Computer Assisted Decision Support CADS for Primary Care of Diabetes

Computer-Assisted Decision Support (CADS) for Primary Care of Diabetes Mellitus Robert Vigersky, M.D., COL, MC1 Robert Galen, M.D., M.P.H.2 David Horne3 Michael Cavotta3 David Rodbard, M.D.4 1Walter Reed Army Medical Center, 2University of Georgia, 3LifeClinic, Inc., 4American Institutes of Research American Telemedicine Association 14 MAY 2007 Problem to be Solved How to improve the long-term outcomes of patients with diabetes given the following:  The vast majority of patients with diabetes are managed by Primary Care Providers (PCPs)  The explosion in number of treatment alternatives (oral agents, combination therapy, and insulins) makes it more difficult for PCPs to be familiar with available options and algorithms Solution Develop a computer-assisted decision support (CADS) program which integrates:  Automated analysis of data from memory meters  Current and past treatment  Laboratory data (e.g. creatinine, liver enzymes) It must have the following features:  Rapid, easy, customization of a treatment plan for each patient, relying on a series of „default‟ plans  Automated interpretation of the “Glucose profile” to identify and prioritize problems  Recommendations to modify the existing treatment regimes  Recommendations for moving to another treatment regimen Architecture of the System Patient Module -Interface with Patient's Meter via PC/modem -Graphs -Statistics -Meal/Med/Exercise Entry -Educational Material Provider Module -Glucose Profile Interpretation -Customize Treatment Plan -Clinical Decision Support -Links CADS AHLTA Clinical Data Repository Administrator Module -Default Treatment Options -Target Ranges - Sequence of doses/regimens -Formulary - Rules for interpretation CADS Module 1. Analysis of the glucose profile (Q Scores) a. Overall level of control - % in target range - Median - % low and “Hypoglycemia score (Mean of (80 - glucose) - % high and “Hyperglycemia score Mean of (glucose - 120) - calculated/predicted A1c b. Overall score: Weighted average using "penalty function" 2. Adequacy of Monitoring 3. Patterns Post-prandial fluctuations Trends during day or night "Somogyi reaction“ Dawn Phenomenon 4. Identification/prioritization of problem 5. Adjustment of doses for current regimen 6. Consideration of alternative regimens 7. Recommendations for laboratory studies and safety alerts c. Hypoglycemia d. Hyperglycemia e. Variability Regimen Type 1 Regimen Type 2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Drug1 diet M S M M S M M S M M S M M S M M S M T T Glargine S T T S T S T T S T S T T S T Glargine Glargine Glargine Glargine Glargine Glargine NPH NPH NPH NPH NPH NPH Drug2 diet Drug3 diet Drug4 Drug5 22 T NPH I-1 I-2 I-3 I-4 I-5 23 24 25 26 27 aspart aspart aspart aspart aspart NPH AM NPH AM aspart aspart aspart aspart PM aspart PM aspart aspart aspart glargine PM NPH PM glargine NPH glargine AM glargine PM Identification of Oral Agents Responsible for Glucose for Specified Times Time of day 1 FBS (BB) 1 Regim en di et m et S M +S M +T S+ T M+ S+T Lant us Lan tus Lan tus L a nt u s M + T M + T M + T M + T M + T M + T M + T Lan tus N P H M NPH NP H NP H NPH 2 3 4 5 6 7 8 9 10 1 1 12 1 3 14 15 16 17 2 AB di et di et di et di et di et di et di et m et m et m et m et m et m et m et S M +S M +S M +S M +S M +S M +S M +S M +T M +T M +T M +T M +T M +T M +T S+ T S+ T S+ T S+ T S+ T S+ T S+ T M+ S+T M+ S+T M+ S+T M+ S+T M+ S+T M+ S+T M+ S+T M S M+ S M+ S M+ S M+ S M+ S M+ S M+ S S+ T S+ T S+ T S+ T S+ T S+ T S+ T S M+ S M+ S M+ S M+ S M+ S M+ S M+ S M+ T M+ T M+ T M+ T M+ T M+ T M+ T S+T 3 BL S M S M S S+T 4 AL S M S M S S+T 5 BD S M S M S S+T 6 AD S M S M S S+T 7 BT S M S M S S+T 8 4AM S M S M S S+T Insulin Responsible for Glucose for Specified Periods Time of day Regimen 1 2 3 4 5 6 7 8 1 FBS (BB) BD “70/30” BT NPH BT Lantus BT NPH BT Lantus BT NPH BT Lantus BT Lantus 2 AB BB “70/30” BB “70/30” BB Aspart BB Aspart BB Aspart BB Aspart BB Aspart BB Aspart 3 BL BB “70/30” BB “70/30” BB Aspart BB Aspart BB Aspart BB Aspart BB Aspart BB Aspart 4 AL BB “70/30” BB “70/30” BB “70/30” BB NPH BB NPH BL Aspart BL Aspart BL Aspart 5 BD BB “70/30” BB “70/30” BB “70/30” BB NPH BB NPH BL Aspart BL Aspart BL Aspart 6 AD BB “70/30” BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart 7 BT BB “70/30” BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart 8 4AM BB “70/30” BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart BD Aspart Algorithm for Treatment of Type 2 Diabetes Diet and Exercise If A1C > 6.5% Monotherapy or Combination Therapy Adequate Not adequate Other Oral Combinations Adequate Follow-up q 3 mo Not adequate Oral Agent Plus Insulin at Bedtime (Glargine or NPH) Adequate Follow-up q 3 mo Not adequate Split-Mixed Insulin or Lispro or Aspart qac + Glargine or NPH qhs Follow-up q 3 mo Validation Strategy De-identified blood glucose and medical records of patients with diabetes on a variety of regimens will be obtained from the Endocrinology Clinic, at Walter Reed Army Medical Center The results of the CADS recommendations for each patient and scored according to the appropriateness of the recommendation by independent PCPs. Testing Phase • PCPs will be randomized into a CADSenabled and usual care group • The primary outcome measures will be improvement in A1c over a 6 month period • Secondary outcomes include: – Medication usage – Adverse events – Usability Summary • We have designed a CADS system which integrates goals of treatment, blood glucose data, current regimen, and laboratory data into a set of recommendations for treatment of DM • The system is customizable on an individual, group, and/or enterprise levels • Preliminary coding of the system has been done and testing will begin this summer Questions CADS DEMO http://ndjh.lifeclinic.com/