Telemonitoring of Hypertensive Patients With Chronic Kidney Disease

Our project is based on implementation of a telemonitoring and case management BP program for remote dwellers with diabetes and CKD in remote regions of Alberta. Our aim is to determine the impact of the intervention on BP control and other CKD risk factors, achievement of guideline-concordant therapeutic targets, evaluate barriers and enablers of implementation of the proposed model of care, patient-centred evaluation and economic modelling.

Objectives:

Objective 1: To determine the effectiveness of telemonitoring and protocol-based case management versus usual care Objective 2: To analyze the usability and acceptability of the intervention Objective 3: To determine the cost-effectiveness and cost utility of the intervention versus usual care

Anticipated outcomes:

(a) improved process of care (control of BP and other CKD risk factors); (b) evidence on user acceptability and safety of the intervention; and (c) better understanding of factors affecting implementation, scale-up and sustainability.

Methods:

Design, Randomization, and Allocation:

The study is a two-arm randomized controlled trial (RCT) comparing the clinical care of remote dwelling hypertensive CKD patients. Independently of the study team, the EPICORE Centre (www.epicore.ualberta.ca) will randomly assign 200 patients to each study arm (1:1): home-based BP telemonitoring plus protocol-based case management versus usual care.

Setting:

This study will be conducted in select remote communities in the Peace Region of Northern Alberta over a period of 2 years.

Inclusion Criteria:

- age ≥ 18 years with a documented diagnosis of established CKD (eGFR < 60 mL/min/1.73m2 and/or proteinuria > 1 g/day), and uncontrolled hypertension (not concordant with the guideline target for CKD)

- English literacy, both verbal and written

- ability and willingness to provide informed consent for participation

Exclusion Criteria:

- systolic BP level > 220 mmHg or diastolic BP level > 110 mmHg on screening BP measurement; immediate consultation with a hypertension specialist will be expedited

- BP within the recommended range

- terminal illness (life expectancy < 1 y)

- participation in a concurrent nephrology or cardiovascular trial

- pregnancy

- patient resides in an area without mobile phone coverage

Intervention:

The patients will be provided with a home BP telemonitoring system whereby a patient pushes a single button on a home BP monitor to initiate measurement, and data are auto-transmitted via Bluetooth to an Android smartphone and relayed to a secured web portal for review. Patients will receive a validated electronic upper arm oscillometric BP device (A&D Ltd. UA-651BLE; San Jose, CA) and an Android smartphone with a prepaid SIM card (enabling wireless data transmission). Patients will take four measurements daily for 1 week using recommended techniques. If BP is uncontrolled (high or low), this 1-week protocol will be followed each month until BP is in the therapeutic range. Once controlled, the 1-week protocol will be repeated every 3 months. Teletransmitted BP readings will be summarized within the health portal using telemonitoring software, which will calculate an overall weekly mean by discarding first-day measurements and averaging the subsequent 24 measurements over the next 6 days. Temporal trends will be plotted. The BP data will be teletransmitted to the study case manager, who will review lifestyle modifications, BP self-monitoring and medication adherence. The case manager will also review telemonitored health portal BP summaries, make protocol-based therapeutic adjustments and send summaries to participants' PCPs to inform them of treatment changes. Medications will be adjusted based on existing treatment protocols.

Control:

For the control arm (usual standard care), participants' home BP series mean, trends, and individual readings will be sent via secure electronic medical records (EMR) to their PCP along with a 1-page summary of Canadian guidelines for BP thresholds, targets, and treatments relevant for CKD.

Baseline data:

- demographics and health behaviours: age, sex, race, smoking, alcohol intake;

- past medical history, particularly significant CVD (coronary artery disease, stroke, peripheral arterial disease, heart failure);

- medication history (BP-lowering drugs): name, type, dosage, frequency, duration;

- anthropometric indices: height, weight, body mass index (BMI), waist circumference;

- upper mid-arm circumference (to determine proper cuff size) using standard measures;

- 24-hr ambulatory BP monitoring (ABPM), home BP series, and pulse rate; BP will be measured according to recommended techniques using validated devices;

- laboratory investigations: serum electrolytes, glycated hemoglobin (A1c), lipids panel, urinary albumin/creatinine ratio;

- health care use in past year: physician and/or emergency room visits, and hospitalizations ascertained through patient self-reports and/or provincial administrative data sources;

- quality of life and utility measurements (Kidney Disease Quality of Life-36 [KDQOL-36] Euro-QOL 5D [EQ-5D] as a preference-based utility measure); and Patient Assessment of Chronic Illness Care [PACIC].

Follow-up data will be collected at 6 months and 1 year, measured as follows:

- clinical endpoints and patient-centred outcomes: 24-hr ABPM, telemonitored home BP, heart rate, medications, anthropometrics, cardiovascular risk factors and markers (A1c, lipids, smoking, urinary albumin), health care use, quality of life and utility, and satisfaction with medical care

- safety endpoints: frequency of syncope and hypotension requiring assistance or medical attention and/or electrolyte disturbances;

- user acceptability: two 10-point Likert scales used to evaluate usability and acceptability

- costing data: trial-based costing using the three-step micro-costing technique of identification, measurement and valuation of relevant health care and non-health care resources using standard methods, with focus on cost of telemonitoring and case management, health care costs (through linkage with Alberta Health data), and patient-borne costs (patient survey).

Data Analysis:

Objective 1: Effectiveness of telemonitoring and protocol-based case management versus usual care:

The primary outcome will be the 1-year difference in the proportion of patients meeting the guideline-concordant BP target between intervention and control groups. We will use 24-hour ABPM because it is the gold standard measurement and the best validated clinical trial BP endpoint. Major secondary outcomes will include the change in mean 24-hour systolic and diastolic BP (overall, daytime and nighttime). First, we will use descriptive statistics to summarize key variables as appropriate, including assessments of temporal trends and tests of normality. Second, we will calculate the 6-month and 1-year mean changes from baseline in each outcome and compare them between study arms using chi-square tests for dichotomous outcomes and unpaired t-tests for continuous outcomes. Third, we will identify multivariable predictors of the 1-year change in a given outcome using appropriately constructed and calibrated logistic regression models for dichotomous outcomes (including the primary outcome) or linear regression models for continuous outcomes. Initial models will be appropriately adjusted for key confounding demographic, clinical and laboratory variables. Missing data will be handled using last observation carried forward.

Objective 2: Usability and acceptability of the intervention:

Assessment of usability and acceptability is critical for all technology-enhanced care interventions to minimize risk of undesired consequences commonly seen post-implementation. End-user input into system design and operation, and rigorous system evaluation is important to reduce the risk that interventions will be ineffective, unusable or unsafe. We will take a qualitative approach to this process by using focus groups. A target sample size of 5 to 8 participants will be recruited from patients in the telemonitoring arm and care providers. The first 5 users typically identify 70% of severe usability problems and the first 8 users typically identify 85%; a higher sample size has low incremental yield because of data saturation.Focus groups will be conducted for each of the group (patients and care providers) at baseline and follow-up. Sessions will be audiotaped, transcribed verbatim and analyzed using the methodology of interpretive description (ID). NVIVO, a qualitative software system, will be used to create a filing system and coding database. The first focus group in each group will be read and re-read to generate an initial codebook. The codebook will be iteratively refined throughout the analysis. Codes will be categorized and analyzed thematically. Patient and clinician data will first be analyzed separately, and then across groups.

Objective 3: Cost-effectiveness and cost utility of the intervention versus usual care:

Cost-effectiveness (Model A) and cost utility (Model B) of the intervention compared with the control arm will be analyzed using a modified version of a validated economic model. Best practices for economic evaluation will be followed. The intervention includes intermittent or one-time costs (e.g., devices, treatment algorithm development, patient and health care professional training) and ongoing costs (e.g., network and data, case manager time, BP medications). Data captured during the RCT will accurately determine the costs of telemonitoring, case management, and health care system use by patients. The change in BP at 1 year, a validated surrogate, is a key variable in both analyses. Model A will calculate the cost per decrement in mean BP using a 1-year time horizon and a public health care payer perspective. Model B will estimate incremental cost per quality-adjusted life year (QALY) gained over a lifetime. It will consider longer-term health outcomes, including the probability of developing ESRD (requiring dialysis or transplant) or major cardiovascular events (i.e., coronary artery disease, heart failure, stroke), all-cause mortality, impacts on quality of life (EQ-5D), and health care costs associated with these outcomes. In each model we will explore distributions using bootstrapping, create acceptability curves, and perform one-way and probabilistic sensitivity analyses to explore the impact of uncertainty in key model parameters. In Model B, the potential impact of reducing complications of treating hypertension in patients with CKD (hypotension, syncope) by a plausible range will be explored in sensitivity analysis using baseline risk determined from a population-based cohort of patients and ambulatory care sensitive conditions for emergency room visits and hospitalizations.