The purpose of this study is to determine whether aggressive (goal pre-prandial blood glucose <110 mg/dl) versus conservative (goal pre-prandial blood glucose <180mg/dl) diabetes treatment of type 2 diabetic patients on the general medical wards has any effect on hospital outcomes.

Objective: The objective of this study is to determine the effect of an "aggressive" blood glucose control (goal pre-prandial CBG <110 mg/dl) versus a "conservative" blood glucose control (goal pre-prandial CBG <180 mg/dl) on length of stay in hospitalized type 2 diabetic patients using levemir (detemir) and novolog (aspart) insulins.

Study Site: Parkland Memorial Hospital, 5201 Harry Hines Blvd, Dallas, Texas 75235. Parkland Hospital is a 720 adult bed public hospital that serves as the main teaching hospital for the University of Texas-Southwestern Medical Center. In addition, Parkland Memorial Hospital possesses the University Diabetes Treatment Center, an eleven bed medical ward devoted to inpatient diabetes management. This study will be conducted throughout all of Parkland's general medical floors, including the University Diabetes Treatment Center.

Patient Population: The population for this study includes men and women with type 2 diabetes admitted to the general medicine wards of Parkland Memorial Hospital.

Study Design and Duration: This is a single-center, randomized, single blind, non-inferiority study design.

Patients admitted to Parkland Hospital with type 2 diabetes will be recruited to the study within 24 hours of admission. Gravid patients, patients in DKA, or HHS will be excluded. All patients will be evaluated for inclusion and exclusion criteria. During the hospitalization, patients will be stratified by age and admitting diagnosis and randomized to either an "aggressive" blood glucose control (goal pre-prandial CBG <110 mg/dl) or a "conservative" blood glucose control (goal pre-prandial CBG <180 mg/dl). Subcutaneous detemir and novolog will be employed according to the enclosed protocol to meet the goal blood glucose level. The insulin will be administered by nursing staff via the FlexPen, a pre-filled pen-like insulin delivery device. CBGs will be obtained before breakfast, before lunch, before supper, and at bedtime. If a patient is NPO, then the CBGs will be obtained at the time that the patient was to have eaten.

Treatment of the patient's primary admitting diagnosis including plan and procedures will be completely at the primary team's discretion.

Should a patient require transfer to an intensive care unit or to a telemetry unit, then the patient's participation in the study will be held until discharge from said unit. While the patient is in an ICU or telemetry unit, the blood glucose goal and management will be completely decided upon by the primary team.

The study will terminate at the end of the hospitalization.

Inclusion Criteria:

• Existing diagnosis of type 2 diabetes.
• Admitted to a non-telemetry, non-ICU medicine inpatient bed.
• Willing and able to give informed consent.
• HgbA1c of any value will be accepted.

Exclusion Criteria:

• Type 1 diabetes
• Admitted to a telemetry, ICU, or non-medicine inpatient bed (e.g., surgery, neurosurgery, obstetrics/gynecology, psychiatry).
• Diabetic ketoacidosis (arterial ph <7.24, serum bicarbonate <15, positive serum ketones, anion gap >12)
• Hyperosmolar hyperglycemic state (blood glucose >200 mg/dl, serum osmolarity >320 mOsm/kg)
• Inability to cooperate with study personnel.
• Known allergy or intolerance to detemir or novolog.
• Admission to the hospital >24 hours from entry into the study.
• Admission to the hospital for inpatient hospice care.
• Admission/continued admission to the hospital for procurement of a guardian.
• Admission/continued admission to the hospital for rehabilitation.
• Patients admitted with the diagnosis of acute coronary syndrome.
• Patients admitted with the diagnosis of acute cerebrovascular accident.
• Patients currently pregnant or breast-feeding.
• Patients not fluent in English or Spanish as it will be difficult to obtained written informed consent in another language.