William R Carter

Evaluating Clinical Accuracy of Systems for Self-Monitoring of Blood Glucose

Although the scientific literature contains numerous reports of the statistical accuracy of systems for self-monitoring of blood glucose (SMBG), most of these studies determine accuracy in ways that may not be clinically useful. We have developed an error grid analysis (EGA), which describes the clinical accuracy of SMBG systems over the entire range of blood glucose values, taking into account 1) the absolute value of the system-generated glucose value, 2) the absolute value of the reference blood glucose value, 3) the relative difference between these two values, and 4) the clinical significance of this difference. The EGA of accuracy of five different reflectance meters (Eyetone, Dextrometer, Glucometer I, Glucometer II, Memory Glucometer II), a visually interpretable glucose reagent strip (Glucostix), and filter-paper spot glucose determinations is presented. In addition, reanalyses of a laboratory comparison of three reflectance meters (Accucheck II, Glucometer II, Glucoscan 9000) and of two previously published studies comparing the accuracy of five different reflectance meters with EGA is described. EGA provides the practitioner and the researcher with a clinically meaningful method for evaluating the accuracy of blood glucose values generated with various monitoring systems and for analyzing the clinical implications of previously published data.

Accuracy of Perceiving Blood Glucose in IDDM

Type I (insulin-dependent) diabetic individuals and health professionals often assume that the symptoms of extremely low or high blood glucose (BG) levels can be recognized and, consequently, appropriate treatment decisions can be based on symptom perception. Because no research has documented the validity of these assumptions, this study tested the ability to perceive BG concentration. Nineteen type I adults, experienced in self-monitoring of BG (SMBG), estimated their BG 40–54 times just before measurement of actual BG. This procedure was repeated under two conditions: (1) in the hospital (hospital condition) while connected to an insulin/glucose infusion system that artificially manipulated BG, leaving subjects only symptomatic, or internal, cues and (2) in the natural environment (home condition), where both internal and external cues, e.g., food and insulin consumption, were available. Estimates significantly correlated with actual BG for 7 of 16 subjects in the hospital condition and for 18 of 19 subjects in the home condition. Believed ability to estimate BG did not predict documented ability in either condition. An evaluation of the treatment significance of estimation errors showed that the majority of errors were relatively benign. The most common error affecting clinical outcome was estimated euglycemia when actual BG was hypoglycemic or hyperglycemic.

Validation of Scale Measuring Environmental Barriers to Diabetes-Regimen Adherence

This study reports on the validation of a diabetesspecific measure of environmental barriers to regimen adherence. The reliability and validity of the environmental barriers to adherence scale (EBAS) were determined for a sample of 214 insulin-dependent and non-insulin-dependent diabetic patients. The scale was shown to be a valid measure of barriers to adherence as assessed by its relationship to the barriers to adherence questionnaire and the barriers to adherence portion of the diabetes-care profile. The medication, testing, exercise, and diet subscales of the EBAS were correlated with four corresponding and three noncorresponding measures of self-care behavior from the diabetes selfcare behaviors scale. Each subscale correlated well with its corresponding self-care behavior and less well with noncorresponding self-care behavior. The internal consistency of the scale and the test-retest reliability were found to be good. The results suggest that the EBAS scale is a valid, reliable measure of barriers to diabetes-regimen adherence.

Effects and Correlates of Blood Glucose Awareness Training Among Patients With IDDM

Whereas self-monitoring of blood glucose (SMBG) is the recommended source of information on which to make self-care decisions, patients frequently use estimates of their own blood glucose (BG). This study evaluated whether patients with insulin-dependent diabetes mellitus (IDDM) could learn to improve accuracy of BG estimations and whether this would lead to improved metabolic control. Subjects in BG awareness training improved both their BG-estimation accuracy and glycosylated hemoglobin (HbA1) compared with the control group. Initial BG-estimation accuracy was marginally associated with pretreatment HbA1 and months of previous SMBG experience. Posttreatment improvement was associated with pretreatment BGestimation accuracy and the ability to counterregulate to insulin-induced hypoglycemia.

Blood glucose discrimination training in insulin-dependent diabetes mellitus (IDDM) patients

Self-management of insulin-dependent diabetes mellitus (IDDM) is dependent on a negative feedback loop of blood glucose (BG) fluctuations, which in turn directs treatment decisions to maintain normal BG. Although this feedback is typically accomplished by self-monitoring of blood glucose (SMBG), SMBG has limitations, and patients often rely on what their BG “feels” like. Two studies were performed to evaluate whether patients could learn to more accurately “feel”/discriminate their BG on the basis of internal cues or internal plus external BG cues. In Study I, BG Awareness Training significantly improved pre- to posttreatment BG estimation accuracy, relative to a control group. Study II replicated BG Awareness Training efficacy in improving BG estimation accuracy. Improvement in estimation accuracy was related only to initial accuracy; those who were initially less accurate improved the most. This improvement was represented in a 31% reduction in dangerous BG estimation errors and a 9% increase in accurate estimates. Resulting estimations were, however, still significantly less accurate than SMBG at the end of training.

Clarification of error-grid analysis.

Koschinsky et al. (1) reviewed the shortcomings of efforts to quantify accuracy of self-monitoring blood glucose (SMBG) devices. We concur with their conclusions that 7) linear regression, percent deviation, and correlational analyses are inappropriate; 2) many referencetest blood glucose readings across the entire range of hypo-, eu-, and hyperglycemia are required; 3) no clear distinctions exist between demarcations of acceptable and unacceptable blood glucose values within a clinically relevant range; and 4) evaluation schemes should be flexible enough to adjust to the needs of specific groups (2-7). Furthermore, as Koschinsky et al. point out with the error-grid analysis and as we point out herein with their acceptance analysis, any criteria will lead to decisions that may be unacceptable to some. Because of our common perspectives, it was surprising to read how these authors misrepresented the errorgrid analysis. They repeatedly state that the error grid considers B-zone (benign) errors as acceptable (Fig. 1). They cite several examples of reference-test B-zone blood glucose values (i.e., 150-40, 240-180, and 400-190 mg/dl). In fact, these examples are unacceptable by the error grid, the American Diabetes Association (ADA) (8), and others (9). The error-grid analysis only considers Azone values as acceptable. All examples given fall in the B zones. Although they are unacceptable, these examples do not put the patient at immediate clinical risk, as with C, D, and E zones. To summarize the error grid, A-zone values are acceptable, and any values outside the A zones are unacceptable. Any values that fall in zones C-E represent potential immediate clinical risk: C zones (correction errors) create a risk of overtreating euglycemia leading to potential hypoor hyperglycemia; D zones (detection errors) represent failure to detect immediate and significant hypoor hyperglycemia; E zones (erroneous treatment errors) may encourage treating hypoglycemia during actual hyperglycemia or vice versa. Unlike any other system, including that of Koschinsky et al. (1), the error-grid analysis differentiates and quantifies the type and significance of unacceptable and acceptable blood glucose tests. In contrast, Koschinsky et al.s acceptance analysis attempts to reduce clinical accuracy into a simplistic dichotomous variable: acceptable/unacceptable. This fails to recognize the clinical significance of different types of unacceptable tests. The analysis considers tests acceptable if they overestimate reference blood glucose as much as 57% in the hypoglycemic range and 85% in the hyperglycemic range, whereas test overestimates are considered good if differing from reference ^20% across the entire blood glucose range, similar to the error-grid analysis and others (9). Consequently their tolerance of acceptable tests is extremely liberal, their criteria for good has been reported repeatedly in the literature, and their scheme reflects basic percent deviation, a method they decry. Additionally, their scheme does not consider the clinical significance of the type of test errors made. They consider an instrument acceptable if 1 time in 20 it reads a reference test of 300-60 mg/dl. Such a 1:20 chance could lead to the lethal decision to inappropriately increase a patients insulin dose. Inversely, their method would consider a reference-test reading of 20-40 mg/dl unacceptable, when this unacceptable test result would lead to an acceptable clinical decision to elevate blood glucose. Koschinsky et al. (1) make a significant statement echoing our call for a standardized system for determining the accuracy of SMBC technology (2-7). We feel that

Metabolic and Cutaneous Events Associated With Hypoglycemia Detected by Sleep Sentry

Eighteen insulin-dependent diabetic subjects [age (mean ± SD) 33.2 ± 10.6 yr] participated in a study designed to determine the metabolic and cutaneous parameters associated with activation of the nocturnal hypoglycemia monitor Sleep Sentry. Plasma glucose, glucagon, epinephrine, norepinephrine, and pancreatic polypeptide concentrations were determined every 10 min during a 2-h constant intravenous insulin infusion (40 mU kg −1 h−1). In addition, skin temperature and electrical conductance were monitored at the same time intervals, and subjects were asked to rate the degree to which they felt cold and/or sweaty. Ten of the subjects (alarmers) activated the device with a mean plasma glucose nadir of 52.8 ± 13.8 mg/dl, whereas eight (nonalarmers) failed to do so despite a mean plasma glucose nadir of 50.5 ± 8.2 mg/dl. There were no significant differences between alarmers and nonalarmers with respect to initial or nadir plasma glucose levels, rate of fall of plasma glucose, or changes in plasma epinephrine, norepinephrine, or pancreatic polypeptide concentrations. In addition, changes in skin temperature and conductance were similar in both groups as were descriptive variables including age, disease duration, gender, and level of glucose control. No subject reported an increase in coldness, whereas 80% of both groups reported an increase in sweatiness. Three subjects studied on more than one occasion over a year failed to exhibit consistent activation of the alarm. This study suggests that it may not be possible to identify patients for whom the Sleep Sentry would be a reliable addition to their self-management regimen and that physicians should exercise caution in recommending its use.

Evidence-Based Treatment and ST-Segment Elevation Myocardial Infarction

1. Najjar SS, Rao SV, Melloni C, et al; REVEAL Investigators. Intravenous erythropoietin in patients with ST-segment elevation myocardial infarction: REVEAL: a randomized controlled trial. JAMA. 2011;305(18):1863-1872. 2. Wu E, Ortiz JT, Tejedor P, et al. Infarct size by contrast enhanced cardiac magnetic resonance is a stronger predictor of outcomes than left ventricular ejection fraction or end-systolic volume index: prospective cohort study. Heart. 2008; 94(6):730-736. 3. de Waha S, Fuernau G, Eitel I, et al. Measuring treatment effects in clinical trials using cardiac MRI. Curr Cardiovasc Imaging Rep. 2011;4(2):98-107. 4. Eitel I, Desch S, Fuernau G, et al. Prognostic significance and determinants of myocardial salvage assessed by cardiovascular magnetic resonance in acute reperfused myocardial infarction. J Am Coll Cardiol. 2010;55(22):24702479.

Behavioral Medicine Update, Vol. 6 (1), 1984, pp 12-16 Use, accuracy, adherence, and im pact

Behavioral Medicine Update, Vol. 6 (1), 1984, pp 17-21 Dietary adherence in patients with diabetes By: Rena R. Wing, Leonard Epstein, and Mary P. Norwalk Poor adherence to dietary management has been a problem for those with either Type I or Type II diabetes. A number of behavioral techniques have been helpful for Type I diabetes, this article reports. Type II diabetics seem to have a much less positive response to any kind of behavioral interventions. This problem is discussed and potential solutions suggested.