A pilot project for improving paediatric diabetes outcomes using a website: the Pediatric Diabetes Education Portal

Introduction

Type 1 diabetes mellitus is a chronic disease that requires lifelong medical management. Education is given to patients and their families at the time of diagnosis, but this is often only basic management (sometimes referred to as ‘survival skills’) and may not be fully understood. ¹ Patients and their families need continuing education as they encounter specific problems in managing their diabetes. It has been suggested that difficulties in accessing appropriately trained medical staff can be a significant barrier to improving patient and family understanding of diabetes management. ² In addition, studies suggest that educational interventions need to continue. ^3–7

We hypothesized that patients would prefer to use an Internet-based system for continuing education. Therefore we created a website, the Pediatric Diabetes Education Portal (PDEP), designed to augment the teaching delivered at clinic visits and improve access to care.

Pediatric diabetes education website

Patients and their families were encouraged to use the website to review their test results, review educational materials and post questions between clinic visits. Patients were shown how to log in to the website at clinic visits but received no additional training on using the site. An example of a family and provider interacting on the website is shown in Figure 1. Parents were given access to their child's case file on the website, and depending on the age of the child and parent preference, the patient was also granted access. In some cases both parents were given separate access to their child's case so both could be notified by email separately when new comments were added.

OPEN IN VIEWER

Figure 1

An example of the provider screen showing comments from a patient, including a blood glucose log and comments from the provider

Results of point-of-care HbA_1c testing and summary graphs of the CGMS readings were printed for patients at clinic visits and were then posted to the website so they could later be viewed from home. When a new comment or document was added to the information on the website, all case participants were notified by email so that they could log in to view it. An example of the test results posted on the website are shown in Figure 2.

OPEN IN VIEWER

Figure 2

Example of HbA_1c and CGMS data posted to the website after a clinic visit. The differently coloured lines on the CGMS graph represent different days. The CGMS recorder measured glucose readings every 5 min and used the patient's glucometer readings to calibrate these points

Postings from the diabetes care team invited patients to review information related to their test results and to review educational materials. Providers were able to switch between cases on the website, use a built-in billing system, and forward the case to other physicians for review if necessary. Patients and their families saw a simplified view of the website without these additional features.

Methods

All children with type 1 diabetes and their families seen at our clinic over one year were offered access to the PDEP. Patients and their families were enrolled and confirmed to have Internet access at home and/or work from which to access the website. The PDEP website allowed families to log in from home to view educational materials related to their test results, post questions and read comments from their physician and diabetes educator. HbA_1c was measured at enrolment and then every three months. Use of a continuous glucose monitoring system (CGMS) was offered at least once to every patient in the study (Medtronic iPro, Medtronic, Minneapolis, USA). The sensor and recorder were worn for 3–5 days and the data were downloaded at clinic visits, when summary graphs were printed for the patient and family.

After six months of use of the PDEP, questionnaires were given to patients and families to assess their satisfaction with the website, point-of-care fingerstick HbA_1c testing and CGMS use. The responses were made on a 5-point Likert scale. All questionnaires were completed by all family members together and either returned by mail or at subsequent clinic visits.

Patients and/or parents were classified as users or non-users of the website. Users were defined as patients and/or parents who (1) logged in to the website four or more times or (2) logged in two or more times if they also posted comments.

To examine improvement in HbA_1c, patients were divided into those whose baseline value was above the ADA age-recommended guideline at enrolment and those whose value was already within the guideline. The change in HbA_1c from baseline was compared between the user and non-user groups after six months with an unpaired t-test.

To determine if use of the website encouraged patients to use other means of accessing care, the number of clinic visits and telephone calls amongst users and non-users related to diabetes management was compared for six months before and six months after enrolment. This analysis excluded telephone calls or clinic visits for prescription refills or for administrative reasons (e.g. school forms).

The study protocol was approved by the appropriate ethics committee. Written informed consent was obtained from all study participants.

Results

A total of 52 patients were enrolled over one year. Twenty patients were excluded for various reasons (see Figure 3), 16 patients had HbA_1c values within ADA guidelines and 16 had HbA_1c values above guidelines. The mean age of the patients who were users of the website was 11.5 years (SD 4.0), and the mean age of the non-users was 10.9 years (SD 4.2). Twenty-five of the 52 patients were female.

OPEN IN VIEWER

Figure 3

Study participants

Of the 16 patients whose baseline HbA_1c was above the guideline value, nine met the definition of being a user of the website. Of these, seven had an HbA_1c measured after six months. The mean HbA_1c decreased from a baseline of 10.5% (SD 2.2) to 9.1% (SD 1.2). This change from baseline was not significant (P = 0.06).

There were seven non-users, of whom only four had a follow-up HbA_1c measured at six months. The mean HbA_1c increased from baseline 9.5% (SD 1.5) to 10.4% (SD 2.5). This change was not significant (P = 0.13).

A between groups comparison (users versus non-users) showed a significant improvement in change in HbA_1c for the user group (P = 0.03). These results are summarized in Table 1.

There was no correlation between CGMS use and the change in HbA_1c for the five patients who tried it in the user group. Only two of the non-users tried CGMS, and neither had a six-month follow-up HbA_1c for comparison.

Both users and non-users showed an increase in provider contacts by clinic appointments and telephone over the six-month period (2.3–7.0 encounters per patient). However, there was no significant difference in the number of contacts between the user and non-user groups (P = 0.42).

Of the 16 patients whose baseline HbA_1c was within guidelines at baseline, all remained so for the duration of the study. Both users and non-users showed a slight increase in mean HbA_1c, but there was no difference between the groups (P = 0.45). These results are summarized in Table 2.

Discussion

The present study suggests ways of improving care for paediatric patients with type 1 diabetes. Previous studies have found that frequent self monitoring of blood glucose (SMBG) is not always helpful and, in isolation, can lead to depressed mood. ¹¹ We hypothesized that when SMBG is linked to educational advice and changes in clinical management, as we did by posting test results and educational material on the PDEP website, patients and their families would more actively manage their diabetes.

It is our contention that patients with type 1 diabetes stand to benefit the most from telemedical intervention. Unlike patients with type 2 diabetes who are often initially managed with oral medications and who only require periodic adjustment of their treatment plan, all of our patients with type 1 diabetes required multiple daily doses of insulin and frequent dose changes. In addition, these patients must constantly adjust their insulin according to their diet plan. Therefore rapid telehealth review of questions to help patients change insulin doses is appealing. While previous studies in type 1 diabetes have shown significant improvement in patient contacts with their provider, they have not yet shown consistent evidence of improvement in metabolic control as measured by HbA_1c and hypoglycaemic and hyperglycaemic episodes. ¹² A meta-analysis of studies using Internet-based blood glucose logs for type 1 diabetes has confirmed these findings. ¹³

Participants in previous diabetes telemedicine intervention studies have reported the need for easier interaction between patients and clinicians, and dietary and nutritional educational material for different groups of patients. ¹⁴ This was also found in the Diabetes-STAR trial which failed to show improvements in self-care behaviour in patients with diabetes who used a website to review their records. Interviews with patients showed the need to: (1) address differences in types of users, (2) send out alerts when new information was available, and (3) provide more oversight of the use of dietary data. ¹⁵ The PDEP was designed to address all these concerns by automatically notifying patients and providers when new information was posted to the website and by providing links to educational material.

In our study both users and non-users increased contact with their diabetes care provider by telephone communication and clinic appointments, but only users of the website showed improvements in HbA_1c. This suggests that it was website use, and not more frequent clinic follow-up, that led to improvement in HbA_1c. We speculate that being able to log in and see clinical test results may have allowed website users to obtain more frequent feedback about managing their diabetes, ultimately leading to improved diabetes management.

We also found that CGMS use was not associated with a change in HbA_1c, probably because of the small number of patients in the study. In addition, users of the website whose HbA_1c started within the recommended range showed no change versus non-users, and on average these patients stayed within the guidelines. This was expected as they had little room for improvement to begin with.