Diabetes Improvement Project


One of the major continuing efforts of the Meharry-Vanderbilt Alliance has been our contact and support of community organizations, especially the Consortium of Safety Net Providers of Middle Tennessee. Both Meharry and Vanderbilt hold positions on the Board of the Consortium, as does the Executive Director of the Alliance.

The Consortium of Safety Net Providers of Middle Tennessee (The Consortium)

The Meharry-Vanderbilt Alliance was a founding member of the Consortium in 2001. The Alliance has provided the central meeting place for the board and related activities of the Consortium on a monthly basis.

The Board of the Consortium initiated the Diabetes Improvement Project in the fall of 2009 and asked Alan Graber, M.D. and  the Executive Director of the Meharry-Vanderbilt Alliance to serve as co-directors of the project. The goal of the project is to normalize as near as possible the clinical outcomes of the 7429 patients with diabetes who attend the 21 Consortium sites (see attached patient demographics of the seven clinic systems serving 21 sites). The standards for clinical outcomes are defined as HbA1c levels <7.0%, LDL cholesterol levels <100 mgm%, and BP < 130/80 (all standards of the ADA).

In summary, the seven clinic systems of the Consortium are:

  • Faith Family Medical Center
  • Matthew Walker Comprehensive Health Center
  • Nashville General Hospital at Meharry
  • Saint Thomas Health Community Clinics (South and West)
  • Siloam Family Health Center
  • United Neighborhoods Health Services
  • University Community Health Services

Patient Demographics of the Consortium Clinics (See Table IV)

These clinics serve 106,544 people; 7429 of whom are diabetic. The racial and ethnic diversity shows 33% are African American, 31% are white, 23% are Hispanic, and 13% are listed as “other.” Gender distribution shows that 42% are male and 58% are female. Payer mix reveals 64% are uninsured, 22 % are TennCAre, 5% are Medicare, and 11% are listed as “other.” Three of the clinic systems are Federal Funded Community Health Centers.

Organization and Governance of the Consortium and the Diabetes Project

The Consortium is under the umbrella of the Saint Thomas Foundation 501c3. This arrangement is financial only. This arrangement has been in effect for over 7 years when the Consortium was moved from the Metro Health Dept. to the Foundation.

The Board of the Consortium is representative of the seven clinic systems and several community organizations. Meharry and Vanderbilt as well as the Alliance have seats on the Board. The Board meets every other month at the Alliance offices. (See attached Board Membership)

IRB Review and Approval of the Diabetes Project

The Diabetes Improvement Project proposal and plans have been reviewed and approved by the IRBs of Vanderbilt, Meharry, and Saint Thomas. The project is classified as a quality improvement project with waiver of informed consent. The project has signed business associate agreements (BAA) with each of the seven clinic systems, allowing Alliance staff full HIPPA protected access to patient data. The collected data is stored in a secure and encrypted computer. Names of patients are given only to the specific clinic directors.

Clinical Outcomes

Each month all of the medical records of the visits of all diabetic patients to all of the clinic sites are reviewed and abstracted in to an Access database which is housed at the Alliance office. Reports by patient name are generated and given to the Clinic Directors at the monthly meetings of the Diabetes Project. Each clinic’s data, without names of patients, are also plotted on a Statistical Process Control (SPC) run chart. (Samples of the run charts are attached.) SPC methods go back to the Bell Laboratories in the 1930s and provide a method for very quick determinations of improvements or lack of improvements in the measurements. Deming was a strong advocate for SPCs, calling it” making work visible,” the first step in continuous quality improvement.

The clinical outcome data can be summarized:

  • HbA1c Levels (2901 patients recorded) = 42% <7.0%, 33 % 7-9, 25% >9.0%
  • LDL Cholesterol (2160 patients recorded) = 49% < 100 mgm %,  and 30% 100-130 mgm %, 21% >130 (The 130 mgm % is a weaker standard than the ADA standard of 100 mgm%)
  • BP (5057 patients recorded) = 46% < 130, 25% > 130-140, 30%>140

If all three measures are considered, we find:

  • All three measurements normal = 9%
  • All three measurements abnormal = 20%

Overall, the majority of the patients are out of clinical control.

Recently, we focused on those patients with HbA1c levels> 9.0%, representing an average blood glucose of over 200 mgm %. We found that 30% have levels greater than 9.0% and that 40% of those are not on insulin. Out major efforts at present are directed at getting those patients started on insulin.

Methods of Implementing Improvements

Each month, the Diabetes Improvement Project team meets at the Alliance office. Each clinic CEO, clinical director, and other staff members serve on the project team. Data are shared but coded, so no clinic’s data are identifiable.  Successes and failures are discussed.

Our present emphasis is directed at asking each clinic system to define its patient flow and systematics of care. Each clinic is autonomous and thus must design its own directions and changes. We have begun staff meetings at each of the clinics, teaching SPC methods, the need for insulin, and the need to involve their entire teams in a defined system of care for the diabetic patients. The chronic care model is stressed.

We now have 12 months of baseline for most of the clinics so we will be able with SPC run charts to determine improvements month by month. We can thus move to quick modifications if there is no improvement and by iteration keep making improvements until we see a statistically valid improvement. This we can do on a month by month basis.

Emergency Department Use by Patients with Diabetes

We have collected aggregated data from the THS state database on the use of EDs of all Davidson County hospitals by diabetic patients who are citizens of Davidson County. The attached chart for 2009 shows that there were 7594 visits by diabetics, generating $64,819,965 in charges. These figures include only the uninsured and TennCare visits and do NOT include commercial or Medicare visits.

In discussions with Davidson County hospital administrators, there is interest in a collaborative effort in reducing the non-emergent use by diabetic patients. The attached pie chart of a sample of uninsured patients with diabetes reveals that the majority of visits were for non-emergent conditions.

With the funded grant from the HCA foundation, we will launch a full scale effort to track ED use by the Consortium clinic patients and feed this information back to each clinic director. We hope to install systems for direct clinic appointments from the EDs and also more intense attention to care at the clinic visits. We hope to demonstrate a reduction in use of EDs and thus seek sustaining funding from the area hospital to sustain the improvement project. A small percentage reduction in $64 million in charges could easily provide funds to sustain the core of the project.

Sources of Funds for the Diabetes Improvement Project

We have received grants from several sources listed below:

  • CTSA provided  $3,200 to assist in data collection and retrieval in the early phases of the project.
  • The Diabetes Research and Training Center provided a stipend for a summer student for additional data collection and for the creation of the Access Database for storage of the data and the generation of the SPC run charts.
  • Amerigroup funded $1,000 for data collection
  • Baptist Healing Trust funded $35,000 that allowed us to hire a full time data analyst and data entry person.
  • HCA recently funded a $35,000 grant that will allow us to track ED use of all area hospitals.
  • The Saint Thomas Foundation provided $18,000 for consultation and construction of the Access Data Base.

We have provided in kind support from the Alliance staff for the project.