Impact of an Intramural Competition on Outcome of Diabetes Care

The translation of research results in diabetes (1) to clinical practice continues to be a major challenge (2,3). Physicians in training present a unique opportunity for shaping practice behavior. We tested the effect of an intramural contest on the outcome of diabetes management by 155 medical residents at a major U.S. teaching hospital. The contest was approved by the house staff leadership and was set for a 6-month period (1 October to 30 April). The residents, organized into four firms (A, B, C, and D), used various strategies (e.g., didactic lectures, case-based learning, and journal reviews) during weekly firm meetings to enhance competitiveness.

The outcome measures were HbA_1c levels (primary) and testing frequency (secondary). These data, extracted from 100 randomly selected charts of diabetic patients from each firm, were compared during and 6 months before the contest. Data from patients who had two or more HbA_1c results during the contest were used for analysis of the primary outcome. The “baseline” HbA_1c was the first measured value during the contest, and the “final” HbA_1c was the last result obtained before conclusion of contest. For comparison, HbA_1c data obtained during the 6 months preceding the contest (“precontest”) were analyzed. Results are means ± SE. Statistical analyses were by paired t test for changes in HbA_1c values, and the χ² test was used for HbA_1c testing frequency. The prizes included a plaque and $1,500 to the firm with the greatest reduction in HbA_1c level, a plaque and $1,000 to the firm with second best HbA_1c result, and $500 to the firm with the third best HbA_1c outcome.

For the primary outcome, the number of patients with evaluable data (i.e., two or more HbA_1c values during the contest) were 35, 31, 32, and 38 in firms A, B, C, and D, respectively. The changes in HbA_1c were −0.70 ± 0.40, −0.12 ± 0.35, −0.54 ± 0.34, and −0.21 ± 0.28% for patients in firms A–D, respectively. Next, we evaluated the frequency of HbA_1c testing: of the 400 patients, 180 (45%) had one HbA_1c measurement during the precontest period as compared with 263 (66%; P < 0.001) postcontest. The number of patients with two HbA_1c tests in 6 months were 57 (14%) precontest and 136 (34%; P < 0.001) postcontest. Analyzing all HbA_1c data, the mean precontest value was 9.3 ± 1.8% (n = 180) vs. 8.8 ± 1.9% (n = 263) postcontest. Among patients with two or more HbA_1c data, the change in HbA_1c over 6 months was 0.01 ± 0.2% precontest and −0.5 ± 0.2% postcontest (P = 0.015).

This study shows that intramural competition can be an effective motivational tool for diabetes care providers. The contest generated an increased tempo of diabetes-related educational activities among house staff, which translated to improved glycemic control. Improvement in diabetes control requires a committed approach: more frequent patient contacts; prompt reaction to test result, including home blood glucose data; dietary reinforcement; and medication adjustments, among others. Because the skills developed during the diabetes contest could have a lasting effect on the practice style of physicians, disease-specific intramural contests might be a useful tool for enhancement of clinical skills during residency training programs. Competitive pressure, desire for recognition, and increased awareness of standards of care probably accounted for physician motivation in this exercise, rather than mere financial reward (4).