Effect of a Nurse-Directed Diabetes Disease Management Program on Urgent Care/Emergency Room Visits and Hospitalizations in a Minority Population

A total of 367 diabetic patients were randomized and followed by a specially trained nurse in the DMCP for 1 year (22). Of those, 331 patients had been followed in this county clinic during the year before entering the DMCP. The algorithms used to treat patients enrolled in the DMCP covered glycemic control including those for diet therapy alone; sulfonylurea agents and metformin, either alone or in combination; a glitazone added to maximal (tolerated) doses of metformin and a sulfonylurea agent; bedtime NPH insulin plus daytime oral antihyperglycemia drugs; and a split-mixed insulin regimen with NPH and regular insulin. There were also algorithms and protocols for evaluating and managing lipid disorders, evaluating nephropathy, and treating microalbuminuria (see online appendix, available at http://dx.doi.org/10.2337/dc06-2022).

Urgent care and emergency room visits were analyzed together because patients access the urgent care center when it is open at the community clinic and use the emergency room at the hospital when the clinic is closed. Thus, reasons for using one or the other is logistical, not medical. The urgent care/emergency room visits and hospitalizations during the DMCP year in these 331 patients were compared with the year before enrolling into the DMCP. This study received approval from the institution review board at Charles R. Drew University. Urgent care/emergency room and hospitalization data for these 331 patients were obtained from the county’s management information system. The first five discharge diagnoses and the total charges specific for each visit or hospitalization (not charges related to diabetes care per se) were recorded. Preventable diabetes-related visits and hospitalizations were defined as either metabolic (diabetic ketoacidosis, hyperglycemia, or hypoglycemia) or infection (cellulitis, foot ulcer, osteomyelitis, fungal infection, or urinary tract infection). Other possible diabetes-related diagnoses that could not be realistically affected by 1 year of appropriate diabetes care, such as angina, myocardial infarction, stroke, and non-diabetes–related causes (e.g., gynecological surgery or psychiatric visits), were excluded from the analysis. The assignment of causes for visits and hospitalizations was carried out by one of the authors (A.A.) who was blinded to when (year prior or DMCP year) the visit or hospitalization occurred. When the primary reason for the urgent care/emergency room visit or hospitalization was not clear, she obtained the chart for review.

Continuous variables were analyzed with the Student’s paired t test. Categorical variables were analyzed with the χ² test. The rates of emergency room/urgent care visits and hospitalizations were analyzed with a Z test for the difference of two Poisson rates. Significance was accepted at a 0.05 level (two tailed).

The baseline demographics of the 331 patients at time of entry into the DMCP are shown in Table 1. Baseline A1C levels were 8.8 ± 2.5% (SD), which fell to 7.1 ± 1.4% after 1 year in the DMCP (P < 0.001). At entry into the DMCP, 28% met the ADA A1C goal of <7.0% compared with 64% after 1 year (P < 0.002). At entry into the DMCP, 37% met the ADA LDL cholesterol goal, whereas 80% met it after 1 year (P < 0.04).

There were 95 total urgent care/emergency room visits and hospitalizations in the year before entering the DMCP compared with 52 during the DMCP year, a 45% reduction. However, there were far fewer visits and hospitalizations for preventable diabetes-related causes (Table 2). During the prior year, 14 patients made 15 emergency room/urgent care visits (11 for metabolic reasons and 4 for infections) and 5 patients had 6 hospitalizations (2 for a metabolic reason and 4 for infection). One patient had two urgent care/emergency room visits, one had two hospitalizations, and a third had one of each.

During the DMCP year, four different patients made five urgent care/emergency room visits (one for a metabolic reason and four for infection) and another one was hospitalized for an infection. Note that there was only 1 episode of hyperglycemia during the DMCP year resulting in an urgent care/emergency room visit compared with the year before entering the DMCP in which 11 episodes of hyperglycemia caused urgent care/emergency room visits and 2 resulted in hospitalizations (Table 2). This difference highlights the marked effect of nurse-directed care on preventable diabetes-related metabolic causes of patients interacting with the medical care system outside of regular clinic hours. The difference between the year prior and the DMCP year for all preventable diabetes-related urgent care/emergency room visits and hospitalizations was significantly different (P < 0.001). Total charges specific for these urgent care/emergency room visits and hospitalizations during the year before entering the DMCP were $129,176 compared with $24,630 during the DMCP year.

Nurse (23–34)- and pharmacist (35–37)-directed care have been shown to yield better process and surrogate outcome measures when compared with standard medical care. Only one study evaluated a clinical outcome. Fewer patients cared for by a nurse developed diabetic retinopathy than those receiving standard care (38). This is the first study to compare urgent care/emergency room use and hospitalizations by patients with diabetes receiving nurse-directed care compared with standard care. The results convincingly demonstrate lower resource use among diabetic patients under nurse-directed care. Although there was no control group followed under standard care for 2 years to rule out the possibility that a learning curve accounted for these results, this seems highly unlikely. The average duration of diabetes was 7.6 years, surely long enough for patients to have learned how to take measures to avoid urgent care/emergency room visits and hospitalizations for preventable diabetes-related causes if they could. More likely, one of the reasons for the success of the DMCP was the self-management skills taught to the patients by the nurse during their year under her care.

There are several limitations to this study. Although it is possible that the patients used other centers that were not part of the county system resulting in an underestimate of actual services used, this is unlikely to have influenced the results for two reasons. First, only 15% of the population cared for in this county clinic have any medical insurance, making it unlikely for them to seek care elsewhere. Second, to account for these results, seeking care outside of the system would have had to occur much more frequently during the DMCP year than in the year prior, also very unlikely. The differences between nurse-directed and standard care may be greater in this minority population when compared with other populations. Finally, total charges do not reflect actual costs of care or reimbursement; therefore the amount of cost savings cannot easily be determined from these data.

In conclusion, nurse-directed diabetes care in this minority population resulted in less use of urgent care/emergency room centers as well as fewer hospitalizations for preventable diabetes-related conditions. Policy makers who seek to improve diabetes care and conserve resources should seriously consider adopting this approach.

This study was funded by the American Diabetes Association; Pfizer Health Solutions, Inc.; and Merck & Co. Dr. Davidson was supported by National Institutes of Health Grant U54-RR014616.

The authors are grateful to Robert Stevens for accessing the county computer system and identifying which patients had urgent care/emergency room visits and hospitalizations and for providing the ICD-9 codes and the total charges for each visit and hospitalization and to Martin L. Lee, PhD, for statistical advice.