office and home SBP targets were

<140 mmHg and <135 mmHg, respectively, in accordance with the then-current European and Finnish hypertension guidelines (5,7).

4.1.3.3.2 Secondary outcomes

Secondary study outcomes included: 1) proportion of study patients achieving the DBP target (office DBP target < 80 mmHg for individuals with diabetes and <

90 mmHg for others.; home DBP targets < 75 mmHg and < 85 mmHg, accordingly), 2) hypertensive patients’ perceived treatment-related information, motivation and behavioral skills at the initiation of antihypertensive medication and at the 12-month follow-up, 3) change in SBP and DBP levels, 4) change in cardiovascular risk factors and perceived HRQoL, 5) use of antihypertensive medication, 6)

hypertension-related use of health care services and 7) hypertensive patients’

experiences and opinions about checklist use and SMS support at the 12-month follow-up.

4.1.3.3.3 Data collection

All study outcomes were collected during, or immediately after the baseline and final follow-up appointments. The study questionnaires were sent by mail to study patients prior to appointments, where they were collected and saved for analyses together with the home BP measurements.

The outcomes of study I were collected with a questionnaire immediately after the appointment, during which the checklist-supported onset of antihypertensive medication took place. The study patients were asked to fill in the study

questionnaires before they received a copy of the checklist. Thus, they could not use the checklist or enclosed written material to, for example, revise the adequate BP target. Instead, they had to trust their memory as is the usual practice in health care.

4.2 INTERVENTIONS

4.2.1 Theoretical basis: Information-motivation-behavioral skills (IMB) model

Considering poor treatment adherence as the most evident reason for poor BP control, we used the Information-Motivation-Behavioral skills (IMB) model as both a theoretical basis and a practical guide in designing and detailing the study interventions (154). The IMB model was originally created by Fisher et al. in the context of human immunodeficiency virus (HIV) medication adherence promotion.

After that, multiple IMB model-based interventions have been proven effective in enhancing adherence to HIV medication (155-157). An IMB-based intervention has

also been shown to have an impact on promoting general adherence after coronary artery bypass grafting (158). A modified IMB model for diabetes medication adherence has also been validated and used, for example, in the assessment of patient-reported barriers to diabetes medication adherence and in the development of a diabetes self-management mobile application, providing evidence of the IMB model’s ability to explain variance in adherence in a disease context quite similar to hypertension (159-161).

The IMB model suggests that disease-specific information is a crucial element for treatment adherence, but not sufficient by itself (154). It has been previously demonstrated that concentrating on patient information alone is not an effective way to improve BP control or to motivate the patient for better medication

adherence (75,118,162). In addition, motivation and adequate behavioral skills are necessary.

In turn, according to the IMB model, the patient may have good motivation but treatment can still fail because of either inaccurate/insufficient information or insufficient behavioral skills (154). Figure 1 illustrates the IMB model in the context of hypertension. In our study, we used the same theoretical framework not only for promoting medication adherence, but also concerning general treatment adherence (monitoring home BP, attending clinical appointments etc.)

Figure 1. An Information-Motivation-Behavioral skills (IMB) model of hypertension medication and treatment adherence, adapted from Fisher et al., Amico et al and Mayberry et al. (154,160,163). Solid lines indicate an effect between constructs of adherence and the dashed line indicates a feedback loop between health

outcomes and future levels of adherence information and motivation. BP = Blood pressure.

4.2.2 Interventions

4.2.2.1 Checklist for initiation of antihypertensive medication

For the patients in the intervention group, the treating physician used a checklist for the initiation of antihypertensive medication The idea of the checklist was to help the patient to obtain all the essential motivational, informational and behavioral elements for successful hypertension treatment. Furthermore, the checklist guided the physician and patient together through the most critical issues concerning the onset of medication: ensuring the need for treatment is

understood, setting a clear BP target, discussing critical matters related to

medication adherence (such as perceived necessity of medication and potential worries about medication) and hypertension-specific behavioral skills (such as agreement on the first follow-up contact. As part of the process, a copy of the

medication adherence (such as perceived necessity of medication and potential worries about medication) and hypertension-specific behavioral skills (such as agreement on the first follow-up contact. As part of the process, a copy of the

In document Check and support : Enhancing the treatment of hypertension in primary care (sivua 63-176)