| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorial |
Departments of
1
Psychiatry and
2
Pathology, Immunology & Laboratory Medicine,, Pediatrics, and Molecular Genetics & Microbiology, University of Florida Health Science Center, Gainesville, FL 32610
a Address correspondence to this author at: University of Florida Health Science Center, Department of Pathology, Immunology & Laboratory Medicine, Box 100275, Gainesville, FL 32610-6249. Fax 352-846-2149;
winter{at}pathology.ufl.edu.
A current definition of evidence-based medicine is given in the British Medical Journal from 1996: evidence-based medicine is "the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients" (1). Interpretation of laboratory data requires that an appropriate reference interval be defined by studying a healthy population. Subjects with laboratory results that consistently fall outside this interval may not be healthy. This is clearly the case in individuals with diabetes mellitus whose mean blood glucose concentrations in the untreated state are above the reference-population random, fasting, and/or post-glucose challenge blood glucose concentrations (2).
In the 1990s, in the Diabetes Control and Complications Trial (DCCT) (3) and United Kingdom Prospective Diabetes Study (UKPDS) (4), an important question was answered: glycemic control does make a difference in the development of microvascular complications. Healthcare professionals are now focusing their efforts on improving the glycemic control of their patients. Diabetes is likely the best example of a disease whose outcome is predominantly determined by the daily actions of the affected individual. The DCCT and UKPDS results demonstrated that improved glycemic control lowers the frequency of retinopathy, nephropathy, and neuropathy and also likely lowers the long-term risk of developing macrovascular disease. The challenge now is to implement improved diabetic therapy that lowers mean blood glucose into or near to the reference interval reflected in a hemoglobin A1c <7% (5).
To achieve improved glycemic control, patients with diabetes must measure their blood glucose several times per day and adjust their antidiabetic medications (e.g., insulin dose, formulation, and timing), diet, and exercise in accordance with their pattern and degree of hyper- (or hypo-) glycemia (6). Whereas other studies have examined the analytical performance of hand-held blood glucose self-monitoring devices, Skeie et al. (7) examine the question of how patients with type 1 diabetes use and interpret such results. From these data, Skeie et al. quantify instrument specifications (e.g., accuracy), based on the patients’ opinions, that are required to provide the desired clinical information. The focus on patient self-report data in this area is important because many professionals in the field have moved from describing their area of study as patient compliance, to patient adherence, to patient self-care. The change in terminology to patient self-care recognizes the responsibility that patients have in managing their diabetes as opposed to obeying the doctor’s orders.
Skie et al. (7) developed a brief self-report measure to examine the patients’ impressions of the need for accuracy in self-monitoring of blood glucose (SMBG). As an initial venture in this area, the measures developed by these investigators are in an early stage of psychometric development. They have done limited preliminary work concerning content validity or establishing the domain of interest, as they informally polled colleagues and patients in developing their patient self-report questionnaire. However, data on parameters of reliability or validity for the questionnaire are yet to be established. This said, the conclusions drawn in the study need to be viewed as tentative until the measure is subjected to further psychometric study.
Acknowledging the preliminary nature of the psychometric measures, as described above, the study has appropriate sample size, design, and statistical analyses to draw some interesting conclusions. The data presented can be summarized as follows: on average, adults with type 1 diabetes test 11 times per week [approximately two-thirds to one-half the frequency recommended by the American Diabetes Association (ADA)], develop hypoglycemic symptoms at or below 3 mmol/L (54 mg/dL), consider blood glucose of 15.5 mmol/L (279 mg/dL) or greater as "threatening", target daily blood glucose concentrations between 4.0 and 10.0 mmol/L (72–180 mg/dL), consider a significant increase in blood glucose to be 3.4 mmol/L (61 mg/dL) or greater, consider a significant decrease in blood glucose to be 3.1 mmol/L (56 mg/dL) or more, and consider a decrease in blood glucose of 3.6 mmol/L (65 mg/dL) to represent a significant reduction in blood glucose after treatment of hyperglycemia. Given that many in the field are recognizing the degree to which patients must assume responsibility for managing their diabetes, it is highly relevant that the typical patient wants "loose" control (e.g., 72–180 mg/dL) relative to ADA goals (5).
To the patient with diabetes, of possibly even greater significance than avoiding hyperglycemia is the avoidance of hypoglycemia. Central nervous system impairment from hypoglycemia can produce decreased consciousness, tiredness or drowsiness, faintness, confusion, blurred or double vision, hemiparesis, behavioral changes, dizziness, paresthesias, incoordination, slurred speech, hunger, and/or headache. In the worst case, hypoglycemia can produce coma or death. When blood glucose is at the highest concentration considered to represent hypoglycemia [e.g., 3.0 mmol/L (54 mg/dL)], decreases of 1.1 mmol/L (20 mg/dL) from this value would trigger action on the part of patients to increase their blood glucose. After patients act to increase their glucose, a significant increase in glucose would represent an increase of 2.6 mmol/L (47 mg/dL).
Skeie et al. (7) state that a small percentage of the patient sample studied had difficulty understanding their questions. Reports of "false" blood glucose values (e.g., values fabricated to "please" the physician) may be inferred from a regularly "reported" pattern of euglycemia in the face of repeated high hemoglobin A1c values. Such false reporting of blood glucose may indicate that patients do not understand that the results are used by their doctors to adjust their insulin dose, exercise, and diet to improve their health.
The majority of the Norwegian subjects studied by Skeie et al. desired
a highly accurate and reliable measure of blood glucose. To
achieve suitable accuracy and precision, meter performance needs to
exceed that proposed by the International Standardization Organization
goals of 95% of results within ± 20% of a comparative method
and within approximately ± 1.1 mmol/L (20 mg/dL) for blood
glucose less than 
5.5 mmol/L (100 mg/dL). Although actions to be
taken for a blood glucose of 11.1 mmol/L (200 mg/dL) vs 12.2 mmol/L
(220 mg/dL) may be similar, at lower blood glucose concentrations, such
as 3.05 mmol/L (55 mg/dL) vs 4.2 mmol/L (75 mg/dL), the need for
accuracy becomes more apparent. This study is also a valuable exercise
in determining how patients use the daily data that they generate. The
user’s perceptions of what is a significant change in blood glucose
and the derived analytical accuracy required to measure that change are
relevant to setting performance standards for hand-held blood
glucose meters. Therefore, the user’s opinions provide one type of
"evidence-based medicine" information.
Skeie et al. (7) studied 201 Norwegian adults with type 1 diabetes. It would be of interest to know whether these results can be generalized to patients with different sample characteristics, such as nationality, socioeconomic status, and age (e.g., adolescence vs younger adulthood vs older adulthood). It also would be of interest to explore the bases of patients’ preference for degree of metabolic control. The present study has not addressed many questions whose answers can assist healthcare professionals in the design and use of hand-held blood glucose meters in the future. Such studies should determine: "Does increased frequency of testing correlate with improved glycemic control?", "Does the use of hand-held devices with improved accuracy lead to improved control and less hypoglycemia?", "How does diabetes education influence the choice of patient cutpoints?", and "How do physicians’ opinions about meter performance compare to the patients’ opinions?". Other important questions include: "Do the findings of Skeie et al. apply to type 2 diabetes?" and "Do the findings of Skeie et al. apply to populations that do not have access to universal healthcare (e.g., the entire United States) who may lack access to diabetes experts and a diabetes specialty team?".
Skeie et al. (7) state that "Education in SMBG should focus on the limitations of portable BG [blood glucose] meters and errors involved in the interpretation of BG results". Although the authors of this editorial do not disagree with this focus, we believe that in the United States, the greater focus should be on educating patients to, indeed, test their blood glucose and then act on those results by altering their medications, diet, and exercise based on an algorithmic approach (8)(9).
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  W. E. Winter A Rosetta Stone for Insulin Treatment: Self-Monitoring of Blood Glucose Clin. Chem., June 1, 2004; 50(6): 985 - 987. [Full Text] [PDF]
|
|
|
|
|
|
S. Skeie, G. Thue, and S. Sandberg Interpretation of Hemoglobin A1c (HbA1c) Values among Diabetic Patients: Implications for Quality Specifications for HbA1c Clin. Chem., July 1, 2001; 47(7): 1212 - 1217. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
