Registry Characteristics

The size of the Diabetes Registry has grown steadily due to an increasingly complete identification of members with diabetes (i.e., increased sensitivity), in addition to substantial increases in Kaiser membership, increased screening and rising incidence and prevalence of diabetes. There is substantial overlap in each source of identification’s ability to identify diabetic patients. However there exist a small proportion of the registry that are identified uniquely by a single source (e.g., laboratory only). These are more likely with incident cases or newly members with pre-existing diabetes because insufficient time has lapsed for other sources to be utilized. Alternatively mild, diet and exercise-controlled patients may only show up with an out-patient diagnosis or an elevated HbA1c test.

The incidence of diabetes was calculated by isolating plan members newly identified as having diabetes who were part of the health plan for at least one year prior with no record consistent with diabetes (as opposed to new joiners with pre-existing diabetes). The size of the registry increased also due to the increase in membership adding prevalent diabetes cases to the registry

Treatment Patterns over Time
We assessed the trends in hypoglycemic agents used within the diabetes population. We have seen a decrease in the use of insulin, and increase in newer agents (thiazolidinediones, Metformin), and greater reliance on combination therapies over time.

Diabetes registry characterization by race/ethnicity
Hispanics were least likely and Asians most likely to be educated past High School. European American subjects were somewhat older than the other ethnic groups. African Americans and Hispanics were more likely to live in a census block where families were in working class occupations or earned below the poverty level. Good glycemic control (<7% HbA1c) was most prevalent among whites and least common among Asians, while poorest control (>10% HbA1c) was most common among African Americans and Hispanics. Obesity was by far least common among Asians and similar in the remaining ethnic groups. Use of insulin was greatest among African Americans followed by Hispanics and lowest among Asian patients. Asians were least likely to be on combination therapy. African Americans were most likely to control their diabetes pharmacologically, while a higher proportion of Asians and European Americans did not use medications to control their diabetes. Hispanics were least likely to report using exercise to control their diabetes. Duration of diabetes was shortest among Asians and longest among African Americans. Relative to the others, Asians reported substantially lower rates of heavy alcohol consumption, and the lowest rates of smoking, as well as the lowest rates of SMBG. Hispanics were most likely to report a first-degree family history of diabetes, followed by African Americans.

In conclusion, African Americans and Hispanics had lower socioeconomic status than Asian and European Americans subjects. Minority subjects were in general, more likely to have a clinical profile consistent with greater diabetes severity (e.g., poorer glycemic control). Asians stood out as having particularly different behavioral characteristics (e.g., smoking, obesity levels) than the other groups.