Adherence to American Diabetes Association Standards of Care by Rural Health Care Providers

This is a descriptive study conducted at multiple rural health clinics in Montana. It was designed to determine rural provider’s baseline compliance with 1999/2000 ADA clinical practice guidelines. Screening for and management of the comorbid conditions of hypertension or dyslipidemia were also included.

Providers practicing in four rural communities in Montana were included in this study. These communities varied in population from 364 to 8,487 and were on average 182.5 miles from the closest urban area that had a full range of medical services including a certified diabetic educator (15). The number of primary health care providers at each rural site ranged from one to eight. Rural clinic sites were well established in their community, and providers had remained consistent at these sites for several years. In three sites all providers were represented. In the fourth site 8 of 17 community providers were represented. Approximately 6 of these 17 providers included specialties other than family practice or internal medicine, such as woman’s health and pediatrics, that were not appropriate for study inclusion. There were no facilities that had computerized records. One facility had a diabetic registry, but it was not in use by all providers. Providers consisted of internal medicine and family practice physicians and physician assistants. No nurse practitioners were currently employed at the rural clinic sites.

A retrospective chart review was conducted on all patients ≥45 years of age diagnosed with diabetes who were seen by providers between 1 January 1999 and 1 August 2000. Medical records from patients with ICD-9 codes between 250.0 and 250.9 were pulled for possible chart review. Charts were reviewed if the rural patient had a definitive diagnosis of diabetes and the rural medical provider was the primary source of diabetes care for the patient. Patient records were excluded if there was steroid, drug, or illness-induced diabetes.

A data collection form was developed for the chart reviews based on standards of care established by the 1999/2000 ADA guidelines (Tables 1 and 2 (8a,8b). The entire medical record was examined for evidence of documentation of testing and treatment. Variables from medical records included patient demographics, chronic conditions, measures of glycemic control (HbA_1c or GHb), microalbumin-to-creatinine ratio, systolic blood pressure (SBP), diastolic blood pressure (DBP), and lipid levels and documentation of or referral for monofilament examinations, immunizations (influenza and pneumococcal), dilated eye examinations, diabetes education, smoking habits, and alcohol use. Diagnosis of and treatment for comorbid conditions of hypertension and dyslipidemia were recorded. Ethnic background, educational level, and socioeconomic status were not consistently recorded in the medical records and consequently not reported in the study results. County census information reports that 97–99% of the residents are Caucasian, 1–4% are Native American, and 1% are black. Mean county incomes ranged from $22,923 to $33,260. It is expected that subject characteristics are consistent with these estimates. There was no patient contact or testing performed as part of this study. Laboratory testing was performed at the laboratory routinely used by the rural health care provider. All information collected was found in medical records at the four rural facilities. Information had been recorded in the records according to the facility’s usual practices. Institutional review board approval was obtained for all study protocols before data collection. The study’s medical and research directors trained all data collectors. Data collectors were registered nurses with several years of diabetic education experience. Inter-rater reliability was established before data collection began. Decisions about patient eligibility due to the possibility of drug- or illness-induced diabetes was made by the medical director, a board-certified internist.

Data were analyzed using SPSS 11.0 statistical software. Data are presented as means ± 1 SD unless otherwise specified. Demographic results were descriptive for age, sex, length of time since diagnosis of diabetes, medication use, and presence of other diagnosis. Frequencies, cross-tabs, and measures of central tendency were used to compare medical record information with ADA standards. Student’s t tests were used to examine differences between sex for treatment and comorbid conditions. There was no statistical difference in performance rates or mean values between sites. GHb values were measured in place of HbA_1c for 70 patients. The laboratory that performed the GHb tests ran control samples for both GHb and HbA_1c. We used these values to establish a regression analysis to allow conversion of GHb values to HbA_1c values. For this study, the conversion formula used was HbA_1c = GHb (0.671) + 1.182. HbA_1c data includes the converted GHb values.

Medical records of 399 diabetic patients met the inclusion criteria and were included in the final analysis. Subjects ranged in age from 45 to 96 years. The mean age was 69 ± 11.5 years, with 64% of the sample being 65 years of age or older. Diabetes had been diagnosed for 3 or more years for 74% (n = 259) of patients. A diagnosis of hypertension was recorded for 59% (n = 236) patients, and 77% (n = 182) of these patients were on antihypertensive medication. Dyslipidemia was recorded in 31% (n = 125) of the subjects with 26% (n = 104) taking a lipid-lowering medication. A total of 85 patients had a diagnosis of both dyslipidemia and hypertension. The majority of the sample was female (59%), took five or more different prescription medications per day (58%), and had three or more diagnosed chronic conditions (63%). The majority of subjects had a current weight recorded, but <30% had heights available. Consequently, BMI could not be calculated.

Documentation of ADA-recommended testing was variable (Fig. 1). HbA_1c was measured in 85% (n = 340) of the patients. Testing was completed within the past year for 81% (n = 322) of the patients and within the last 6 months for 60% (n = 241) of the patients. A total of 58 patients had not had a HbA_1c value recorded for the past 18 months, and one patient had an HbA_1c ordered, but no value was recorded in the medical record. Of 399 patients, there were 21 patients without a recorded blood pressure, and 61% had lipid panels results. LDL values were reported for 228 patients (57%). An additional 16 patients had lipid panels recorded, but their triglyceride levels were too high to calculate an LDL value. Microalbumin-to-creatinine ratio was reported for 15% (n = 61) of the patients.

Few dilated eye examinations, monofilamant tests, influenza and pneumoococcal vaccinations, or smoking and alcohol patterns were documented in the chart. Influenza vaccines and pneumococcal vaccines were recorded for 27 (n = 108) and 30% (n = 121) of the patients, respectively. There were eight people (of 399) that had both a documented dilated eye examination and a monofilament test. Two of these eight patients also had a current flu and pneumococcal vaccine. None had the recommended diagnostic tests (HbA_1c, blood pressure, and LDL cholesterol) plus a documented monofilament test, dilated eye examination, and flu vaccine in the past year (Fig. 2). Any type of information about smoking habits was recorded for 20% and alcohol use for 10% of the sample.

There were few referrals to a nurse educator or dietitian documented in the medical records by the rural health care providers. Five patients (1.3%) were referred to a diabetes educator and 16 (4%) to a dietitian. None of the rural communities included in this study had a certified diabetes or health educator available.

Mean HbA_1c was 7.43 ± 1.7%, and 47% (n = 160) of patients with HbA_1c values met the ADA-recommended goal of <7%. In addition, 30% (n = 102) needed additional treatment for their diabetes based on HbA_1c levels >8%, and 25% of these subjects had not had an HbA_1c test in the past 6 months. Therapeutic management included use of insulin by 22% (n = 89), oral hypoglycemic agents by 77% (n = 306), and combined therapy (insulin and oral hypoglycemic agents) by 11.5% (n = 46) of the patients. Twenty-five subjects were controlled by diet alone. No difference in the pattern of medication management was detected between the controlled and uncontrolled groups.

Optimal blood pressure and lipid control had not been achieved in these rural diabetic patients. Mean SBP and DBP were 139 ± 18.8 and 75 ± 11.5 mmHg, respectively. Based on the 1999/2000 ADA blood pressure goal of 130/85, 28.6% (n = 108) of the patients had controlled SBP, 79.3% (n = 300) had controlled DBP, and 27% (n = 103) had both. Of the 226 patients with a diagnosis of hypertension, 51 were treated to the ADA goal of 130/85 and an additional 40 patients were <140/90. Ninety-eight patients with a diagnosis of hypertension, whose systolic blood pressure was >130 or diastolic blood pressure was ≥85, were not taking an antihypertensive medication. Seven percent (n = 11) of the patients without a diagnosis of hypertension had a blood pressure ≥140/90.

The mean total cholesterol level was 206 ± 45 mg/dl, mean LDL was 119 ± 33 mg/dl, and mean HDL was 45 ± 12 g/dl in the 228 patients with lipid profiles. Twenty-six percent (n = 60) had the ADA recommended LDL values of <100 mg/dl. Forty percent (n = 91) had LDL values ≥130 g/dl, which is the suggested level for treatment of all diabetic individuals, regardless of vascular disease history. Of these 91 patients who needed additional treatment, there were 67 patients not on lipid-lowering medications. Twenty-five percent of the patients (n = 125) with a diagnosis of dyslipidemia had LDL levels <100 mg/dl.

Optimal disease management includes adequate glycemic, blood pressure, and lipid control. Eleven patients of the total sample of 399 met the recommended ADA treatment values of HbA_1c level <7%, a blood pressure <130/85 mmHg, and an LDL level <100 mg/dl (Fig. 2).

Several sex differences were noted in this study. Women were slightly older (70 ± 11 vs. 67 ± 11 years, P < 0.01), had lower HbA_1c values (7.2 ± 1.6 vs. 7.7 ± 1.8%, P< 0.02), and higher HDL levels (49 ± 12 vs. 39 ± 9 mg/dl, P < 0.001). No sex difference in LDL level was found. Woman tended to trend toward more diagnoses of hypertension, dyslipidemia, congestive heart failure, other heart problems, and depression. Only hypertension (66 vs. 49%, P < 0.001) and depression (13 vs. 7%, P < 0.045) were significantly different. Despite a higher percentage of diagnoses of hypertension in woman, no significant sex difference in SBP or DBP was found.

Study results indicate that rural health care providers are achieving better glycemic control in their patients with diabetes than reported national and state averages (13,14,16,24–26). Previous research found mean HbA_1c values between 8.6 and 8.9% and annual HbA_1c testing rates ranging from 16.3 to 90% (13–16,26). In this study, the mean HbA_1c value was 7.43%, less than the ADA recommendation for additional treatment needed. However, >50% of the patients had HbA_1c levels above the 1999/2000 ADA-recommended HbA_1c level of 7%. Using the recent American College of Endocrinology target for diabetes control of HbA_1c <6.5%, an even higher percentage would need more intensive management (19).

Rural Montana providers were ordering HbA_1c levels for their patients with diabetes comparable with those of university-based endocrinologists (26). Eighty-five percent of diabetic patients had a recorded HbA_1c in the past 18 months. In contrast, results from the 1993–1995 Montana Medicare claims data show that HbA_1c claims were only filed for 35% of diabetic seniors in Montana (20,21). This suggests that rural Montana providers are using HbA_1c to monitor glycemic control and have improved compliance with guidelines. Frequency of testing, especially for patients with suboptimal control, needs improvement. Evidence suggests that intensive blood glucose control in type 2 diabetic subjects reduces microvascular complications (4). Thus, the efforts of multiple national health organizations in addition to the ADA to encourage primary health care providers to optimize glycemic control in all diabetic patients may help explain improved testing and glycemic control by these rural providers.

Management of the comorbid conditions for hypertension and dyslipidemia was suboptimal but better than that reported in other studies (26). Rural providers were better at documenting blood pressure than lipid levels. According to ADA recommendations, patients with diabetes should have a blood pressure measurement at the initial and each subsequent follow-up examination. Only 5% (n = 20) of these diabetic patients did not have a blood pressure recorded. However, 10 of these diabetic patients with a diagnosis of hypertension did not have a blood pressure recorded. Research has shown that lowering blood pressure to a mean of 144/82 mmHg in hypertensive diabetic patients significantly reduced complications of macrovascular and microvascular disease and diabetes-related deaths (5,7,8a,8b). Epidemiological data have shown a continuous relationship between these outcomes and SBP >130 mmHg. In addition, 1999/2000 ADA guidelines recommend lowering blood pressure in diabetic patients to <130/85 mmHg. The 2001 ADA recommendations have a more stringent requirement of 130/80 mmHg. Although the majority of patients in the study had blood pressure measured, only 27% of the total sample met criteria for control. In the patients with a diagnosis of hypertension, only 22% were treated to the ADA goal.

Dyslipidemia is prevalent in diabetic patients and increases risk for atherosclerotic vascular disease (5,8a,8b). ADA guidelines recommend a yearly cholesterol profile unless the patient is at low risk. In this study, 39% of the patients did not have documented lipid profiles in the 18-month period studied, a value which is down from 51% reported in 1995 by Stolar (26). There was no indication in the medical records that previous testing placed these patients at low risk, thus eliminating the need for yearly lipid screening. Dyslipidemia increases risk for artherosclerotic vascular disease in diabetic patients, and effective management has been shown to reduce secondary vascular events (8a,8b,13). Management of elevated cholesterol and triglycerides in diabetic patients has become increasingly more stringent. ADA guidelines recommend a LDL level ≤100 mg/dl. In this study, only 60 of 399 patients had documented LDL levels ≤100 mg/dl. Furthermore, only 25% of the 132 patents with a diagnosis of dyslipidemia were at ADA goals. More importantly, 74% (n = 67) of patients whose LDL was ≥130 mg/dl were not on lipid-lowering medication. Strategies to improve rural provider compliance with documenting screening and management of dyslipidemia and hypertension in diabetic patients are needed.

Few rural patients had referrals recorded for diabetes education or dietary counseling. This is problematic because research has shown that a multidisciplinary approach to comprehensive disease management improves provider and patient compliance (8a,8b). Ninety-eight percent had no diabetes education and 96% no referral to a registered dietitian. This was anticipated due to the rural location of the study sites. Small community hospitals and clinics have limited resources beyond the health care providers, so interdisciplinary teams are not possible. Certified diabetic educators are located only in Montana’s urban areas, thus the distance becomes a barrier for patients to receive effective diabetic education.

Rural provider’s adherence to ADA guidelines was better than previous research studies have indicated. However, significant improvement is still needed. To achieve better compliance, it is important to understand the barriers faced by the rural providers. Barriers to successful management by the rural providers is not known but could include lack of knowledge of current recommendations, ageism, provider inertia, patient compliance issues including medication cost and side effects, and lack of available educators (24). Further research is needed to determine barriers to successful diabetes management by rural health care providers so that alternative strategies can be implemented and tested.

A limitation of this study is that all data were collected from medical record review of diabetic patients primarily managed by rural providers. Information collected at the office visits, such as laboratory test results, blood pressure, monofilament examination, and presence of comorbid conditions should be complete and reliable. However, the rate of vaccination and dilated exe examinations may be inadequately documented, since these are performed outside of the rural office practice. Information or education that was given by either the provider or patient informally cannot be accounted for in a medical record review. While information that is not documented does not exist by the legal definition, we found that small rural communities often have informal networks in place. In some cases this means everyone at the health care facility knows multiple generations of the family and knows if they had an immunization or eye examination by another provider or facility. This may indicate that rates of dilated eye examinations and vaccinations are much higher than indicated by the study results. This is supported by the 1993–1995 Medicare claims data, which showed higher rates for eye examinations and immunizations (20).

Results of this study indicate that rural health care practitioners are not adequately following the 1999/2000 ADA standards for comprehensive disease management of their patients with diabetes. Glycemic testing is being ordered, but HbA_1c values indicate that patients do not have their diabetes under optimal control. The comorbid conditions of hypertension and dyslipidemia are not managed according to the ADA guidelines. Other preventative services including monofilament testing, dilated eye examinations, microalbumin-to-creatinine ratio, and immunizations are not being documented. Findings suggest a need for interventions, such as a decision support system with prompted reminders to improve provider compliance with ADA guidelines. In addition, improved patient access to diabetic educators, either through telemedicine or certification of rural office nurses, could benefit rural patients and practitioners.

This study was funded by Administration on Aging Grant no. 90-AM-2304.

The authors thank the Deaconess Billings Clinic Center on Aging, including Connie Koch, Brenda Hellyer, and Karen Granbery