Many Hospitals Fail to Achieve Routine Inpatient Glycemic Control

BOSTON – Hyperglycemia was found in nearly one-third of patients at some time during their hospital stay, and another 6% of inpatients met criteria for hypoglycemia, an analysis of glucose data from the largest number of U.S. hospitals to date has shown.

In addition, the prevalence of hyperglycemia varied significantly according to hospital size, type, and region. These findings were detected in both intensive care unit and non-ICU patients, Dr. Christine Swanson said at the annual meeting of the Endocrine Society.

"This information is important for hospital administrators as well as providers of care, both hospitalists and endocrinologists, who need to know what blood sugars to target, why, and how their hospitals’ glucose control compares to other hospitals nationwide. Higher glucose levels can complicate hospital course, cause longer hospital stay, more complications, and slower recovery from surgeries or infections. Hypoglycemia can also be related to greater morbidity and mortality," Dr. Swanson said in an interview.

For the analysis, she used an automated laboratory data management system (the Remote Automated Laboratory System–Plus) to extract results of point-of-care bedside glucose (POC-BG) tests during January to December 2009. The identity of the hospitals was kept confidential, and no patient identifying information or characteristics were made available.

From the 575 participating hospitals, more than 49 million measurements were analyzed for nearly 3.5 million patients. The study was a joint venture by the Mayo Clinic in Arizona and the Epsilon Group, a research organization.

Hospital data were obtained from across the United States. Hospitals varied by the number of hospital beds, by type (academic vs. rural community vs. urban community), and by region (Northeast, Midwest, South, and West).

The analysis revealed that about one-third of both ICU and non-ICU patient-days were characterized by hyperglycemia (more than 180 mg/dL), thus exceeding the glucose threshold currently recommended. In addition 6% of patient-days in each group met criteria for hypoglycemia (less than 70 mg/dL).

For both ICU and non-ICU patients, there was a significant relationship between patient-day weighted mean POC-BG levels and hospital characteristics. Larger (more than 400 beds), academic, and Western hospitals were able to achieve the lowest glycemic levels. "This analysis confirms the previous results ... [showing] that hyperglycemia is common in U.S. hospitals" (J. Hosp. Med. 2009;4:E7-14), said Dr. Swanson, an internist at the Mayo Clinic in Scottsdale, Ariz.

Among ICU patients, low blood sugar rates were lowest among hospitals in the Northeast and were significantly higher among larger and academic hospitals.

"Inpatient glycemic control will be important to track because it can be used as a marker to see how hospitals are doing with their inpatient diabetes quality improvement programs," said Dr. Swanson. No hospital received feedback about its performance in this study, since the source of the data was kept blinded.

Dr. Swanson said she hopes that increased hospital participation in data warehousing may support a national benchmarking process for the development of best practices to manage inpatient hyperglycemia.

In an interview, Dr. Thomas W. Barrett said that tight glycemic control in the ICU has become the poster child of quality improvement gone wrong. The clinical question was logical and ambitious: Would tight glycemic control using targets of normoglycemia reduce complications and mortality in the sickest patients in the hospital?

But we know where the path paved with good intentions leads, said Dr. Barrett of the Portland (Ore.) VA Medical Center and the division of hospital medicine at the Oregon Health and Science University, Portland.

After initial small single institutions demonstrated decreased mortality in ICU patients with insulin infusions using targets of normoglycemia in 2001, many centers began doing the same. However, 8 years later, the appropriately powered randomized controlled trial clearly demonstrated that efforts at normo-glycemia produced increased hypoglycemia, which was associated with increased mortality, compared with a looser target of less than 180 mg/dL.

These data come from the critical care world, not the everyday floor-patient world, he said. However, these data remain the focus as there have been very few published studies on clinical outcomes with glycemic control in floor patients. Consensus opinion recommendations do exist for floor patients from the American Diabetes Association and the American College of Endocrinology that endorse the goal of 140-180 mg/dL in hospitalized patients. Clinical inertia–the lack of knowledge or lack of effort at reaching these glucometrics–has been well documented. But the one thing that overcomes inertia is data. We are in need of similar quality research on the topic of glycemic control in floor patients that is correlated to hard clinical outcomes.

To succeed at glycemic control at the whole-hospital level, a multidisciplinary team of dedicated clinical nurse specialists, floor nurses, diabetes educators, information technology staff, pharmacists, and physician champions from hospital medicine and endocrinology is needed. This kind of work is time consuming, moves at its own pace, and generally involves all aspects of the hospital. For example, the nutrition staff that brings the patients’ food tray are crucial members of the glycemic team because the patient’s nurse needs to know when to give the patient insulin to cover that meal. Designing studies in this complex environment will require collaborations between implementation and health services researchers, Dr. Barrett said.

Hospitalist physicians are best positioned to lead these multidisciplinary glycemic groups using hospital level data, such as the aggregate RALS [Remote Automated Laboratory System] glucose results, to minimize hypoglycemia and meet consensus glucometric targets. However, tying glucometrics to patient outcomes instead of consensus opinions is the answer that our patients deserve, he said.

No relevant financial disclosures were reported by Dr. Swanson or Dr. Barrett.

BOSTON – Hyperglycemia was found in nearly one-third of patients at some time during their hospital stay, and another 6% of inpatients met criteria for hypoglycemia, an analysis of glucose data from the largest number of U.S. hospitals to date has shown.

In addition, the prevalence of hyperglycemia varied significantly according to hospital size, type, and region. These findings were detected in both intensive care unit and non-ICU patients, Dr. Christine Swanson said at the annual meeting of the Endocrine Society.

"This information is important for hospital administrators as well as providers of care, both hospitalists and endocrinologists, who need to know what blood sugars to target, why, and how their hospitals’ glucose control compares to other hospitals nationwide. Higher glucose levels can complicate hospital course, cause longer hospital stay, more complications, and slower recovery from surgeries or infections. Hypoglycemia can also be related to greater morbidity and mortality," Dr. Swanson said in an interview.

For the analysis, she used an automated laboratory data management system (the Remote Automated Laboratory System–Plus) to extract results of point-of-care bedside glucose (POC-BG) tests during January to December 2009. The identity of the hospitals was kept confidential, and no patient identifying information or characteristics were made available.

From the 575 participating hospitals, more than 49 million measurements were analyzed for nearly 3.5 million patients. The study was a joint venture by the Mayo Clinic in Arizona and the Epsilon Group, a research organization.

Hospital data were obtained from across the United States. Hospitals varied by the number of hospital beds, by type (academic vs. rural community vs. urban community), and by region (Northeast, Midwest, South, and West).

The analysis revealed that about one-third of both ICU and non-ICU patient-days were characterized by hyperglycemia (more than 180 mg/dL), thus exceeding the glucose threshold currently recommended. In addition 6% of patient-days in each group met criteria for hypoglycemia (less than 70 mg/dL).

For both ICU and non-ICU patients, there was a significant relationship between patient-day weighted mean POC-BG levels and hospital characteristics. Larger (more than 400 beds), academic, and Western hospitals were able to achieve the lowest glycemic levels. "This analysis confirms the previous results ... [showing] that hyperglycemia is common in U.S. hospitals" (J. Hosp. Med. 2009;4:E7-14), said Dr. Swanson, an internist at the Mayo Clinic in Scottsdale, Ariz.

Among ICU patients, low blood sugar rates were lowest among hospitals in the Northeast and were significantly higher among larger and academic hospitals.

"Inpatient glycemic control will be important to track because it can be used as a marker to see how hospitals are doing with their inpatient diabetes quality improvement programs," said Dr. Swanson. No hospital received feedback about its performance in this study, since the source of the data was kept blinded.

Dr. Swanson said she hopes that increased hospital participation in data warehousing may support a national benchmarking process for the development of best practices to manage inpatient hyperglycemia.

In an interview, Dr. Thomas W. Barrett said that tight glycemic control in the ICU has become the poster child of quality improvement gone wrong. The clinical question was logical and ambitious: Would tight glycemic control using targets of normoglycemia reduce complications and mortality in the sickest patients in the hospital?

But we know where the path paved with good intentions leads, said Dr. Barrett of the Portland (Ore.) VA Medical Center and the division of hospital medicine at the Oregon Health and Science University, Portland.

After initial small single institutions demonstrated decreased mortality in ICU patients with insulin infusions using targets of normoglycemia in 2001, many centers began doing the same. However, 8 years later, the appropriately powered randomized controlled trial clearly demonstrated that efforts at normo-glycemia produced increased hypoglycemia, which was associated with increased mortality, compared with a looser target of less than 180 mg/dL.

These data come from the critical care world, not the everyday floor-patient world, he said. However, these data remain the focus as there have been very few published studies on clinical outcomes with glycemic control in floor patients. Consensus opinion recommendations do exist for floor patients from the American Diabetes Association and the American College of Endocrinology that endorse the goal of 140-180 mg/dL in hospitalized patients. Clinical inertia–the lack of knowledge or lack of effort at reaching these glucometrics–has been well documented. But the one thing that overcomes inertia is data. We are in need of similar quality research on the topic of glycemic control in floor patients that is correlated to hard clinical outcomes.

To succeed at glycemic control at the whole-hospital level, a multidisciplinary team of dedicated clinical nurse specialists, floor nurses, diabetes educators, information technology staff, pharmacists, and physician champions from hospital medicine and endocrinology is needed. This kind of work is time consuming, moves at its own pace, and generally involves all aspects of the hospital. For example, the nutrition staff that brings the patients’ food tray are crucial members of the glycemic team because the patient’s nurse needs to know when to give the patient insulin to cover that meal. Designing studies in this complex environment will require collaborations between implementation and health services researchers, Dr. Barrett said.

Hospitalist physicians are best positioned to lead these multidisciplinary glycemic groups using hospital level data, such as the aggregate RALS [Remote Automated Laboratory System] glucose results, to minimize hypoglycemia and meet consensus glucometric targets. However, tying glucometrics to patient outcomes instead of consensus opinions is the answer that our patients deserve, he said.

No relevant financial disclosures were reported by Dr. Swanson or Dr. Barrett.

BOSTON – Hyperglycemia was found in nearly one-third of patients at some time during their hospital stay, and another 6% of inpatients met criteria for hypoglycemia, an analysis of glucose data from the largest number of U.S. hospitals to date has shown.

In addition, the prevalence of hyperglycemia varied significantly according to hospital size, type, and region. These findings were detected in both intensive care unit and non-ICU patients, Dr. Christine Swanson said at the annual meeting of the Endocrine Society.

"This information is important for hospital administrators as well as providers of care, both hospitalists and endocrinologists, who need to know what blood sugars to target, why, and how their hospitals’ glucose control compares to other hospitals nationwide. Higher glucose levels can complicate hospital course, cause longer hospital stay, more complications, and slower recovery from surgeries or infections. Hypoglycemia can also be related to greater morbidity and mortality," Dr. Swanson said in an interview.

For the analysis, she used an automated laboratory data management system (the Remote Automated Laboratory System–Plus) to extract results of point-of-care bedside glucose (POC-BG) tests during January to December 2009. The identity of the hospitals was kept confidential, and no patient identifying information or characteristics were made available.

From the 575 participating hospitals, more than 49 million measurements were analyzed for nearly 3.5 million patients. The study was a joint venture by the Mayo Clinic in Arizona and the Epsilon Group, a research organization.

Hospital data were obtained from across the United States. Hospitals varied by the number of hospital beds, by type (academic vs. rural community vs. urban community), and by region (Northeast, Midwest, South, and West).

The analysis revealed that about one-third of both ICU and non-ICU patient-days were characterized by hyperglycemia (more than 180 mg/dL), thus exceeding the glucose threshold currently recommended. In addition 6% of patient-days in each group met criteria for hypoglycemia (less than 70 mg/dL).

For both ICU and non-ICU patients, there was a significant relationship between patient-day weighted mean POC-BG levels and hospital characteristics. Larger (more than 400 beds), academic, and Western hospitals were able to achieve the lowest glycemic levels. "This analysis confirms the previous results ... [showing] that hyperglycemia is common in U.S. hospitals" (J. Hosp. Med. 2009;4:E7-14), said Dr. Swanson, an internist at the Mayo Clinic in Scottsdale, Ariz.

Among ICU patients, low blood sugar rates were lowest among hospitals in the Northeast and were significantly higher among larger and academic hospitals.

"Inpatient glycemic control will be important to track because it can be used as a marker to see how hospitals are doing with their inpatient diabetes quality improvement programs," said Dr. Swanson. No hospital received feedback about its performance in this study, since the source of the data was kept blinded.

Dr. Swanson said she hopes that increased hospital participation in data warehousing may support a national benchmarking process for the development of best practices to manage inpatient hyperglycemia.

In an interview, Dr. Thomas W. Barrett said that tight glycemic control in the ICU has become the poster child of quality improvement gone wrong. The clinical question was logical and ambitious: Would tight glycemic control using targets of normoglycemia reduce complications and mortality in the sickest patients in the hospital?

But we know where the path paved with good intentions leads, said Dr. Barrett of the Portland (Ore.) VA Medical Center and the division of hospital medicine at the Oregon Health and Science University, Portland.

After initial small single institutions demonstrated decreased mortality in ICU patients with insulin infusions using targets of normoglycemia in 2001, many centers began doing the same. However, 8 years later, the appropriately powered randomized controlled trial clearly demonstrated that efforts at normo-glycemia produced increased hypoglycemia, which was associated with increased mortality, compared with a looser target of less than 180 mg/dL.

These data come from the critical care world, not the everyday floor-patient world, he said. However, these data remain the focus as there have been very few published studies on clinical outcomes with glycemic control in floor patients. Consensus opinion recommendations do exist for floor patients from the American Diabetes Association and the American College of Endocrinology that endorse the goal of 140-180 mg/dL in hospitalized patients. Clinical inertia–the lack of knowledge or lack of effort at reaching these glucometrics–has been well documented. But the one thing that overcomes inertia is data. We are in need of similar quality research on the topic of glycemic control in floor patients that is correlated to hard clinical outcomes.

To succeed at glycemic control at the whole-hospital level, a multidisciplinary team of dedicated clinical nurse specialists, floor nurses, diabetes educators, information technology staff, pharmacists, and physician champions from hospital medicine and endocrinology is needed. This kind of work is time consuming, moves at its own pace, and generally involves all aspects of the hospital. For example, the nutrition staff that brings the patients’ food tray are crucial members of the glycemic team because the patient’s nurse needs to know when to give the patient insulin to cover that meal. Designing studies in this complex environment will require collaborations between implementation and health services researchers, Dr. Barrett said.

Hospitalist physicians are best positioned to lead these multidisciplinary glycemic groups using hospital level data, such as the aggregate RALS [Remote Automated Laboratory System] glucose results, to minimize hypoglycemia and meet consensus glucometric targets. However, tying glucometrics to patient outcomes instead of consensus opinions is the answer that our patients deserve, he said.

No relevant financial disclosures were reported by Dr. Swanson or Dr. Barrett.