Table 1.

Randomized controlled trials examining patient portal for diabetes management.

Authors, country	Study aims, design, and level of evidence	Sample and retention	Patient portal features	Intervention	Outcomes (portal related)	Findings
van Vugt et al (2016) [11], Netherlands	2-group study, 6-month randomized controlled trial (RCT) to study the uptake and effects of e-Vita with a self-management support program (SSP) and personalized coaching for patients (Ps) with type 2 diabetes mellitus (T2DM); Evidence: Grade A	N=132; males: 59.1%; white: 91%; age: 67.9 (SD 10.4) years; body mass index (BMI): 30.2 (SD 5.2); glycated hemoglobin (HbA1c): 6.6%; retention: Coaching group (CG): 43.9%; noncoaching group (NCG): 59.1%	e-Vita (diabetes mellitus [DM]-specific) by VU University Medical Center allows Ps to access diabetes education; access data from electronic medical records (EMRs) of primary care physicians (PCPs); receive messages from providers; receive SSP	CG (n=66): Personal health record (PHR)+SSP+coaching; NCG (n=66): PHR+SSP	HbA1c, BMI, systolic blood pressure (SBP), diastolic blood pressure (DBP), cholesterol, diabetes self-care, diabetes-related distress, and PHR and SSP use	Intention-to-treat (ITT) was applied. PHRs were assessed by 128 Ps, of which 59 Ps never returned to the PHR. The use declined over time. The SSP was used by 5 Ps in the CG and 1 patient in the NCG group, 3 of whom asked a coach for feedback. Ps recently diagnosed actively used the SSP; no differences were observed on outcome measures between baseline (BSL) and 6 months for the 2 groups.
Tang et al (2013) [12], United States	2-group study, 12-month RCT to evaluate an Web-based disease management system by Ps with uncontrolled T2DM; Evidence: Grade A	N=415; Intervention (Int) vs Control (Con): males: 58.9% vs 61%; white: 60% vs 58%; age: 54 (SD 10.7) vs 53.5 (SD 10.2) years; weight: 215.3 (SD 49.4) vs 218.4 (SD 51.3) pounds; HbA1c: 9.24 (SD 1.59) vs 9.28 (SD 1.74); Retention: 87%	Web-based diabetes management system (DM specific) by Palo Alto Medical Foundation allows Ps to monitor glucose remotely; view summary report; document nutrition and exercise; record insulin; communicate with the health team; receive advice; personalized education	Int (n=202): access to Web-based disease management system for diabetes; Con (n=213): usual care	HbA1c, BP, low-density lipoprotein (LDL), health care utilization, diabetes knowledge, diabetes treatment satisfaction, and depression screening	ITT was applied. Int had reduced HbA1c at 6 months (−1.32% Int vs −0.66 Con, P<.001), but not at 12 months. The Int had better LDL control at 12 months (P=.001), but no difference for BP or weight. Ps in the Int had a lower distress score (P<.001), better knowledge of glucose testing (P=.004), better understanding of diabetes (P<.001), greater treatment satisfaction (P<.001). No differences were noted in the depression screening or health care utilization.
Fonda et al (2009) [13], United States	2-group study, 12-month RCT to examine changes in Problem Areas in Diabetes (PAID), and its association with use of an internet-based diabetes care management (IBCM) program; Evidence: Grade A	N=104; males: 99%; white: 76.7%; age: 60.9 (SD 10.3) years; HbA1c: 9.9 (SD 0.9%); Retention not reported	IBCM (DM specific) by VA Boston Healthcare System allows Ps to transmit BP and glucose data from devices; view BP and glucose data; message care managers; access diabetes education	Int (n=52): access to the IBCM program; Con (n=52): usual care	Diabetes distress (PAID), and pattern of usage	The decline in PAID score was significant for sustained users of the portal but not for nonusers in the Int group. Sustained users (n=27) had lower PAID scores at baseline.
McCarrier et al (2009) [14], United States	2-group study, 12-month RCT to test whether a diabetes case management program can improve glycemic control and self-efficacy in adults with T1DM; Evidence: Grade A	N=77; males: 67.5%; white: 96.1%; age: 37.3 (SD 8.09) years; HbA1c: 8%; Retention: 83%	Web-based program (DM specific) by University of Washington (UW) General Internal Medicine Clinic allows Ps to view EHR data; upload glucose readings; enter medication, nutrition, and exercise; create action plans; access education	Int (n=41): usual care+Web-based case management program; Con (n=36): usual care	HbA1c, diabetes-related self-efficacy, and usage	ITT was applied. A nonsignificant decrease in HbA1c in the Int compared with the Con group (−0.48%, 95% CI −1.22 to 0.27) between groups. The Int group had an increase in self-efficacy compared with the Con group (95% CI 0.01 to 0.59, P=.04). The log-in rate was 61%, and averaged 3.3 log-ins per patient. Emails were sent by 44% users, with a mean of 5.0 messages.
Ralston et al (2009) [15], United States	2-group study, 12-month RCT to test Web-based care management of glycemic control using a shared EMR in Ps with T2DM; Evidence: Grade A	N=83; Int vs Con: females: 47.6% vs 51.2%; white: 89.7% vs 73% (P=.06); age: 57 vs 57.6; Glycohemoglobin (GHb): 8.2% vs 7.9%; Retention: 89.2%	Web-based diabetes support program (DM specific) by UW General Internal Medicine Clinic allows Ps to access EHR data; communicate with providers; send glucose readings; enter exercise, diet, and medication data; access education	Int (n=42): usual care+Web-based case management program; Con (n=41): usual care	GHb, total cholesterol, SBP, DBP, health care utilization, and usage	ITT was applied. More change in GHb among the Int group compared with the Con group at 12 months (change −0.7%, P=.01). SBP, DBP, total cholesterol levels, and use of in-person health care services did not differ between groups. EHR was accessed 76%, 69% emailed, and 33% entered data. Number of page views was not associated with GHb improvement.
Grant et al (2008) [16], United States	2-group study, 12-month RCT to evaluate the impact of a PHR for T2DM; Evidence: Grade A	N=244; Int vs Con: females: 43% vs 56% (P=.04); white: 93% vs 84% (P=.04); age: 58.8 vs 53.3 years (P<.001); HbA1c: 7.3% vs 7.4%; Retention: 50.4%	Patient Gateway by Partners Health care system allows Ps to update registration information; send messages; confirm appointments; request prescription refills; access DM modules	Int (n=126): access to a DM-specific PHR (ie, review mediations, and access decision support and care plans); Con (n=118): non-DM-specific PHR	HbA1c, BP, and LDL	ITT was applied. More Ps in the Int group had DM treatment adjusted compared with the Con group (53% vs 15%; P<.001). There was no difference in HbA1c between groups (Int vs Con: 7.1% vs 7.2%) after 1 year. BP and LDL showed similar patterns at BSL and follow-up between groups.