Insulin Orders and Algorithm Improve Glycemic Control

SAN DIEGO — Insulin-use patterns and glycemic control improved, while episodes of hypoglycemia declined, among non-critical care patients with diabetes after structured insulin orders and an insulin management algorithm were added to the hospital order set at the University of California, San Diego Medical Center.

After the structured insulin order set was added, the relative risk of an inpatient stay decreased to 0.84 when the mean blood glucose level was greater than 180 mg/dL. With the order set plus the algorithm in place, the relative risks of an uncontrolled patient day and an uncontrolled patient stay declined to 0.77 and to 0.73, respectively. Furthermore, the relative risks of hypoglycemia per patient day and patient stay declined to 0.80 and 0.92, respectively, and the percentage of patient days with hypoglycemia decreased from 3.7% to 2.6%.

The findings are based on data from more than 5,500 subjects who had at least seven glucose measures during stays at the UCSD center. Data were compared for the year before the institution of the order set, for 18 months after introduction of the order set alone, and for the subsequent 8 months after the algorithm was added to the order set.

The success of the order set and the one-page algorithm—which recommended treatment pathways for adult medical and surgical inpatients who had a diagnosis of diabetes or documented hyperglycemia—were reported at the annual meeting of the Society of Hospital Medicine by Dr. Greg Maynard, chief of the division of hospital medicine at UCSD. The pathways were tailored to whether inpatients were receiving regular meals, were under NPO (nothing by mouth) orders, or were tube fed.

The outcomes were a “win-win situation,” he said. The “chaotic swings” in glycemia seen before the order set was instituted gave way to better control, and no special team was needed to get the results. He attributed most of the good results to a shift to basal/bolus insulin regimens from sliding-scale insulin regimens. After the order set and algorithm were instituted, sliding-scale-only insulin regimens decreased from 72% of 477 insulin regimens to 26% of 499 insulin regimens.

The next step for inpatient diabetes research, according to Dr. Maynard, is to develop reaction and prevention protocols. Suboptimal response to episodes of hyper- and hypoglycemia is the norm in hospitals, he said. “We don't take action when it should be taken, and opportunities to prevent hypoglycemia are missed.”

As an example, Dr. Maynard discussed the results of a study published last year. Within 48 hours after they were given an antihyperglycemic agent, 10% of 2,174 hospitalized patients with diabetes had hypoglycemia. Of those 206 patients, 44% had more than one hypoglycemic event. No adverse events accompanied 464 of the hypoglycemic episodes; there were 20 adverse events and 10 of these resulted in seizures or loss of consciousness. None of the adverse events was attributable to medication errors, and just 11% of the patients were on oral-only regimens.

About half of the hypoglycemic events were associated with reductions in enteral intake, but the precipitating factors were unclear in the other half. Just one-third of patients had their blood glucose rechecked within 60 minutes and fewer than half had documented euglycemia within 2 hours of their hypoglycemia; the average time to a documented resolution was 4 hours (J. Hosp. Med. 2007;2:234-40).

Dr. Maynard also advised examining patients' outpatient diabetes regimens, and imparted some recommendations for transitioning patients at hospital discharge.

The patient who was taking metformin and was admitted with a hemoglobin A_1c level greater than 9% and a baseline glucose measure of more than 350 mg/dL clearly needs her regimen adjusted, he said. But any recommendations need to consider her physical limitations; any new comorbidities; her willingness and ability to self-monitor; treatment goals; hypoglycemia risk factors; and the patient's financial situation. “You're not going to shoot for [a blood glucose level of] 110 mg/dL in a hospice patient, or encourage self-monitoring four times per day in a patient who can't afford it or is unwilling to do the tests,” he said. Those considerations weigh into treatment selection.

Dr. Maynard offered the following general observations to consider when you select a discharge therapy.

▸ Once HbA_1c exceeds 8.5%, additional oral agents are unlikely to achieve goals.

▸ Insulin at bedtime is a good initial strategy.

▸ Testing drives the cost of therapy. Testing four times per day is more expensive than insulin.

▸ Elderly patients are at higher risk for hypoglycemia. Decrease their insulin as they get better, and make sure they have good follow-up and aren't being overtreated.

▸ Glyburide has been linked with a higher risk of hypoglycemia than has glipizide.

▸ The evidence on risk factors is imperfect, but be hesitant to start glitazones de novo.

As a treatment reference, Dr. Maynard suggested algorithms available at the Web site of the American Academy of Clinical Endocrinologists (www.aace.com/resources/igcrcwww.hospitalmedicine.org/ResourceRoomRedesign/GlycemicControl.cfm

Dr. Maynard reported that he and his coinvestigators had no disclosures related to the study.

SAN DIEGO — Insulin-use patterns and glycemic control improved, while episodes of hypoglycemia declined, among non-critical care patients with diabetes after structured insulin orders and an insulin management algorithm were added to the hospital order set at the University of California, San Diego Medical Center.

After the structured insulin order set was added, the relative risk of an inpatient stay decreased to 0.84 when the mean blood glucose level was greater than 180 mg/dL. With the order set plus the algorithm in place, the relative risks of an uncontrolled patient day and an uncontrolled patient stay declined to 0.77 and to 0.73, respectively. Furthermore, the relative risks of hypoglycemia per patient day and patient stay declined to 0.80 and 0.92, respectively, and the percentage of patient days with hypoglycemia decreased from 3.7% to 2.6%.

The findings are based on data from more than 5,500 subjects who had at least seven glucose measures during stays at the UCSD center. Data were compared for the year before the institution of the order set, for 18 months after introduction of the order set alone, and for the subsequent 8 months after the algorithm was added to the order set.

The success of the order set and the one-page algorithm—which recommended treatment pathways for adult medical and surgical inpatients who had a diagnosis of diabetes or documented hyperglycemia—were reported at the annual meeting of the Society of Hospital Medicine by Dr. Greg Maynard, chief of the division of hospital medicine at UCSD. The pathways were tailored to whether inpatients were receiving regular meals, were under NPO (nothing by mouth) orders, or were tube fed.

The outcomes were a “win-win situation,” he said. The “chaotic swings” in glycemia seen before the order set was instituted gave way to better control, and no special team was needed to get the results. He attributed most of the good results to a shift to basal/bolus insulin regimens from sliding-scale insulin regimens. After the order set and algorithm were instituted, sliding-scale-only insulin regimens decreased from 72% of 477 insulin regimens to 26% of 499 insulin regimens.

The next step for inpatient diabetes research, according to Dr. Maynard, is to develop reaction and prevention protocols. Suboptimal response to episodes of hyper- and hypoglycemia is the norm in hospitals, he said. “We don't take action when it should be taken, and opportunities to prevent hypoglycemia are missed.”

As an example, Dr. Maynard discussed the results of a study published last year. Within 48 hours after they were given an antihyperglycemic agent, 10% of 2,174 hospitalized patients with diabetes had hypoglycemia. Of those 206 patients, 44% had more than one hypoglycemic event. No adverse events accompanied 464 of the hypoglycemic episodes; there were 20 adverse events and 10 of these resulted in seizures or loss of consciousness. None of the adverse events was attributable to medication errors, and just 11% of the patients were on oral-only regimens.

About half of the hypoglycemic events were associated with reductions in enteral intake, but the precipitating factors were unclear in the other half. Just one-third of patients had their blood glucose rechecked within 60 minutes and fewer than half had documented euglycemia within 2 hours of their hypoglycemia; the average time to a documented resolution was 4 hours (J. Hosp. Med. 2007;2:234-40).

Dr. Maynard also advised examining patients' outpatient diabetes regimens, and imparted some recommendations for transitioning patients at hospital discharge.

The patient who was taking metformin and was admitted with a hemoglobin A_1c level greater than 9% and a baseline glucose measure of more than 350 mg/dL clearly needs her regimen adjusted, he said. But any recommendations need to consider her physical limitations; any new comorbidities; her willingness and ability to self-monitor; treatment goals; hypoglycemia risk factors; and the patient's financial situation. “You're not going to shoot for [a blood glucose level of] 110 mg/dL in a hospice patient, or encourage self-monitoring four times per day in a patient who can't afford it or is unwilling to do the tests,” he said. Those considerations weigh into treatment selection.

Dr. Maynard offered the following general observations to consider when you select a discharge therapy.

▸ Once HbA_1c exceeds 8.5%, additional oral agents are unlikely to achieve goals.

▸ Insulin at bedtime is a good initial strategy.

▸ Testing drives the cost of therapy. Testing four times per day is more expensive than insulin.

▸ Elderly patients are at higher risk for hypoglycemia. Decrease their insulin as they get better, and make sure they have good follow-up and aren't being overtreated.

▸ Glyburide has been linked with a higher risk of hypoglycemia than has glipizide.

▸ The evidence on risk factors is imperfect, but be hesitant to start glitazones de novo.

As a treatment reference, Dr. Maynard suggested algorithms available at the Web site of the American Academy of Clinical Endocrinologists (www.aace.com/resources/igcrcwww.hospitalmedicine.org/ResourceRoomRedesign/GlycemicControl.cfm

Dr. Maynard reported that he and his coinvestigators had no disclosures related to the study.

SAN DIEGO — Insulin-use patterns and glycemic control improved, while episodes of hypoglycemia declined, among non-critical care patients with diabetes after structured insulin orders and an insulin management algorithm were added to the hospital order set at the University of California, San Diego Medical Center.

After the structured insulin order set was added, the relative risk of an inpatient stay decreased to 0.84 when the mean blood glucose level was greater than 180 mg/dL. With the order set plus the algorithm in place, the relative risks of an uncontrolled patient day and an uncontrolled patient stay declined to 0.77 and to 0.73, respectively. Furthermore, the relative risks of hypoglycemia per patient day and patient stay declined to 0.80 and 0.92, respectively, and the percentage of patient days with hypoglycemia decreased from 3.7% to 2.6%.

The findings are based on data from more than 5,500 subjects who had at least seven glucose measures during stays at the UCSD center. Data were compared for the year before the institution of the order set, for 18 months after introduction of the order set alone, and for the subsequent 8 months after the algorithm was added to the order set.

The success of the order set and the one-page algorithm—which recommended treatment pathways for adult medical and surgical inpatients who had a diagnosis of diabetes or documented hyperglycemia—were reported at the annual meeting of the Society of Hospital Medicine by Dr. Greg Maynard, chief of the division of hospital medicine at UCSD. The pathways were tailored to whether inpatients were receiving regular meals, were under NPO (nothing by mouth) orders, or were tube fed.

The outcomes were a “win-win situation,” he said. The “chaotic swings” in glycemia seen before the order set was instituted gave way to better control, and no special team was needed to get the results. He attributed most of the good results to a shift to basal/bolus insulin regimens from sliding-scale insulin regimens. After the order set and algorithm were instituted, sliding-scale-only insulin regimens decreased from 72% of 477 insulin regimens to 26% of 499 insulin regimens.

The next step for inpatient diabetes research, according to Dr. Maynard, is to develop reaction and prevention protocols. Suboptimal response to episodes of hyper- and hypoglycemia is the norm in hospitals, he said. “We don't take action when it should be taken, and opportunities to prevent hypoglycemia are missed.”

As an example, Dr. Maynard discussed the results of a study published last year. Within 48 hours after they were given an antihyperglycemic agent, 10% of 2,174 hospitalized patients with diabetes had hypoglycemia. Of those 206 patients, 44% had more than one hypoglycemic event. No adverse events accompanied 464 of the hypoglycemic episodes; there were 20 adverse events and 10 of these resulted in seizures or loss of consciousness. None of the adverse events was attributable to medication errors, and just 11% of the patients were on oral-only regimens.

About half of the hypoglycemic events were associated with reductions in enteral intake, but the precipitating factors were unclear in the other half. Just one-third of patients had their blood glucose rechecked within 60 minutes and fewer than half had documented euglycemia within 2 hours of their hypoglycemia; the average time to a documented resolution was 4 hours (J. Hosp. Med. 2007;2:234-40).

Dr. Maynard also advised examining patients' outpatient diabetes regimens, and imparted some recommendations for transitioning patients at hospital discharge.

The patient who was taking metformin and was admitted with a hemoglobin A_1c level greater than 9% and a baseline glucose measure of more than 350 mg/dL clearly needs her regimen adjusted, he said. But any recommendations need to consider her physical limitations; any new comorbidities; her willingness and ability to self-monitor; treatment goals; hypoglycemia risk factors; and the patient's financial situation. “You're not going to shoot for [a blood glucose level of] 110 mg/dL in a hospice patient, or encourage self-monitoring four times per day in a patient who can't afford it or is unwilling to do the tests,” he said. Those considerations weigh into treatment selection.

Dr. Maynard offered the following general observations to consider when you select a discharge therapy.

▸ Once HbA_1c exceeds 8.5%, additional oral agents are unlikely to achieve goals.

▸ Insulin at bedtime is a good initial strategy.

▸ Testing drives the cost of therapy. Testing four times per day is more expensive than insulin.

▸ Elderly patients are at higher risk for hypoglycemia. Decrease their insulin as they get better, and make sure they have good follow-up and aren't being overtreated.

▸ Glyburide has been linked with a higher risk of hypoglycemia than has glipizide.

▸ The evidence on risk factors is imperfect, but be hesitant to start glitazones de novo.

As a treatment reference, Dr. Maynard suggested algorithms available at the Web site of the American Academy of Clinical Endocrinologists (www.aace.com/resources/igcrcwww.hospitalmedicine.org/ResourceRoomRedesign/GlycemicControl.cfm

Dr. Maynard reported that he and his coinvestigators had no disclosures related to the study.