Skip to main content
  • Research article
  • Open access
  • Published:

A web-based intervention to support self-management of patients with type 2 diabetes mellitus: effect on self-efficacy, self-care and diabetes distress



Management of diabetes mellitus is complex and involves controlling multiple risk factors that may lead to complications. Given that patients provide most of their own diabetes care, patient self-management training is an important strategy for improving quality of care. Web-based interventions have the potential to bridge gaps in diabetes self-care and self-management. The objective of this study was to determine the effect of a web-based patient self-management intervention on psychological (self-efficacy, quality of life, self-care) and clinical (blood pressure, cholesterol, glycemic control, weight) outcomes.


For this cohort study we used repeated-measures modelling and qualitative individual interviews. We invited patients with type 2 diabetes to use a self-management website and asked them to complete questionnaires assessing self-efficacy (primary outcome) every three weeks for nine months before and nine months after they received access to the website. We collected clinical outcomes at three-month intervals over the same period. We conducted in-depth interviews at study conclusion to explore acceptability, strengths and weaknesses, and mediators of use of the website. We analyzed the data using a qualitative descriptive approach and inductive thematic analysis.


Eighty-one participants (mean age 57.2 years, standard deviation 12) were included in the analysis. The self-efficacy score did not improve significantly more than expected after nine months (absolute change 0.12; 95% confidence interval −0.028, 0.263; p = 0.11), nor did clinical outcomes. Website usage was limited (average 0.7 logins/month). Analysis of the interviews (n = 21) revealed four themes: 1) mediators of website use; 2) patterns of website use, including role of the blog in driving site traffic; 3) feedback on website; and 4) potential mechanisms for website effect.


A self-management website for patients with type 2 diabetes did not improve self-efficacy. Website use was limited. Although its perceived reliability, availability of a blog and emailed reminders drew people to the website, participants’ struggles with type 2 diabetes, competing priorities in their lives, and website accessibility were barriers to its use. Future interventions should aim to integrate the intervention seamlessly into the daily routine of end users such that it is not seen as yet another chore.

Peer Review reports


Management of diabetes mellitus is complex, and involves controlling multiple risk factors that may lead to complications. However, care gaps exist: the Behavioral Risk Factor Surveillance System has estimated that only 68% of patients with type 1 or type 2 diabetes had HbA1c measured at least twice in the previous year [1], despite a recommendation from the American Diabetes Association that it be measured at least two to four times per year [2]. Given that patients provide most of their own diabetes care, patient self-management training is an important strategy for improving quality of care [3], particularly in the current era of patient-centred outcomes and comparative clinical effectiveness research [4]. Patient self-management interventions have demonstrated benefits in terms of both quality of life [5] and glycemic control [6], but participation is low [7], effectiveness wanes over time [6], and access to trained professionals to support self-management is limited [8]. Web-based self-management interventions are promising because they offer ease of access for patients who are computer-literate, and they can be scaled up with little cost [9]. Web-based media have improved patient knowledge, the extent of behaviour change, and clinical outcomes for a range of conditions [10]. However, principles of effective education, self-management support, and behaviour change have not been incorporated into current diabetes-related websites [11]-[13]. Reviews of existing diabetes websites showed that they presented didactic information of variable quality, they required advanced reading levels, and they followed a static, newspaper-format display, rather than harnessing the inherent advantages of websites, such as interactive technology, social support, and problem-solving assistance [11],[13]. A systematic review of electronic diabetes-related tools found that they had moderate but inconsistent effects on a variety of psychological and clinical outcomes, including HbA1c and weight; tools that were more interactive tools were associated with continued website use and greater clinical improvement [10]. In addition, greater website use was correlated with greater clinical improvements: regular website users had greater reductions in HbA1c compared with intermittent users. Although this finding could be a consequence of the healthy user effect [14], addressing usability issues to increase the proportion of regular users may increase the effectiveness of interventions.

In a previous study, we developed an approach to address many of these limitations of existing web-based interventions [15]. In the current study, we tested the impact of this approach on self-efficacy, quality of life, self-care, blood pressure, cholesterol, glycemic control, and exercise promotion amongst people with type 2 diabetes.


Study overview

This study consisted of five phases: 1) development of the intervention, 2) feasibility testing, 3) usability testing; 4) refinement of the intervention, and 5) evaluation of the intervention using a cohort study and individual interviews. The study protocol and results of the first four phases are reported elsewhere [15],[16]. We report here the results of Phase 5.

Diabetes online companion: a web-based self-management intervention

The Diabetes Online Companion is a self-contained diabetes self-management website that was systematically developed according to self-efficacy theory. Self-efficacy refers to “beliefs in one’s capabilities to organize and execute the courses of action required to produce given attainments” [17]. Randomized controlled trials have shown that diabetes self-management education programs incorporating principles of self-efficacy are associated with improvements in knowledge [18], health behaviours [18],[19], self-efficacy [18]-[20], HbA1c [18]-[21], weight [18], and microvascular complications [19].

Our intervention incorporated evidence-based content and behaviour-change strategies and followed the principles of user-centred design [15]. The website had four main components: 1) general information (static), 2) tailored information (interactive), 3) self-monitoring logs (interactive), and 4) a blog (interactive) (see Additional file 1 for sample screenshots). We posted a total of 53 blog posts over the intervention period, initially at a frequency of one per week. After four weeks of limited user activity, we increased the frequency of blog posts to two per week and added email prompts with each new posting. The topics, which covered medical content, diabetes-related news items, and practical issues, were selected on the basis of our feasibility and usability testing [15]. In addition, participants received weekly email reminders to visit the site or complete their self-management trackers, as well as notices of any new content [15].

Cohort study


We conducted a single-arm pre-post cohort study. Consecutive series of individuals with diabetes were recruited from two family practice units and two endocrinology clinics in Toronto (one each from two academic health science centres). Those eligible for inclusion were aged ≥ 25 years with at least one of HbA1c > 7.0% (53 mmol/mol), systolic blood pressure > 130 mmHg, low-density-lipoprotein cholesterol (LDL-C) > 2.0 mmol/L, or body mass index (BMI) > 25 kg/m2. We excluded those who had Canadian Cardiovascular Society class 3 or 4 angina, did not speak English, were not available for follow-up, or had no regular access to the telephone and internet.


Website usage: We analyzed logs for the web server to assess the frequency and duration of specific components of the intervention [16]. Specifically, we collected data for the following variables: duration of use by individual users, frequency of use, site penetration, most frequently accessed tools and pages, and patterns of use over time.

Patient-centred outcomes: We assessed self-efficacy, our primary outcome, with the Modified Grossman Self-efficacy for Diabetes Scale, which has moderate to high reliability (Cronbach’s alpha = 0.51 to 0.86; Additional file 1) [22],[23]. We selected self-efficacy because not only has it been validated in predicting and promoting patient behaviour change, but it also has been demonstrated to improve clinical outcomes [18],[20],[24],[25]. We assessed self-care behaviour with the Summary of Diabetes Self-Care Activities Measure – Revised [26] and diabetes-specific quality of life with the Diabetes Distress Scale [27]. These patient-based outcomes were selected because they are relevant measures of knowledge use by patients.

Clinical outcomes: We collected data on HbA1c, systolic and diastolic blood pressure, LDL-C, and weight every three months. These outcomes were chosen to inform the sample size calculations in future trials.

Data collection

We obtained data for age, sex, ethnicity, education, self-reported health literacy, employment, duration of diabetes, complications, smoking status, medications, HbA1c, systolic blood pressure, LDL-C, weight, current use of and comfort with a computer and the internet, self-care score, self-efficacy score, and quality-of-life score at baseline. Outcome data were collected by means of patient-completed questionnaires. For the pre- and post-implementation phases, aggregates of patient-completed questionnaires were obtained every three weeks for nine months through web-based surveys, resulting in 12 data points for each phase. Health literacy was measured by a three-item validated questionnaire completed by the patients [28],[29]. HbA1c and LDL-C were collected from medical records via chart audit. Systolic and diastolic blood pressures were measured by the research coordinator and were recorded as the average of three readings. Weight was also measured by the research coordinator. At the end of the study, each participant was asked to disclose whether he or she had used other web-based interventions and if so, whether those interventions employed text- or image-based didactic materials, interactive technology, or behavioural strategies. To assess for threats to validity from historical effects, we recorded secular events that might have affected our outcomes (such as diabetes-related news reports).

Sample size calculation

Using a range of correlations from 0.2 to 0.8, a significance level of 0.05, and a power of 80%, we calculated that a sample of at most 52 participants was required to detect a change of 0.5 units in self-efficacy score after the intervention (relative to the score before implementation). Differences of 0.1 to 0.5 in self-efficacy score have been correlated with metabolic control, eating behaviour, exercise behaviour, and other self-management behaviours [22],[23]. A formula for paired mean comparisons was applied [30], and the longitudinal nature of the study increased its power [31]. A previous analysis reported a dropout rate of 20%–51% in studies of self-management [32]; we further adjusted the sample size to account for an expected dropout rate of 40%.

Data analysis

Linear mixed models were used to examine the effect of the intervention and time (intervention × time interaction) on self-efficacy, self-care, and diabetes distress. We selected these models to accommodate the complexities of typical longitudinal data sets for continuous outcomes; specifically, they allowed us to account properly for both within- and between-participant variability [33],[34] and have been used in previous studies for similar analyses [35]-[37]. The models were also adjusted for age, sex, ethnicity, income (above or below Can $30, 000), education, employment, and health literacy, as each of these variables could affect the study outcomes [21],[38],[39]. The model examining the self-care outcome was also adjusted for interaction terms of the aforementioned variables with time. No additional interaction terms with time were included for other outcome models, because all additional interaction terms examined were non-significant. To avoid inflation of R2, all variables were specified a priori, and all interactions were tested simultaneously using a cut-off value of 0.30 [40]. Models were assessed by means of residual plots.

To assess the potential effect of missing income data for three of the participants, a sensitivity analysis (imputing income as both high and low) was performed. Missing health literacy data for 15 of the participants were imputed using the mode of the distribution, because 95% of the remaining participants were health literate.

Linear mixed models were also used to examine the effect of the intervention and time (intervention × time interaction) on secondary outcomes. These models were adjusted for age, self-efficacy score, income, ethnicity, and insulin use (for HbA1c and weight only). We also compared the effect of the intervention between users and non-users of the website. Finally, we used descriptive statistics to analyze website usage. R software version 2.1.15 was used for all analyses [41].


Individual interviews were conducted 2 to 21 weeks after completion of quantitative data collection. We used a purposive sampling strategy to recruit participants with a range of experiences and characteristics [42] (sex, age, ethnicity, duration of diabetes, educational attainment, income) from the broader pool of cohort study participants. We developed a semi-structured interview guide to elicit participants’ views regarding the following website features: acceptability, usability, strengths and weaknesses of the intervention, facilitators and barriers to its use, user satisfaction, and sustainability of use (Additional file 1). We made the website available during each interview, in case the interviewee wanted to show the interviewer something on the website.

All interviews were audiotaped and transcribed verbatim [43]. Transcripts were inductively analyzed to identify emergent categories and themes using a constant comparative approach [44]. Coding was conducted independently by three team members with expertise in qualitative research methods (CHY, JAP, SH) [44]. After coding an initial subset of interviews, a preliminary coding framework was developed on the basis of the emerging analysis, with discussion and consensus amongst the analysts [45]; the framework was then iteratively tested and refined with subsequent interviews [44]. Thematic saturation was attained with 21 interviews [42]. NVivo software (version 9) was used to assist with data management and retrieval. Techniques to ensure analytic rigour included use of multiple analysts, negative case analysis, and triangulation of the qualitative findings with the quantitative results [42],[44],[46]. Triangulation consisted of 1) examining the interview data through the lens of “effect on self-efficacy”, 2) corroborating qualitative findings with quantitative data, and 3) interrogating how the Diabetes Online Companion affected self-efficacy [46].

Research ethics

The study was approved by the Research Ethics Boards of St. Michael’s Hospital (reference number 09–091) and Sunnybrook Health Sciences Centre (reference number 177–2009). All participants gave written and verbal informed consent.


Cohort study

Of the 98 participants recruited, 81 had complete data collection for at least two time points (one before and one after the intervention was implemented) and were included in the analysis. The questionnaire response rate for these 81 participants was 83%. Patients’ characteristics and baseline self-efficacy, self-care, and diabetes distress are reported in Table 1 (Demographic characteristics and baseline values of observational cohort and qualitative study).

Table 1 Demographic characteristics and baseline values of observational cohort and qualitative study

Website use

The mean number of days on which users logged in during the study period was 8.2 days (standard deviation 13); the median was three days. The average frequency of use was 0.7 logins/month, or one visit every 5.8 weeks, distributed as follows: non-user: 11 participants (14%); infrequent user (<2 times/month): 61 participants (75%); frequent user (>2 times/month): seven participants (9%); heavy user (>1 time/week): two participants (2%). Website usage across all users ranged from 4 to 50 logins/week (median 14.5/week), with peaks of 50 logins in week 10 and 37 logins in week 27. Increased use of the website during those weeks appeared to be driven by the blog. In general, website use appeared to parallel blog use, with users visiting the blog repeatedly during the same login or visit (Figure 1). The most-accessed pages during week 10 were the blog (regarding medication log, supplements, and insulin) (34% of hits) and the blood pressure (8%) and medication (9%) logs. For week 27, the most-accessed pages were the blog (regarding foot care) (44% of hits), “My blood glucose log” (32%), and “7 steps to take care of your feet” (3%).

Figure 1
figure 1

Website login and blog use by week. Black bar: Number of logins per week. Grey bar: Number of blog views per week.

Overall, the most frequently accessed tools, for both first-time and return users, were the blog, followed by “My blood glucose log”, “My medication log”, and “My activity log”. Regarding site penetration, users viewed 6.6 pages per session, spending an average of 5 minutes 43 seconds on the site, and 1 minute 39 seconds per page.

Blog use

Within the blog section of the website, there were a total of 569 page views by 35 participants over the study period, with peaks at week 10 (54 views), week 27 (43 views), and week 30 (53 views), corresponding to blog entries about the medication log, supplements and insulin, and foot and kidney care, respectively. A total of 13 comments responding to the blog postings were submitted by five participants. These comments took the following forms: 1) responding to the blog (agreement or disagreement); 2) requesting help with or providing feedback on the website; 3) requesting help with self-management; 4) offering assistance, empowerment, and their own solutions (including food recipes); 5) self-reporting behaviour change; 6) sharing responses to medication; and 7) warning others about interactions with health care providers.

Use of interactive and static tools

Overall, 47 (67%), 63 (90%), and 43 (60%) of 70 users visited static, interactive, and log pages, respectively, at least once. These users had a mean of 3.4, 4.5, and 9.3 visits/user to each of these page types, respectively.

Patient-centred outcomes

Self-efficacy: Despite a significant short-term increase in self-efficacy score immediately after implementation of the intervention (0.13; 95% confidence interval [CI]: 0.06, 0.20; p < 0.0004), by nine months, this outcome had not increased significantly more than expected from its pre-implementation trajectory (effect: 0.12; 95% CI: −−0.028, 0.263; p = 0.11; Figure 2 and Table 2).

Figure 2
figure 2

Self-efficacy, self-care, and diabetes distress nine months before and nine months after intervention implementation. Reference categories used in the plot were as follows: female, mean age 57.34 years, employed, university education, income > Can$30,000, adequate health literacy, white.

Table 2 Summary statistics of psychological and clinical outcomes (with 95% confidence interval), at implementation of intervention and 9 months later

Self-care: The self-care score improved by 0.44 (95% CI: 0.23, 0.63; p < 0.0001) beyond what was expected at nine months (Figure 2 and Table 2).

Diabetes distress: Despite an immediate short-term decrease in diabetes distress score (−−2.29; 95% CI: −3.76, −0.81; p = 0.002), by nine months, this outcome had not decreased significantly over what was expected (effect: 1.84; 95% CI: −4.81, 1.12; p = 0.22; Figure 2 and Table 2).

There was no interaction with insulin use by time or intervention for any of these outcomes.

Self-care scores were positively correlated with age (0.04/year, 95% CI: 0.02, 0.06), p <0.001).

Diabetes distress varied with age and sex: younger female participants had greater diabetes distress.

When we conducted the sensitivity analysis with missing values for income assumed to be below $15,000 and missing values for health literacy assumed to be the mode, there were no changes in results for self-efficacy, self care, or diabetes distress.

Clinical outcomes

Seventy-three of the participants were included in the analysis of clinical outcomes. The other eight participants were excluded because of missing data for HbA1c, blood pressure, LDL-C, or weight within 90 days of the self-efficacy data or because no data were obtained after implementation of the intervention. The intervention had no effect on HbA1c, blood pressure, LDL-C, or weight in either the unadjusted or the adjusted models (Table 2).

Comparison of users and non-users:

A total of 70 participants (86%) used the website at least once, whereas 11 (14%) did not use the site at all. At the nine-month follow-up after implementation of the intervention, there was no difference between users and non-users in terms of self-efficacy (0.15 vs. 0.13, p = 0.35) or self-care (0.18 vs. 0.13, p = 0.21). Users had a greater reduction in diabetes distress than did non-users (−4.7 vs. -0.9, p < 0.0001). There was no difference in the effect of using the intervention on any secondary outcomes, with the exception of diastolic blood pressure (users: +3.27 mm Hg; non-users: −1.6 mm Hg; p = 0.014; Additional file 1).


Twenty-one individuals (Table 1) participated in an interview. The sample consisted of White and Asian men and women of various ages, duration of diabetes, educational attainment, and employment status, who used computers frequently and were comfortable with using the internet. Analysis of the interviews revealed numerous themes, four of which were most relevant to interpretation of the cohort study’s negative results, in particular, exploration of why participants used the website to only a limited extent. Additional themes will be the focus of future publications. The following four themes are considered here: 1) barriers and facilitators of website use; 2) patterns of website use, including the role of the blog in driving site traffic; 3) general feedback on website characteristics; and 4) potential mechanisms for the effect of the website on self-efficacy, behaviour change, and diabetes distress. Representative quotes for each theme appear in Table 3.

Table 3 Themes identified and representative quotes from in-depth interviews

1) Barriers and facilitators to use: Participants stated that they struggled with competing health and life concerns. They reported that it was “not just diabetes” that they dealt with (Table 3; 1a) and that they had to manage other concurrent medical conditions (Table 3; 1b). They spoke about their attempts to balance illness work with everyday life work; they found that after completing the latter, “there wasn’t a lot” of time or energy left for self-management of their disease, much less to use the website (Table 3; 1b). Some participants identified lack of motivation as “a me thing, as opposed to a site thing”, while others commented that “laziness” (Table 3; 1c) was a barrier to use.

Participants’ attitudes toward diabetes also coloured their approach to self-management and thus their use of the site. In particular, participants reported feeling frustrated with the uncontrolled nature of their disease, and the collection of self-monitoring information that showed a lack of metabolic control exacerbated this frustration (Table 3; 1d). Similarly, some participants said that they were sometimes overcome with a sense of futility. They perceived that regardless of their actions, some outcomes such as dialysis were inevitable (Table 3; 1e); hence, they saw no value in learning about the disease or in trying to self-manage the disease or use the website.

Eleven (52%) of the 21 interview participants said the requirement for a login and password prevented them from using the website because they often forgot their password (Table 3; 1f). Others were limited by poor computer or internet access and said they would prefer a mobile solution (Table 3; 1g). Finally, some participants noted that the onerous process for correcting error in log entries discouraged them from using the self-management tools (Table 3; 1h).

In contrast, other website characteristics appeared to encourage users to visit and return. The perceived reliability of the website’s information and the perception that it was an “authoritative source” drew users (Table 3; 1i). In addition, email reminders prompted them to return; such prompts seemed well-suited to what users characterized as a “fast-paced world” and served as effective reminders to make diabetes self-care a high priority (Table 3; 1j). Similarly, routinization of the online experience appeared to routinize use of the internet for certain aspects of health care. For example, participants reported that increasing their use of the internet and expanding their scope of internet activities created a new norm for internet usage, such that it became more commonplace for them to “look up whatever is interesting to me”, “track stuff online”, and embrace “going paperless” (Table 3; 1 k).

2) Patterns of website use, including role of the blog: Participants said that use of the website was driven by their individual context and circumstances. Rather than browsing at random, users said they were goal-directed: when they had a specific concern, they focused on that area of the website (Table 3; 2a). For example, one participant was initially motivated to visit (and subsequently continued to visit) the foot care section of the website because of her foot symptoms (Table 3; 2b). Participants also commented that they used the website to gauge the urgency of their concerns and to try to obtain immediate answers to their questions (Table 3; 2b).

We explored potential reasons for the unexpected finding that the blog was the most frequently accessed tool and appeared to drive website usage. Participants identified dual roles of the blog in providing a forum for both “expert” and “fellow patient” advice (Table 3; 2c). “Fellow patient” advice consisted of submitting suggested content or commenting on blog posts (i.e. participating in a discussion thread). For patients who were otherwise uncomfortable with asking questions of a health care provider, the blog afforded them the opportunity to obtain “expert” medical answers anonymously (Table 3; 2c,i). The provision of “fellow patient” advice was characterized as promoting a sense of community that some participants felt might combat their sense of isolation (Table 3; 2c,ii). However, this was not a universal sentiment, and some participants felt uncomfortable with and disconnected from the blog (Table 3; 2c,ii). In addition, there was a tension between a desire for online community and a fear of “looking foolish”. For example, some participants said they were afraid of “putting [up] a stupid question” or surmised that others were “shy about how they write” (Table 3; 2c,iii). Other reasons offered for not contributing to the blog included participants’ perceptions that they had nothing to offer, that the blog was not their preferred mode of communication, and that they preferred face-to-face communication (Table 3; 2c,iv). There were mixed views regarding the reliability of the blog content: one user commented that a blog represented “the blind leading the blind”, but others said they were reassured by the fact it was moderated by an expert (Table 3; 2c,iv).

3) General feedback on website characteristics: Participants commented on their impressions of the website overall and provided feedback concerning general features that were appreciated. Participants perceived that the website was accurate, comprehensive (Table 3; 3a), and easy to navigate (Table 3; 3b). While they appreciated the website’s provision of “evidence-based medical content”, they reported a desire for more practical solutions and “real life answers” (Table 3; 3c) and spoke about finding a balance between these two characteristics.

4) Mechanisms of effect of the website on self-efficacy, behaviour change, and diabetes distress: Despite apparently limited use of the website, the intervention appeared to have an effect on self-efficacy, behaviour change, and diabetes distress. Deeper exploration of the data regarding patterns of use and website features uncovered factors that might account for these quantitative findings. For example, rather than returning to the site to revisit and review items, some participants reported that they printed items of interest from the website and subsequently referred to these paper copies (Table 3; 4a). The use of reminder emails also had an effect. Participants reported that these emails not only prompted them to return to and log into the website, but also encouraged them in their own self-management (Table 3; 4b). Finally, one participant noted that the website had kick-started her self-management behaviours, by initially “guilting” her into recording them online. Although she did not subsequently login to the website to record these behaviours, she did continue to record them on paper.


We found that a self-management website for patients with type 2 diabetes led to no improvement in self-efficacy, diabetes distress, or clinical outcomes over the study period. However, there was an improvement in self-care (a secondary outcome), and the group that used the website experienced significantly lower diabetes distress than those who did not use it. Despite a user-centred design process and an increase in the frequency of blog posting from weekly to twice weekly, use of the website (as ascertained by login records) was limited. Our interviews revealed that both patient-related factors (e.g. competing health and life concerns, a sense of futility) and website-related factors (e.g. requirement for login, limited computer or internet access) limited use of the website. We also found that participants were motivated to access the website on the basis of their current needs and concerns, as well as new blog postings, with the blog fulfilling a need for both “expert” medical content and peer support and a sense of community.

These qualitative findings have confirmed the importance of website features such as the reliability and authoritativeness of information [47], as well as the use of blogs [15] and reminders [48] for continued engagement of users. Our findings also emphasize the need to provide a greater proportion of “practical” patient-centred content. A recent qualitative study analyzing 3005 diabetes-related blog posts showed that the most influential blogs were those written by patients and that only 10% of blogs cited biomedical literature [49], highlighting the dual needs for reliable, evidence-based content and engaging patient-based content.

Our data also suggest that mobile devices are a potential avenue through which to improve accessibility and use of a self-management site. A Cochrane review of computer-based diabetes self-management interventions identified 16 randomized controlled trials, which showed a small effect on glycemic control (−2.3 mmol/mol or −0.2%, p = 0.009), with the mobile phone subgroup experiencing a greater effect (−5.5 mmol/mol or −0.5%, p < 0.00001) [50]. Given the increasing preference for mobile devices over desktop computers [51] for health information resources [52], mobile technology may overcome the barriers to website access and use that we identified, through greater integration with patients’ existing routine, such that self-management is no longer seen as additional “illness work”. However, as with web-based technology, a systematic approach to development, testing, implementation, and evaluation of mobile health technology is warranted. Although such technologies are proliferating, with over 736 applications related to diabetes alone, their usability and clinical effectiveness are variable [53], and concerns exist regarding their effectiveness and safety, as well as the security of personal health information [54]. Our findings regarding user engagement with web-based technology echo those for mobile technology: an evaluation of 10 mobile diabetes applications emphasized the importance of user-centred design, an engaging interface, and context-driven use [55].

Competing health concerns were identified as a barrier to web-based self-management. Patients’ adherence to diabetes care is affected by multimorbidity (e.g. depression), which in turn directly affects self-management ability and competes for time and attention [56]. For example, patients with a greater number of comorbidities placed a lower priority on diabetes and had worse diabetes self-management ability [57]. Future interventions should consider strategies, such as shared decision-making and priority-setting, to empower patients with multiple comorbidities to optimize their self-care [58]. For example, a patient may identify mood management as a priority, which is key to subsequent self-care. Thus, greater integration with the person’s cognitive, emotional, and health information-seeking preferences, daily living routine, and health context through the use of patient-based content, mobile devices, and individualized decision-making, may be further strategies to maximize website use and reduce intervention attrition.

Finally, our results may be extrapolated to other chronic diseases. In particular, our finding of the need for tailored content and peer support, balanced with concerns regarding information reliability and confidentiality, is applicable to other strategies for managing chronic disease. For example, a systematic review of the benefits and limitations of social media in the context of chronic disease identified benefits (increased interaction and social support, tailored and accessible information) and limitations (quality concerns and lack of reliability, confidentiality, and privacy) [59] to those we identified. Similarly, our finding of a reduction in diabetes distress in conjunction with no improvements in clinical outcomes echoes findings from intervention strategies targeting other chronic diseases. For example, another systematic review examining the effect of social media on psychological and physical outcomes in chronic disease found a relatively large body of evidence demonstrating psychological benefit (19 identified studies) but limited evidence for physical outcomes (4 identified studies) [60].

This study was limited by its non-randomized design. However, we employed a repeated-measures design that permitted reliable assessment of baseline self-efficacy. Although our primary outcome (self-efficacy) was a non-clinical outcome, it is a validated predictor of patient behaviour change and clinical outcomes [18],[20],[24],[25]. The infrequency of website use likely limited the effect of this intervention, but we obtained valuable insights regarding mediators of website use through our individual interviews. The qualitative evaluation was conducted by individuals who were also involved in developing the intervention, which created a potential for bias; however, we guarded against this bias by including individuals who were not involved in designing the website as members of the qualitative analytic team and by having three coders. As such, we were able to obtain and report critical feedback that participants openly shared. Study strengths include the use of multiple repeated measures, the use of validated outcomes, dual coding of all transcripts, and triangulation of the qualitative findings with the quantitative results [42],[44],[46].


Increasing use of the World Wide Web by consumers for health information and ongoing revolutions in social media are strong indicators that consumers are welcoming and demanding a new era of technology in health care. However, full potential of this technology is hindered by limited uptake and high attrition rates. Use of the Diabetes Online Companion may be optimized by integrating a mobile interface, emphasizing “practical” patient-centred content such as a patient-led blog, and including a “prioritization” feature to help users with competing concerns. Our research findings have shed light on these limitations by identifying characteristics associated with website use and attrition and suggesting strategies to reduce website attrition as a way to potentially optimize clinical outcomes.

Additional file


  1. Centers for Disease Control and Prevention: Diabetes Report Card 2012. Atlanta, GA: Centers for Disease Control and Prevention, US Department of Health and Human Services; 2012.

  2. American Diabetes Association: Standards of medical care in diabetes 2013. Diabetes Care. 2013, 36: S11-S66. 10.2337/dc13-S011.

    Article  Google Scholar 

  3. Haas L, Maryniuk M, Beck J, Cox CE, Duker P, Edwards L, Fisher EB, Hanson L, Kent D, Kolb L, McLaughlin S, Orzeck E, Piette JD, Rhinehart AS, Rothman R, Sklaroff S, Tomky D, Youssef G: National standards for diabetes self-management education and supportdiabetes. Diabetes Care. 2014, Supplement 1: S144-S153. 10.2337/dc14-S144.

    Article  Google Scholar 

  4. Selby JV, Beal AC, Frank L: The Patient-Centered Outcomes Research Institute (PCORI) national priorities for research and initial research agenda. J Am Med Assoc. 2012, 307: 1583-1584. 10.1001/jama.2012.500.

    Article  CAS  Google Scholar 

  5. Cochrane J, Conn VS: Meta-analysis of quality of life outcomes following self-management training. Diabetes Educ. 2008, 34: 815-823. 10.1177/0145721708323640.

    Article  Google Scholar 

  6. Minet L, Møller S, Vach W, Wagner L, Henriksen JE: Mediating the effect of self-care management intervention in type 2 diabetes: a meta-analysis of 47 randomised controlled trials. Patient Educ Couns. 2010, 80: 29-41. 10.1016/j.pec.2009.09.033.

    Article  PubMed  Google Scholar 

  7. Ruppert K, Uhler A, Siminerio L: Examining patient risk factors, comorbid conditions, participation, and physician referrals to a rural diabetes self-management education program. Diabetes Educ. 2010, 36: 603-612. 10.1177/0145721710369705.

    Article  PubMed  Google Scholar 

  8. Shaw K, Killeen M, Sullivan E, Bowman P: Disparities in diabetes self-management education for uninsured and underinsured adults. Diabetes Educ. 2011, 37: 813-819. 10.1177/0145721711424618.

    Article  PubMed  Google Scholar 

  9. Shekelle PG, Morton SC, Keeler EB: Costs and Benefits of Health Information Technology. 2006, Agency for Healthcare Research and Quality (US), Rockville

    Google Scholar 

  10. Yu CH, Bahniwal R, Laupacis A, Leung E, Orr MS, Straus SE: Systematic review and evaluation of web-accessible tools for management of diabetes and related cardiovascular risk factors by patients and healthcare providers. J Am Med Inform Assoc. 2012, 19: 514-522. 10.1136/amiajnl-2011-000307.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Bull SS, Gaglio B, McKay HG, Glasgow RE: Harnessing the potential of the internet to promote chronic illness self-management: diabetes as an example of how well we are doing. Chronic Illness. 2005, 1: 143-155. 10.1177/17423953050010021101.

    Article  PubMed  Google Scholar 

  12. van Vugt M, de Wit M, Cleijne WH, Snoek FJ: Use of behavioral change techniques in web-based self-management programs for type 2 diabetes patients: systematic review. J Med Internet Res. 2013, 15: e279-10.2196/jmir.2800.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Seidman JJ, Steinwachs D, Rubin HR: Design and testing of a tool for evaluating the quality of diabetes consumer-information Web sites. J Med Internet Res. 2003, 5: e30-10.2196/jmir.5.4.e30.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Shrank WH, Patrick AR, Brookhart MA: Healthy user and related biases in observational studies of preventive interventions: a primer for physicians. J Gen Intern Med. 2011, 26: 546-550. 10.1007/s11606-010-1609-1.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Yu CH, Parsons J, Hall S, Newton D, Jovicic A, Lottridge D, Shah BR, Straus SE: User-centered design of a web-based self-management site for individuals with type 2 diabetes – providing a sense of control and community. BMC Med Inform Decis Mak. 2014, 14: 60-10.1186/1472-6947-14-60.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Yu C, Parsons J, Mamdani M, Lebovic G, Shah BR, Bhattacharyya O, Laupacis A, Straus SE: Designing and evaluating a web-based selfmanagement site for patients with type 2 diabetes - systematic website development and study protocol. BMC Med Inform Decis Mak. 2012, 12: 57-10.1186/1472-6947-12-57.

    Article  PubMed  PubMed Central  Google Scholar 

  17. Bandura A: Self-Efficacy: the Exercise of Control. 1997, W.H. Freeman, New York

    Google Scholar 

  18. Glasgow RE, Toobert DJ, Hampson SE, Brown JE, Lewinsohn PM, Donnelly J: Improving self-care among older patients with type II diabetes: the “sixty something…” study. Patient Educ Couns. 1992, 19: 61-74. 10.1016/0738-3991(92)90102-O.

    Article  CAS  PubMed  Google Scholar 

  19. Reichard P, Toomingas B, Rosenqvist U: Changes in conceptions and attitudes during five years of intensified conventional insulin treatment in the stockholm diabetes intervention study (SDIS). Diabetes Educ. 1994, 20: 503-508. 10.1177/014572179402000607.

    Article  CAS  PubMed  Google Scholar 

  20. Anderson RM, Funnell MM, Butler PM, Arnold MS, Fitzgerald JT, Feste CC: Patient empowerment: results of a randomized controlled trial. Diabetes Care. 1995, 18: 943-949. 10.2337/diacare.18.7.943.

    Article  CAS  PubMed  Google Scholar 

  21. Trief PM, Teresi JA, Eimicke JP, Shea S, Weinstock RS: Improvement in diabetes self-efficacy and glycaemic control using telemedicine in a sample of older, ethnically diverse individuals who have diabetes: the IDEATel project. Age Ageing. 2009, 38: 219-225. 10.1093/ageing/afn299.

    Article  PubMed  Google Scholar 

  22. Aljasem LI, Peyrot M, Wissow L, Rubin RR: The impact of barriers and self-efficacy on self-care behaviors in type 2 diabetes. Diabetes Educ. 2001, 27: 393-404. 10.1177/014572170102700309.

    Article  CAS  PubMed  Google Scholar 

  23. Grossman HY, Brink S, Hauser S: Self-efficacy in adolescent girls and boys with insulin-dependent diabetes mellitus. Diabetes Care. 1987, 10: 324-329. 10.2337/diacare.10.3.324.

    Article  CAS  PubMed  Google Scholar 

  24. Nakahara R, Yoshiuchi K, Kumano H, Hara Y, Suematsu H, Kuboki T: Prospective study on influence of psychosocial factors on glycemic control in Japanese patients with type 2 diabetes. Psychosomatics. 2006, 47: 240-245. 10.1176/appi.psy.47.3.240.

    Article  PubMed  Google Scholar 

  25. Lorig KR, Ritter PL, Jacquez A: Outcomes of border health Spanish/English chronic disease self-management programs. Diabetes Educ. 2005, 31: 401-409. 10.1177/0145721705276574.

    Article  PubMed  Google Scholar 

  26. Toobert DJ, Hampson SE, Glasgow RE: The summary of diabetes self-care activities measure: results from 7 studies and a revised scale. Diabetes Care. 2000, 23: 943-950. 10.2337/diacare.23.7.943.

    Article  CAS  PubMed  Google Scholar 

  27. Polonsky WH, Fisher L, Earles J, Dudl RJ, Lees J, Mullan J, Jackson RA: Assessing psychosocial distress in diabetes: development of the Diabetes Distress Scale. Diabetes Care. 2005, 28: 626-631. 10.2337/diacare.28.3.626.

    Article  PubMed  Google Scholar 

  28. Chew LD, Bradley KA, Boyko EJ: Brief questions to identify patients with inadequate health literacy. Fam Med. 2004, 36: 588-594.

    PubMed  Google Scholar 

  29. Chew LD, Griffin JM, Partin MR: Validation of screening questions for limited health literacy in a large VA outpatient population. J Gen Intern Med. 2008, 23: 561-566. 10.1007/s11606-008-0520-5.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Zar JH: Biostatistical Analysis. 1996, Upper Saddle River, N.J, Prentice Hall

    Google Scholar 

  31. Liu G, Liang KY: Sample size calculations for studies with correlated observations. Biometrics. 1997, 53: 937-947. 10.2307/2533554.

    Article  CAS  PubMed  Google Scholar 

  32. Norris SL, Engelgau MM, Narayan KM: Effectiveness of self-management training in type 2 diabetes: a systematic review of randomized controlled trials. Diabetes Care. 2001, 24: 561-587. 10.2337/diacare.24.3.561.

    Article  CAS  PubMed  Google Scholar 

  33. Diggle P, Heagerty P, Liang KY, Zeger S: Analysis of Longitudinal Data. 2002, Oxford University Press, Oxford, New York

    Google Scholar 

  34. Verbeke G, Molenberghs G: Linear Mixed Models for Longitudinal Data. 2009, Springer, Verlag, New York

    Google Scholar 

  35. Laird NM, Donnelly C, Ware JH: Review papers: longitudinal studies with continuous responses. Stat Methods Med Res. 1992, 1: 225-247. 10.1177/096228029200100302.

    Article  CAS  PubMed  Google Scholar 

  36. Cnaan A, Laird NM, Slasor P: Tutorial in biostatistics: using the general linear mixed model to analyse unbalanced repeated measures and longitudinal data. Stat Med. 1997, 16: 2349-2380. 10.1002/(SICI)1097-0258(19971030)16:20<2349::AID-SIM667>3.0.CO;2-E.

    Article  CAS  PubMed  Google Scholar 

  37. Sherrill DL, Holberg CJ, Enright PL, Lebowitz MD, Burrows B: Longitudinal analysis of the effects of smoking onset and cessation on pulmonary function. Am J Respir Crit Care Med. 1994, 149: 591-597. 10.1164/ajrccm.149.3.8118623.

    Article  CAS  PubMed  Google Scholar 

  38. Figaro MK, Elasy T, BeLue R: Exploring socioeconomic variations in diabetes control strategies: impact of outcome expectations. J Natl Med Assoc. 2009, 101: 18-23.

    Article  PubMed  Google Scholar 

  39. Osborn CY, Cavanaugh K, Wallston KA, Rothman RL: Self-efficacy links health literacy and numeracy to glycemic control. J Health Commun. 2010, 15: 146-158. 10.1080/10810730.2010.499980.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Harrell FE, Lee KL, Mark DB: Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996, 15: 361-387. 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4.

    Article  PubMed  Google Scholar 

  41. R Team: R: A Language and Environment for Statistical Computing. 2006, R Foundation for Statistical Computing, Vienna

    Google Scholar 

  42. Kuzel AJ: Sampling in Qualitative Inquiry. Doing Qualitative Research. Edited by: Crabtree BF, Miller W. 1999, Sage Publications, Thousand Oaks, 33-45.

    Google Scholar 

  43. Poland BP: Transcription quality as an aspect of rigor in qualitative research. Qual Inq. 1995, 1: 290-310. 10.1177/107780049500100302.

    Article  Google Scholar 

  44. Kvale S: InterViews: An Introduction to Qualitative Research Interviewing. 1996, Sage Publications, Thousand Oaks

    Google Scholar 

  45. Pope C, Ziebland S, Mays N: Qualitative research in health care: analysing qualitative data. BMJ. 2000, 320: 114-116. 10.1136/bmj.320.7227.114.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Bryman A: Quantity and Quality. Social Research. 1988, Routledge, London

    Google Scholar 

  47. Lemire M, Paré G, Sicotte C, Harvey C: Determinants of internet use as a preferred source of information on personal health. Int J Med Inform. 2008, 77: 723-734. 10.1016/j.ijmedinf.2008.03.002.

    Article  PubMed  Google Scholar 

  48. Nijland N, van Gemert-Pijnen JE, Kelders SM, Brandenburg BJ, Seydel ER: Factors influencing the use of a web-based application for supporting the self-care of patients with type 2 diabetes: a longitudinal study. J Med Internet Res. 2011, 13: e71-10.2196/jmir.1603.

    Article  PubMed  PubMed Central  Google Scholar 

  49. Gruzd A, Black FA, Le Yen TN, Amos K: Investigating biomedical research literature in the blogosphere: a case study of diabetes and glycated hemoglobin (HbA1c). J Med Libr Assoc. 2012, 100: 34-42. 10.3163/1536-5050.100.1.007.

    Article  PubMed  PubMed Central  Google Scholar 

  50. Pal K, Eastwood SV, Michie S, Farmer AJ, Barnard ML, Peacock R, Wood B, Inniss JD, Murray E: Computer-based diabetes self-management interventions for adults with type 2 diabetes mellitus.Cochrane Database Syst Rev 2010, 3.,

  51. Smith A: A: 35% of American Adults Own a Smartphone: One Quarter of Smartphone Owners Use their Phone for Most of their Online Browsing. 2011, Pew Research Center, Washington

    Google Scholar 

  52. Manhattan Research: U.S. mobile health audience jumps to 95 million adults. Cybercitizen Health® U.S. 2013. Available at

  53. Demidowich AP, Lu K, Tamler R, Bloomgarden Z: An evaluation of diabetes self-management applications for Android smartphones. J Telemed Telecare. 2012, 18: 235-238. 10.1258/jtt.2012.111002.

    Article  PubMed  Google Scholar 

  54. Eng DS: The promise and peril of mobile health applications for diabetes and endocrinology. Pediatr Diabetes. 2013, 14: 231-238. 10.1111/pedi.12034.

    Article  PubMed  Google Scholar 

  55. Årsand E, Frøisland DH, Skrøvseth SO, Chomutare T, Tatara N, Hartvigsen G, Tufano JT: Mobile health applications to assist patients with diabetes: lessons learned and design implications. J Diabetes Sci Technol. 2012, 6: 1197-1206. 10.1177/193229681200600525.

    Article  PubMed  PubMed Central  Google Scholar 

  56. Ciechanowski PS, Katon WJ, Russo JE: Depression and diabetes: impact of depressive symptoms on adherence, function, and costs. Arch Intern Med. 2000, 160: 3278-3285. 10.1001/archinte.160.21.3278.

    Article  CAS  PubMed  Google Scholar 

  57. Kerr E, Heisler M, Krein S, Kabeto M, Langa K, Weir D, Piette J: Beyond comorbidity counts: how do comorbidity type and severity influence diabetes patients’ treatment priorities and self-management?. J Gen Intern Med. 2007, 22: 1635-1640. 10.1007/s11606-007-0313-2.

    Article  PubMed  PubMed Central  Google Scholar 

  58. Joosten EA, DeFuentes-Merillas L, de Weert GH, Sensky T, van der Staak CP, de Jong CA: Systematic review of the effects of shared decision-making on patient satisfaction, treatment adherence and health status. Psychother Psychosom. 2008, 77: 219-226. 10.1159/000126073.

    Article  CAS  PubMed  Google Scholar 

  59. Moorhead SA, Hazlett DE, Harrison L, Carroll JK, Irwin A, Hoving C: A new dimension of health care: systematic review of the uses, benefits, and limitations of social media for health communication. J Med Internet Res. 2013, 15: e85-10.2196/jmir.1933.

    Article  PubMed  PubMed Central  Google Scholar 

  60. Merolli M, Kathleen G, Martin-Sanchez F: Health outcomes and related effects of using social media in chronic disease management: A literature review and analysis of affordances. J Biomed Inform. 2013, 46: 957-969. 10.1016/j.jbi.2013.04.010.

    Article  PubMed  Google Scholar 

Download references


We thank Jovita Sundaramoorthy and Carolyn Gall-Casey of the Canadian Diabetes Association for their input and feedback related to project design. We also thank Kevin Thorpe for his invaluable input into the choice of methods for the statistical analysis.


The study was funded by the Canadian Institutes of Health Research (CIHR) Knowledge to Action Operating Grant (funding reference number KAL 94484). Dr. SE Straus is supported by a Tier 1 Canada Research Chair.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Catherine H Yu.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

CHY conceived of the study, collected and analyzed the data, and wrote the manuscript. JAP contributed to the study design, analyzed the data, contributed to the interpretation of the data, and critically revised the manuscript for important intellectual content. MM, BRS, AL, and SES contributed to the study design and to interpretation of the data and critically revised the manuscript for important intellectual content. GL analyzed the data and wrote part of the manuscript. SH conducted the interviews and qualitative analyses and drafted portions of the manuscript. DN conducted the computer programming to develop and refine the website, collected data on web usage and critically revised the manuscript for important intellectual content. OB contributed to interpretation of the data and critically revised the manuscript for important intellectual content. CHY is guarantor for this article. All authors have given final approval of the version to be published, and all agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Electronic supplementary material


Additional file 1:Appendix A: Representative Screenshots of the Diabetes Online Companion. Appendix B: Description of primary outcome scales. Appendix C: Semistructured interview guide. Appendix D: Effect of Intervention comparing users and non-users (after adjustment for time). Appendix E: References.(PDF 1 MB)

Authors’ original submitted files for images

Below are the links to the authors’ original submitted files for images.

Authors’ original file for figure 1

Authors’ original file for figure 2

Rights and permissions

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Yu, C.H., Parsons, J.A., Mamdani, M. et al. A web-based intervention to support self-management of patients with type 2 diabetes mellitus: effect on self-efficacy, self-care and diabetes distress. BMC Med Inform Decis Mak 14, 117 (2014).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: