HbA1c is the glycosylated fraction of hemoglobin A. HbA1c is measured to identify average plasma glucose concentration over prolonged periods of time.

Least Squares (LS) mean was determined by analysis of covariance (ANCOVA) model with last observation carried forward (LOCF) and with terms for change from baseline in HbA1c as response, treatment as fixed effect and baseline HbA1c as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for HbA1c.

HbA1c is the glycosylated fraction of hemoglobin A. HbA1c is measured to identify average plasma glucose concentration over prolonged periods of time.

Least Squares (LS) mean was determined by ANCOVA model with LOCF and with terms for change from baseline in HbA1c as response, treatment as fixed effect and baseline HbA1c as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for HbA1c.

Hemoglobin A1c (HbA1c) is the glycosylated fraction of hemoglobin A. HbA1c is measured to identify average plasma glucose concentration over prolonged periods of time.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for HbA1c <7%.

Hemoglobin A1c (HbA1c) is the glycosylated fraction of hemoglobin A. HbA1c is measured to identify average plasma glucose concentration over prolonged periods of time.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for HbA1c <7%.

Fasting Plasma glucose (FPG) is a test to determine how much glucose (sugar) is in a plasma sample after an overnight fast. Least Squares (LS) means was determined by ANCOVA model with LOCF and with terms for change from baseline in Venous FPG as response, treatment as fixed effect and baseline Venous FPG as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for Venous Fasting Plasma Glucose.

Fasting Plasma glucose (FPG) is a test to determine how much glucose (sugar) is in a plasma sample after an overnight fast. Least Squares (LS) means was determined by ANCOVA model with LOCF and with terms for change from baseline in Venous FPG as response, treatment as fixed effect and baseline Venous FPG as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for Venous Fasting Plasma Glucose.

Fasting blood glucose (FBG) and post prandial glucose (PPG) [pre-breakfast (fasting) and post-breakfast (approximately 2 hours after breakfast)] was measured using FSBG. LS Mean was measured with ANCOVA model with LOCF and with terms for change from baseline in FSBG-based FBG/PPG as response, treatment as fixed effect and baseline FSBG-based FBG/PPG as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for FBG and PPG.

Fasting blood glucose (FBG) and post prandial glucose (PPG) [pre-breakfast (fasting) and post-breakfast (approximately 2 hours after breakfast)] was measured using FSBG. LS Mean was measured with ANCOVA model with LOCF and with terms for change from baseline in FSBG-based FBG/PPG as response, treatment as fixed effect and baseline FSBG-based FBG/PPG as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for FBG and PPG.

Total daily insulin dose in the basal insulin analog and in Insulin Analog Mid Mixture group at week 24 and 48.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for Daily Insulin Dose.

LS means were calculated using ANCOVA model with LOCF and with terms for change from baseline in bodyweight as response, treatment as fixed effect and baseline bodyweight as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for Body Weight.

LS means were calculated using ANCOVA model with LOCF and with terms for change from baseline in bodyweight as response, treatment as fixed effect and baseline bodyweight as covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for Body Weight.

Hypoglycemic episodes are defined as events that are associated with reported signs and symptoms of hypoglycemia and/or documented blood glucose (BG) concentrations of ≤70 mg/dL (3.9 mmol/L). The overall yearly rates (events/participant/year) of those hypoglycemic events, calculated as, for each participant, the number of episodes times 365.25 and then divided by the participants treatment duration, will be summarized, and analyzed by a Negative-binomial regression model with treatment as fixed effects and log of (patient's treatment duration/365.25) as an offset variable.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for Hypoglycemia.

Insulin treatment change can be insulin treatment discontinuation, switch, intensification or reduction in frequency.

1. Discontinuation: Defined as stopping insulin treatment for 30 days or more.
2. Switch: Defined as stop the initial insulin therapy and started another insulin therapy of different class.
3. Intensification: Defined as any of the following: adding meal time insulin in basal insulin analog QD group; changing from BID to TID (Three times a day) in insulin analog mid mixture BID group
4. Reduction in frequency: Defined as any of the following: changing from BID to QD; changing from TID to BID or QD.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for Insulin Treatment Change.

Hemoglobin A1c (HbA1c) is the glycosylated fraction of hemoglobin A. HbA1c is measured to identify average plasma glucose concentration over prolonged periods of time. Percentages of participants who achieved HbA1c levels of <7% were analyzed using a logistic regression model with logic link function, treatment as fixed effect and baseline HbA1c as continuous covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for HbA1c <7%.

Hemoglobin A1c (HbA1c) is the glycosylated fraction of hemoglobin A. HbA1c is measured to identify average plasma glucose concentration over prolonged periods of time. Percentages of participants who achieved HbA1c levels of <7% were analyzed using a logistic regression model with logic link function, treatment as fixed effect and baseline HbA1c as continuous covariate.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for HbA1c <7%.

The SEITQ is designed to measure an individual's self-efficacy related to insulin therapy. The SEITQ consists of 5 items (that is, statements). The first 4 statements imply confidence in completing the tasks needed to take insulin correctly and avoid both hyperglycemia and hypoglycemia, whereas the last statement is an outcome expectation and implies that performance of these tasks will lead to avoidance of complications. Each item score ranges from 1 (strongly disagree) to 7 (strongly agree). The total SEITQ score is the sum of each item scores, with the range of 5 to 35. Higher SEITQ score indicates better outcome (higher self-efficacy). LS Mean was calculated using Mixed Models Analysis (MMRM) for repeated measures with all post-baseline SEITQ as responses, baseline SEITQ as a continuous covariate, treatment group, Visits, and treatment by visit interaction as fixed effects and participant as a random effect.

All randomized participants who received at least one dose of study drug and had evaluable baseline and post-baseline data for SEITQ Score.