In total we extracted data from 52 RCTs (10,388 patients): 14 HF trials (4,264 patients); 14 OAT trials (3,049 patients); 18 BP trials (1,714 patients) and 6 SMBG trials (1,361 patients). Remote monitoring in heart failure (HF) is effective in reducing all cause mortality (Relative Risk = 0.62 (95%CI 0.45 to 0.85) [
2]. Self-monitoring of INR for patients on oral anticoagulation therapy leads to fewer thromboembolic events (Odds Ratio = 0.27 (0.12 to 0.59)), and lower mortality (OR = 0.37 (0.16 to 0.85)) [
3]. Self-monitoring of blood pressure leads to reductions in systolic BP of 4.2 mmHg (95% CI, 1.5 to 6.9) as well as diastolic pressure 2.4 mmHg (95% CI, 1.2 to 3.5) [
4]. Self-monitoring of blood glucose (SMBG) which has been found to be effective for patients with type 1 diabetes and type 2 diabetes if they are taking insulin appears not be effective in improving HbA1c in with patients with type 2 diabetes using oral hypoglycaemic drugs [
5].
Components of self-monitoring
We have summarised the data extracted into four main component areas forming the self-monitoring interventions described in the trials: a) education b) self-measurement c) adjustment of (or adherence to) medication and/or behaviour d) contact with health professionals. Not all components were evident in every trial, and in some cases the intervention was too poorly described to be clear whether or not a component was present.
1) Education
The first component in a self-monitoring intervention is patient (and sometimes health professional) education. At its most basic, once patients have been identified as suitable for self-monitoring they need to be taught how to use the self-monitoring equipment [
22]. However, the education can also be used to provide patients with information on disease, management and lifestyle [
23]. Initial training may be necessary for the health professionals involved [
24] to deliver an effective education session. A theoretical basis for the type of training provided has also been suggested [
25,
26].
The intensity and structure of education varied considerably even for the same clinical problem. In all the heart failure trials, provision of education was part of the interventions with ongoing phone contact by study staff for counselling and education in addition to the monitoring of signs and symptoms. Nonetheless, the amount of education varied. One study [
22] did not mention education about heart failure or counselling during monthly phone calls at all and only reported education on the operation of the telemonitoring equipment. In contrast another trial provided one-to-one or group counselling covering disease, management, lifestyle and monitoring as well as providing extensive educational materials. The patients were also assessed before and after each session and were able to contact medical staff for advice and help [
23]. No trials mentioned any theoretical basis for their educational strategies.
In self-monitoring of INR studies, most trials gave two to three educational sessions, including an assessment of competency. One trial [
27] based their educational component on Social Learning theory [
16,
28] and many trials referred to each other to determine the education strategy [
29‐
32].
In BP self-monitoring trials, education varied from very comprehensive (education provided to both the patients and health care providers)[
24] to patients simply being instructed in the use of the monitor[
25]. Training and education was based on a theoretical model in only one trial [
26]; the model adopted was the Health Belief Model [
33].
For the self-monitoring of blood glucose trials, education ranged from ongoing counselling and education using an algorithm delivered monthly by a trained nurse[
34] to instructions in the use of a monitor and renewal of dietary recommendations twice [
35]. Again no theoretical basis was given for delivery of the education provided, though one study used behavioural rather than didactic education as a means of increasing compliance [
36].
2) Self-measurement
The obvious defining feature of self-monitoring is self-measurement. Hence a major focus of the initial education was on the "how to" and the interpretation of self-measurements. Self-measuring regimes have to take into account the accuracy of the monitoring device, the run in period required to ensure patients are safe and effective at self-measuring, the quality assurance of the monitoring device and the frequency with which patients are required to self-measure.
Ignoring the evidence underpinning self-measurement regimes led to poor ongoing research. For example, in an early trial of blood pressure self-monitoring, daily self-measurement of blood pressure proved too much for most patients [
37] with only 1/5
th of patients allocated to the intervention completing the trial. Yet 11 of the 18 hypertension trials asked the patients to measure daily or more frequently without any rationale given for doing so [
4].
Considerable variation occurred in what patients were asked to monitor and the extent of external monitoring in the trials. The minimum heart failure patients were asked to do was measure their weight, record their medications daily and monitoring by a monthly phone call by a research coordinator [
38]. In contrast, other patients were asked to electronically measure weight, blood pressure, heart rate and rhythm twice a day and transmit results immediately to the study centre[
22].
In the INR trials, only one study mentioned a rationale for the frequency of testing. This trial suggested testing twice a week was optimal to keep patients within the target therapeutic range based on previous research [
39]. There was considerable variation between what patients were told to do and what they actually did in terms of frequency of INR measurements. For example, in one trial they were instructed to take 11 measures over the 6 months and whilst the median number of measures performed was 17, the range was 2 to 39 [
27]. Only 6 trials (43%) mentioned external quality control of the tests [
29,
30,
40‐
43].
In the BP trials, self-measurement varied considerably as well as the mode of recording and the responses to the readings (self-management or other). This ranged from patients taking electronic BP self-measures three times in the morning and evening at least three times a week with the readings being automatically transmitted to the study centre, providing patients and physicians with weekly averages, but compliance with this regime was not reported [
44]. In contrast, in another trial patients were asked to take two consecutive readings twice a week and mail results to the study centre once a month but only 50% of patients were compliant[
45]. No trial gave a rationale for the monitoring regime the patients were asked to perform which ranged between 1[
46] and 21[
47] blood pressure readings per week. Overall compliance with the measuring strategies ranged from 15% to above 90%.
The trials of self-monitoring of blood glucose in patients with type-2 diabetes varied in the number of tests patients performed and in whether tests were pre or post-prandial. Whilst some authors specifically hypothesised that postprandial changes in glucose were important [
34,
35,
48,
49], none gave any rationale for their decision on how often patients were asked to test which ranged from 24 [
50] to 48 [
34,
48] times per month. Reported compliance with testing ranged from 45% to above 90%, however, in some cases patients tested twice as often as requested, 25 times a week instead of 12 [
34]. Quality assurance of measurements was mentioned in only two of the six trials [
34,
50].
Adjustment/Adherence
In self-monitoring some subsequent action needs to occur to lead to a clinical change: either adjustment of treatment or better adherence to treatment. If the purpose of the self-monitoring is to increase motivation or reinforce behaviour then there appears to be no point in re-testing before the behaviour has had time to produce a meaningful change in what is being measured. For example, the lack of effective blood glucose self-monitoring [
5] where the results of behaviour changes to diet and exercise do not have time to effect what is being monitored. An example of where this does work well is self-management of INR where the test results inform medication dose adjustments, the effects of which can be seen in the next test result [
3]. The relatively small effect size found in the blood pressure trials may also be due to the fact that the patients do not make adjustments in response to their test results or their tests are too frequent for meaningful changes to have taken place.
The heart failure trials differed in their aims and consequently in what the patients were asked to do. In more intense interventions the aim was to provide clinicians with diagnostic information once or twice daily to improve titration of medications [
22]. In direct contrast other trials were designed to do less self monitoring and more self-care, while adjustments of therapy were minimal[
38]. Patients carried out their own dose adjustments in two-thirds (64%) of the INR trials but in only one of the blood pressure trials [
51] where adherence to medication was more often the purpose of the self-monitoring rather than adjustment of therapy. In the diabetes trials the patient self-adjustments were to lifestyle and nutrition with any medication adjustments being carried out by an external health professional [
35,
36,
48,
50].
Health care professionals
It is clear that the purpose of the contact with health professionals in many of the trials was to increase compliance with medication and measurement and provide physicians with information for therapy adjustment. Contact with health professionals was also a way of periodically reinforcing or updating education. However, it is not clear how much contact is optimal and whether contact is best by phone, home visit, clinic visit or computer.
In most of the heart failure trials the purpose of the contact was mainly for monitoring symptoms, medication, adherence and education and advice. The health professionals contacting the patients were mainly nurses with only two trials using pharmacists[
52,
53] and one a research co-ordinator[
38]. The mode of contact was phone in all but two trials where videoconferencing was used[
54,
55]. The amount of contact varied from three calls by a pharmacist over 6 months [
52] to 17 calls from a nurse (median 14 range 11 to 22) over 6 months[
56].
In the INR trials the ongoing contact was mainly for safety. Contact was either with clinicians or nurses and in one study a pharmacist also contacted patients monthly [
42]. Some patients were able to contact a 24 hour help desk,[
29]or a clinician available 24 hours [
27] or during work hours [
40] though in most trials patients were given instructions on when they needed to contact a health professional.
In the BP trials the purpose of the contact with health professionals varied from motivation[
57] to medication adjustment during phone calls or clinic visits[
44]. Most of the contacts were with doctors or nurses during clinic visits but in one case the contact was with a study co-ordinator who had no health professional training[
57], in another a pharmacist contacted the patients[
58] and in one the contact was with a telephone linked computer system[
46]. The frequency of contact varied from mean of 1.5 clinic visits (SE 0.1) over 12 months[
59] to weekly phone calls from a nurse for counselling[
26].
The purpose of the contact with health professionals in the blood glucose trials was both for education and motivation. Dieticians[
35,
48,
49], doctors and nurses [
34,
36,
49,
50]contacted the patients. The trials were 24 to 28 weeks with five[
35] to at least 15[
49] contacts over that period.
Finally, few trials mention a theoretical basis for the mediating effect of the results of the self-measurement on the patients' behaviour. For example, what is the theoretical basis for the assumption that the blood pressure self-monitoring results will increase compliance with hypertension medication?
Based on the gaps identified we summarised the factors to be considered at each stage in the design of a self-monitoring intervention in table
1. These factors cover aspects related to purpose, people, content and timing of the four main components of self-monitoring interventions. Potentially these factors could be used as a checklist when designing self-monitoring interventions for CVD disease (table
1). Answering the relevant questions in the table could not only improve the evidence base of the interventions but may also make them more reproducible.
Table 1
Self-monitoring components - factors to be considered at each stage in the design of a self-monitoring intervention
Purpose
|
What is the purpose?
• To increase knowledge • To provide skills • To increase compliance • To increase motivation |
What is the purpose?
• To provide information • To increase compliance • To increase motivation |
What is the purpose?
• Titration of medication • Titration of behaviour • Adherence to medication regime • Adherence to behaviour regime |
What is the purpose?
• To provide education • To increase compliance • To increase motivation • Safety |
People
|
Who receives the education?
• Patients • Health care providers |
Who is the information for?
• Patients • Health care providers |
Who adjusts/adheres?
• Patients • Health care provider |
Which health professionals?
• Doctors • Nurses • Other |
Content
|
What type of education?
• Theoretical basis • Content • Mode of delivery • Support materials |
What test is to be used?
• Accuracy of the test • Feasibility in this setting |
What is adjusted/adhered to?
• Medication • Behaviour |
What is the format of the contact?
• Effectiveness • Cost |
Timing
|
What timing is optimal?
• Should it be once off or repeated • How long between sessions • Is it sustainable |
What timing is optimal?
• How long should the run in be
What is the frequency of measurements and does it take account of:
• The signal to noise ratio • Fatigue factor and compliance |
What is the frequency of adjustments considering:
• The signal to noise ratio • Fatigue factor and compliance |
What is the timing of the contact?
• Feasibility • Compliance |
Other
|
Should the learning be assessed?
• How many assessments • What level of success before allowing self-measurement • How much re-training |
What quality assurance is required?
• Internal QA and External QA • How often should QA be conducted?
What is recorded?
• How reliable is the recording method • How accurate is the recording method • Is electronic recording available and feasible |
What guidance is provided?
• Algorithm • Web based guidance • Clinician guidance |
Other?
• Algorithm based contact • Video conference |