Investigating the behavioural effects of a mobile-phone based home telehealth intervention in people with insulin-requiring diabetes: Results of a randomized controlled trial with patient interviews

Abstract

Introduction

Evidence supporting home telehealth effects on clinical outcomes in diabetes is available, yet mechanisms of action for these improvements remain poorly understood. Behavioural change is one plausible explanation. This study investigated the behavioural effects of a mobile-phone based home telehealth (MTH) intervention in people with diabetes. It was hypothesized that MTH would improve self-efficacy, illness beliefs, and diabetes self-care.

Methods

A randomized controlled trial compared standard care to standard care supplemented with MTH (self-monitoring, data transmission, graphical and nurse-initiated feedback, educational calls). Self-report measures of self-efficacy, illness beliefs, and self-care were repeated at baseline, three months, and nine months. MTH effects were based on the group by time interactions in hierarchical linear models and effect sizes with 95% confidence intervals (CIs). Interviews with MTH participants explored the perceived effects of MTH on diabetes self-management.

Results

Eighty-one participants were randomized to the intervention (n = 45) and standard care (n = 36). Significant group by time effects were observed for five out of seven self-efficacy subscales. Effect sizes were large, particularly at nine months. Interaction effects for illness beliefs and self-care were non-significant, but effect sizes and confidence intervals suggested MTH may positively affect diet and exercise. In interviews, MTH was associated with increased awareness, motivation, and a greater sense of security. Improved self-monitoring and diet were reported by some participants.

Discussion

MTH empowers people with diabetes to manage their condition and may influence self-care. Future MTH research would benefit from investigating behavioural mechanisms and determining patient profiles predictive of greater behavioural effectiveness.

Keywords

Telehealth ehealth self-care home telecare

Introduction

Suboptimal adherence to recommended diabetes self-care behaviours is a well-recognized problem.^1–4 Adherence rates have been found to be 10–80% for long-term exercise,^1,5 35–37% for diet,⁵ 64–70% for self-monitoring of blood glucose (BG),⁵ and 71–72% for appointment keeping.⁵ This has implications for the successful achievement of clinical targets, which for glycosylated hemoglobin (HbA1c) and blood pressure (BP), for example, are not reached by 40–60% of people with diabetes.^6,7

The type of home telehealth intervention examined in this study involved the person with diabetes using technology to record and transmit data to health care professionals and receive feedback. Home telehealth provides health care professionals with the ability to provide timely care, detect early signs of deterioration, and identify/monitor closely high risk patients. Reviews of these interventions in people with diabetes suggest that they result in improvements in HbA1c^8–10 and that they are well received by patients.¹¹

In addition to supporting clinical care, home telehealth aims to improve treatment adherence.¹² Self-monitoring and feedback are the hallmark of home telehealth.¹³ These behaviour change strategies have successfully changed peoples’ behaviours^14–18 by improving their confidence in managing their health (i.e. self-efficacy) and the way they think about it (i.e. self-regulation).¹⁹ Education/information is another behaviour change strategy integrated into many telehealth interventions. It is sometimes considered a secondary feature of home telehealth²⁰ as real-time education via feedback²¹ is dependent on clinical data transmission. Despite the potential for home telehealth to influence diabetes self-care, few studies measure these effects. Only three (17.7%) of 26 ‘telediabetes’ studies were found to report on the effects of telehealth on self-care (e.g. carbohydrate counting, physical activity) in a recent review.¹²

Studying the effects of home telehealth on self-care would not be complete without understanding pathways of change. Research recommendations have underlined the need to understand these causal mechanisms.²² Such investigations are likely to be facilitated by use of theory. Theories of health behaviour propose that beliefs (e.g. beliefs/knowledge about the condition or the performance of the behaviour), attitude, and expectations (e.g. efficacy or outcome expectations) drive behaviour²³ and act as mediators of behaviour change. So far, little research has investigated pathways of behavioural change in home telehealth. One review²⁴ examined the effects of home telehealth on two mediators of behaviour change, i.e. knowledge and self-efficacy. The small number of studies reviewed, however, combined with their poor methodological quality, made conclusions difficult to reach.

We have reported the effects of a randomized controlled trial (RCT) based evaluation of home telehealth on clinical outcomes in diabetes.²⁵ Primary (HbA1c) and secondary clinical outcomes (BP, insulin dose, diabetes outpatient appointments) were not significantly affected by the intervention. The current paper reports on the more proximal effects of this intervention on mediators of behaviour change (self-efficacy and illness beliefs) and diabetes self-care. The selection of these behavioural outcomes was guided by social cognitive theory²⁶ and self-regulation theory.²⁷ It was hypothesized that home telehealth would result in improvements in self-efficacy, illness beliefs, and diabetes self-care. Interview data collected on participants’ perception of the behavioural effects of home telehealth is also included in this paper to supplement the quantitative findings.

Methods

Study design

This study used a RCT design to compare standard care supplemented with a mobile-phone home telehealth intervention, i.e. mobile telehealth (MTH), to standard care alone. Semi-structured interviews were conducted with participants after the MTH intervention. Ethical approval was gained from the University College London/University College London Hospitals Committees on the Ethics of Human Research (09/H0715/69).

Setting and participants

The study was conducted in the diabetes unit of a secondary health centre in a multiethnic borough of London, United Kingdom (UK). Based on a power calculation detailed in the published protocol,²⁸ the aim was to recruit 248 participants. Participants were adults with poorly controlled type 1 or type 2 diabetes (most recent HbA1c had to be taken in the last 12 months and above 7.5%) and adequate English language. Patients requiring district nurse home visits for BG monitoring and insulin administration were excluded, as well as those with prior MTH experience, visual or dexterity problems, or those who travelled regularly outside the UK (≥3 weeks). A diagnosis of kidney failure, sickle cell disease, or severe mental illness resulted in exclusion from the study, as did pregnancy.

Recruitment and randomization

Potentially eligible participants were met by a researcher after their outpatient appointment. Full eligibility was verified. Interested participants signed a consent form, and received a baseline questionnaire and a pre-stamped envelope for return. Once completed, participants were randomized to the intervention or control group (standard care) using an online sequence generator that generated blocks of 20. A separate consent form was signed prior to the interview.

The MTH intervention

Intervention participants received the MTH equipment and training to transmit diabetes-related data (i.e. BG and BP readings, time since last meal, level of physical activity performed, insulin dose, and weight) to a MTH nurse for feedback. Automated and colour-coded graphical feedback on clinical readings (BG and BP) was displayed on the mobile phone whenever new entries were recorded. These graphs compared the most recent clinical reading to those recorded in the last five and 30 days. Each colour represented a different glycaemic state (blue: hypoglycaemia, green: normoglycaemia, amber: borderline hyperglycaemia, red: hyperglycemia). The monitoring protocol²⁸ included feedback on out of range clinical readings, support for insulin titration, and six-weekly educational calls. Participants requiring substantial medication adjustments were encouraged to schedule an appointment with their diabetes specialist nurse. MTH participants continued to receive standard care which consisted of one follow-up appointment with the diabetes specialist nurse every 3–4 months and 1–2 annual appointments with the diabetes consultant. A diabetes specialist nurse was available during working hours to respond to patients’ urgent queries.

Measures

Behavioural outcome data were collected at baseline, three, and nine months using measures with adequate psychometric properties.^29–33 Follow-up questionnaires were mailed to participants.

Self-efficacy

Two subscales of the health education impact questionnaire (HeiQ, v3.0)³⁴ were used to measure empowerment. The self-monitoring and insight subscale measures beliefs in one’s ability to monitor the physical and/or emotional responses for appropriate self-management. The skills and technique acquisition subscale measures one’s beliefs in the knowledge-based skills and techniques acquired to self-manage health. Respondents indicated their level of agreement using a six point Likert scale. Higher mean subscale scores indicated greater self-efficacy.

The insulin management diabetes self-efficacy scale (IMDSES)³¹ measures diabetes-specific self-efficacy relating to general management, insulin, diet, exercise, and foot care. A six point Likert scale was used, with higher scores indicating greater self-efficacy. One item on food exchange was removed as this concept is not commonly used in the UK (‘I can correctly exchange one food for another in the same food group’).

Illness beliefs

The personal models of diabetes (PMD) scale assesses beliefs on perceived seriousness of diabetes as well as beliefs on perceived effectiveness of treatment to control diabetes and prevent complications.³⁵ A five point Likert scale was used. Higher mean subscale scores indicated stronger endorsement of these beliefs.

Diabetes self-care

The summary of diabetes self-care activities (SDSCA)³³ assesses the number of days in a week (0–7) on which participants engaged in general diet, fruit and vegetable intake, fat intake, exercise, foot care, and BG testing. Higher mean subscale scores indicated greater frequency of self-care behaviours. The SDSCA does not differentiate between participants who self-monitor BG once or more each day. An item to capture this information as free text was added into the survey (‘How many times approximately do you measure your blood glucose levels in one week?’).

Interview topic guide

The interview guide included a question on perceived effects of the MTH on diabetes self-management. Participants were asked: ‘To what extent has the use of MTH influenced the way you self-manage your diabetes?’ Other areas of interest were covered;²⁸ findings for these will be published separately.

Data analysis

Quantitative data were analysed using an intention-to-treat principle. Baseline differences were checked using independent Student t-tests and chi-square tests. Intervention effects were examined using hierarchical linear models (HLM). Random effects accounted for within-participant correlation. A first order autoregressive covariance structure was used. Baseline differences were adjusted for (fixed effects). Interactions between group and time (fixed effects) were used to identify differential treatment effectiveness. Hedges’ g standardized effect sizes for mean group differences at follow-up were computed with 95% confidence intervals (CIs). Significance was set at p < 0.05.

Audio-recordings of interviews were transcribed verbatim, de-identified, and imported into NVivo (v10.0). Recommendations for thematic analysis were followed.³⁶ One researcher coded all transcripts. A second researcher independently coded two transcripts to help discuss emerging themes. Once coding of all transcripts was complete, these two researchers met again to finalize the qualitative analysis.

We intended to combine the results of the quantitative and qualitative analyses in our discussion of the behavioural effects of MTH. Our intention was to use the interview data to verify and validate the results of the quantitative analyses. The interviews were also anticipated to allow for a richer interpretation of quantitatively measured MTH effects, and to help identify MTH effects not assessed in questionnaires.

Results

Sample characteristics

Eighty-one participants were randomized to the intervention (n = 45) and control (n = 36) groups. Just over half (56.8%) were male. Mean age was 57.2 ± 13.6 years and the majority of participants (65.4%) had no formal education or no further than GCSE/O-levels. Mean number of years with diabetes was 16.6 ± 7.7. The majority (87.7%) had type 2 diabetes. Intensive insulin therapy (≥3 daily injections) was prescribed to half (53.1%) of participants. Mean HbA1c (%) was 9.1 ± 1.8 in the intervention group and 8.9 ± 1.7 in controls. The majority were overweight (25.9% with a body mass index (BMI) ≥25 kg/m²) or obese (56.8% with a BMI ≥ 30 kg/m²). Gender was the only statistically significant difference (p = 0.013) between interventions (68.9% male) and controls (42.7% male).

Attrition

Data were not imputed for missing questionnaires. Survey data were available for 81 participants at baseline (45 interventions, 36 controls), 74 at three months (41 interventions, 33 controls), and 70 at nine months (39 interventions, 31 controls).

Intervention effects on self-efficacy and illness beliefs

Table 1 presents the adjusted estimated marginal means from the HLM models to examine the effects of MTH on self-efficacy and illness beliefs. Group by time interactions effects were significant for four out of five diabetes-specific self-efficacy subscales (general management, p = 0.010; diet, p = 0.049; insulin, p = 0.017; foot care, p = 0.017) and for one of two HeiQ subscales measuring a more generic health self-efficacy (self-monitoring and insight, p = 0.010). Effect sizes for these mean group differences at three and nine months were moderate to large (–0.27 and −0.65), with magnitudes generally increasing over time. Whilst some of the respective 95% CIs crossed zero for three month differences, none of them did at nine months, lending further support to the existence of an intervention effect on self-efficacy at this time point.

Table 1.

Changes in self-efficacy and illness beliefs in the intervention and control groups.

Measures	Intervention (mean ± SD)	Control (mean ± SD)	F-statistic, p-value for group (2) × time (3)	Effect size for group differences at 3/9 months (95% CIs)
IMDSES –General management			F_2,120.38 = 4.79, p = 0.010*
Baseline	27.78 ± 5.98	27.58 ± 5.85		–
3 months	29.79 ± 5.89	26.31 ± 5.80		−0.58 (−1.04, −0.12)
9 months	29.66 ± 5.81	25.83 ± 5.68		−0.65 (−1.13, −0.17)
IMDSES – Diet			F_2,112.38 = 3.10, p = 0.049*
Baseline	23.44 ± 6.76	23.84 ± 6.61		–
3 months	24.79 ± 6.64	22.92 ± 6.49		−0.28 (−0.73, 0.18)
9 months	25.26 ± 6.57	22.11 ± 6.43		−0.47 (−0.94, 0.00)
IMDSES – Insulin			F_2,115.43 = 4.19, p = 0.017*
Baseline	40.67 ± 8.70	40.50 ± 8.50		–
3 months	41.43 ± 8.52	39.14 ± 8.32		−0.27 (−0.72, 0.19)
9 months	44.81 ± 8.40	39.77 ± 8.32		−0.59 (−1.06, −0.11)
IMDSES – Exercise			F_2,115.51 = 0.19, p = 0.827
Baseline	7.28 ± 2.78	6.54 ± 2.72		–
3 months	7.36 ± 2.74	6.25 ± 2.68		−0.40 (−0.86, 0.06)
9 months	7.36 ± 2.52	6.37 ± 2.67		−0.37 (−0.84, 0.10)
IMDSES – Foot care			F_2,122.617 = 4.24, p = 0.017*
Baseline	9.42 ± 2.56	9.91 ± 2.51		–
3 months	10.05 ± 2.53	9.34 ± 2.49		−0.28 (−0.73, 0.18)
9 months	10.36 ± 2.52	8.94 ± 2.48		−0.55 (−1.03, −0.08)
HeiQ – Self-monitoring and insight			F_2,140.31 = 4.77, p = 0.010*
Baseline	4.77 ± 0.75	4.74 ± 0.74		–
3 months	5.00 ± 0.74	4.76 ± 0.72		−0.32 (−0.78, 0.14)
9 months	5.05 ± 0.72	4.57 ± 0.71		−0.65 (−1.13, −0.17)
HeiQ – Skills and technique acquisition			F_2,126.08 = 1.92, p = 0.151
Baseline	4.52 ± 0.98	4.40 ± 0.96		–
3 months	4.56 ± 0.97	4.45 ± 0.95		−0.11 (−0.57, 0.34)
9 months	4.72 ± 0.96	4.23 ± 0.94		−0.50 (−0.98, −0.03)
PMD – Seriousness			F_2,97.26 = 0.59, p = 0.558
Baseline	3.35 ± 0.84	3.22 ± 0.82		−
3 months	3.13 ± 0.82	3.11 ± 0.80		−0.02 (−0.48, 0.43)
9 months	3.16 ± 0.82	2.97 ± 0.80		−0.23 (−0.70, 0.24)
PMD – Treatment effectiveness			F_2,144.91 = 0.36, p = 0.698
Baseline	4.29 ± 0.66	4.29 ± 0.65		–
3 months	4.31 ± 0.66	4.21 ± 0.64		−0.15 (−0.60, 0.30)
9 months	4.38 ± 0.64	4.27 ± 0.63		−0.17 (−0.64, 0.30)

Notes: Higher scores indicate greater self-efficacy or endorsement of beliefs; * indicates significant interaction effects.

IMDSES: insulin management diabetes self-efficacy scale; HeiQ: health education impact questionnaire; PMD: personal models of diabetes; SDSCA: summary of diabetes self-care activities; CI: confidence intervals.

Although the group by time interaction effect for the second HeiQ subscale on skills and technique acquisition was not significant, the effect size for this mean group difference at nine months was large (−0.50) and the 95% CI did not cross zero. This suggests the intervention may have had an effect on this subscale. The lack of significance may be related to lack of power.

In contrast, none of the group by time interaction effects for illness beliefs were statistically significant. Effect sizes for mean differences at three and nine months were small (−0.02 and −0.23) and crossed zero.

Intervention effects on diabetes self-care

Table 2 presents the adjusted estimated marginal means from the HLM models to examine the effects of MTH on diabetes self-care. None of the group by time interaction effects were significant. Effect sizes for mean group differences were large for the exercise subscale at three months and the general diet subscale at nine months (−0.47 and −0.54, respectively), with 95% CIs not crossing zero. The intervention may have influenced these self-care activities, but lack of power may have prevented p-values for these effects from reaching significance.

Table 2.

Changes over time in diabetes self-care in the intervention and control groups.

Measures	Intervention (mean ± SD)	Control (mean ± SD)	F-statistic, p-value for group (2) × time (3) interaction	Effect size estimate for group differences at 3/9 months with 95% CIs
SDSCA – General diet			F_2,108.63 = 2.78, p = 0.067
Baseline	4.63 ± 1.83	4.65 ± 1.81		–
3 months	5.24 ± 1.81	5.23 ± 1.78		−0.01 (−0.46, 0.45)
9 months	5.31 ± 1.82	4.31 ± 1.78		−0.54 (−1.02, −0.07)
SDSCA – Fruit and vegetable			F_2,105.59 = 1.61, p = 0.204
Baseline	4.57 ± 2.17	4.79 ± 2.12		–
3 months	4.71 ± 2.15	4.62 ± 2.09		−0.04 (−0.50, 0.41)
9 months	4.84 ± 1.89	4.14 ± 2.07		−0.35 (−0.82, 0.12)
SDSCA – Fat intake			F_2,118.78 = 0.06, p = 0.939
Baseline	4.49 ± 1.82	4.40 ± 1.79		–
3 months	4.58 ± 1.82	4.62 ± 1.77		0.02 (−0.43, 0.48)
9 months	4.86 ± 1.79	4.77 ± 1.76		−0.05 (−0.52, 0.42)
SDSCA – BG monitoring			F_2,113.49 = 0.61, p = 0.826
Baseline	4.86 ± 2.04	5.62 ± 2.00		–
3 months	4.67 ± 2.01	5.13 ± 1.97		0.23 (−0.23, 0.68)
9 months	4.60 ± 1.99	4.82 ± 1.95		0.11 (−0.36, 0.58)
SDSCA – Exercise			F_2,125.01 = 2.03, p = 0.135
Baseline	2.78 ± 2.10	2.73 ± 2.06		–
3 months	3.09 ± 2.07	2.10 ± 2.03		−0.47 (−0.93, −0.01)
9 months	3.12 ± 2.05	2.38 ± 2.02		−0.35 (−0.82, 0.12)
SDSCA – Foot care			F_{2, 102.76} = 0.36, p = 0.701
Baseline	3.43 ± 2.58	3.76 ± 2.52		–
3 months	3.73 ± 2.52	3.69 ± 2.46		−0.02 (−0.47, 0.44)
9 months	3.79 ± 2.50	3.77 ± 2.44		−0.01 (−0.47, 0.46)
Weekly frequency BG monitoring^a			F₂,_134.68 = 1.04, p = 0.312
Baseline	10.31 ± 15.76	9.19 ± 15.55		–
3 months	9.74 ± 15.71	11.39 ± 15.51		0.10 (−0.35, 0.56)
9 months	13.68 ± 16.66	11.29 ± 15.52		−0.14 (−0.61, 0.32)

Notes: Higher scores indicate greater frequency of self-care behaviours. ^aIndividual item added into SDSCA survey.

SDSCA: summary of diabetes self-care activities; BG: blood glucose.

Themes from qualitative data

Forty MTH participants completed the intervention and 26 (65%) agreed to be interviewed. Four themes on the perceived effects of MTH on diabetes self-management emerged (Table 3 presents supporting quotations).

Table 3.

Themes from the qualitative enquiry on perceived effects of mobile home telehealth (MTH).

Theme 1: Increased awareness
… of level of diabetes control	‘Actually it was really good because then it gives me an idea of what group that I fall in. Mostly I see that I am in the yellow, but I should be green and sometimes I am in the red which is not too good for me, but um … It really does help for you to see’. (Female patient, 33 years)
… of the factors influencing diabetes control	‘I mean I can see things and change, I get quite annoyed with myself when I see the reading go up, so I start investigating what have I been doing that’s done that. Now this is the opportunity that you’ve got with this system, that you wouldn’t have any other way’. (Male patient, 73 years)
Theme 2: Increased motivation
MTH is a motivational tool/personal challenge	‘So let me tell you on the positive side, this is my health. Somebody is helping me to do it. I need to prove, so it became a challenge to me’. (Male patient, 48 years)
Theme 3: Influence on diabetes self-care
Increased monitoring of blood glucose	‘Sometimes I do, I did, but since I’ve got this [patient shows the phone], I do it more often. I do it more often because sometimes I go … I’m going out and I don’t bother, but with this because I know it will be recorded’. (Male patient, 71 years)
Dietary changes	‘Well the feedback I’m getting now … Before I just couldn’t be bothered, I would eat anything. Now I think, I think I’m more cautious of what I can and cannot eat, or the portion that I can have’. (Female patient, 33 years)
Improved insulin intake and adjustments	‘It does make a difference; it’s made me think that I’ve got to keep to it … I’ve got to keep it down, I’ve got to keep it even. No ups and downs. You know if it goes up I give a bit more insulin, see if I can make it come down’. (Male patient, 69 years)
Theme 4: Perceived sense of security
Someone there to monitor my clinical readings and prevent complications	‘You know where you have regular readings going through to somebody and uh … you know at least, if something happens, you know that somebody is monitoring, you know. Though I didn’t have much contact with [name of home telehealth nurse], at the back of my mind I had that peace of mind, which I knew that if something was really drastically wrong, they would get in touch with me because somebody was looking at my readings’. (Male patient, 53 years)

Theme 1: Increased awareness

Participants reported an increased awareness of the extent to which they controlled their diabetes. Visualizing the graphical feedback increased participants’ awareness of trends in BG. This allowed them to see ‘the bigger picture’ and to position their overall level of diabetes control on a continuum from poor to good. The colour-codes were reported to facilitate the identification of time periods when tighter control of BG readings was needed. They also prompted participants to think about the reasons behind fluctuations in BG and participants reported a heightened awareness of the factors influencing diabetes control.

Theme 2: Increased motivation

Participants reported that MTH was an opportunity to think of making changes to their diabetes self-management. The graphical feedback in particular was considered to act as an incentive to improve BG readings. Combined with the awareness that a MTH nurse was available to monitor their readings, participants felt that controlling their BG had become a personal challenge. Some used the average BG reading provided in the automated feedback to try ‘to beat that average for a better average next time’. MTH acted as a reminder to self-care in moments of behavioural slippage.

Theme 3: Influence on diabetes self-care

Self-monitoring of BG was the behaviour most frequently reported to have increased. This was because receipt of graphical feedback, and knowledge that someone reviewed their data, was considered to add ‘purpose’ to this behaviour. Some participants also reported improvements in dietary intake through reported reductions in portion sizes and snack consumption. In fewer cases, MTH was reported to increase adherence to insulin (i.e. fewer missed dosages) or to insulin adjustments.

A minority of participants reported no benefits from MTH on self-care. They tended to be dissatisfied with the MTH nurse feedback, to not use the graphical feedback, or to consider their diabetes control as adequate.

Theme 4: Perceived sense of security

Most participants experienced a heightened feeling of security and safety because they believed that someone to whom they were ‘directly in some mysterious way connected with’ was watching over their BG readings. The knowledge that someone was ‘in the background’ to check BG readings and prevent any complications created a peace of mind. Two participants compared MTH with ‘Big Brother’, a reality television show involving continuous monitoring by cameras.

Discussion

Intervention effects on self-efficacy

Results of the RCT suggest that MTH improved self-efficacy. These effects were consistent across several generic and disease-specific subscales, and increased over time. Only five of 21 papers included in a review of mobile-phone interventions in diabetes³⁷ were reported to measure self-efficacy. Where evaluated, it has often been positively influenced by MTH.^38–42 These results are in line with the suggestion that self-monitoring and feedback influence self-efficacy. Self-efficacy is considered to determine or reflect levels of motivation.^43,44 The increased motivation reported by MTH participants in interviews therefore supports the quantitative findings. This complementarity of the qualitative and quantitative data suggests that findings on MTH effects on self-efficacy are robust.

Intervention effects on illness beliefs

Few MTH studies have assessed illness beliefs. Results from this RCT suggest that MTH did not influence beliefs on the perceived seriousness of diabetes, or on treatment effectiveness to control diabetes and prevent complications. These quantitative findings were supported by the absence of themes relating to these beliefs in the interview data. Another mobile-phone intervention also failed to influence beliefs on the likelihood for long-term complications.⁴⁵ These beliefs have been found to be relatively stable over time⁴⁶ and this may limit any intervention effects. The intervention in this study did not include behavioural skills training, a strategy suggested to influence illness beliefs.⁴⁷ Finally, it is possible that the lack of intervention effects on personal models of diabetes is related to a ceiling effect. Means at baseline indicated that participants considered diabetes to be ‘fairly’ to ‘very’ serious and the prescribed treatment regimen to be ‘very’ to ‘extremely’ effective. In contrast, the qualitative data suggested that MTH increased participants’ awareness of their level of diabetes control and factors associated with it. MTH may positively influence beliefs relating to these aspects of self-management, rather than those assessed in this study. Future research may benefit from measuring these beliefs found in the qualitative study.

Another belief that MTH is likely to influence according to the qualitative data relates to perceived security. This belief was not measured as an outcome in the RCT. The interview data indicated that MTH made participants feel safer because clinical staff monitored their clinical readings and could help prevent acute complications. The concept of perceived security is related to perceived susceptibility (one’s perception of the risk or the chances of contracting a health disease or condition). Perceived susceptibility, a component of the health belief model,⁴⁸ is considered one of the health beliefs with the greatest influence on behaviour. The qualitative findings from our study suggest that MTH may result in a decrease in perceived susceptibility through increased perceptions of security.

Intervention effects on diabetes self-care

Results from the group by time interactions suggest that MTH did not influence diabetes self-care. Quantitative evidence on the impact of telehealth on diabetes self-care is mixed. Some studies have resulted in improvements in self-care;^39,49 others have not.^41,50–52 Several reasons for this conflicting evidence are possible. First, inconsistencies in findings could be due to the measurement of adherence, which is customarily self-report.⁵³ Self-report measures of adherence differ in content and response options.⁵⁴ They are prone to bias, and none of the studies mentioned above, including ours, used multiple adherence measures or cross-validated self-report measures with objective data.⁵³ Novel approaches such as mobile-phone and computerized logbooks for ecological momentary assessments are superior to self-reported adherence⁵⁵ and may be particularly feasible to implement in MTH.⁵⁴ Physical activity data in particular is becoming easier to capture with Smartphones and wearable sensors.⁵⁶ These measurement methods were not easily available or affordable when this study was designed.

Despite the lack of statistically significant group by time interaction effects on self-care, several participants reported improvements in their diet and BG self-monitoring frequency when interviewed. Such effects have been reported in qualitative studies.^57–60 They suggest that benefits from MTH may have been experienced by some participants. Importantly though, in line with these qualitative data, the effect sizes and CIs from the quantitative analyses did suggest a MTH effect on general diet at nine months. Together, these findings suggest that effect sizes and CIs should be considered alongside p-values when interpreting quantitative data.⁶¹ The variation in MTH effects on self-care across participants supports the need for further research to identify the combinations in which context, mechanisms, and outcomes interact to promote effectiveness,⁵¹ and patient characteristics associated with greater behaviour change.

The relationship between self-efficacy and self-care

Social cognitive theory postulates that there is a linear relationship between self-efficacy and behaviour. Although used to design mobile-phone interventions,⁶² few studies have assessed both these constructs. Those that have either found improvements in self-efficacy only,^38,41 or in both constructs.³⁹ These inconsistencies have led some researchers to question the value of self-efficacy as a health outcome.⁶³ The protocol for this study included a statistical analysis plan to examine mediating effects of self-efficacy on self-care.²⁸ Significant changes in self-care scores were required for such analyses; however, as these were not observed, they were not undertaken.

A possible reason why improvements in self-efficacy did not translate to self-care is duration of follow-up. The time period necessary for these translations may vary according to a range of factors, including MTH intervention content (e.g. intensity of health care professional support, behaviour change techniques involved), participants’ characteristics (e.g. baseline self-efficacy and self-care, disease severity), and technology usage (e.g. transmission frequency). The nine month follow-up in this study may have been insufficient for improvements in self-care to occur across a majority of intervention participants. Behaviour change may occur more rapidly in some participants, as the qualitative data on dietary and BG self-monitoring changes suggest. Also, other factors not accounted for in social cognitive theory may influence the likelihood of self-efficacy translating into behaviour change. The decrease in MTH participants’ perceived susceptibility mentioned above is one example and others include personality, subjective norms, and cultural factors.

Limitations

The main drawback of this study is the sample size for the RCT. The challenges experienced during recruitment⁶⁴ led to lack of power, underlining the need for RCT findings to be interpreted with caution and limiting the number of statistical analyses. A larger sample size may have yielded different results and greater precision of estimates. The parallels observed in the quantitative and qualitative findings do, however, suggest that the findings are reliable. They also emphasize the value of combining these two research methods to provide richer interpretations.

Recommendations for future research

Future MTH research would benefit from continuing to explore the mechanisms through which MTH results in clinical improvements, including a behavioural pathway of change. Identifying mediators of change will help build the science of MTH and design effective interventions. Greater efforts to reach adequate sample sizes for quantitative evaluations and to use a combination of objective and subjective measure of behaviour are also necessary.

Footnotes

Acknowledgements

The authors would like to thank participants who agreed to take part in this study and the clinical staff at the clinic who offered their help during the study.

All authors were equally involved in the design of the study. JB was responsible for study implementation and data collection under the supervision of SN and SH. JB is the primary author of this paper, and revisions were made by SN and SH. JB, SH, and SN approved the final version of the paper and take responsibility for the content of this paper.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We would like to thank the Policy Research Programme of the Department of Health for England for funding this study. The views expressed are not necessarily those of the Department.

References

Garcia-Perez

Alvarez

Dilla

. Adherence to therapies in patients with type 2 diabetes. Diabetes Ther 2013; 4(2): 175–194.

Hillson

. Adherence to diabetes care. Practical Diabetes 2014; 31(9): 357–358.

Larkin

Hoffman

Stevens

. Determinants of adherence to diabetes treatment. J Diabetes 2015; 7(6): 864–871.

Price

. How can we improve adherence? Diabetes Metab Res Rev 2016; 32(1): 201–205.

Peyrot

Rubin

Lauritzen

. Psychosocial problems and barriers to improved diabetes management: Results of the cross-national diabetes attitudes, wishes and needs (DAWN) study. Diabet Med 2005; 22(10): 1379–1385.

Health and Social Care Information Centre. National Diabetes Audit – 2012–2013: Report 1, care processes and treatment targets, www.hscic.gov.uk/catalogue/PUB14970 (2014, accessed 4 December 2015).

Neumark

Brudin

Neumark

. Adherence to national diabetes guidelines through monitoring quality indicators–A comparison of three types of care for the elderly with special emphasis on HbA1c. Prim Care Diabetes 2015; 9(4): 253–260.

Huang

Tao

Meng

. Management of endocrine disease. Effects of telecare intervention on glycemic control in type 2 diabetes: A systematic review and meta-analysis of randomized controlled trials. Eur J Endocrinol 2015; 172(3): R93–101.

Marcolino

Maia

Alkmim

. Telemedicine application in the care of diabetes patients: Systematic review and meta-analysis. PLoS One 2013; 8(11): e79246–e79246.

10.

Garabedian

Ross-Degnan

Wharam

. Mobile phone and smartphone technologies for diabetes care and self-management. Curr Diabetes Rep 2015; 15(12): 109–109.

11.

Harrison

Stadler

Ismail

. Are patients with diabetes mellitus satisfied with technologies used to assist with diabetes management and coping? A structured review. Diabetes Tech Therapeut 2014; 16(11): 771–783.

12.

Bashshur

Shannon

Smith

. The empirical evidence for the telemedicine intervention in diabetes management. Telemed J E Health 2015; 21(5): 321–354.

13.

Balkhi

Reid

Westen

. Telehealth interventions to reduce management complications in type 1 diabetes: A review. World J Diabetes 2015; 6(3): 371–379.

14.

Ashford

Edmunds

French

. What is the best way to change self-efficacy to promote lifestyle and recreational physical activity? A systematic review with meta-analysis. Br J Health Psychol 2010; 15(Pt 2): 265–288.

15.

Dombrowski

Sniehotta

Avenell

. Identifying active ingredients in complex behavioural interventions for obese adults with obesity-related co-morbidities or additional risk factors for co-morbidities: A systematic review. Health Psychol Rev 2012; 6(1): 7–32.

16.

Golley

Hendrie

Slater

. Interventions that involve parents to improve children's weight-related nutrition intake and activity patterns – What nutrition and activity targets and behaviour change techniques are associated with intervention effectiveness? Obes Rev 2011; 12(2): 114–130.

17.

Michie

Abraham

Whittington

. Effective techniques in healthy eating and physical activity interventions: A meta-regression. Health Psychol 2009; 28(6): 690–701.

18.

van Achterberg

Huisman-de Waal

Ketelaar

. How to promote healthy behaviours in patients? An overview of evidence for behaviour change techniques. Health Promot Int 2011; 26(2): 148–162.

19.

Abraham

Kok

Schaalma

Health promotion. In: Martin

Cheung

Kyrios

(eds). IAAP handbook of applied psychology, 1st ed. Oxford: Wiley-Blackwell, 2011, pp. 83–111.

20.

El-Gayar

Timsina

Nawar

. Mobile applications for diabetes self-management: Status and potential. J Diabetes Sci Technol 2013; 7(1): 247–262.

21.

Suter

Johnston

. Theory-based telehealth and patient empowerment. Popul Health Manag 2011; 14(2): 87–92.

22.

Craig

Dieppe

Macintyre

. Developing and evaluating complex interventions: The new Medical Research Council guidance. Br Med J 2008; 337: 979–983.

23.

Abraham C, Norman P, and Conner M (eds) Understanding and changing health behaviour: From health beliefs to self-regulation. Amsterdam: Harwood Academic Publishers, 2000.

24.

Ciere

Cartwright

Newman

. A systematic review of the mediating role of knowledge, self-efficacy and self-care behaviour in telehealth patients with heart failure. J Telemed Telecare 2012; 18(7): 384–391.

25.

Baron JS, Hirani S and Newman SP. A randomised, controlled trial of the effects of a mobile telehealth intervention on clinical and patient-reported outcomes in people with poorly controlled diabetes. J Telemed Telecare. Epub ahead of print 15 February 2016. Pii: 1357633X16631628.

26.

Bandura

. Self-efficacy: Toward a unifying theory of behavioral change. Psychol Rev 1977; 84(2): 191–215.

27.

Leventhal

Diefenbach

Leventhal

. Illness cognition: Using common sense to understand treatment adherence and affect cognition interactions. Cognit Ther Res 1992; 16(2): 143–163.

28.

Baron

Hirani

Newman

. A mobile telehealth intervention for adults with insulin-requiring diabetes: Early results of a mixed-methods randomized controlled trial. JMIR Res Protoc 2015; 4(1): e27–e27.

29.

Glasgow

Hampson

Strycker

. Personal-model beliefs and social-environmental barriers related to diabetes self-management. Diabetes Care 1997; 20(4): 556–561.

30.

Hampson

Glasgow

Strycker

. Beliefs versus feelings: A comparison of personal models and depression for predicting multiple outcomes. Br J Health Psychol 2010; 5(1): 27–40.

31.

Hurley

Harvey

. The insulin management diabetes self-efficacy scale. Measurement of nursing outcomes (2nd ed., volume 3 ed), New York: Springer Publishing Company, 2003, pp. 52–64.

32.

Nolte

Elsworth

Sinclair

. Tests of measurement invariance failed to support the application of the “then-test”. J Clin Epidemiol 2009; 62(11): 1173–1180.

33.

Toobert

Hampson

Glasgow

. The summary of diabetes self-care activities measure: Results from 7 studies and a revised scale. Diabetes Care 2000; 23(7): 943–950.

34.

Osborne

Elsworth

Whitfield

. The health education impact questionnaire (heiQ): An outcomes and evaluation measure for patient education and self-management interventions for people with chronic conditions. Patient Educ Counsel 2007; 66(2): 192–201.

35.

Skinner

Hampson

. Personal models of diabetes in relation to self-care, well-being, and glycemic control. A prospective study in adolescence. Diabetes Care 2001; 24(5): 828–833.

36.

Braun

Clarke

. Using thematic analysis in psychology. Qual Res Psychol 2006; 3(2): 77–101.

37.

Holtz

Lauckner

. Diabetes management via mobile phones: A systematic review. Telemed J E Health 2012; 18(3): 175–184.

38.

Faridi

Liberti

Shuval

. Evaluating the impact of mobile telephone technology on type 2 diabetic patients' self-management: The NICHE pilot study. J Eval Clin Pract 2008; 14(3): 465–469.

39.

Quinn

Clough

Minor

. WellDoc mobile diabetes management randomized controlled trial: Change in clinical and behavioral outcomes and patient and physician satisfaction. Diabetes Tech Therapeut 2008; 10(3): 160–168.

40.

Trief

Teresi

Izquierdo

. Psychosocial outcomes of telemedicine case management for elderly patients with diabetes: The randomized IDEATel trial. Diabetes Care 2007; 30(5): 1266–1268.

41.

Wakefield

Holman

Ray

. Outcomes of a home telehealth intervention for patients with diabetes and hypertension. Telemed J E Health 2012; 18(8): 575–579.

42.

Young

Miyamoto

Ward

. Sustained effects of a nurse coaching intervention via telehealth to improve health behavior change in diabetes. Telemed J E Health 2014; 20(9): 828–834.

43.

Bandura

. Toward a psychology of human agency. Perspect Psychol Sci 2006; 1(2): 164–180.

44.

Williams

Rhodes

. The confounded self-efficacy construct: Conceptual analysis and recommendations for future research. Health Psychol Rev 2016; 10(2): 113–128.

45.

Nundy

Mishra

Hogan

. How do mobile phone diabetes programs drive behavior change? Evidence from a mixed methods observational cohort study. Diabetes Educat 2014; 40(6): 806–819.

46.

Lawson

Bundy

Harvey

. The development of personal models of diabetes in the first 2 years after diagnosis: A prospective longitudinal study. Diabet Med 2008; 25(4): 482–490.

47.

McAndrew

Musumeci-Szabo

Mora

. Using the common sense model to design interventions for the prevention and management of chronic illness threats: From description to process. Br J Health Psychol 2008; 13(Pt 2): 195–204.

48.

Rosenstock

. Why people use health services. Milbank Q 1966; 44(3): 94–127.

49.

Greenwood

Blozis

Young

. Overcoming clinical inertia: A randomized clinical trial of a telehealth remote monitoring intervention using paired glucose testing in adults with type 2 diabetes. J Med Internet Res 2015; 17(7): e178–e178.

50.

Holmen

Torbjørnsen

Wahl

. A mobile health intervention for self-management and lifestyle change for persons with type 2 diabetes, part 2: One-year results from the Norwegian randomized controlled trial RENEWING HEALTH. JMIR Mhealth Uhealth 2014; 2(4): e57–e57.

51.

Kenealy

Parsons

Rouse

. Telecare for diabetes, CHF or COPD: Effect on quality of life, hospital use and costs. A randomised controlled trial and qualitative evaluation. PLoS One 2015; 10(3): e0116188–e0116188.

52.

Torbjørnsen

Jenum

Småstuen

. A low-intensity mobile health intervention with and without health counseling for persons with type 2 diabetes, part 1: Baseline and short-term results from a randomized controlled trial in the Norwegian part of RENEWING HEALTH. JMIR Mhealth Uhealth 2014; 2(4): e52–e52.

53.

DiMatteo

. Variations in patients' adherence to medical recommendations: A quantitative review of 50 years of research. Med Care 2004; 42(3): 200–209.

54.

Clifford

Perez-Nieves

Skalicky

. A systematic literature review of methodologies used to assess medication adherence in patients with diabetes. Curr Med Res Opin 2014; 30(6): 1071–1085.

55.

Marszalek

Morgulec-Adamowicz

Rutkowska

. Using ecological momentary assessment to evaluate current physical activity. Biomed Res Int 2014; 2014: 915172–915172.

56.

Savage

. Mobile data: Made to measure. Nature 2015; 527(7576): S12–S13.

57.

Carlisle

Warren

. A qualitative case study of telehealth for in-home monitoring to support the management of type 2 diabetes. J Telemed Telecare 2013; 19(7): 372–375.

58.

Lyles

Harris

. Qualitative evaluation of a mobile phone and web-based collaborative care intervention for patients with type 2 diabetes. Diabetes Tech Therapeut 2011; 13(5): 563–569.

59.

Pecina

Vickers

Finnie

. Telemonitoring increases patient awareness of health and prompts health-related action: Initial evaluation of the TELE-ERA study. Telemed J E Health 2011; 17(6): 461–466.

60.

Watson

Kvedar

Rahman

. Diabetes connected health: A pilot study of a patient- and provider-shared glucose monitoring web application. J Diabetes Sci Technol 2009; 3(2): 345–352.

61.

Sullivan

Feinn

. Using effect size—Or why the P value is not enough. J Grad Med Educ 2012; 4(3): 279–282.

62.

Riley

Rivera

Atienza

. Health behavior models in the age of mobile interventions: Are our theories up to the task? Transl Behav Med 2011; 1(1): 53–71.

63.

Griffiths

Foster

Ramsay

. How effective are expert patient (lay led) education programmes for chronic disease? Br Med J 2007; 334(7606): 1254–1256.

64.

Baron

Hirani

Newman

. Challenges in patient recruitment, implementation, and fidelity in a mobile telehealth study. Telemed J E Health 2016; 22(5): 400–409.