jimmys devoted
11-13-2006, 08:27 PM
07 Nov 2006
When persons with diabetes use miscoded blood glucose meters to determine how much insulin to take, significant errors in insulin dose can result that may potentially lead to short- and long-term health complications, according to findings of a new study presented at the Sixth Annual Diabetes Technology Meeting in Atlanta, Georgia.
The American Diabetes Association estimates that there are 14.6 million children and adults diagnosed with diabetes in the United States(1), of which an estimated 4.4 million, or 30%, require insulin to manage their disease(2). Those who require insulin must closely monitor their blood sugar with a blood glucose meter to plan their meals, exercise regimens and insulin dosage.
In this study, for certain miscoded meters, the probability of insulin error of plus or minus 2 units of insulin was 50% as compared to 8% for correctly, manually coded meters. The probability of insulin dose error of plus or minus 3 units of insulin was 23% for the miscoded meters but only 0.5% for the manually correctly coded meters.
Coding is the process by which a blood glucose meter is matched to each new box of test strips being used. This is done either by inserting a code strip or code chip into the meter, or by entering a code number into the meter. If this step is not performed, the meter may give inaccurate results leading to wrong therapy. For example, relying on a miscoded blood glucose meter to determine how much insulin to take can result in a potentially harmful overdose. Insulin overdose may cause dangerously low blood sugar (hypoglycemia) leading to behavioral changes, confusion, loss of consciousness and, if untreated, seizure, coma and even death. Chronic under-dosing of insulin may contribute to the long-term health problems associated with high blood sugar including kidney disease, nerve disease, eye problems, and heart disease.
"When dealing with patients with diabetes we've observed that many either do not understand what proper coding is, or do not realize its importance. Patients sometimes use expired test strips and/or fail to properly code their blood glucose meters to the lot of test strips they are using," said Dr. Steven Edelman, an author on the study and Professor of Medicine, division of Endocrinology and Metabolism, University of California, San Diego, and the Veterans Affairs Medical Center and founder of TAKING CONTROL OF YOUR DIABETES (http://www.tcoyd.org).
The study findings also showed that auto-coded meters (meters that automatically set the correct code anytime a test strip is inserted) gave more accurate blood glucose values than meters that had been correctly coded manually. This also translated into a lower probability of insulin dose error. For auto-coded meters, the probability of plus or minus 1 unit and plus or minus 2 units of insulin could be as high as 35.4% and 1.4% respectively. However, with the auto-coded meters, there were no calculated insulin dose errors above plus or minus 2 units.
"These findings are significant because studies have shown that approximately 16% - or one out of six - persons failed to properly manually code their blood glucose meters to the lot of test strips being used(3). Understanding the potentially serious consequences of relying on a meter that is not properly coded - is essential for every person with diabetes, especially those who need to take insulin," said Linda Schrock, a nurse and certified diabetes educator, who was also an investigator on the study, at Elkhart General Hospital, Elkhart, Indiana.
The study authors concluded that to avoid insulin dosing errors, people should be carefully instructed how to correctly code their meters or be advised to use an auto-coded meter.
Study Design
The study involved 116 patients at three clinical centers. The blood glucose values for patients in this study ranged from 52 - 498 mg/dL. After fasting, the patients were given a two-hour meal tolerance test. At zero, 60 and 120 minutes the study subjects' fingerstick blood was tested on five different popular blood glucose meters (two were auto-coded meters). Some of the meters were purposely miscoded to the lot of test strips. The auto-coded meters were always properly coded due to their inherent design. The values from all the meters were compared with blood glucose values measured on a laboratory glucose analyzer to determine how accurate (inaccurate) the meters were.
Glucose values obtained from some of the miscoded meters used for this study showed an average error ranging between plus 29% and minus 37%.
A Monte Carlo simulation, (a statistical method that uses existing data sets to forecast performance in the field) was conducted on the data from the clinical trial to generate 'ideal' and 'simulated-meter' glucose values, and subsequent insulin doses. This simulation was based on various assumptions such as, one unit (1U) of insulin covers 50mg/dL blood glucose.* From these calculations, the probability of insulin dose errors for the three types of blood glucose meters (miscoded, manually correctly coded and autocoded) were determined.
The probability for an error of plus or minus one unit of insulin was 44.6% for correctly coded meters compared to 49.6% for incorrectly coded meters. The probability for an error with a miscoded meter of plus or minus four units of insulin was 2.8% and for plus or minus five units of insulin was 0.06%. There was no instance of a plus or minus four or five unit error with correctly, manually coded meters. For auto-coded meters there were no calculated insulin dose errors above plus or minus two units.
Bayer HealthCare, Diabetes Care
When persons with diabetes use miscoded blood glucose meters to determine how much insulin to take, significant errors in insulin dose can result that may potentially lead to short- and long-term health complications, according to findings of a new study presented at the Sixth Annual Diabetes Technology Meeting in Atlanta, Georgia.
The American Diabetes Association estimates that there are 14.6 million children and adults diagnosed with diabetes in the United States(1), of which an estimated 4.4 million, or 30%, require insulin to manage their disease(2). Those who require insulin must closely monitor their blood sugar with a blood glucose meter to plan their meals, exercise regimens and insulin dosage.
In this study, for certain miscoded meters, the probability of insulin error of plus or minus 2 units of insulin was 50% as compared to 8% for correctly, manually coded meters. The probability of insulin dose error of plus or minus 3 units of insulin was 23% for the miscoded meters but only 0.5% for the manually correctly coded meters.
Coding is the process by which a blood glucose meter is matched to each new box of test strips being used. This is done either by inserting a code strip or code chip into the meter, or by entering a code number into the meter. If this step is not performed, the meter may give inaccurate results leading to wrong therapy. For example, relying on a miscoded blood glucose meter to determine how much insulin to take can result in a potentially harmful overdose. Insulin overdose may cause dangerously low blood sugar (hypoglycemia) leading to behavioral changes, confusion, loss of consciousness and, if untreated, seizure, coma and even death. Chronic under-dosing of insulin may contribute to the long-term health problems associated with high blood sugar including kidney disease, nerve disease, eye problems, and heart disease.
"When dealing with patients with diabetes we've observed that many either do not understand what proper coding is, or do not realize its importance. Patients sometimes use expired test strips and/or fail to properly code their blood glucose meters to the lot of test strips they are using," said Dr. Steven Edelman, an author on the study and Professor of Medicine, division of Endocrinology and Metabolism, University of California, San Diego, and the Veterans Affairs Medical Center and founder of TAKING CONTROL OF YOUR DIABETES (http://www.tcoyd.org).
The study findings also showed that auto-coded meters (meters that automatically set the correct code anytime a test strip is inserted) gave more accurate blood glucose values than meters that had been correctly coded manually. This also translated into a lower probability of insulin dose error. For auto-coded meters, the probability of plus or minus 1 unit and plus or minus 2 units of insulin could be as high as 35.4% and 1.4% respectively. However, with the auto-coded meters, there were no calculated insulin dose errors above plus or minus 2 units.
"These findings are significant because studies have shown that approximately 16% - or one out of six - persons failed to properly manually code their blood glucose meters to the lot of test strips being used(3). Understanding the potentially serious consequences of relying on a meter that is not properly coded - is essential for every person with diabetes, especially those who need to take insulin," said Linda Schrock, a nurse and certified diabetes educator, who was also an investigator on the study, at Elkhart General Hospital, Elkhart, Indiana.
The study authors concluded that to avoid insulin dosing errors, people should be carefully instructed how to correctly code their meters or be advised to use an auto-coded meter.
Study Design
The study involved 116 patients at three clinical centers. The blood glucose values for patients in this study ranged from 52 - 498 mg/dL. After fasting, the patients were given a two-hour meal tolerance test. At zero, 60 and 120 minutes the study subjects' fingerstick blood was tested on five different popular blood glucose meters (two were auto-coded meters). Some of the meters were purposely miscoded to the lot of test strips. The auto-coded meters were always properly coded due to their inherent design. The values from all the meters were compared with blood glucose values measured on a laboratory glucose analyzer to determine how accurate (inaccurate) the meters were.
Glucose values obtained from some of the miscoded meters used for this study showed an average error ranging between plus 29% and minus 37%.
A Monte Carlo simulation, (a statistical method that uses existing data sets to forecast performance in the field) was conducted on the data from the clinical trial to generate 'ideal' and 'simulated-meter' glucose values, and subsequent insulin doses. This simulation was based on various assumptions such as, one unit (1U) of insulin covers 50mg/dL blood glucose.* From these calculations, the probability of insulin dose errors for the three types of blood glucose meters (miscoded, manually correctly coded and autocoded) were determined.
The probability for an error of plus or minus one unit of insulin was 44.6% for correctly coded meters compared to 49.6% for incorrectly coded meters. The probability for an error with a miscoded meter of plus or minus four units of insulin was 2.8% and for plus or minus five units of insulin was 0.06%. There was no instance of a plus or minus four or five unit error with correctly, manually coded meters. For auto-coded meters there were no calculated insulin dose errors above plus or minus two units.
Bayer HealthCare, Diabetes Care