Type 2 Diabetes: Patient Outcomes

Summary

There is strong evidence that home telehealth can significantly improve glycemic control in people with type 2 diabetes.  Furthermore, this finding appears remarkably robust; over two-thirds of the studies that reported on hemoglobin A1c (HbA1c) found significant improvements in patients using home telehealth.  In most cases, these improvements are significantly greater than those seen in control groups using usual care.  It is also worth noting that higher use of home telehealth services is associated with greater improvements in HbA1c.  While few will be surprised by this finding, it testifies to the importance of incorporating sound patient engagement strategies into interventions.

There is strong evidence that home telehealth has the potential to significantly improve cholesterol and lipid levels.  However, with almost half of interventions finding no significant change, it is evident that achieving these improvements is highly dependent on context.  The same is true of reductions in body mass index (BMI).  Careful assessment of whether the findings of successful studies are applicable to one’s target population is advisable.  Improvements in blood pressure appear more readily achievable.  There is strong evidence of home telehealth’s ability to improve blood pressure significantly more than usual care in a range of patient populations.

There is strong evidence that patients using home telehealth experience significantly greater improvements in self-efficacy and quality of life than those under usual care.  The evidence consistently finds improvement in both measures, suggesting high levels of generalizability.

There is strong evidence that home telehealth can significantly improve diet and eating habits in type 2 diabetes patients.  While this is not universally the case, it does appear to be achievable in a majority of interventions that target these outcomes.  There is also strong evidence supporting the ability of home telehealth to increase levels of physical activity to a significantly greater degree than usual care.  This finding is strikingly consistent, again suggesting unusually high generalizability.

Other self-care activities appear to be more resistant to change.  There is moderate evidence that home telehealth significantly increases the frequency of blood glucose testing, but it does not inevitably do so.

In general, patients report high levels of satisfaction with home telehealth programs.  While this satisfaction often translates into willingness to continue using the program, one study found that this was not the case; patients were highly satisfied, but few wished to carry on with the intervention beyond the scheduled end date.  The authors of the study in question suggest that high-intensity interventions may be best used to address urgent short-term needs of specific patient sub-groups – pregnant patients, for example.

The level of patient engagement in home telehealth interventions varies widely.  High uptake is achievable in some cases, but technical problems and data entry requirements have been seen to lead to high levels of attrition.  Disease status and the prevalence of co-morbidities might be expected to have a strong influence on ability to use a system unassisted, particularly when manual dexterity or vision have been compromised.

The extent to which susceptibility to intervention effect differs by patient sub-group is an interesting question. There is conflicting evidence on the role of baseline HbA1c in determining the extent of intervention-related improvements.  We encourage further investigation of this. Several recent studies provided fascinating analyses of the interactions between non-clinical patient characteristics and intervention uptake. As these studies were retrieved in a final scan of 2011-2012 literature and were not subject to the same level of analysis as the other studies included in this review, we refer the reader to Arora et al. (2012), Glasgow et al. (2011), and Sarker et al. (2011) for more details.

Little information on patient time or cost savings was available.  At present, there is insufficient evidence of benefit in these areas.

 

Details

Uptake and Use of Technology

Uptake and use of technology outcomes fell into 2 general areas: satisfaction with technology and perceived ease of technology use. Readers should bear in mind the limitations of research in this area. Due to the voluntary nature of study recruitment, participants may have been self-selected for comfort with technology or general enthusiasm for innovative methods of disease management. The Hawthorne effect should also be considered: that is, the possibility that satisfaction with care or adherence to treatment improved not due to the design or content of the intervention, but simply because of its novelty.

 

Satisfaction

Summary: Patients’ satisfaction with telehealth was reported in 11 studies (Boaz et al., 2009; Buckley et al., 2008; Bujnowska-Fedak et al., 2006; Cho et al., 2009; Kim et al., 2005, 2006; Luzio et al., 2007; Robertson et al., 2007; Shea et al., 2007; Tjam et al., 2006; Watson et al., 2009). In some cases, this measurement took the form of an assessment of patient satisfaction with the intervention as a whole; in others, researchers assessed satisfaction with specific aspects of the technology. Measurement instruments were not always reported, and, when reported, were always validated. Some results appear to be based on anecdotal evidence.

In general, patients report high levels of satisfaction with home telehealth programs. While this satisfaction often translates into willingness to continue using the program, 1 study found that this was not the case; patients were highly satisfied, but few wished to carry on with the intervention beyond the scheduled end date (Luzio et al., 2007). The authors of the study in question suggest that high-intensity interventions may be best used to address urgent short-term needs of specific patient sub-groups – pregnant patients, for example.

Study Details: Longitudinal measurements of satisfaction were carried out in 4 studies. Significant increases in satisfaction with care and/or technology were reported in a single group study by Kim et al. (2005), with mean scores on a 100-point visual analogue scale increasing from 68.6±19.3 to 79±20.0. Tjam et al. (2006) also found significant improvement in this measure, with the intervention group’s average score on a 5-point scale increasing from 3.19 at baseline to 3.57 at 3 months (p=.0150) and 3.68 at 6 months (p=.0138). Improvements in the control group were also present, but failed to reach significance. In contrast, Cho et al. (2009) reported no difference between intervention and control groups in degree of satisfaction with medical service. Buckley et al. (2008) found that perceptions of telehealth, as measured by the Telemedicine Perception Questionnaire, were significantly improved after the intervention (T=3.24, p<.05).

Post-intervention measurements of satisfaction with care were conducted in 6 studies. Willingness or desire to continue using the intervention program was a common indicator. In Robertson et al. (2007), 100% of intervention completers gave positive responses when asked whether they wanted to program to continue. A majority of patients surveyed in Shea et al. (2007) also expressed generally high levels of satisfaction, while the 7 participants in Watson et al. (2009) both found the program satisfactory, and stated that they would continue to use it and recommend its use. In Bujnowska-Fedak et al. (2006), 75% of intervention patients wanted to continue receiving telehealth support, while over 90% of those in Boaz et al. (2009) stated that the system was ‘valuable’. In Luzio et al. (2007), on the other hand, despite reportedly high levels of patient satisfaction, only a few intervention participants were willing to carry on with the intervention beyond the study end date. Few additional details on methods or results were provided.

Kim et al. (2006) reported on patient satisfaction with specific aspects of the intervention: “Common features [the patients] liked included promptness of feedback and information, ability to follow up on their trends of physical activity, availability of updated results, and the specific and selective nature of individualized information about physical activity. Among the Web features, participants particularly favored graphic assessment tools for physical activity and up-to-date activity sheets” (p. 342-3).  Patients in the mobile phone/SMS arm of Cho et al. (2009), on the other hand, were dissatisfied with the brevity of the disease management recommendations sent through text messaging.

 

Ease of Use

Summary: Ease of technology use is a common concern for those considering adopting telehealth, particularly when working with populations that are elderly and physically and/or cognitively impaired. Given this, it is perhaps surprising that few studies explicitly address this issue. The 5 exceptions include Boaz et al. (2009), Istepanian et al. (2009a), Bujnowska-Dedak et al. (2006), Forjuoh et al. (2007, 2008), and Trudel et al. (2007). Though some studies found few problems, other findings indicate that patients’ functional status and comfort with technology should be a key consideration when selecting home telehealth technology.

Study Details: In Boaz et al. (2009), over 95% of patients agreed that the technology was easy to use. For some, it also improved confidence in disease management. Trudel et al. (2007), though acknowledging that some features of their system may be challenging for users with low dexterity or visual impairments, stated that few problems were reported.

Istepanian (2009a), did not report quantitative measures, but wrote that patient satisfaction was negatively affected by connectivity problems. Bujnowksa-Dedak et al. (2006), though reporting that only 9% of the intervention group experienced technical problems, noted that two-thirds of patient required family support to use the telehealth system. A significant correlation was found between patients’ HbA1c levels and their ability to use the system on their own. Patients in Forjuoh et al. (2007, 2008) also had difficulty meeting the demands of the system. High drop-out rates (58%) were attributed to the high burden of daily data entry, despite the fact that 75% of the intervention group rated their comfort level as 4 or greater on a 7-point scale.

 

Recent Developments: A scan of material from 2011-2012, a time period not covered by our initial searches, found 3 additional articles that addressed uptake and use of technology outcomes (Arora et al., 2012; Glasgow et al., 2011; Sarker et al., 2011). All 3 studies report on interactions between demographic factors and uptake. Arora et al. (2012) found that a mobile intervention for diabetes management could achieve high uptake and satisfaction in a low-income inner-city population. Sarker et al. (2011), reporting on use of an online patient portal that was offered as part of regular care, found that Latinos and African-Americans were less likely to log on. An analysis by Glasgow et al. (2011), in contrast, found few significant associations between use of a diabetes website and patient characteristics. As these studies were not subject to the same level of analysis as the other studies included in this review, we will not attempt further analysis. See Arora et al. (2012), Glasgow et al. (2011), and Sarker et al. (2011) for more details.

And on the qualitative side . . .Hopp et al. (2007) noted that the home telehealth unit used in their study was not suitable for all patients. Both clinical and non-clinical reasons were at play. Some patients’ diabetes control was very poor and demanded more specialized care; others had impaired vision so the technology was not appropriate. Patients had irregular uptake of the technology.Trief et al. (2008) reported on 3 sets of interview data (baseline, 6 months, 12 months) gathered from 25 patients participating in the IDEAtel project. Overall, patients had a positive attitude throughout the 1-year study period and had a strong desire to get their diabetes care under control. Recurring themes in post-intervention interviews included appreciation of on-demand access to providers and the sense of accountability that arose from the daily monitoring requirement.

 

Self-management, Self-Efficacy, and Behaviour Change

Outcomes around self-management, self-efficacy, and behaviour change were reported in 21 studies.

 

Diet and Eating Patterns

Summary: Intervention effects on diet and eating patterns were examined in 6 studies: three Level 2 (Boaz et al., 2009; Glasgow et al., 2010; Izquierdo et al., 2010), one Level 3 (Robertson et al., 2007), and two Level 4 (Forjuoh et al., 2008; Kim et al., 2006). There is strong evidence associating home telehealth with significant improvements in diet and eating habits, although this was not universally the case.

Study Details: Significant improvements in intervention group patients were seen in three Level 2 studies and one Level 4 study (Boaz et al., 2009; Glasgow et al., 2010; Izquierdo et al., 2010; and Forjuoh et al., 2008, respectively).

In Glasgow et al. (2010), intervention group patients had significantly greater improvements (p<.01) in eating habits and fat intake. Ammerman’s ‘Starting the Conversation’ scale was used to measure eating habits (Ammerman, 2004, in Glasgow et al., 2010). Patients’ mean total scores increased from 2.12±0.31 to 2.19±0.28 in the intervention group vs. 2.18±0.30 to 2.32±0.28 in the control group (effect size 0.32). Estimated fat intake (% energy from fat) decreased from 35.21±4.70% to 34.95±4.93% in the intervention group vs. 34.85±5.12 to 33.51±5.20 in the control group.

Izquierdo et al. (2010), reporting 2-year results from the IDEATel program, found a significant intervention effect, favouring the treatment group, in a combined measure of diet and exercise knowledge (p=.002). Adjusted scores at baseline, 1 year, and 2 years – with lower scores indicating greater knowledge – were 21.64 [SE=0.65], 19.08 [SE=0.58], and 16.48 [SE=0.70] in the intervention group vs. 21.79 [SE=0.65], 20.46 [SE=0.58], and 19.12 [SE=0.70] in the control group. Improvements in diet and exercise scores were significantly associated with reduced BMI and waist circumference (p<.001).

Boaz et al. (2009) found that the odds of a patient requiring family assistance for diet did not differ significantly by group assignment. However, intervention group patients were significantly more likely (p=.008) to report that they felt in control of their weight management (53% vs. 11%).

Forjuoh et al. (2008) found a significant improvement in participants’ scores in the general diet subscales of the Summary of Diabetes Self-Care Activities measure. This improvement (+13.60%, 95% CI=0.19-27.10) occurred between months 3 and 6 of the intervention. No significant changes were found in the specific diet subscales.

No significant changes were reported Robertson et al. (2007) (Level 3 evidence), which found no significant change in treatment subjects in self-reported consumption of calories, grams of carbohydrate, grams of protein, or grams of fat. Nor were any significant changes found in a Level 4 study by Kim et al. (2006), although a non-significant decline in diet adherence was reported at the post-test of the 3-month intervention.

 

Activity Levels

Summary: Changes in activity levels and exercise habits were examined in five Level 2 studies (Boaz et al., 2009; Glasgow et al., 2010; Izquierdo et al., 2010; Kim & Kang, 2006; Weinstock et al., 2011a) and one Level 4 study (Kim et al., 2006). There is strong evidence that home telehealth can lead to significantly greater improvements in activity levels than usual care. The striking consistency with which this result is reported suggests unusually high generalizability.

Study Details: Izquierdo et al. (2010), reporting 2-year results from the IDEATel program, found a significant intervention effect favouring the treatment group on a combined measure of diet and exercise knowledge (p=.002). Adjusted scores at baseline, 1 year, and 2 years – with lower scores indicating greater knowledge – were 21.64 [SE=0.65], 19.08 [SE=0.58], and 16.48 [SE=0.70] in the intervention group vs. 21.79 [SE=0.65], 20.46 [SE=0.58], and 19.12 [SE=0.70] in the control group. Improvements in diet and exercise scores were significantly associated with reduced BMI and waist circumference (p<.001).

In a multi-year study, again of the IDEATel program, mean number of days per week in which physical activity was undertaken was significantly higher in intervention patients at the endpoint (estimated means 2.47 [SE=0.09] vs. 2.09 [SE=0.09] in control group; p=.003) (Weinstock et al., 2011a). In addition, rates of decline in physical activity were significantly lower in the intervention group (mean decline of 0.49 days in intervention group vs. 0.83 days in control group; p=.0128). Physical activity levels within the intervention group were also significantly associated with use of a pedometer (p=.006).

Glasgow et al. (2010) found significantly better outcomes (p=.042) in change in physical activity over time in the intervention group, with the average number of calories expended in exercise changing from 3981±3019 to 3923±3431 in the intervention group vs. 3979±3292 to 3241±3221 in the control group. Boaz et al. (2009) reports that intervention group patients were significantly more likely than control group patients to report that they never required family assistance in managing exercise (77% vs. 22%; p=.001). In Kim & Kang (2006), the intervention group had access to a web-based disease management intervention. Another group of participants – Control Group 2 – received the same intervention in a paper-based form, while the remainder of the sample – Control Group 1 – were offered usual care. Participants in the intervention group and in Control Group 2 had significantly greater improvements in physical activity levels than those in Control Group 1 (p<.01 and p<.001, respectively). There were no significant differences between the groups receiving the paper-based and the web-based interventions.

In an uncontrolled study by Kim et al. (2006), adherence to exercise recommendations improved significantly over the course of the 12-week intervention (p=.036). According to self-reports, the number of days per week in which participants engaged in the recommended thirty minutes of physical exercise rose from 2.5±2.2 to 3.4±2.2.

 

Disease Knowledge

Summary: Changes in diabetes knowledge were measured in 3 studies: one Level 3 (Song et al., 2009) and two Level 4 (Buckley et al., 2008; Sun et al., 2010). Evidence that home telehealth interventions can improve disease knowledge is currently insufficient. While 2 of 3 studies did find significant improvement, their design limits the strength of the conclusions that can be drawn.

Study Details: In a 1-arm study by Buckley et al. (2008), significant post-intervention improvements in disease knowledge (p=.10) were found in a sub-sample (n=10) of patients with diabetes (t=3.25). Sun et al. (2010) found significantly increased diabetes knowledge in intervention group patients with a lesser increase in the control group (p<.05; no additional details provided). Song et al. (2009) reported increased knowledge in both intervention and control group patients, but no significant differences over time between groups.

 

Self-Efficacy

Summary: Indicators of self-efficacy were tracked in 3 studies: two Level 2 (Boaz et al., 2009; Bond et al. 2010) and one Level 4 (Buckley et al., 2008). There is strong evidence that home telehealth interventions can improve self-efficacy significantly more than usual care.

Study Details: In Bond et al. (2010), scores on the Diabetes Empowerment Scale improved significantly more in intervention patients (p<.05). Unadjusted means at baseline and 6 months were 2.2 [0.42] to 2.0 [0.35] in the intervention group vs. 2.1 [0.48] to 2.2 [0.45] in the control group.

In Boaz et al. (2009), intervention patients were significantly more likely than control group patients to report that they had no difficulty with disease management (53% vs. 17%; p=.01). They were also significantly more likely to report that they felt in control of managing their glucose levels (53% vs. 6%; p=.002) and diabetes (94% vs. 28%; p=.0001) and that they never needed family assistance with managing medication (88% vs. 50%; p=.01).

Buckley et al. (2008), using a single-arm study design, found significant improvements in self-efficacy (7.7 to 8.25 on Chronic Disease Self-Efficacy Inventory; t=2.29; p=.024).

And on the qualitative side . . .Nurses in Hopp et al. (2007) observed that some people liked entering the data but did not connect it to lifestyle decisions. Interestingly, entering all the data on a regular basis did not necessarily result in better overall self-management. However, the system was thought to facilitate appropriate adjustments in insulin dose.Some patients in Hopp et al. (2007) were motivated by the knowledge that they were being monitored. Trief et al. (2008) reported similar feedback from patients participating in the IDEAtel project. This comfort from being watched through technology was a common theme across the qualitative articles for all chronic diseases reviewed.Hopp et al. (2007) found that patients who entered the study with an interest in working on their diabetes management benefited the most from the intervention, and recommended that these patients be the focus of home telehealth programs for diabetes management.

 

Blood Glucose Testing and Self-Care Activities

Summary: Changes in diabetes self-care activities were reported in 8 studies. Two of these were Level 2 evidence (Cho et al., 2009; Rodriguez-Idigoras et al., 2009), one was Level 3 (Song et al., 2009), and five were Level 4 (Chang et al., 2007; Forjuoh et al., 2008; Kim et al., 2006; Levine et al., 2009; Sun et al., 2010). The most common indicator used was frequency of blood glucose testing. There is moderate evidence that home telehealth can significantly increase the frequency of blood glucose testing, although one Level 2 study found no intervention effect. Evidence of home telehealth-related improvements in other diabetes self-care behaviours is weak or insufficient. This is partly attributable to a gap in the literature; although there are several well-executed case series studies, the strength of the evidence suffers from a lack of controlled studies.

Study Details: Cho et al. (2009) found no significant difference between groups in frequency of blood glucose monitoring. Rodriguez-Idigoras et al. (2009), in contrast, found a significant difference between the intervention and control groups in number of blood glucose tests per month (7.37 vs. 5.85, respectively; p=.02). A 1-arm study by Levine et al. (2009) suggested that the frequency of blood glucose monitoring significantly increased with the number of messages from patient to provider (r=0.237, p<.05) and from provider to patient (r=0.350, p<.01). When messages were further broken down by content, the proportion of ‘person-centred’ messages was also found to be significantly associated with more frequent blood glucose testing (r=0.0339, p<.01).

 

Other

Summary: Few other measures were used in enough studies to allow for meaningful synthesis.

Study Details: Chang et al. (2007) found no significant difference between intervention and control patients in time to meet a combined endpoint of glycemic goal or consistent non-compliance. Cho et al. (2009) found no significant difference between groups in self-reported adherence to physician recommendations; however, in Kim et al. (2006), self-reported adherence to medication taking improved significantly. The mean number of days per week in which medication was taken as prescribed increased from 4.8±2.6 to 5.9±1.9 (p=.032). Forjuoh et al. (2008) reported significant improvements in the foot care subscales of the Summary of Diabetes Self-Care Activities measure (+25.60%, 95% CI=10.60-40.50).

Song et al. (2009) noted significant increases in intervention patient’s scores for diabetes care behaviour from baseline to 3 months (46.4 to 57.7, p=.004) and from 3 months to 6 months (to 62.4, p<.001). The control group experienced a significant increase from baseline to 3 months only (45.1 to 53.6, p=.013; 53.3 at 6 months). Sun et al. (2010) also reports significantly increased self-management scores in the intervention group.

 

Clinical Outcomes, Symptoms, and Health Status

Clinical outcomes, symptoms, and health status outcomes were reported in virtually all studies.

 

Hemoglobin A1c and Other Glycemic Outcomes

Summary: Hemoglobin A1C was perhaps the most studied outcome in the entirety of this review. It was primary or secondary outcome in 32 studies, half of which qualified as Level 2 evidence. There is strong evidence associating home telehealth with significant improvements in HbA1c. This finding appears remarkably robust. Of the 30 studies that provided p-values, significant improvements were seen in 23.[1] Moreover, of the 17 Level 2 studies that used a control group, 12 found that improvements were significantly greater in intervention patients than in controls.[2]

A number of interesting observations emerged from the body of evidence on the effect of home telehealth interventions on HbA1c. A selection follows:

 

1. Higher uptake was associated with greater improvements in HbA1c (Forjuoh et al., 2007, 2008; Istepanian et al., 2009a,b; McMahon et al., 2005). While few will be surprised by this, these findings are a testament to the importance of incorporating patient engagement strategies into interventions.

 

2. Several studies noted that mean HbA1c initially improved regardless of group assignment, but that this trend persisted only in the intervention group (Cho et al., 2006; Kim & Jeong, 2007; Kim & Kim, 2008). This may be attributable to the Hawthorne effect: that is, improvements that arise solely from the knowledge that one is under observation. Regression to the mean is another possibility.

 

3. There were conflicting reports on a possible interaction effect between baseline HbA1c and intervention effect. A strongly executed randomized controlled trial by Stone et al. (2010) found no significant interaction. However, this is contradicted by some intriguing findings in a 30-month study by Cho et al. (2006).

Authors of Cho et al. (2006) observed the effect described in (2), but noted that patterns of change differed significantly when the sample was stratified by HbA1c. Among patients with a baseline HbA1c <7.0%, those in the control group experienced a slight decrease followed by an increase. HbA1c levels in intervention patients were stable. When baseline HbA1c was greater or equal to 7.0%, the control group saw a slow increase while the intervention group showed a steeper decrease.

There are a number of possible explanations for this phenomenon. First, the intervention might have a ceiling effect – successful to a certain point, but less effective thereafter. This may be due in part to a sense of diminishing returns among study participants, or it may simply be that certain factors in patient health are beyond the reach of the intervention. A second possibility is that regression to the mean is acting as a confounder. If this were in fact the case, we might expect to see an inverse correlation between intervention effect and baseline HbA1c – exactly the kind of effect that Stone et al. (2010) did not find.

The kind of formal meta-analysis that would shed light on this question is beyond the scope of this narrative synthesis. However, we encourage further investigation of this area.

 

Other tests of glycemic control included mean blood glucose, fasting blood glucose, 2-hour post-prandial blood glucose, and incidence of hypoglycemia.

Study Details: For a complete summary of findings on glycemic control, see Table C.8.3.1: Patient Outcomes – Measures of Glycemic Control.

 

Non-Glycemic Outcomes

Common non-glycemic outcomes included lipids, blood pressure, and body mass index (BMI).

Summary: Lipids were measured in 19 studies. Approximately half were Level 2 evidence[3], with the remainder divided between Levels 3, 4, and 5.[4] There is strong evidence that home telehealth can significantly improve cholesterol and lipid levels[5]. There is also strong evidence that this improvement can be significantly greater than that seen with usual care.[6] However, with almost half of the Level 2 studies[7] and multiple lower-level studies[8] finding no significant change, it is evident that achieving this outcome is highly dependent on context.

Blood Pressure was measured in 13 studies. Roughly half were Level 2 evidence[9], with the remainder divided between Levels 3, 4, and 5.[10] There is strong evidence that home telehealth can lead to significantly greater improvements in blood pressure. This improvement was detected in 6 of 8 Level 2 studies, and therefore appears to be relatively generalizable. In 4 cases, improvement was also significantly greater than that seen with usual care. However, as with BMI, uncontrolled studies were less likely to find significant improvements in this measure.

BMI was reported in 10 studies: six Level 2 evidence (Bond et al., 2007; Glasgow et al., 2010; Izquierdo et al., 2010; Rodriguez-Idrigoras et al., 2009; Stone et al., 2010; Yoo et al., 2009), three Level 4 evidence (Chumbler et al., 2005a; Forjuoh et al., 2007, 2008; Sun et al., 2010), and one Level  5 evidence ( Luzio et al., 2007). There is strong evidence associating home telehealth interventions with significant reductions in BMI. However, in only 2 of the 6 controlled studies was this improvement significantly greater than that seen with usual care. Furthermore, although controlled studies found significant reductions in BMI, 1 reported a significant increase (Stone et al., 2010) and a number of Level 4 and 5 studies found no significant change (Chumbler et al., 2005a; Forjuoh et al., 2007, 2008; Luzio et al., 2007). Home telehealth appears to have the potential to be at least as effective as usual care in catalyzing significant reductions in BMI, but does not invariably do so. Careful assessment of whether these findings are likely to be applicable to one’s target population is advisable. For additional details, see Table C.8.3.1: Patient Outcomes – Non-Glycemic Measures (below).

Among the rarer measures were renal function (Luzio et al., 2007; Cho et al., 2006), waist circumference (Cho et al., 2006), weight (Bond et al., 2007), mortality (Jia et al., 2009; Shea et al., 2009), and electrolytes (Cho et al., 2006). As cross-study comparisons were not possible in these cases, they have been omitted from the tables below. Interested parties are advised to refer to the citations provided.

Study Details: See Table C.8.3.2: Patient Outcomes – Non-Glycemic Measures (above) for additional details on lipids, blood pressure, BMI, renal function, waist circumference, weight, mortality, and electrolytes.

 

Recent Developments: A scan of material from 2011-2012, a time period not covered by our initial searches, found 2 additional articles that addressed clinical outcomes, symptoms, and health status outcomes (Logan et al., 2011; Weinstock et al., 2011b). Logan et al. (2011) found that an association between home monitoring and significant reductions in blood pressure in patients with hypertension and diabetes. However, this association disappeared when home monitoring system did not include a feedback element. Weinstock et al. (2011b), in another analysis of the IDEATel study, reports that improvements in glycemic control were especially marked in Hispanic patients. The authors further note that ‘[Hispanic patients] had the highest baseline [Hb]A1c levels, suggesting that telemedicine has the potential to help reduce disparities in diabetes management’. As these study studies were not subject to the same level of analysis as the other studies included in this review, we will not attempt further analysis. See Logan et al. (2011) and Weinstock et al. (2011b) for more details.

 

Quality of Life

Summary: Quality of life was examined in 5 studies: two Level 2 evidence (Boaz et al., 2009; Bond et al., 2010) and three Level 4 evidence (Bujnowksa-Fedak et al., 2006; Chumbler et al., 2005a; Forjuoh et al., 2008). There is strong evidence associating home telehealth interventions with significantly greater improvements in quality of life than that seen with usual care, although neither Level 4 study found an intervention effect.

Study Details: In a 6-month study by Boaz et al. (2009), patients in the intervention group were significantly more likely not to experience anxiety (53% vs. 17%, p=.02), depression (59% vs. 11%, p=.003), disease-associated difficulty (53% vs. 17%, p=.02), disease-associated life complications (65% vs. 11%, p=.003), feelings of impotence (88% vs. 11%, p<.001), and feelings of ineptitude (88% vs 11%, p=.001).

Bond et al. (2010) used the Centre for Epidemiological Studies Depression Scale (CES-D) and the Problem Areas in Diabetes Scale (PAID) to assess a range of outcomes within the spectrum of emotional and psychological well-being. Analysis found significant intervention effects for depression (unadjusted means 12.0 [10.4] to 9.8 [7.9] in intervention group vs. 22.1 [8.7] to 12.1 [8.5] in control group; effect size 0.7; p<.05) and quality of life (unadjusted means 2.3 [0.88] to 2.0 [0.67] in intervention group vs. 2.1 [0.84] to 2.2 [0.91] in control group; effect size 0.6; p<.05). Treatment regime had a moderating effect on changes in quality of life and depression. Patients with a no-medication regime had much lower improvements in quality of life and depression outcomes than patients taking oral medication and/or insulin.

Chumbler et al. (2005a) measured health-related quality of life at the 1 year endpoint of an intervention. Significant improvements were found in 3 of 8 sub-scales: Role-Physical (unadjusted means: 42.1 [1.7] to 47.1 [1.5]; p=.0064; adjusted: 40.6 [2.1] to 46.1 [2.2]; p=.0165), Bodily Pain (unadjusted means: 58.0 [1.6] to 66.5 [1.6]; p<.0001; adjusted: 53.1 [2.1] to 60.8 [2.1]; p=.0005), and Social Functioning (unadjusted means: 56.6 [1.7] to 61.0 [1.5]; p=.0130; adjusted: 56.5 [2.1] to 61.0 [2.2]; p=.0498). No significant changes were found for the subscales General Health, Vitality, Role-Emotional, or Mental Health.

The two Level 4 studies found no significant changes. However, Forjuoh et al. (2008) collected quality of life data at 3 months and 6 months only, rendering comparison to baseline impossible. Patients in both studies appear to have had low health status at baseline. In Forjuoh et al. (2008), mean baseline HbA1c was 9.7, and mean BMI 35.40. The other did not provide clinical characteristics for the sample; however, a number of co-morbidities were mentioned, and 2/3 patients were unable to use the system without family assistance (Bujnowkska-Fedak et al., 2006). While conclusions would be premature, it is possible that poor baseline health and/or high levels of co-morbidities inhibit improvements in quality of life.

 

Cost and Time Savings

Summary: Cost and time savings were examined in 2 studies. The effects of home telehealth appear to vary with the intervention. More research is needed before a meaningful analysis of the factors involved in cost and time savings will be possible.

Study Details: In Stone et al. (2010), there were no time savings in the time spent on nurse-participant telephone contact; each intervention group participant spent an average of roughly 1.3 hours per month, whereas the control group averaged 0.3 hours per participant per month. Cho et al. (2006), on the other hand, reports that physician-patient contact time was ‘relatively reduced’ with the telehealth intervention when compared with in-person appointments.

 


[1] Level 2: Bond et al., 2007; Cho et al., 2006, 2009; Kim & Jeong, 2007 ; Kim & Kang, 2006; Istepanian et al., 2009a (*in sub-group analysis of study completers); Lorig et al., 2010; McMahon et al., 2005; Ralston et al., 2009; Rodriguez-Idrigoras et al., 2009; Shea et al., 2009; Stone et al., 2010; Yoo et al., 2009. Level 3: Robertson et al., 2007; Song et al., 2009; Tildesley et al., 2010. Level 4. Chang et al., 2007; Chumbler et al., 2005a; Forjuoh et al., 2007, 2008; Kim et al., 2006; Kim & Kim, 2008; Watson et al., 2009. Level 5: Luzio et al., 2007

[2] Bond et al., 2007; Cho et al., 2006; Kim & Jeong, 2007 ; Kim & Kang, 2006; Istepanian et al., 2009a (*in sub-group analysis of study completers); Lorig et al., 2010; McMahon et al., 2005; Ralston et al., 2009; Rodriguez-Idrigoras et al., 2009 (*at 6 month measurements only); Shea et al., 2009; Stone et al., 2010; Yoo et al., 2009

[3] Cho et al., 2006, 2009; Istepanian et al., 2009a; Kim & Song, 2007; McMahon et al., 2005; Ralston et al., 2009; Rodriguez-Idrigoras et al., 2009; Shea et al., 2006, 2009; Glasgow et al., 2010; Bond et al., 2007; Stone et al., 2010; Yoo et al., 2009.

[4] Level 3: Lee et al., 2007; Tildesley et al., 2010; Grant et al., 2008; Level 4: Chumbler et al., 2005a; Kim & Kim, 2008; Kim et al., 2005; Level 5: Luzio et al., 2007.

[5] Seven Level 2 studies (Bond et al., 2007; Cho, 2006 (triglycerides only); Rodriguez-Idrigoras et al., 2009; Shea et al., 2006, 2009; Stone et al., 2010; Yoo et al., 2009), 2 Level 3 studies (Lee et al., 2007; Tildesley et al., 2010), and one Level 5 study (Luzio et al., 2007).

[6] Four Level 2 studies (Bond et al., 2007; Cho, 2006 (triglycerides only); Shea et al., 2006, 2009)

[7] Cho et al., 2009; Istepanian et al., 2009a; Glasgow et al., 2010; Kim & Song, 2007; Ralston et al., 2009

[8] Chumbler et al., 2005a; Kim & Kim, 2008; Kim et al., 2005; Luzio et al., 2007

[9] Glasgow et al., 2010; Rodriguez-Idrigoras et al., 2009; Shea et al., 2006; Yoo et al., 2009; McMahon et al., 2005; Bond et al., 2007; Istepanian et al., 2009a; Ralston et al., 2009

[10] Level 3: Grant et al., 2008; Level 4: Chumbler et al., 2005a; Forjuoh et al., 2007, 2008; Trudel et al., 2007; Level 5: Luzio et al., 2007

 

 

Leave a Reply