Special Series: AAFP Award-winning Research Papers Vol. 31, No. 5 331 Using a Flow Sheet to Improve Performance in Treatment of Elderly Patients With Type 2 Diabetes Gary Ruoff, MD; Lynn S. Gray, MD, MPH Background and Objectives: Numerous studies have shown that physicians do not provide all the preventive and therapeutic care recommended for patients with diabetes. This study determined if use of a medical record flow sheet could increase compliance with seven quality-of-care indicators developed by the American Diabetes Association. Methods: Subjects included Medicare enrollees with type 2 diabetes. Following an analysis of baseline data, physicians in the practice used a flow sheet that contained recommended guidelines for diabetes care. Staff inserted the flow sheet into the records of patients included in the baseline sample. Physicians and staff also received education about use of the flow sheet. The postintervention sample consisted of the same subjects, if they had been seen by the practice during a 3-month period. Results: The records of 114 subjects were reviewed at baseline. Of these subjects, 109 received care during the study period. Improvement was shown in six of the seven quality indicators and was also observed in the performance of post-intervention rates for patients whose flow sheet was used, compared with those for whom it was not used. Conclusions: The results indicate that education and performance in diabetes care can improve with the use of a flow sheet. (Fam Med 1999;31(5):331-6.) More than 11 million Americans are afflicted with diabetes mellitus, 5% to 10% of whom are insulin dependent. The remainder have non-insulin dependent (type 2) diabetes mellitus, and most of those individuals are in older age groups. In fact, by age 70, up to 10% of the population has diabetes. With more than 1.5 million Medicare beneficiaries in the state of Michigan, more than 150,000 have diabetes, and of those, 135,000 145,000 have type 2 diabetes. Complications of diabetes lead to significant morbidity. For example, 20% of patients with type 2 diabetes develop nephropathy; more than 36,000 amputations per year are attributable to diabetes neuropathy. Diabetic retinopathy is the leading cause of blindness in the elderly, with 8,000 new cases of diabetes-related blindness in the United States annually. 1 In 1992, more than $100 billion was spent in the United States on medical care for patients with diabetes. In the Medicare population, $30 billion, or about 27% of their total budget, is spent annually on diabetes care. 2 The cost of such care is largely due to preventable complications. From the Department of Family Practice, Michigan State University, Kalamazoo. The Diabetes Control and Complications Trial demonstrated the effects of good glycemic control in the prevention of diabetic complications in insulindependent (type 1) diabetes. 3 Based on the results of the United Kingdom Prospective Diabetes Study, 4 these results also appear to be applicable to patients with type 2 diabetes. Advances in diabetes therapy include the control of hyperglycemia and other risk factors, as well as surveillance for and treatment of diabetes-related complications. Unfortunately, surveys of US physicians have consistently shown poor performance in surveillance for and prevention of complications. 5,6 As a consequence of these findings, we developed an intervention to demonstrate that the use of a flow sheet (Appendix 1) that contained seven practice guidelines for the management of patients with type 2 diabetes would improve performance of family physicians in their care of patients with diabetes. The study was carried out in the context of a quality improvement project in collaboration with the State of Michigan s Peer Review Organization/Quality Improvement Organization (PRO/QIO), which implements quality improvement projects for Medicare beneficiaries in the state under contract with the Health Care Finance Administration (HCFA).
332 May 1999 Family Medicine Methods Overview This project followed the typical Health Care Quality Improvement Project design as implemented by HCFA and used by most state PRO/QIOs. After selection of clinical quality indicators from accepted clinical practice guidelines, baseline data were collected and analyzed by a chart review. After reviewing the data, the practice developed an intervention to improve performance. Following the intervention, data were again collected to measure changes in the physicians performance. Subjects The study was conducted in a 17-physician private community-based family practice group in western Michigan. Patients to be included in the study were selected by the practice staff, who identified Medicare patients with type 2 diabetes who were seen multiple times at the practice site over a 3-month period. This methodology was chosen because of the need to identify patients who were being seen on a relatively regular basis. All patients were age 65 or older. Baseline Data The medical records of each patient were reviewed to determine if the patient had received interventions recommended in the American Diabetes Association guidelines for patients with type 2 diabetes. 7 These interventions included: 1) conduct an evaluation with a glycosylated hemoglobin (HbAlC) every 6 months, 2) perform annual urine protein dipstick, 3) if the dipstick is negative, obtain urine for micro albuminuria, 4) perform retinoscopy annually, 5) if abnormalities are noted in the retinoscopy, refer the patient to an eye care professional, and 6) perform a foot exam every 6 months. Additional American Diabetes Association guideline indicators were assessed by determining if the record contained evidence of regular home blood glucose monitoring and diabetic education. Flow Sheet Intervention Following collection and analysis of the baseline data, the practice developed and implemented a flow sheet to be used in the medical records of these patients, beginning in November 1996. The flow sheet is shown in Appendix 1. Staff at the practice inserted the flow sheet into the records of all patients who had been seen during the baseline period. All physicians in the practice were instructed in the purpose and use of the flow sheets. Post-intervention Data The post-intervention sample consisted of all patients originally identified in the baseline analysis who were seen in the practice during a 3-month period after we began using the flow sheet. Records of these patients were examined to determine compliance with the quality indicators specified in the flow sheet. Data Analysis The project design enabled two comparisons. The first comparison was between the baseline sample and the post-project sample (Table 1). The records were analyzed to determine compliance before and after the intervention, with the established quality indicators outlined earlier. The second comparison was based on the fact that while the flow sheet was inserted into all intended records, it was not always used. Thus, we examined differences in indicator rates between the records of patients in the post-intervention sample for whom a flow sheet had been used (n=36) and those for whom it was not used (n=73). Two-tailed Pearson chi-square tests were used for all comparisons except when cell sizes were less than five. A Fisher exact test was used for comparisons involving small cell sizes. Table 1 Differences in Indicator Rates: Comparison Between Baseline Sample and Post-project Sample Baseline Post Project Total patients 114 109 1. HbA1C every 6 months 11.4% 19.3% 2. Annual urine protein dipstick 40.4% 14.7%* 3. Test for micro albuminuria 0% 25.0%** if dipstick negative*** (0/39 cases) (1/4 cases) 4. Annual retinoscopy 26.3% 31.2% 5. Refer to ophthalmologist if abnormal*** 27.2% 63.6% (3/11 cases) (7/11 cases) 6. Foot exam every 6 months 10.5% 11.1% 7. Home blood glucose monitoring 38.6% 45.9% HbAlC glycosylated hemoglobin * P<.01 ** P<.1 *** Two-tailed Fisher s exact test was used to make comparison due to small cell sizes ( 5).
Special Series: AAFP Award-winning Research Papers Vol. 31, No. 5 333 Results At baseline, 128 patients were identified by staff as having diabetes. Of these, only 114 had type 2 diabetes and were included in the study. Their mean age was 72, 44% were male, and the diagnosis of diabetes had been made an average of 8 years previously. Of these 114 patients, 109 received care during the study period and had medical records that could be evaluated after the intervention. Data in Tables 1 and 2 indicate improvement in compliance with the quality indicators between the baseline and follow-up chart review. Even though the flow sheet was not used in the charts of 73 of 109 patients in the study, improvement was still noted in six of the seven indicators. Prior to implementation of the flow sheet, diabetic education was documented only in 10.5% of patient records. After we began using the flow sheet, the overall rate of diabetic education by physicians was 38.5%. On the other hand, rates of documentation of diabetic education by registered nurses, physician assistants, licensed practical nurses, and registered dieticians were unchanged following implementation of the flow sheet. Evidence of blood glucose monitoring at home demonstrated an improvement. Prior to implementation of the flow sheet, home blood glucose monitoring was noted in 38.6% of the flow sheets; the rate of documentation of home glucose monitoring was 45.9% following implementation. On the other hand, the percentage of subjects whose records indicated that the patient s urine had been dipstick tested for protein decreased from 40.4% to 14.7%. However, this decrease was caused by the fact that we instituted microalbuminuria testing between the pre- and post-implementation assessment. Prior to that time, we used conventional urine protein dipsticks. If one examines Table 1, it is clear that the rate of checking either urine protein dipstick or urine microalbuminuria (items 2 and 3 on Table 1) actually remained stable. Discussion Use of a flow sheet was associated with an increase in documentation of performance of the identified qualityindicator behaviors. However, the flow sheet serves as both a documentation tool and a reminder system. For some items, one cannot be certain if the higher rate of performance recorded in the charts represented a true increase in the frequency of performance or simply an increase in the frequency with which performance was documented in the record. However, the fact there was improvement in all indicators in Tables 1 and 2, even those that could easily be assessed regardless of chart documentation (eg, HbAlC results are available from lab test results), suggests that improvement was related to improved performance and not just improved documentation. Limitations One important limitation of our study is that it was conducted in a single family practice clinic in which staff had experience in research projects and continuous quality improvement methods. This precludes generalizing the results beyond the setting in which the study was conducted. The patients also consisted only of Medicare enrollees with type 2 diabetes, so we cannot determine if flow sheets are associated with change in behavior in management of other diseases. A second limitation is that all data abstraction was conducted onsite at the clinic, and medical records were not available for further review during data analysis. This approach did not permit identification or correction of data entry errors or omissions detected during data analysis. A third limitation is that the study results could have been influenced during the study period by changes in staff, changes in referral patterns, seasonal variation, etc, that might have explained changes in physician behavior. This possibility cannot be elimi- Table 2 Differences in Indicator Rates When Flow Sheet Was Used Versus Not Used Flow Sheet Flow Sheet Not Used Used Total patients 73 36 1. HbA1C every 6 months 15.1% 27.8% 2. Annual urine protein dipstick*** 0% 44.4%* 3. Test for micro albuminuria 0% 25.0% if dipstick negative**** (0 cases) (1/4 cases) 4. Annual retinoscopy 16.4% 61.1%* 5. Refer to ophthalmologist if abnormal*** 33.3% 100.0%** (2/16 cases) (5/5 cases) 6. Foot exam every 6 months*** 1.4% 30.6%* 7. Blood glucose monitoring at home 34.3% 69.4%* HbAlC glycosylated hemoglobin * P<.01 ** P<.1 *** Two-tailed Fisher s exact test was used to make comparison due to small cell sizes ( 5). **** No comparison was made due to zero cases in the flow sheet not used group.
334 May 1999 Family Medicine nated because we did not monitor changes in these parameters. Finally, the most important limitation is that the results do not permit us to ascribe improvement in physician performance to the flow sheets, per se. Rather, it is possible that there was an increase in physicians awareness of issues surrounding diabetes management because of educational activities in the clinic surrounding implementation of the project. This is a likely possibility because prior to implementation of the project, site visits were conducted to both educate clinic staff regarding project design and to provide clarification of the overall study design. Nonetheless, because performance for patients whose charts contained the flow sheet (Table 2) improved, the flow sheet does seem to have been independently related to improved performance. Implications The project provided several important insights into cooperative improvement projects in the ambulatory setting and, based on the data, resulted in significant improvement in the quality of care for Medicare patients with type 2 diabetes seen by this private practice group. As a consequence, the study shows that practices can identify methods to improve care and develop a means to implement these methods. In addition, the flow sheet provides a method for improving documentation, while also serving as a functional reminder system. This project also provided an opportunity for a state PRO/QIO to work with a primary care practice to modify the delivery of care. The resulting improvement, in compliance with practice guidelines, suggests that tools such as the flow sheet we used can be effective in improving quality of care. Acknowledgment: This manuscript was presented at the 1998 American Academy of Family Physicians Annual Scientific Assembly in San Francisco. Disclaimer: The analyses on which this publication is based were performed under contract number 500-96P542, titled Utilization and Quality Control Peer-review Organization for the State of Michigan, sponsored by the Health Care Financing Administration, Department of Health and Human Services. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. The authors assume full responsibility for the accuracy and completeness of the ideas presented. This article is a direct result of the Health Care Quality Improvement Program initiated by the Health Care Financing Administration, which has encouraged identification of quality improvement projects derived from analysis of patterns of care and therefore required no special funding on the part of this contractor. Ideas and contributions to the authors concerning experience in engaging with the issues presented are welcomed. Corresponding Author: Address correspondence to Dr Ruoff, Michigan State University, College of Medicine, Department of Family Practice, 6565 West Main, Kalamazoo, MI 49009. 616-375-0400. Fax: 616-372-4205. REFERENCES 1. Rubin RJ, Altman WM, Mendelson DN. Health care expenditures for people with diabetes mellitus, 1992. J Clin Endocrinol Metab 1994;78: 809 A-F. 2. Lebovitz HE. Introduction: goals of treatment. In: Lebovitz HE, ed. Therapy for diabetes mellitus and related disorders, second edition. Alexandria, Va: American Diabetes Association, 1994. 3. Diabetes Control and Complication Trial. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993:329:977-86. 4. United Kingdom Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulfonylureas or insulin, compared with conventional treatment and risk of complications in patients with type 2 diabetes. Lancet 1998;352:839-55. 5. O Connor PJ, Rush WA, Peterson J, et al. Continuous quality improvement can improve glycemic control for HMO patients with diabetes. Arch Fam Med 1996;3:502-6. 6. Nathan DM. Monitoring diabetes mellitus. In: Lebovitz HE, ed. Therapy for diabetes mellitus and related disorders, second edition. Alexandria, Va: American Diabetes Association, 1994. 7. American Diabetes Association. Medical management of non-insulindependent (type 2) diabetes, third edition. Alexandria, Va: American Diabetes Association, 1994. This paper received the first-place award for original research by a family physician in practice at the 1998 American Academy of Family Physicians Annual Scientific Assembly in San Francisco.
Special Series: AAFP Award-winning Research Papers Vol. 31, No. 5 335 Appendix 1 Diabetic Care Flow Sheet (continued on next page)
336 May 1999 Family Medicine Appendix 1 (continued)