Telemedicine in Paediatric Patients with Poorly Controlled Type 1 Diabetes

Introduction

Telemedicine for management of diabetes has been well described in the literature over the last decade and increasing data exists for its use in type 1 diabetes. Reviews of randomized controlled trials have demonstrated the ability of diabetes telemedicine to provide effective education, reduce overall costs and patient time, and to be equivalent to standard care regarding HbA_1c levels.^1–4 Among paediatric patients, telemedicine protocols vary but typically include transmission of data with follow-up telephone counselling. ⁵ Moreover, they also involve transmission of data with computer-based follow-up; ⁶ videoconferencing between school nurse, child and diabetes team; ⁷ text-messaging;^8,9 and a website portal with real time results and educational materials. ¹⁰

One particular benefit of telemedicine could be for patients living in remote locations and/or unable to attend appointments. At our institution, access to specialist paediatric diabetes care was lacking. While all of our patients were military dependants who received primary care from a military treatment facility, most received diabetes care from specialists located over 100 km away and none received regular telemedicine support. None of these outside supporting institutions provided regular feedback and adjustment of insulin regimens unless patients were seen in their respective clinics.

We hypothesized that telemedicine could achieve improvement in HbA_1c control. A paediatric diabetes clinic was established that relied almost exclusively on telemedicine to make changes to insulin regimens. This study reports the initial results of the clinic.

Methods

Case managers and primary care providers at Womack Army Medical Center (WAMC), Fort Bragg, North Carolina, were invited to enrol their type 1 paediatric diabetes patients into the clinic. The staff comprised a diabetes provider, diabetes educator, insulin pump trainer and dietician. All patients were seen at an initial intake visit. This was followed by a group diabetic clinic meeting that occurred approximately three months after the initial visit. Immediately following the group appointment patients were able to meet with a diabetic team member as needed.

Clinic visits were designed to be entirely educational in nature and interventional changes in insulin regimen were not performed except for rare occasions, e.g. when significant lows were identified by the diabetic team. All patients and family members were advised that adjustments of insulin regimens would be performed remotely. They were provided with a messaging service that consisted of a group email address sent to the diabetic team and paging access for emergencies. Patients and parents were encouraged to make all routine contact through email; this included requests for school notes, prescription and supply refills, and routine questions. Text messaging and videoconferencing were not used. Clinic appointments were scheduled as needed.

Thirty-one patients with type 1 diabetes (14 male) with a mean age of 12 years (range 6–19) were admitted to the clinic. Patients were maintained on their insulin delivery system (standard injection or insulin pump therapy). The primary insulin pump (MiniMed Paradigm (Medtronic, Northridge, CA, USA) allowed wireless transmission of data to a website that could be viewed by the clinic staff. The data included glucose values and continuous glucose sensor readings, quantity of carbohydrates counted, amount of meal and correction boluses given, insulin pump settings and the exact time the data were collected.

For patients not on an insulin pump, a spreadsheet was provided that allowed them to enter time, glucose values, quantity of carbohydrates ingested and quantity of basal or bolus insulin given. These values were transmitted to the clinic via the Internet (all patients authorized the sending and receiving of medical information by email). All insulin adjustments were made remotely to the patient by email and occasionally by telephone.

Baseline data of HbA_1c were collected before starting the clinic and after three months; a subsequent follow up visit was also recorded. Statistical analyses were conducted using a standard package (SPSS version 18, IBM Corporation, Armonk, NY, USA). The study was approved by the appropriate ethics committee.

Results

At baseline 9 patients were receiving insulin injection therapy and 22 patients were on insulin pump therapy (see Table 1). Of the 31 total enrolled patients, 45% were male; 59% were white, 23% black and 14% Hispanic. For all patients, the HbA_1c compared to baseline was significantly lower at 3 months (9.1% vs. 11.1%, P < 0.05) and 6.5 months (9.5% vs. 11.1%, P < 0.05). There were no significant differences in HbA_1c levels when comparing age groups (6–11 y vs. 12–19 y) or insulin delivery methods (standard insulin injections vs. insulin pump therapy). The number of telemedicine encounters was not a factor in HbA_1c; the mean number of email messages for HbA_1c decreases of <1.0%, 1.0–2.9 and >3.0% was 5.0 (SD 2.9), 4.5 (SD 4.3) and 4.4 (SD 2.3), respectively. A total of five urgent pager messages was received; all were for sustained hyperglycaemia and none required admission.

OPEN IN VIEWER

Table 1

HbA_1c values (%) at baseline, 3 months and 6 months

	Initial			3 months			6 months
	No	Mean	Range	No	Mean	Range	No	Mean	Range
Insulin injection	9	12.3	(7.4–16.6)	9	9.5	(7.1–13.9)	7	10.7	(7.4–15.4)
Pump therapy	22	10.6	(7.6–17.8)	22	9.0	(6.6–11.9)	20	9.2	(6.2–12.0)
Total	31	11.1	(7.4–77.8)	31	9.1	(6.6–13.9)	27	9.6	(6.2–15.4)

Discussion

In our cohort of paediatric patients with poorly controlled diabetes, the use of telemedicine demonstrated a significant improvement in HbA_1c of 1.5% from baseline to the 3 month follow-up. However, acceptable target levels were not reached and the improvement stopped after 3 months. Our results are similar to other reported trials that showed the average of all end-of-study HbA_1c values in both the control and telemedicine arms was greater than 9%.^11–14 This suggests that for paediatric patients with initially high HbA_1c values, telemedicine is an effective means to achieve improvement, but not necessarily control.

The present study had several limitations, including the use of a retrospective design, small sample size and a lack of matched controls. It should be noted that of the paediatric randomized controlled trials reporting HbA_1c data^11–16 only one has shown an improvement in HbA_1c compared to controls. ¹⁶

Although its generalizability is limited, our study offers some insight into the use of telemedicine. For instance, patients with diabetes who live in rural areas and/or are unable to travel may have an effective way to demonstrate that improvement is possible. The lack of association between improvement and the number of telemedicine contacts suggests that some kind of Hawthorne effect may exist, i.e. part of the benefit is due to a new form of care with a different diabetic team. Telemedicine may thus need to focus on benefits after the initial period of novelty. Alternatively, for those in whom less aggressive control is required – such as those with macrovascular disease or decreased life expectancy – telemedicine may be a useful alternative to standard care. ¹⁷

In summary, the present study demonstrated the effectiveness of telemedicine in reducing HbA_1c levels at 3 months and 6 months in poorly controlled paediatric patients with type 1 diabetes. Further research into sustaining such benefits beyond 6 months and the ability to reach target HbA_1c levels would be beneficial.