Homko 2007.

Methods	Randomised controlled trial
Participants	63 women randomised Setting: prenatal clinics at Temple University Hospital, Philadelphia, USA, or one of its satellites; from September 2004 to May 2006 Inclusion criteria: women with GDM (3‐hour OGTT using Carpenter and Coustan 1982 criteria), between 18 and 45 years, at 33 weeks' gestation or less Exclusion criteria: women with prior history of glucose intolerance, or with multiple gestations
Interventions	Telemedicine (n = 34) Women were asked to transmit information via a diabetes health network at least 3 times per week to their healthcare provider. Women without access to the internet received a refurbished computer with free telephone‐based web access to be used during the study; they received a 1‐hour training session by graduate/undergraduate students on how to use a computer, how to access websites, how to set up an email address and receive/send emails. The intervention used 'ITSMyHealthfile', a web‐based disease management interactive healthcare delivery system, with a secure internet server and database which allowed women to send blood glucose and other health data directly to their care provider (blood glucose; fetal movement counts; insulin doses; episodes of hypoglycaemia; ask questions/messages), and receive information and advice from healthcare providers. It required a log‐on ID and password. Standard care (n = 29) Women were asked to record information in a log‐book, which was reviewed by the medical team at prenatal visits. All women All received standard care in the ‘diabetes‐in‐pregnancy program’: were seen for clinical evaluation every 2 weeks until 36 weeks, after which they were seen weekly. Care was provided by a team of maternal‐fetal medicine specialists, residents, diabetes educator, and nutritionists. All received individualised dietary counselling and diabetes education and were instructed in glucose self‐monitoring with portable reflectance meters – women were asked to monitor 4 times per day (before breakfast; 2 hours after meals), and were treated to maintain: fasting glucose ≤ 95 mg/dL, and 2‐hour ≤ 120 mg/dL; women who failed to meet the targets > 90% of the time were started on glyburide or insulin. Women were also asked to perform fetal movement counting 3 times per day and record insulin doses and episodes of hypoglycaemia.
Outcomes	Review outcomes reported: hypertensive disorders of pregnancy (pre‐eclampsia/gestational hypertension); caesarean section; placental abruption; use of additional pharmacotherapy (glyburide; insulin); glycaemic control (fasting blood sugar; blood glucose 2 hours post breakfast, lunch, dinner; mean; HbA1c at birth); adherence to intervention (frequency of monitoring; appointment adherence); sense of well‐being and quality of life (maternal feelings of diabetes self‐efficacy); large for gestational age; perinatal mortality; neonatal mortality or morbidity composite ('composite outcome'); stillbirth; neonatal mortality; preterm birth; respiratory distress syndrome/respiratory complications; hypoglycaemia; hyperbilirubinaemia/jaundice; neonatal intensive care unit admission; gestational age at birth; birthweight
Notes	Funding: "This study was supported by grant RO3 NR008776‐01 from the National Institute of Nursing Research, National Institutes of Health". Declarations of interest: not reported Dates: September 2004 to May 2006
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Women were randomized into one of two groups."
Allocation concealment (selection bias)	Unclear risk	As above; no further details provided.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not feasible to blind participants and personnel.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No detail provided; unclear how lack of blinding would have affected outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	29 women randomised to control group; 2 lost to follow‐up, 1 excluded (twin pregnancy), 1 did not meet criteria for GDM; therefore 25 included in analyses; 34 women randomised to intervention group, 2 formally withdrew; therefore 32 included in analyses.
Selective reporting (reporting bias)	Unclear risk	No access to trial protocol to enable confident assessment of selective reporting.
Other bias	Low risk	Comparable groups at baseline; no other obvious sources of bias apparent.