Patient Perceptions and Provider Documentation of Diabetes Care in Rural Areas
This study was funded by an Administration on Aging Grant, No. 90-AM-2304.
Diabetes is a chronic disease that requires continuous treatment, including adequate glycemic, blood pressure, and lipid control, to prevent complications. Studies have repeatedly demonstrated that adequate management reduces the risks of macrovascular and microvascular complications.1,2 These complications (eg, retinopathy, nephropathy, neuropathy, and cardiovascular disease) significantly increase cost of care and result in higher mortality and morbidity for people with diabetes.3,4
Recent advances in diabetes care and treatment of complications are frequently reported in the literature. Best practice guidelines have been developed.5,6 Consequently, healthcare providers need to stay up-to-date and apply current guidelines to their medical practice. People with diabetes also must assume responsibility for day-to-day control of the disease, making dietary and lifestyle changes, obtaining needed preventive services, being aware of the signs and symptoms of possible complications, and understanding the purpose and use of their prescribed medications.7 For rural persons with diabetes, this may be more complicated because endocrinologists, dietitians, diabetic educators, and other sources of medical information may be geographically distant.8
Measures of effective diabetes management have included either provider adherence to best practice guidelines or clinical indicators such as laboratory values.9 These offer insight into the clinical picture from the provider’s viewpoint but leave out the patient perspective. Adherence to care needs to be examined as a multivariate construct that looks globally at behavioral components, such as knowledge and beliefs held by the person with diabetes and provider components that focus on use of current treatment guidelines.5,6 To address this shortfall of information, a study was designed to 1) determine congruence between rural patient self-reported and provider-documented information on American Diabetes Association (ADA) recommended guidelines for measurement and control of HgA1c, blood pressure, lipid levels, and appropriate preventive services (see Figure 1,5,6 2) examine patient-identified barriers to care, including knowledge deficits that prevent effective disease management, and 3) describe health and lifestyle characteristics of rural persons with diabetes.
A descriptive study containing baseline data from a larger, ongoing intervention study designed to improve provider adherence with ADA guidelines and patient knowledge was conducted by Deaconess Billings Clinic Center on Aging researchers in rural Montana communities.10 Provider-documented data abstracted from outpatient medical records was paired with patient’s self-reported information obtained from a questionnaire.
Patient inclusion criteria. All persons 45 years of age or older, diagnosed as having type 2 diabetes and managed by a healthcare provider at one of four rural healthcare clinics between January 1, 1999 and August 1, 2000 were eligible to participate. Diagnosis was established through provider documentation of type 2 diabetes in the medical record.
Provider background. The study included healthcare providers practicing in four rural communities in Montana. Rural clinic sites were well established in their communities and had remained consistent at these sites for several years. Each rural site had one to eight primary care providers, including internal medicine and family practice physicians and physician assistants. No nurse practitioners were employed at the rural clinic sites.
Patients and providers were from rural communities that varied in population size from 364 to 8,487 residents and were, on average, 183 miles from the closest urban area that had a full range of medical services, including endocrinologists, diabetic educators and dietitians.
Patient questionnaire. The patient questionnaire included self-reported health information (physical and lifestyle health habits), diabetes knowledge, and barriers to care. Health information tools included in the questionnaire have established reliability and validity in a variety of research settings. These included the Lawton Independent Activities of Daily Living Scale,11 Nutrition Screening Initiative (NSI),12 SF-12, and the Quality of Life Rating Scale.13,14 Questions about diabetes knowledge and satisfaction with care were adapted from the Michigan Diabetes Research and Training Center Diabetes Care Profile.15 Two questions about barriers to care were developed for this study: “How much of a problem have you had receiving diabetes care?” and “How much of a problem have you had getting lab work or medical tests?” Answers included quantitative responses of a “big problem,” “small problem,” and “no problem,” as well as qualitative responses explaining the problems experienced.
Provider-documented information from the outpatient medical record. A data collection form developed for medical record review was based on standards of care established by the 1999/2000 ADA guidelines.5,6 The entire patient medical record from January 1, 1999 to August 1, 2000 was examined for evidence of documentation of testing and treatment. Variables obtained from medical records included demographic information, documentation of medications prescribed, chronic conditions, glycemic control (HbA1c or glycosolated hemoglobin), systolic and diastolic blood pressure, lipid levels, monofilament examinations, immunizations (influenza and pneumococcal), dilated eye examinations, smoking habits, alcohol use, and referral for diabetes education. Information collected in this study was recorded in the patients’ medical records according to the facility’s usual medical practices. Laboratory values were obtained from the patient’s medical record and all laboratory testing was performed at the facility routinely used by the rural facility. Glycosolated hemoglobin (gHb) values were measured in place of HbA1c for 70 patients. The laboratory performing the gHb tests ran control samples for both gHb and HbA1c. These values were used to establish a regression analysis to allow conversion of glycosolated hemoglobin values to HbA1c values. For this study, the conversion formula used was HbA1c = gHb (.671)+1.182. HbA1c data includes the converted gHb values.
Ethnic background, educational level, and socioeconomic status were not consistently recorded in the medical records nor were they collected from the patients. Consequently, they are not reported in the study results. According to county census information reports, 97% to 99% of the residents are Caucasian, 1% to 4% are Native American, and 1% are African American. Mean county incomes ranged from $22,923 to $33,260.16 Subject characteristics were consistent with these estimates.
Data analysis. Analysis of data was conducted using SPSS 11.5 statistical software. Patient and provider information were matched and then merged into one data set to allow for statistical comparison. Congruence data is based on an exact pairing of provider-documented information and patient self-reported information. Descriptive statistics were used for health and diabetes knowledge questions. Measures of central tendency, cross tabs, matched t-tests, ANOVA, and chi-squared analysis were used to examine differences between patient and provider information. All data are mean ±SD unless otherwise specified. Internal Review Board approval was obtained for all study protocols before data collection.
Sample. Questionnaires were sent to 243 rural community-dwelling persons with type 2 diabetes. The total response rate was 65% (n = 149) after two mailings. Reasons given for not responding included cognitive problems, time constraints, nursing home placement, and travel plans out of area. Participating patients were significantly younger (67 ± 10 versus 70 ± 12 years) and had a lower HgA1c (7.1 ± 1.4 versus 7.6 ± 1.8%) compared to non-respondents (P <0.05). No significant difference was found in gender distribution, number of chronic medical conditions, blood pressure, cholesterol, and LDL-C values between the groups. The remaining information is for participating patients only.
Provider-documented diabetes/comorbid disease management. Patients were predominately female (86, 58%), took an average of 5.5 ± 3.2 prescription medications, and had 3.6 ± 2.5 chronic conditions (see Table 1). On medical record review, patients had a mean systolic blood pressure (SBP) of 139 ± 18 mm Hg, diastolic blood pressure (DBP) of 75 ± 11 mm Hg, and LDL-C value of 106 ± 42 mg/dL. A diagnosis of hypertension was documented for 60% of the patients and dyslipidemia for 40%. An additional 13 patients were on lipid-lowering medications but did not have a diagnosis of dyslipidemia documented in their medical records.
Measurement and documentation of ADA recommended physiologic variables and preventive services were variable (Table 2). Providers documented HgA1c results for 57% and blood pressure values for 80% of the patients in the previous 6 months; lipid results were recorded for 60% of patients in the previous 12 months. For preventive services, 13 (9%) of patients had documented monofilament testing; 19 (13%), dilated eye examination; and 40 (27%), influenza and 49 (33%), pneumococcal vaccinations. Providers documented smoking habits for 33 (22%), alcohol use for 13 (9%), and referrals for diabetes or nutritional education for five (3%) of the patients.
Of patients achieving the ADA recommended physiologic goals, 51% of patients had a HgA1c value <7.0%; 40%, a SBP level <130 mm Hg; 81%, a DBP level <85 mm Hg; and 32%, a LDL-C value <100 mg/dL.
Patient self-reported information.
Diabetes/comorbid disease management. The majority of patients (68%) had diabetes for 3 years or longer. Type 2 diabetes were reported by 76% and 5% of patients, respectively; 24% did not know their type of diabetes. The majority of patients (54%) felt they did a good to excellent job of managing their diabetes. With regard to diabetes monitoring, 54% of patients reported having a HbA1c measured in the past 6 months; 96%, a blood pressure reading taken in the past 6 months; and 69%, a lipid level measured in the past 12 months. In the 34 patients reporting a HbA1c result, answers ranged from 5% to 150% (mean 22 ± 36). LDL-C values were reported by 15% of patients and ranged from 70 to 162 mg/dL (mean 110 ± 23).
Of the 159 patient participants, 105 (71%) reported having an influenza vaccination in the past year and 87 (58%) received a pneumococcal vaccination; 65 (44%) never smoked, 65 (44%) were prior smokers, and 10% currently smoked; 31 (21%) used alcoholic beverages — four (3%) consumed three or more alcoholic beverages per day.
Physical health habits. General health screening results were variable (see Table 1). A large majority (130, 90%) of the patients reported having a yearly physical examination and 55 (38%) had a yearly dental examination. For male patients, 51 (81%) reported having a prostate examination as part of their yearly physical and 18 (30%) performed testicular self-examination on a regular basis. For female patients, 54 (63%) reported having a yearly mammogram and 48 (56%) performed regular breast self-examination.
Lifestyle health habits. Study patients self-reported high functional ability on the Lawton Instrumental Activities of Daily Living Scale; 65 (13%) rated themselves totally independent in all activities. Few people had a planned exercise program; 109 (73%) stated they either did not exercise or exercised two times or less per week. Most people (75%) stated that they had 7 or 8 hours of sleep a night, had no difficulty staying asleep (53%), and were not experiencing any stress (69%).
Using the NSI scale, the majority of the patients scored at some risk for nutritional problems (see Table 1) — 87 (60%) scored at moderate risk and 39 (30%) scored at high risk for malnutrition. Individual items from the NSI showed that 99 (66%) of the patients took three or more medications per day; 70 (47%) ate few fruits, vegetables, or milk a day; 57 (38%) had lost 10 or more pounds in the past 6 months; and 43 (29%) ate alone. No gender differences were noted.
Quality of life. Patients expressed positive feelings about their health (see Table 1). Their overall quality of life was rated 7.3 ± 2.0 out of 10 on a Likert scale, where 1 = lowest/worst and 10 = highest/best. Although 131 (90%) said they had some health concerns in the past year, 77 (52%) rated their health about the same as others their age. The mean score on the physical health subscale of the SF-12 was 42 ± 10, and the mental health subscale mean was 58 ± 6. Physical or mental subscale scores were not significantly different for persons under versus over age 65. No gender differences were noted.
Congruence between patient and provider information.
Congruency in testing between patient and provider. A lack of congruence was noted between patient and provider information in multiple areas (see Table 2). Statistically significant differences (P <0.05) were noted between patient’s self-report and provider documentation of whether a blood pressure and lipid level had been measured, eye examination performed, influenza and pneumococcal vaccination given, and diabetic and nutrition education received. The percentage of patients with a HbA1c measurement was similar (57% versus 54%) between provider documentation and patient self-report. However, when patient and provider data were matched by individual, percent agreement as to whether a test was actually done was 47%. Congruency in testing ranged from 79% for blood pressure measurement to 60% for cholesterol check, 53% for pneumovax, 44% for influenza vaccination, and 42% for eye examination. Congruence between patient and provider for diabetes or nutritional education was 80% or above but the number of patients receiving education was too small for this to be meaningful.
Congruence of disease control between patient and provider. Patient perception of diabetes and blood pressure control did not relate to documented values. Patients reporting fair/poor diabetes control had HgA1c values (54, 6.9% ± 1.0%) comparable to those reporting good/excellent control (69, 7.3% ± 1.7%). In the 28 patients who recorded a HgA1c level, four (3% of the total patient sample) gave a value within ± 0.5 of the documented laboratory result. No significant difference was found in mean blood pressure values for patients rating their blood pressure as low (21, 136/76 ± 17/12 mm Hg) compared to patients who rated it as high (56, 137/77 ± 17/12 mm Hg). In the 22 patients reporting LDL-C values, four were within ± 10 mg/dL of documented levels.
A lack of agreement was noted regarding medication information. On medical chart review, 136 (93%) were on medication for diabetes management — according to self-report, 110 (81%) of the patients agreed; 53 (36%) of the patients were on a lipid-lowering medication — 17 (11%) patients agreed; 40 (45%) persons not on a documented lipid-lowering agent thought they were; 72 (81%) of patients with a diagnosis of hypertension were on a medication to lower their blood pressure — by self-report, 44 (61%) agreed.
A lack of congruence was also evident in preventive services (see Table 2). Patients reported a significantly higher (P <0.05) percentage of influenza vaccinations (71% to 27%) and pneumovax rates (58% to 33%) compared to provider information. Medical record review documented 13% of patients had a dilated eye exam in the past year but 82% of the patients reported having one and could name the examining ophthalmologist. Providers documented monofilament testing for 9% of the patients. No monofilament testing was documented for eight of nine patients who reported foot ulcers and for 50 of 57 who reported tingling and numbness in their hands or feet.
Patient-identified barriers to care.
Knowledge. Knowledge about symptoms of low blood sugar was inconsistent. Patients were asked two separate questions about symptoms of low blood sugar within the past 6 months. When asked if they had sweating, weakness, trembling, or an insulin reaction, 87 (67%) said they had experienced these symptoms. Patients were subsequently asked if they had a low blood sugar reaction with no symptoms listed. In the second question, 42 (28%) of the subjects responded affirmatively; 99 (66%) could not identify how they would treat symptoms of low blood sugar; and 102 (68%) did not carry anything with them to treat symptoms of low blood sugar.
Although all of the patients had a diagnosis of diabetes, 81 (54%) did not feel that they had an illness or condition that made them change the kind or amount of food they ate. A diabetic or combination diet (low salt, low fat and/or low sugar) was followed by 70% of the patients but 21% reported that they followed no special diet.
Receipt of diabetes services. Rural patients did not perceive barriers to receiving adequate medical care for their diabetes management — 99% and 97% reported no problems receiving diabetes care and no problems obtaining laboratory testing, respectively. Comments from patients who perceived barriers to care included “I couldn’t get in to see my provider” and “I had to drive 1 hour for care.”
Satisfaction with care was not a perceived barrier. On a 10-point scale where 1 = lowest and 10 = highest, patients rated their overall satisfaction with care at 8.6 ± 1.6 and satisfaction with the office staff as 8.8 ± 1.4. The majority of patients felt their provider usually or always listened to them and showed respect for what they had to say (85%), explained care in a way they could understand and spent enough time with them during office visits (88%), and felt providers worked with them to develop a plan of care for treatment of their diabetes (62%). However, 75% of the patients reported that they did not have a copy of their plan of care in writing. Few people had received reminders about their appointments by letter and/or postcard (40%) or by phone call (19%) from their providers.
Although the patients did not perceive barriers to medical care, most had not seen a diabetic educator (83%) or dietitian (88%). The majority of patients had obtained little information about their diabetes and associated risk factors from either their provider or other sources — 33 (22%) had received information on cholesterol management; 43 (29%) on foot care; 51 (34%) on nutrition; 55 (37%) on medication; 56 (38%) on exercise; and 71 (48%) on how to monitor their blood glucose.
This study underscores the fact that multiple barriers to comprehensive diabetes management exist in rural areas. Provider adherence to ADA clinical guidelines was suboptimal. Rural patients showed limited knowledge about their disease, its management, and its associated comorbid conditions. A lack of congruence in multiple areas of diabetes care was noted between provider information documented in outpatient medical records and patients’ self-reported information.
Tight glycemic, blood pressure, and lipid control; diet; and exercise are important components in the diabetic patient’s plan of care.6,17 Research has examined the relationship between glycemic control and diabetes knowledge, perceived quality of life, and barriers to control.17-20 However, these studies examined subjects of various age ranges, had different measurements of knowledge, and were conducted in predominately urban settings, making it difficult to compare results. Studies that examine diabetes care from both the provider and patient perspective are limited.
Rural healthcare providers’ adherence to ADA guidelines was not optimal in this study. This may be due to both lack of timely measurement and poor medical record documentation of recommended guidelines. Measurement and documentation of physiologic variables and preventive services ranged from 9% to 80% in these patients. These results are similar to what the authors and others have reported in the past.10,21 Providers and office staff are directly responsible for ordering HbA1c and lipid levels and performing blood pressure and monofilament testing. It is unlikely that patient refusal accounted for the 20% to 43% of patients who did not have a HbA1c or blood pressure documented in the last 6 months of chart review or lipid level in the last 12 months. These results indicate that physiologic variables and monofilament testing may not be measured/performed in accordance with ADA guidelines by rural healthcare providers.
In contrast, the low rates of immunizations and dilated eye examination in medical records may be due to poor documentation by healthcare providers when in fact these preventive services may have been obtained. A significant lack of agreement was found in rates of immunization and dilated eye examination rates between what was documented in the medical records and patient self-reported information. By self-report, 58% to 82% of patients received these services in accordance to ADA guidelines — significantly higher than the 13% to 33% indicated by medical record review. These differences may reflect the ability of patients to obtain these services outside of the usual office visit or the healthcare clinic. Limitations in the use of information from patient self-report should not account for the degree of disparity between the two sets of results.
In this study, rural patients displayed significant knowledge gaps about their disease management — eg, 19% of patients did not know the type of diabetes they had. A number of patients did not understand the purpose and use of their medications and inaccurately reported HbA1c and LDL-C results. Patients’ perceived levels of glycemic and blood pressure control was not confirmed by documented physiologic levels.
This study also demonstrates that health habits of rural patients with diabetes were problematic. With the exception of prostate examinations, testicular exam by men and preventive breast services in women were not done in accordance with national guidelines.22 Nutritional risk was a concern for the majority of patients (90%) — 12 of the subjects responded that they did not always have enough money for food, 29% ate alone, and 8% ate two or fewer meals per day. These findings are consistent with other rural research findings.21
Although disease control and patient knowledge was not optimal, rural patients did not identify physical barriers to obtaining diabetes care. Although some diabetic patients lived more than 180 miles from a healthcare provider, few patients reported difficulty obtaining diabetes care or laboratory testing. Satisfaction with the rural provider and office staff was high. From the rural diabetic patient’s perspective, a good provider-patient relationship seems to exist.
Type 2 diabetes is a complex disease to manage. Glycemic, blood pressure, and lipid control can be achieved and preventive services offered according to best practice guidelines to prevent adverse health consequences. To accomplish this, adherence to care should be examined as a multivariate construct that looks globally at behavioral components, such as knowledge, health habits of the person with diabetes, and provider components that focus on use of current treatment guidelines.5,6 Results of this study show that gaps in rural patients’ knowledge and healthcare providers’ adherence to ADA guidelines, as well as the disconnect or lack of communication between provider and patient, are potential barriers to effective disease management. The latter was demonstrated by the limited agreement between patient and provider on whether a physiologic variable or preventive service had or had not been done. With the exception of blood pressure measurement, congruence between the two ranged from 40% to 60%.
Patients taking an active role in their diabetes management may be one way to improve communication with their provider. To accomplish this, patients need to understand their disease and its treatment. The ADA recommends referral to diabetic educators and dietitians for patient education. However, few patients in this study reported receiving education. This may be due to a lack of diabetes services in rural Montana communities, where only 30% report having a diabetic educator available.21 Providers also must take an active role in educating patients in their disease control, medication use, and need for preventive services.
Diabetes requires a lifelong commitment to maintain control. To achieve this outcome, providers must implement clinical practice guidelines into patient care. Patients must take an active role in their disease management. All entities must communicate effectively. In this study, multiple barriers to successful diabetes management were identified. These included limited patient knowledge, suboptimal provider implementation of ADA recommended guidelines, and a lack of congruency between provider and patient. Changes in healthcare delivery that decrease barriers are needed in rural areas.
The authors would like to acknowledge the staff at Deaconess Billings Clinic Center on Aging, including Connie Koch, Karen Gransbery, Betty Mullette, and Brenda Hellyer.