Susan Birk

CHICAGO — When testing patients for celiac disease, physicians can no longer rely on a single paradigm for both overtly symptomatic patients and asymptomatic but genetically at-risk patients, according to Dr. Edwin Liu.

These two categories of patients require different approaches, said Dr. Liu, who spoke at a meeting on celiac disease sponsored by the American Gastroenterological Association.

Most symptomatic patients need only one antibody test, transglutaminase IgA (IgA-TGA) and an IgA antibody level to assess for celiac disease. But genetically at-risk patients may need multiple tests over time to screen for the presence of celiac autoimmunity and to determine if a biopsy is needed. Patients considered at risk for celiac disease include first-degree relatives of those with celiac disease or type 1 diabetes, and patients with type 1 diabetes.

The patient with classic symptoms and an abnormal TGA result usually can be biopsied immediately with a greater than 90% likelihood that intestinal lesions will be found, but TGA predicts disease in only about 75% of asymptomatic patients at genetic risk.

Patients with very elevated blood TGA levels are more likely to have more severe intestinal injury, so “in screening those at genetic risk, we have to understand our own lab tests well,” Dr. Liu said in an interview. Therefore, in deciding when to perform a biopsy, physicians should interpret tests in a quantitative fashion. This interpretation should consider changes in TGA values over time, because a single positive result may not provide enough information to make a decision to proceed with biopsy.

“In the case of a symptomatic person, [a single positive result] is probably okay, because you are looking for the presence or absence of disease. However, in the case of a person who's at risk for celiac disease, multiple tests over time may be needed due to the potential for disease. In addition, “we really need to understand what is a very high level,” he said, “because higher TGA levels are more likely to lead to findings of intestinal lesions.”

Complicating this diagnostic picture is the wide variability of currently available IgA-TGA assays, said Dr. Liu of the Barbara Davis Center for Childhood Diabetes and the section of gastroenterology, hepatology, and nutrition at the Children's Hospital and the University of Colorado at Denver. The definition of what constitutes a high TGA value differs depending on the laboratory and the assay used.

Asymptomatic individuals may need to be tested several times before deciding whether to proceed with biopsy. This is because a biopsy done too soon could produce normal histologic findings that suggest the absence of disease, but these normal findings do not necessarily rule out the possibility that disease will develop, Dr. Liu said.

He cited an example of the patient with type 1 diabetes who has an abnormal TGA and whose small intestinal biopsy is normal. The finding is not necessarily a “false positive” TGA level, but could be caused instead by the underlying biology of celiac disease. “If we biopsy patients too early they may not have had time to develop intestinal lesions,” he said. “If we believe that the paradigm for most autoimmunity also applies to celiac disease—that autoantibodies precede the development of actual disease—then performing intestinal biopsy in the early stages of autoimmunity might lead to findings of normal histology.”

Although some clinicians prefer to perform a biopsy at the first sign of abnormality on TGA because they do not want to miss a case of disease, Dr. Liu said the approach to diagnosis at his institution differs somewhat. “We don't want to biopsy more than once,” he said, adding that this strategy is viable because “the kids we follow are in the medical system, so we can afford to be patient and try to determine the optimal time for biopsy” with serial serologic tests.

CHICAGO — When testing patients for celiac disease, physicians can no longer rely on a single paradigm for both overtly symptomatic patients and asymptomatic but genetically at-risk patients, according to Dr. Edwin Liu.

These two categories of patients require different approaches, said Dr. Liu, who spoke at a meeting on celiac disease sponsored by the American Gastroenterological Association.

Most symptomatic patients need only one antibody test, transglutaminase IgA (IgA-TGA) and an IgA antibody level to assess for celiac disease. But genetically at-risk patients may need multiple tests over time to screen for the presence of celiac autoimmunity and to determine if a biopsy is needed. Patients considered at risk for celiac disease include first-degree relatives of those with celiac disease or type 1 diabetes, and patients with type 1 diabetes.

The patient with classic symptoms and an abnormal TGA result usually can be biopsied immediately with a greater than 90% likelihood that intestinal lesions will be found, but TGA predicts disease in only about 75% of asymptomatic patients at genetic risk.

Patients with very elevated blood TGA levels are more likely to have more severe intestinal injury, so “in screening those at genetic risk, we have to understand our own lab tests well,” Dr. Liu said in an interview. Therefore, in deciding when to perform a biopsy, physicians should interpret tests in a quantitative fashion. This interpretation should consider changes in TGA values over time, because a single positive result may not provide enough information to make a decision to proceed with biopsy.

“In the case of a symptomatic person, [a single positive result] is probably okay, because you are looking for the presence or absence of disease. However, in the case of a person who's at risk for celiac disease, multiple tests over time may be needed due to the potential for disease. In addition, “we really need to understand what is a very high level,” he said, “because higher TGA levels are more likely to lead to findings of intestinal lesions.”

Complicating this diagnostic picture is the wide variability of currently available IgA-TGA assays, said Dr. Liu of the Barbara Davis Center for Childhood Diabetes and the section of gastroenterology, hepatology, and nutrition at the Children's Hospital and the University of Colorado at Denver. The definition of what constitutes a high TGA value differs depending on the laboratory and the assay used.

Asymptomatic individuals may need to be tested several times before deciding whether to proceed with biopsy. This is because a biopsy done too soon could produce normal histologic findings that suggest the absence of disease, but these normal findings do not necessarily rule out the possibility that disease will develop, Dr. Liu said.

He cited an example of the patient with type 1 diabetes who has an abnormal TGA and whose small intestinal biopsy is normal. The finding is not necessarily a “false positive” TGA level, but could be caused instead by the underlying biology of celiac disease. “If we biopsy patients too early they may not have had time to develop intestinal lesions,” he said. “If we believe that the paradigm for most autoimmunity also applies to celiac disease—that autoantibodies precede the development of actual disease—then performing intestinal biopsy in the early stages of autoimmunity might lead to findings of normal histology.”

Although some clinicians prefer to perform a biopsy at the first sign of abnormality on TGA because they do not want to miss a case of disease, Dr. Liu said the approach to diagnosis at his institution differs somewhat. “We don't want to biopsy more than once,” he said, adding that this strategy is viable because “the kids we follow are in the medical system, so we can afford to be patient and try to determine the optimal time for biopsy” with serial serologic tests.

CHICAGO — When testing patients for celiac disease, physicians can no longer rely on a single paradigm for both overtly symptomatic patients and asymptomatic but genetically at-risk patients, according to Dr. Edwin Liu.

These two categories of patients require different approaches, said Dr. Liu, who spoke at a meeting on celiac disease sponsored by the American Gastroenterological Association.

Most symptomatic patients need only one antibody test, transglutaminase IgA (IgA-TGA) and an IgA antibody level to assess for celiac disease. But genetically at-risk patients may need multiple tests over time to screen for the presence of celiac autoimmunity and to determine if a biopsy is needed. Patients considered at risk for celiac disease include first-degree relatives of those with celiac disease or type 1 diabetes, and patients with type 1 diabetes.

The patient with classic symptoms and an abnormal TGA result usually can be biopsied immediately with a greater than 90% likelihood that intestinal lesions will be found, but TGA predicts disease in only about 75% of asymptomatic patients at genetic risk.

Patients with very elevated blood TGA levels are more likely to have more severe intestinal injury, so “in screening those at genetic risk, we have to understand our own lab tests well,” Dr. Liu said in an interview. Therefore, in deciding when to perform a biopsy, physicians should interpret tests in a quantitative fashion. This interpretation should consider changes in TGA values over time, because a single positive result may not provide enough information to make a decision to proceed with biopsy.

“In the case of a symptomatic person, [a single positive result] is probably okay, because you are looking for the presence or absence of disease. However, in the case of a person who's at risk for celiac disease, multiple tests over time may be needed due to the potential for disease. In addition, “we really need to understand what is a very high level,” he said, “because higher TGA levels are more likely to lead to findings of intestinal lesions.”

Complicating this diagnostic picture is the wide variability of currently available IgA-TGA assays, said Dr. Liu of the Barbara Davis Center for Childhood Diabetes and the section of gastroenterology, hepatology, and nutrition at the Children's Hospital and the University of Colorado at Denver. The definition of what constitutes a high TGA value differs depending on the laboratory and the assay used.

Asymptomatic individuals may need to be tested several times before deciding whether to proceed with biopsy. This is because a biopsy done too soon could produce normal histologic findings that suggest the absence of disease, but these normal findings do not necessarily rule out the possibility that disease will develop, Dr. Liu said.

He cited an example of the patient with type 1 diabetes who has an abnormal TGA and whose small intestinal biopsy is normal. The finding is not necessarily a “false positive” TGA level, but could be caused instead by the underlying biology of celiac disease. “If we biopsy patients too early they may not have had time to develop intestinal lesions,” he said. “If we believe that the paradigm for most autoimmunity also applies to celiac disease—that autoantibodies precede the development of actual disease—then performing intestinal biopsy in the early stages of autoimmunity might lead to findings of normal histology.”

Although some clinicians prefer to perform a biopsy at the first sign of abnormality on TGA because they do not want to miss a case of disease, Dr. Liu said the approach to diagnosis at his institution differs somewhat. “We don't want to biopsy more than once,” he said, adding that this strategy is viable because “the kids we follow are in the medical system, so we can afford to be patient and try to determine the optimal time for biopsy” with serial serologic tests.

CHICAGO – The cooperative pooling and exchange of data on underinsured and uninsured patients has enabled a local alliance of health care providers to identify and stratify frequent users of emergency department services and create interventions targeted to the needs of subpopulations and individuals.

The group's longitudinal study using a master index of 750,000 patients called the I-Care database offers an example of how health information exchange at the grassroots level can be successfully used to pinpoint problems, devise solutions, and reduce overutilization, according to Anjum Khurshid, Ph.D., director of clinical research and evaluation for the nonprofit Integrated Care Collaboration (ICC), an Austin-based safety net consortium.

The ICC includes 24 major providers in central Texas, and a total of 70 sites, including hospitals, clinics, academic institutions, jail health services, mental health agencies, and public health departments.

According to Dr. Khurshid, “Our data analysis is immediately picked up by people who are decision makers who want to translate that research into action. We are making decisions based on data, and at the same time, we have a mechanism by which we can collaborate and do something at the community level.”

An analysis of I-Care data for May 2006-June 2008 on frequent ED users, defined as patients who made six or more ED visits in a quarter, showed these patients to be a heterogeneous and dynamic group, Dr. Khurshid reported in a session on developments in health information exchange at the annual research meeting of AcademyHealth.

The consortium has used the findings as a basis for efforts to improve the delivery of care, he said. “Once we've identified patients, we're not just starting programs blindly, but are focusing on subpopulations so that we get the maximum effect.”

The study calculated changes in patient lists from one quarter to the next. Only 20%-26% of frequent users in one quarter were frequent users in another quarter, and 2% of patients were frequent users in all eight quarters. In all, 1,348 unduplicated patients were frequent users during the 2-year period. The number of frequent users in a quarter ranged from 178 to 251 (mean 215).

The pooling of information among providers in the I-Care database allowed the identification of many more frequent ED users than would have been possible had individual providers or hospital systems conducted independent analyses, because many frequent users visited more than one location.

For May-July 2008, for example, 205 frequent users were identified in the I-Care database, but the number of frequent users identified at individual sites reached a combined total of only 128 (range 2-55).

An analysis of visits by patients who were frequent users every quarter revealed a staggering number of ED visits among a very small group, Dr. Khurshid said.

The study identified nine individuals who, on further investigation, were found to have made a total of 2,678 ED visits between 2003 and 2009. One patient had made more than 100 ED visits per year over a 4-year span.

“There is a small number of [patients] whose frequent use of the system takes up a disproportionate amount of resources,” said Diana Resnik, senior vice president of community care at the Seton Family of Hospitals, Austin, an ICC member. “It stood out pretty quickly that if we focused on those patients we were going to have an impact.”

The data analysis identified three primary subpopulations of frequent users: patients with chronic medical needs who did not have access to primary care, individuals with behavioral health and/or chemical dependence diagnoses, and homeless individuals. The analysis offered the numbers needed for consortium members to develop plans to address these patients' needs.

The patient who had been visiting the ED 100 times yearly, for example, was found to have a diagnosis of Aspberger's syndrome. Intervention included helping this individual find an appropriate group home where she would receive social support. After entering the group home in late 2008, the patient made only two unnecessary ED visits during 2009.

Similarly, a diabetic patient who had made multiple visits to the ED and had been hospitalized was connected for the first time with a primary care clinic and provided with glucose testing strips and a glucose meter for home use.

In an initiative known as High Alert, the consortium flags frequent ED users and posts their care plans to a shared site. The information sharing among members facilitates continuity of care and reduces duplication of efforts.

The ICC is also developing parameters to identify frequent users proactively “before they've been through the system 50 times” and to generate daily reports from the I-Care database, Ms. Resnik said.

For more information about the Integrated Care Collaboration, go to www.icc-centex.org

CHICAGO – The cooperative pooling and exchange of data on underinsured and uninsured patients has enabled a local alliance of health care providers to identify and stratify frequent users of emergency department services and create interventions targeted to the needs of subpopulations and individuals.

The group's longitudinal study using a master index of 750,000 patients called the I-Care database offers an example of how health information exchange at the grassroots level can be successfully used to pinpoint problems, devise solutions, and reduce overutilization, according to Anjum Khurshid, Ph.D., director of clinical research and evaluation for the nonprofit Integrated Care Collaboration (ICC), an Austin-based safety net consortium.

The ICC includes 24 major providers in central Texas, and a total of 70 sites, including hospitals, clinics, academic institutions, jail health services, mental health agencies, and public health departments.

According to Dr. Khurshid, “Our data analysis is immediately picked up by people who are decision makers who want to translate that research into action. We are making decisions based on data, and at the same time, we have a mechanism by which we can collaborate and do something at the community level.”

An analysis of I-Care data for May 2006-June 2008 on frequent ED users, defined as patients who made six or more ED visits in a quarter, showed these patients to be a heterogeneous and dynamic group, Dr. Khurshid reported in a session on developments in health information exchange at the annual research meeting of AcademyHealth.

The consortium has used the findings as a basis for efforts to improve the delivery of care, he said. “Once we've identified patients, we're not just starting programs blindly, but are focusing on subpopulations so that we get the maximum effect.”

The study calculated changes in patient lists from one quarter to the next. Only 20%-26% of frequent users in one quarter were frequent users in another quarter, and 2% of patients were frequent users in all eight quarters. In all, 1,348 unduplicated patients were frequent users during the 2-year period. The number of frequent users in a quarter ranged from 178 to 251 (mean 215).

The pooling of information among providers in the I-Care database allowed the identification of many more frequent ED users than would have been possible had individual providers or hospital systems conducted independent analyses, because many frequent users visited more than one location.

For May-July 2008, for example, 205 frequent users were identified in the I-Care database, but the number of frequent users identified at individual sites reached a combined total of only 128 (range 2-55).

An analysis of visits by patients who were frequent users every quarter revealed a staggering number of ED visits among a very small group, Dr. Khurshid said.

The study identified nine individuals who, on further investigation, were found to have made a total of 2,678 ED visits between 2003 and 2009. One patient had made more than 100 ED visits per year over a 4-year span.

“There is a small number of [patients] whose frequent use of the system takes up a disproportionate amount of resources,” said Diana Resnik, senior vice president of community care at the Seton Family of Hospitals, Austin, an ICC member. “It stood out pretty quickly that if we focused on those patients we were going to have an impact.”

The data analysis identified three primary subpopulations of frequent users: patients with chronic medical needs who did not have access to primary care, individuals with behavioral health and/or chemical dependence diagnoses, and homeless individuals. The analysis offered the numbers needed for consortium members to develop plans to address these patients' needs.

The patient who had been visiting the ED 100 times yearly, for example, was found to have a diagnosis of Aspberger's syndrome. Intervention included helping this individual find an appropriate group home where she would receive social support. After entering the group home in late 2008, the patient made only two unnecessary ED visits during 2009.

Similarly, a diabetic patient who had made multiple visits to the ED and had been hospitalized was connected for the first time with a primary care clinic and provided with glucose testing strips and a glucose meter for home use.

In an initiative known as High Alert, the consortium flags frequent ED users and posts their care plans to a shared site. The information sharing among members facilitates continuity of care and reduces duplication of efforts.

The ICC is also developing parameters to identify frequent users proactively “before they've been through the system 50 times” and to generate daily reports from the I-Care database, Ms. Resnik said.

For more information about the Integrated Care Collaboration, go to www.icc-centex.org

CHICAGO – The cooperative pooling and exchange of data on underinsured and uninsured patients has enabled a local alliance of health care providers to identify and stratify frequent users of emergency department services and create interventions targeted to the needs of subpopulations and individuals.

The group's longitudinal study using a master index of 750,000 patients called the I-Care database offers an example of how health information exchange at the grassroots level can be successfully used to pinpoint problems, devise solutions, and reduce overutilization, according to Anjum Khurshid, Ph.D., director of clinical research and evaluation for the nonprofit Integrated Care Collaboration (ICC), an Austin-based safety net consortium.

The ICC includes 24 major providers in central Texas, and a total of 70 sites, including hospitals, clinics, academic institutions, jail health services, mental health agencies, and public health departments.

According to Dr. Khurshid, “Our data analysis is immediately picked up by people who are decision makers who want to translate that research into action. We are making decisions based on data, and at the same time, we have a mechanism by which we can collaborate and do something at the community level.”

An analysis of I-Care data for May 2006-June 2008 on frequent ED users, defined as patients who made six or more ED visits in a quarter, showed these patients to be a heterogeneous and dynamic group, Dr. Khurshid reported in a session on developments in health information exchange at the annual research meeting of AcademyHealth.

The consortium has used the findings as a basis for efforts to improve the delivery of care, he said. “Once we've identified patients, we're not just starting programs blindly, but are focusing on subpopulations so that we get the maximum effect.”

The study calculated changes in patient lists from one quarter to the next. Only 20%-26% of frequent users in one quarter were frequent users in another quarter, and 2% of patients were frequent users in all eight quarters. In all, 1,348 unduplicated patients were frequent users during the 2-year period. The number of frequent users in a quarter ranged from 178 to 251 (mean 215).

The pooling of information among providers in the I-Care database allowed the identification of many more frequent ED users than would have been possible had individual providers or hospital systems conducted independent analyses, because many frequent users visited more than one location.

For May-July 2008, for example, 205 frequent users were identified in the I-Care database, but the number of frequent users identified at individual sites reached a combined total of only 128 (range 2-55).

An analysis of visits by patients who were frequent users every quarter revealed a staggering number of ED visits among a very small group, Dr. Khurshid said.

The study identified nine individuals who, on further investigation, were found to have made a total of 2,678 ED visits between 2003 and 2009. One patient had made more than 100 ED visits per year over a 4-year span.

“There is a small number of [patients] whose frequent use of the system takes up a disproportionate amount of resources,” said Diana Resnik, senior vice president of community care at the Seton Family of Hospitals, Austin, an ICC member. “It stood out pretty quickly that if we focused on those patients we were going to have an impact.”

The data analysis identified three primary subpopulations of frequent users: patients with chronic medical needs who did not have access to primary care, individuals with behavioral health and/or chemical dependence diagnoses, and homeless individuals. The analysis offered the numbers needed for consortium members to develop plans to address these patients' needs.

The patient who had been visiting the ED 100 times yearly, for example, was found to have a diagnosis of Aspberger's syndrome. Intervention included helping this individual find an appropriate group home where she would receive social support. After entering the group home in late 2008, the patient made only two unnecessary ED visits during 2009.

Similarly, a diabetic patient who had made multiple visits to the ED and had been hospitalized was connected for the first time with a primary care clinic and provided with glucose testing strips and a glucose meter for home use.

In an initiative known as High Alert, the consortium flags frequent ED users and posts their care plans to a shared site. The information sharing among members facilitates continuity of care and reduces duplication of efforts.

The ICC is also developing parameters to identify frequent users proactively “before they've been through the system 50 times” and to generate daily reports from the I-Care database, Ms. Resnik said.

For more information about the Integrated Care Collaboration, go to www.icc-centex.org

CHICAGO — In the fall of 2006, a vendor accidentally cut the wrong cable in the computer room at Children's Specialized Hospital, New Brunswick, N.J., leaving the large pediatric rehabilitation provider's eight facilities without computers or phones for 3 days.

Staff continued to care for patients with semimanual systems, and no adverse events occurred as a result of the power loss. But the sobering experience sparked a comprehensive overhaul of the organization's communications downtime procedures, using a Six Sigma risk reduction method known as FMEA (failure modes and effects analysis). Widely used in manufacturing, the Six Sigma system was first created by Motorola as a method for improving quality and efficiency. The FMEA component involves rating on a numerical scale the risks associated with components in a process, then prioritizing corrective actions based on risk levels.

A root cause analysis and a review of policies and procedures revealed serious shortcomings, including gaps between the administrative policy and the emergency operations plan, inconsistencies across some departments, no policies and procedures at all in other departments, and critical steps that were missing, including a formal process for communicating to staff that systems were down, said Lorraine Quatrone, medical administrator at Children's Specialized Hospital.

“The staff was completely out of the loop,” she said at the Joint Commission national conference on quality and patient safety. “We thought we had a plan in place,” but “we were operating in silos.”

The hospital tightened computer room security, revised the administrative policy, and developed a flowchart showing who should notify whom after a communications failure. For example, the chief information officer was instructed to notify the chief executive officer and the chief safety officer.

Following these quick fixes, “we could have sat back and said we're prepared,” Ms. Quatrone said. Instead, the hospital decided to “drill down and look behind doors” using FMEA methodology.

The hospital identified potential areas of vulnerability and prioritized areas for improvement. “Our analysis told us we were weak in communication of unplanned downtime and the implementation of procedures,” Ms. Quatrone said. In other words, staff needed to be alerted about a communications failure, and they needed to know what to do to ensure patient safety after they were informed about the situation.

The hospital developed a template for downtime policies and procedures for every department. In addition to asking directors and managers what their departments needed in order to continue to function without computers and telephones, the hospital asked them to look at their departments as suppliers of information to the organization and to indicate how they could help other departments.

“We wanted to make this an organizationwide commitment to helping each other,” Ms. Quatrone said. Facilities management, for example, is now responsible for immediately distributing two-way radios to patient areas, making hourly rounds to check for emergency issues, and monitoring the hospital's energy management system. All nursing units are required to immediately begin recording the administration of all medications on a written worksheet.

The hospital also addressed procedures that would govern how each department would continue to function after systems were working again, including how information from the interim paper process would get entered into the electronic system. Once the system has returned to operational status, for example, pharmacy staff are required to enter all new medication orders electronically.

Following the revision of policies and procedures, department directors and managers were asked to educate their staff and to decide with them which electronic forms would be needed in paper form and where information should be kept. Information about the emergency plan became an integral part of new employee orientations as well, she said.

To measure the initiative's success, the hospital conducted a series of simulated downtime drills and asked a sample of employees six questions about the emergency plan, including “How would you complete an event report if the system went down?” and “Can you show me your department's downtime policy?”

Awareness “seemed pretty low at the beginning, but as time went on and we did drills to reinforce our commitment to the process, we started to see the results edge up,” Ms. Quatrone said.

The need for emergency plans will become even more crucial as more providers move toward electronic medical records and continue to automate other systems. “You can't fall short of recognizing the value and importance of having a backup within your organization,” she said.

CHICAGO — In the fall of 2006, a vendor accidentally cut the wrong cable in the computer room at Children's Specialized Hospital, New Brunswick, N.J., leaving the large pediatric rehabilitation provider's eight facilities without computers or phones for 3 days.

Staff continued to care for patients with semimanual systems, and no adverse events occurred as a result of the power loss. But the sobering experience sparked a comprehensive overhaul of the organization's communications downtime procedures, using a Six Sigma risk reduction method known as FMEA (failure modes and effects analysis). Widely used in manufacturing, the Six Sigma system was first created by Motorola as a method for improving quality and efficiency. The FMEA component involves rating on a numerical scale the risks associated with components in a process, then prioritizing corrective actions based on risk levels.

A root cause analysis and a review of policies and procedures revealed serious shortcomings, including gaps between the administrative policy and the emergency operations plan, inconsistencies across some departments, no policies and procedures at all in other departments, and critical steps that were missing, including a formal process for communicating to staff that systems were down, said Lorraine Quatrone, medical administrator at Children's Specialized Hospital.

“The staff was completely out of the loop,” she said at the Joint Commission national conference on quality and patient safety. “We thought we had a plan in place,” but “we were operating in silos.”

The hospital tightened computer room security, revised the administrative policy, and developed a flowchart showing who should notify whom after a communications failure. For example, the chief information officer was instructed to notify the chief executive officer and the chief safety officer.

Following these quick fixes, “we could have sat back and said we're prepared,” Ms. Quatrone said. Instead, the hospital decided to “drill down and look behind doors” using FMEA methodology.

The hospital identified potential areas of vulnerability and prioritized areas for improvement. “Our analysis told us we were weak in communication of unplanned downtime and the implementation of procedures,” Ms. Quatrone said. In other words, staff needed to be alerted about a communications failure, and they needed to know what to do to ensure patient safety after they were informed about the situation.

The hospital developed a template for downtime policies and procedures for every department. In addition to asking directors and managers what their departments needed in order to continue to function without computers and telephones, the hospital asked them to look at their departments as suppliers of information to the organization and to indicate how they could help other departments.

“We wanted to make this an organizationwide commitment to helping each other,” Ms. Quatrone said. Facilities management, for example, is now responsible for immediately distributing two-way radios to patient areas, making hourly rounds to check for emergency issues, and monitoring the hospital's energy management system. All nursing units are required to immediately begin recording the administration of all medications on a written worksheet.

The hospital also addressed procedures that would govern how each department would continue to function after systems were working again, including how information from the interim paper process would get entered into the electronic system. Once the system has returned to operational status, for example, pharmacy staff are required to enter all new medication orders electronically.

Following the revision of policies and procedures, department directors and managers were asked to educate their staff and to decide with them which electronic forms would be needed in paper form and where information should be kept. Information about the emergency plan became an integral part of new employee orientations as well, she said.

To measure the initiative's success, the hospital conducted a series of simulated downtime drills and asked a sample of employees six questions about the emergency plan, including “How would you complete an event report if the system went down?” and “Can you show me your department's downtime policy?”

Awareness “seemed pretty low at the beginning, but as time went on and we did drills to reinforce our commitment to the process, we started to see the results edge up,” Ms. Quatrone said.

The need for emergency plans will become even more crucial as more providers move toward electronic medical records and continue to automate other systems. “You can't fall short of recognizing the value and importance of having a backup within your organization,” she said.

CHICAGO — In the fall of 2006, a vendor accidentally cut the wrong cable in the computer room at Children's Specialized Hospital, New Brunswick, N.J., leaving the large pediatric rehabilitation provider's eight facilities without computers or phones for 3 days.

Staff continued to care for patients with semimanual systems, and no adverse events occurred as a result of the power loss. But the sobering experience sparked a comprehensive overhaul of the organization's communications downtime procedures, using a Six Sigma risk reduction method known as FMEA (failure modes and effects analysis). Widely used in manufacturing, the Six Sigma system was first created by Motorola as a method for improving quality and efficiency. The FMEA component involves rating on a numerical scale the risks associated with components in a process, then prioritizing corrective actions based on risk levels.

A root cause analysis and a review of policies and procedures revealed serious shortcomings, including gaps between the administrative policy and the emergency operations plan, inconsistencies across some departments, no policies and procedures at all in other departments, and critical steps that were missing, including a formal process for communicating to staff that systems were down, said Lorraine Quatrone, medical administrator at Children's Specialized Hospital.

“The staff was completely out of the loop,” she said at the Joint Commission national conference on quality and patient safety. “We thought we had a plan in place,” but “we were operating in silos.”

The hospital tightened computer room security, revised the administrative policy, and developed a flowchart showing who should notify whom after a communications failure. For example, the chief information officer was instructed to notify the chief executive officer and the chief safety officer.

Following these quick fixes, “we could have sat back and said we're prepared,” Ms. Quatrone said. Instead, the hospital decided to “drill down and look behind doors” using FMEA methodology.

The hospital identified potential areas of vulnerability and prioritized areas for improvement. “Our analysis told us we were weak in communication of unplanned downtime and the implementation of procedures,” Ms. Quatrone said. In other words, staff needed to be alerted about a communications failure, and they needed to know what to do to ensure patient safety after they were informed about the situation.

The hospital developed a template for downtime policies and procedures for every department. In addition to asking directors and managers what their departments needed in order to continue to function without computers and telephones, the hospital asked them to look at their departments as suppliers of information to the organization and to indicate how they could help other departments.

“We wanted to make this an organizationwide commitment to helping each other,” Ms. Quatrone said. Facilities management, for example, is now responsible for immediately distributing two-way radios to patient areas, making hourly rounds to check for emergency issues, and monitoring the hospital's energy management system. All nursing units are required to immediately begin recording the administration of all medications on a written worksheet.

The hospital also addressed procedures that would govern how each department would continue to function after systems were working again, including how information from the interim paper process would get entered into the electronic system. Once the system has returned to operational status, for example, pharmacy staff are required to enter all new medication orders electronically.

Following the revision of policies and procedures, department directors and managers were asked to educate their staff and to decide with them which electronic forms would be needed in paper form and where information should be kept. Information about the emergency plan became an integral part of new employee orientations as well, she said.

To measure the initiative's success, the hospital conducted a series of simulated downtime drills and asked a sample of employees six questions about the emergency plan, including “How would you complete an event report if the system went down?” and “Can you show me your department's downtime policy?”

Awareness “seemed pretty low at the beginning, but as time went on and we did drills to reinforce our commitment to the process, we started to see the results edge up,” Ms. Quatrone said.

The need for emergency plans will become even more crucial as more providers move toward electronic medical records and continue to automate other systems. “You can't fall short of recognizing the value and importance of having a backup within your organization,” she said.

ROSEMONT, ILL. — A multifaceted intervention enabled a large health system to increase compliance with evidence-based guidelines for venous thromboembolism prophylaxis, according to Dr. Valerie Allusson, director of inpatient medicine services at Atlantic Health, Morristown, N.J.

Although the health system has not yet reached all of its quality benchmarks, compliance has risen substantially as the result of measures such as the creation of a physician order set and daily monitoring of compliance, Dr. Allusson reported at the Joint Commission national conference on quality and patient safety.

Atlantic Health's 504-bed Overlook Hospital was one of 41 hospitals that completed a 6-month pilot study of VTE quality measures sponsored by the Joint Commission in 2006–2007. Since then, Atlantic Health has spent 2 years focusing on improving VTE prophylaxis for medical and surgical patients at Overlook Hospital and the 629-bed Morristown Memorial Hospital.

In a baseline study of 100 randomly selected charts in one of Atlantic Health's medical units, only 39% of patients received VTE prophylaxis. The system implemented a quality improvement initiative based on recommendations from the American College of Chest Physicians (Chest 2004;126:338S–400S) and the National Consensus Standards for the Prevention and Care of Deep Vein Thrombosis developed by the National Quality Forum and the Joint Commission.

Areas of particular focus were VTE risk assessment/prophylaxis within 24 hours of hospital admission and VTE written discharge instructions for patients on warfarin addressing follow-up monitoring, compliance issues, dietary restrictions, and potential drug reactions or interactions.

The system set a 6-month goal to conduct a VTE risk assessment and provide appropriate prophylaxis within 24 hours of hospital admission or surgery end time for 95% of all patients. A second 6-month goal was to reach 95% of patients who had “fallen through the cracks” and had been admitted without prophylaxis.

A multidisciplinary steering committee developed a VTE prophylaxis order set addressing risk assessment, contraindications, and management options. A prototype daily VTE prophylaxis outlier list tracked patients who were and were not receiving acceptable medications (including argatroban, fondaparinux, heparin, and low-molecular-weight heparin, and warfarin). A sticker at the front of outlier charts alerted physicians about patients not receiving prophylaxis.

As of June 2009, the system had surpassed its target of 75% for prophylaxis in ICU patients (90%) and overlap therapy (82%), and was continuing to work on the remaining target of 95% for prophylaxis in medical/surgical patients (68%) and discharge instructions (79%).

In an interview, Dr. Allusson called the 95% target for prophylaxis in medical/surgical patients ambitious, considering the 39% baseline rate. Reaching 68% within 6 months represented significant progress, she said.

The system also achieved a 5%–7.5% reduction in in-hospital mortality due to VTE during this time, but whether the decline was due to the VTE quality improvement project is not known.

Dr. Allusson attributed the progress to date in part to the frequent and routine sharing of data at every level of the organization, and to the multidisciplinary collaboration. Efforts to systematize and streamline procedures related to VTE prophylaxis also helped. The new VTE prophylaxis order set, for example, allows physicians to document medications administered simply by checking the appropriate box. In addition, collaboration with information technology helped reduce the likelihood of human error.

VTE is 100 times more common in hospitalized patients than in the general population (Mayo Clin. Proc. 2001;76:1102), Dr. Allusson noted. Up to 2 million Americans experience VTE each year, and of these, 800,000 develop pulmonary thromboembolic syndrome (PTS), 600,000 develop pulmonary embolism (PE), and 300,000 die from PE (Lancet 1999;353:1386–9).

In the future, the health system aims to develop a business plan for inpatient and outpatient anticoagulation management, include pharmacists in rounds to discuss anticoagulation and discharge instructions, create a unit performance tracking system, and establish mandatory prophylaxis order forms for all admissions.

ROSEMONT, ILL. — A multifaceted intervention enabled a large health system to increase compliance with evidence-based guidelines for venous thromboembolism prophylaxis, according to Dr. Valerie Allusson, director of inpatient medicine services at Atlantic Health, Morristown, N.J.

Although the health system has not yet reached all of its quality benchmarks, compliance has risen substantially as the result of measures such as the creation of a physician order set and daily monitoring of compliance, Dr. Allusson reported at the Joint Commission national conference on quality and patient safety.

Atlantic Health's 504-bed Overlook Hospital was one of 41 hospitals that completed a 6-month pilot study of VTE quality measures sponsored by the Joint Commission in 2006–2007. Since then, Atlantic Health has spent 2 years focusing on improving VTE prophylaxis for medical and surgical patients at Overlook Hospital and the 629-bed Morristown Memorial Hospital.

In a baseline study of 100 randomly selected charts in one of Atlantic Health's medical units, only 39% of patients received VTE prophylaxis. The system implemented a quality improvement initiative based on recommendations from the American College of Chest Physicians (Chest 2004;126:338S–400S) and the National Consensus Standards for the Prevention and Care of Deep Vein Thrombosis developed by the National Quality Forum and the Joint Commission.

Areas of particular focus were VTE risk assessment/prophylaxis within 24 hours of hospital admission and VTE written discharge instructions for patients on warfarin addressing follow-up monitoring, compliance issues, dietary restrictions, and potential drug reactions or interactions.

The system set a 6-month goal to conduct a VTE risk assessment and provide appropriate prophylaxis within 24 hours of hospital admission or surgery end time for 95% of all patients. A second 6-month goal was to reach 95% of patients who had “fallen through the cracks” and had been admitted without prophylaxis.

A multidisciplinary steering committee developed a VTE prophylaxis order set addressing risk assessment, contraindications, and management options. A prototype daily VTE prophylaxis outlier list tracked patients who were and were not receiving acceptable medications (including argatroban, fondaparinux, heparin, and low-molecular-weight heparin, and warfarin). A sticker at the front of outlier charts alerted physicians about patients not receiving prophylaxis.

As of June 2009, the system had surpassed its target of 75% for prophylaxis in ICU patients (90%) and overlap therapy (82%), and was continuing to work on the remaining target of 95% for prophylaxis in medical/surgical patients (68%) and discharge instructions (79%).

In an interview, Dr. Allusson called the 95% target for prophylaxis in medical/surgical patients ambitious, considering the 39% baseline rate. Reaching 68% within 6 months represented significant progress, she said.

The system also achieved a 5%–7.5% reduction in in-hospital mortality due to VTE during this time, but whether the decline was due to the VTE quality improvement project is not known.

Dr. Allusson attributed the progress to date in part to the frequent and routine sharing of data at every level of the organization, and to the multidisciplinary collaboration. Efforts to systematize and streamline procedures related to VTE prophylaxis also helped. The new VTE prophylaxis order set, for example, allows physicians to document medications administered simply by checking the appropriate box. In addition, collaboration with information technology helped reduce the likelihood of human error.

VTE is 100 times more common in hospitalized patients than in the general population (Mayo Clin. Proc. 2001;76:1102), Dr. Allusson noted. Up to 2 million Americans experience VTE each year, and of these, 800,000 develop pulmonary thromboembolic syndrome (PTS), 600,000 develop pulmonary embolism (PE), and 300,000 die from PE (Lancet 1999;353:1386–9).

In the future, the health system aims to develop a business plan for inpatient and outpatient anticoagulation management, include pharmacists in rounds to discuss anticoagulation and discharge instructions, create a unit performance tracking system, and establish mandatory prophylaxis order forms for all admissions.

ROSEMONT, ILL. — A multifaceted intervention enabled a large health system to increase compliance with evidence-based guidelines for venous thromboembolism prophylaxis, according to Dr. Valerie Allusson, director of inpatient medicine services at Atlantic Health, Morristown, N.J.

Although the health system has not yet reached all of its quality benchmarks, compliance has risen substantially as the result of measures such as the creation of a physician order set and daily monitoring of compliance, Dr. Allusson reported at the Joint Commission national conference on quality and patient safety.

Atlantic Health's 504-bed Overlook Hospital was one of 41 hospitals that completed a 6-month pilot study of VTE quality measures sponsored by the Joint Commission in 2006–2007. Since then, Atlantic Health has spent 2 years focusing on improving VTE prophylaxis for medical and surgical patients at Overlook Hospital and the 629-bed Morristown Memorial Hospital.

In a baseline study of 100 randomly selected charts in one of Atlantic Health's medical units, only 39% of patients received VTE prophylaxis. The system implemented a quality improvement initiative based on recommendations from the American College of Chest Physicians (Chest 2004;126:338S–400S) and the National Consensus Standards for the Prevention and Care of Deep Vein Thrombosis developed by the National Quality Forum and the Joint Commission.

Areas of particular focus were VTE risk assessment/prophylaxis within 24 hours of hospital admission and VTE written discharge instructions for patients on warfarin addressing follow-up monitoring, compliance issues, dietary restrictions, and potential drug reactions or interactions.

The system set a 6-month goal to conduct a VTE risk assessment and provide appropriate prophylaxis within 24 hours of hospital admission or surgery end time for 95% of all patients. A second 6-month goal was to reach 95% of patients who had “fallen through the cracks” and had been admitted without prophylaxis.

A multidisciplinary steering committee developed a VTE prophylaxis order set addressing risk assessment, contraindications, and management options. A prototype daily VTE prophylaxis outlier list tracked patients who were and were not receiving acceptable medications (including argatroban, fondaparinux, heparin, and low-molecular-weight heparin, and warfarin). A sticker at the front of outlier charts alerted physicians about patients not receiving prophylaxis.

As of June 2009, the system had surpassed its target of 75% for prophylaxis in ICU patients (90%) and overlap therapy (82%), and was continuing to work on the remaining target of 95% for prophylaxis in medical/surgical patients (68%) and discharge instructions (79%).

In an interview, Dr. Allusson called the 95% target for prophylaxis in medical/surgical patients ambitious, considering the 39% baseline rate. Reaching 68% within 6 months represented significant progress, she said.

The system also achieved a 5%–7.5% reduction in in-hospital mortality due to VTE during this time, but whether the decline was due to the VTE quality improvement project is not known.

Dr. Allusson attributed the progress to date in part to the frequent and routine sharing of data at every level of the organization, and to the multidisciplinary collaboration. Efforts to systematize and streamline procedures related to VTE prophylaxis also helped. The new VTE prophylaxis order set, for example, allows physicians to document medications administered simply by checking the appropriate box. In addition, collaboration with information technology helped reduce the likelihood of human error.

VTE is 100 times more common in hospitalized patients than in the general population (Mayo Clin. Proc. 2001;76:1102), Dr. Allusson noted. Up to 2 million Americans experience VTE each year, and of these, 800,000 develop pulmonary thromboembolic syndrome (PTS), 600,000 develop pulmonary embolism (PE), and 300,000 die from PE (Lancet 1999;353:1386–9).

In the future, the health system aims to develop a business plan for inpatient and outpatient anticoagulation management, include pharmacists in rounds to discuss anticoagulation and discharge instructions, create a unit performance tracking system, and establish mandatory prophylaxis order forms for all admissions.

CHICAGO — The cooperative pooling and exchange of data on underinsured and uninsured patients has enabled a local alliance of health care providers to identify and stratify frequent users of emergency department (ED) services and create interventions targeted to the needs of subpopulations and individuals.

The group's longitudinal study using a master index of 750,000 patients called the I-Care database offers an example of how health information exchange at the grassroots level can be successfully used to pinpoint problems, devise solutions, and reduce overutilization, according to Anjum Khurshid, Ph.D., director of clinical research and evaluation for the nonprofit Integrated Care Collaboration (ICC), an Austin-based safety net consortium.

The ICC includes 24 major providers in central Texas, and a total of 70 sites, including hospitals, clinics, academic institutions, jail health services, mental health agencies, and public health departments.

According to Dr. Khurshid, “Our data analysis is immediately picked up by people who are decision makers who want to translate that research into action. We are making decisions based on data, and at the same time, we have a mechanism by which we can collaborate and do something at the community level.”

An analysis of I-Care data for May 2006 to June 2008 on frequent ED users, defined as patients who made six or more ED visits in a quarter, showed these patients to be a heterogeneous and dynamic group, Dr. Khurshid reported in a session on developments in health information exchange at the annual research meeting of AcademyHealth.

The consortium has used the findings as a basis for efforts to improve the delivery of care, he said. “Once we've identified patients, we're not just starting programs blindly, but are focusing on subpopulations so that we get the maximum effect.”

The study calculated changes in patient lists from one quarter to the next. Only 20%–26% of frequent users in one quarter were frequent users in another quarter, and 2% of patients were frequent users in all eight quarters. In all, 1,348 unduplicated patients were frequent users during the 2-year period. The number of frequent users in a quarter ranged from 178 to 251 (mean 215).

The pooling of information among providers in the I-Care database allowed the identification of many more frequent ED users than would have been possible had individual providers or hospital systems conducted independent analyses, because many frequent users visited more than one location.

For May-July 2008, for example, 205 frequent users were identified in the I-Care database, but the number of frequent users identified at individual sites reached a combined total of only 128 (range 2–55).

An analysis of visits by patients who were frequent users every quarter revealed a staggering number of ED visits among a very small group, Dr. Khurshid said.

The study identified nine individuals who, on further investigation, were found to have made a total of 2,678 ED visits between 2003 and 2009. One patient had made more than 100 ED visits per year over a 4-year span.

“There is a small number of [patients] whose frequent use of the system takes up a disproportionate amount of resources,” said Diana Resnik, senior vice president of community care at the Seton Family of Hospitals, Austin, an ICC member. “It stood out pretty quickly that if we focused on those patients we were going to have an impact.”

The data analysis identified three primary subpopulations of frequent users: patients with chronic medical needs who did not have access to primary care, individuals with behavioral health and/or chemical dependence diagnoses, and homeless individuals. The analysis offered the numbers needed for consortium members to develop plans to address these patients' needs.

The patient who had been visiting the ED 100 times yearly, for example, was found to have a diagnosis of Aspberger's syndrome. Intervention included helping this individual find an appropriate group home where she would receive social support. After entering the group home in late 2008, the patient made only two unnecessary ED visits during 2009.

Similarly, a diabetic patient who had made multiple visits to the ED and had been hospitalized was connected for the first time with a primary care clinic and provided with glucose testing strips and a glucose meter for home use.

In an initiative known as High Alert, the consortium flags frequent ED users and posts their care plans to a shared site. The information sharing among members facilitates continuity of care and reduces duplication of efforts.

The ICC is also developing parameters to identify frequent users proactively “before they've been through the system 50 times” and to generate daily reports from the I-Care database, Ms. Resnik said.

Added Steve Conti, director of disease management at Seton, “the challenge we often face in the ER is that patients with complex chronic diseases don't always present with the same diagnosis. A diabetic may come in one day with a glucose issue, but then another day with foot pain, so we may not recognize the same person coming in.” High Alert prompts providers that the patient has an ongoing medical condition so that “we can recognize and deal with that rather than work through this as a whole new case,” he said.

For more information about the Integrated Care Collaboration, go to www.icc-centex.org

CHICAGO — The cooperative pooling and exchange of data on underinsured and uninsured patients has enabled a local alliance of health care providers to identify and stratify frequent users of emergency department (ED) services and create interventions targeted to the needs of subpopulations and individuals.

The group's longitudinal study using a master index of 750,000 patients called the I-Care database offers an example of how health information exchange at the grassroots level can be successfully used to pinpoint problems, devise solutions, and reduce overutilization, according to Anjum Khurshid, Ph.D., director of clinical research and evaluation for the nonprofit Integrated Care Collaboration (ICC), an Austin-based safety net consortium.

The ICC includes 24 major providers in central Texas, and a total of 70 sites, including hospitals, clinics, academic institutions, jail health services, mental health agencies, and public health departments.

According to Dr. Khurshid, “Our data analysis is immediately picked up by people who are decision makers who want to translate that research into action. We are making decisions based on data, and at the same time, we have a mechanism by which we can collaborate and do something at the community level.”

An analysis of I-Care data for May 2006 to June 2008 on frequent ED users, defined as patients who made six or more ED visits in a quarter, showed these patients to be a heterogeneous and dynamic group, Dr. Khurshid reported in a session on developments in health information exchange at the annual research meeting of AcademyHealth.

The consortium has used the findings as a basis for efforts to improve the delivery of care, he said. “Once we've identified patients, we're not just starting programs blindly, but are focusing on subpopulations so that we get the maximum effect.”

The study calculated changes in patient lists from one quarter to the next. Only 20%–26% of frequent users in one quarter were frequent users in another quarter, and 2% of patients were frequent users in all eight quarters. In all, 1,348 unduplicated patients were frequent users during the 2-year period. The number of frequent users in a quarter ranged from 178 to 251 (mean 215).

The pooling of information among providers in the I-Care database allowed the identification of many more frequent ED users than would have been possible had individual providers or hospital systems conducted independent analyses, because many frequent users visited more than one location.

For May-July 2008, for example, 205 frequent users were identified in the I-Care database, but the number of frequent users identified at individual sites reached a combined total of only 128 (range 2–55).

An analysis of visits by patients who were frequent users every quarter revealed a staggering number of ED visits among a very small group, Dr. Khurshid said.

The study identified nine individuals who, on further investigation, were found to have made a total of 2,678 ED visits between 2003 and 2009. One patient had made more than 100 ED visits per year over a 4-year span.

“There is a small number of [patients] whose frequent use of the system takes up a disproportionate amount of resources,” said Diana Resnik, senior vice president of community care at the Seton Family of Hospitals, Austin, an ICC member. “It stood out pretty quickly that if we focused on those patients we were going to have an impact.”

The data analysis identified three primary subpopulations of frequent users: patients with chronic medical needs who did not have access to primary care, individuals with behavioral health and/or chemical dependence diagnoses, and homeless individuals. The analysis offered the numbers needed for consortium members to develop plans to address these patients' needs.

The patient who had been visiting the ED 100 times yearly, for example, was found to have a diagnosis of Aspberger's syndrome. Intervention included helping this individual find an appropriate group home where she would receive social support. After entering the group home in late 2008, the patient made only two unnecessary ED visits during 2009.

Similarly, a diabetic patient who had made multiple visits to the ED and had been hospitalized was connected for the first time with a primary care clinic and provided with glucose testing strips and a glucose meter for home use.

In an initiative known as High Alert, the consortium flags frequent ED users and posts their care plans to a shared site. The information sharing among members facilitates continuity of care and reduces duplication of efforts.

The ICC is also developing parameters to identify frequent users proactively “before they've been through the system 50 times” and to generate daily reports from the I-Care database, Ms. Resnik said.

Added Steve Conti, director of disease management at Seton, “the challenge we often face in the ER is that patients with complex chronic diseases don't always present with the same diagnosis. A diabetic may come in one day with a glucose issue, but then another day with foot pain, so we may not recognize the same person coming in.” High Alert prompts providers that the patient has an ongoing medical condition so that “we can recognize and deal with that rather than work through this as a whole new case,” he said.

For more information about the Integrated Care Collaboration, go to www.icc-centex.org

CHICAGO — The cooperative pooling and exchange of data on underinsured and uninsured patients has enabled a local alliance of health care providers to identify and stratify frequent users of emergency department (ED) services and create interventions targeted to the needs of subpopulations and individuals.

The group's longitudinal study using a master index of 750,000 patients called the I-Care database offers an example of how health information exchange at the grassroots level can be successfully used to pinpoint problems, devise solutions, and reduce overutilization, according to Anjum Khurshid, Ph.D., director of clinical research and evaluation for the nonprofit Integrated Care Collaboration (ICC), an Austin-based safety net consortium.

The ICC includes 24 major providers in central Texas, and a total of 70 sites, including hospitals, clinics, academic institutions, jail health services, mental health agencies, and public health departments.

According to Dr. Khurshid, “Our data analysis is immediately picked up by people who are decision makers who want to translate that research into action. We are making decisions based on data, and at the same time, we have a mechanism by which we can collaborate and do something at the community level.”

An analysis of I-Care data for May 2006 to June 2008 on frequent ED users, defined as patients who made six or more ED visits in a quarter, showed these patients to be a heterogeneous and dynamic group, Dr. Khurshid reported in a session on developments in health information exchange at the annual research meeting of AcademyHealth.

The consortium has used the findings as a basis for efforts to improve the delivery of care, he said. “Once we've identified patients, we're not just starting programs blindly, but are focusing on subpopulations so that we get the maximum effect.”

The study calculated changes in patient lists from one quarter to the next. Only 20%–26% of frequent users in one quarter were frequent users in another quarter, and 2% of patients were frequent users in all eight quarters. In all, 1,348 unduplicated patients were frequent users during the 2-year period. The number of frequent users in a quarter ranged from 178 to 251 (mean 215).

The pooling of information among providers in the I-Care database allowed the identification of many more frequent ED users than would have been possible had individual providers or hospital systems conducted independent analyses, because many frequent users visited more than one location.

For May-July 2008, for example, 205 frequent users were identified in the I-Care database, but the number of frequent users identified at individual sites reached a combined total of only 128 (range 2–55).

An analysis of visits by patients who were frequent users every quarter revealed a staggering number of ED visits among a very small group, Dr. Khurshid said.

The study identified nine individuals who, on further investigation, were found to have made a total of 2,678 ED visits between 2003 and 2009. One patient had made more than 100 ED visits per year over a 4-year span.

“There is a small number of [patients] whose frequent use of the system takes up a disproportionate amount of resources,” said Diana Resnik, senior vice president of community care at the Seton Family of Hospitals, Austin, an ICC member. “It stood out pretty quickly that if we focused on those patients we were going to have an impact.”

The data analysis identified three primary subpopulations of frequent users: patients with chronic medical needs who did not have access to primary care, individuals with behavioral health and/or chemical dependence diagnoses, and homeless individuals. The analysis offered the numbers needed for consortium members to develop plans to address these patients' needs.

The patient who had been visiting the ED 100 times yearly, for example, was found to have a diagnosis of Aspberger's syndrome. Intervention included helping this individual find an appropriate group home where she would receive social support. After entering the group home in late 2008, the patient made only two unnecessary ED visits during 2009.

Similarly, a diabetic patient who had made multiple visits to the ED and had been hospitalized was connected for the first time with a primary care clinic and provided with glucose testing strips and a glucose meter for home use.

In an initiative known as High Alert, the consortium flags frequent ED users and posts their care plans to a shared site. The information sharing among members facilitates continuity of care and reduces duplication of efforts.

The ICC is also developing parameters to identify frequent users proactively “before they've been through the system 50 times” and to generate daily reports from the I-Care database, Ms. Resnik said.

Added Steve Conti, director of disease management at Seton, “the challenge we often face in the ER is that patients with complex chronic diseases don't always present with the same diagnosis. A diabetic may come in one day with a glucose issue, but then another day with foot pain, so we may not recognize the same person coming in.” High Alert prompts providers that the patient has an ongoing medical condition so that “we can recognize and deal with that rather than work through this as a whole new case,” he said.

For more information about the Integrated Care Collaboration, go to www.icc-centex.org

ROSEMONT, ILL. — Despite major patient safety strides during the past decade, health care providers need to create more accountability for medical errors and patient safety lapses in order to continue improving, according to Dr. Robert M. Wachter, professor and associate chairman of medicine at the University of California, San Francisco.

At the Joint Commission national conference on quality and patient safety, Dr. Wachter offered his perspectives on the status of patient safety in health care 10 years after the publication of the first Institute of Medicine report on the subject ("To Err Is Human: Building a Safer Health System"). He and Dr. Peter J. Pronovost of Johns Hopkins University, Baltimore, published an editorial on the topic shortly after the conference (N. Engl. J. Med. 2009;361:1401–6).

Balancing a culture of "no blame" with a culture of accountability remains a key challenge for providers.

"'No blame' is the right response for innocent slips and mistakes, which turn out to be most errors. But there need to be clear demarcations of blameworthy acts—not just gross incompetence, not just disruptive behavior, but also failure to follow reasonable safety standards," said Dr. Wachter.

Dr. Wachter cited the fact that average hand-washing compliance rates continue to hover at only about 50% as an example of the need for more accountability. "I don't believe that is fully a systems problem," he said. Part of the problem is that "there have been no penalties for transgressions."

Although accountability is essential, "my guess is it will go too far," Dr. Wachter added. "We've created an environment where people don't want to have to talk about errors. We're probably going to have to go too far and then come up with a sweet spot."

Dr. Wachter also commented on other aspects of patient safety:

▸ Regulation. Health care organizations need regulators to set standards, but the challenge is ensuring that these standards truly help organizations improve safety. Until the Joint Commission developed standards for reading back instructions, "virtually none of us thought of doing that on our own," he said.

At the same time, "it is extraordinarily difficult to have a set of rules and standards that apply equally in nuanced areas to organizations that are incredibly different in the way they do business, their financial resources, and their capacity," he said.

For that reason, "regulation is extraordinarily useful to get people moving, but it tends to run out of gas over time," Dr. Wachter said. To illustrate, he cited the Joint Commission's recent decision to remove adherence to medication reconciliation standards as a requirement for accreditation because organizations struggled to develop appropriate processes.

However, having an "outside organization creating rules and standards we must abide by was extraordinarily important in the first 5 years" after the IOM report, he said. Despite some glitches, "the Joint Commission has improved its processes tremendously."

▸ Reporting. "The admonition to report everything is silly," Dr. Wachter said. "Our mistake here was to not be thoughtful about what we are going to do with all of these reports" before requiring them.

However, providers have learned from this experience and begun to think more critically about what should be reported and how the data should be used, he said.

State reporting requirements on the 27 "never events" put forth by the National Quality Forum have led to more focused patient safety efforts.

He added, however, that labeling these adverse occurrences "never" events was a misnomer that put unfair pressure on providers because many of these problems, such as decubitus ulcers, are not even largely preventable based on present science. "It's a politically driven misnomer," he said.

ROSEMONT, ILL. — Despite major patient safety strides during the past decade, health care providers need to create more accountability for medical errors and patient safety lapses in order to continue improving, according to Dr. Robert M. Wachter, professor and associate chairman of medicine at the University of California, San Francisco.

At the Joint Commission national conference on quality and patient safety, Dr. Wachter offered his perspectives on the status of patient safety in health care 10 years after the publication of the first Institute of Medicine report on the subject ("To Err Is Human: Building a Safer Health System"). He and Dr. Peter J. Pronovost of Johns Hopkins University, Baltimore, published an editorial on the topic shortly after the conference (N. Engl. J. Med. 2009;361:1401–6).

Balancing a culture of "no blame" with a culture of accountability remains a key challenge for providers.

"'No blame' is the right response for innocent slips and mistakes, which turn out to be most errors. But there need to be clear demarcations of blameworthy acts—not just gross incompetence, not just disruptive behavior, but also failure to follow reasonable safety standards," said Dr. Wachter.

Dr. Wachter cited the fact that average hand-washing compliance rates continue to hover at only about 50% as an example of the need for more accountability. "I don't believe that is fully a systems problem," he said. Part of the problem is that "there have been no penalties for transgressions."

Although accountability is essential, "my guess is it will go too far," Dr. Wachter added. "We've created an environment where people don't want to have to talk about errors. We're probably going to have to go too far and then come up with a sweet spot."

Dr. Wachter also commented on other aspects of patient safety:

▸ Regulation. Health care organizations need regulators to set standards, but the challenge is ensuring that these standards truly help organizations improve safety. Until the Joint Commission developed standards for reading back instructions, "virtually none of us thought of doing that on our own," he said.

At the same time, "it is extraordinarily difficult to have a set of rules and standards that apply equally in nuanced areas to organizations that are incredibly different in the way they do business, their financial resources, and their capacity," he said.

For that reason, "regulation is extraordinarily useful to get people moving, but it tends to run out of gas over time," Dr. Wachter said. To illustrate, he cited the Joint Commission's recent decision to remove adherence to medication reconciliation standards as a requirement for accreditation because organizations struggled to develop appropriate processes.

However, having an "outside organization creating rules and standards we must abide by was extraordinarily important in the first 5 years" after the IOM report, he said. Despite some glitches, "the Joint Commission has improved its processes tremendously."

▸ Reporting. "The admonition to report everything is silly," Dr. Wachter said. "Our mistake here was to not be thoughtful about what we are going to do with all of these reports" before requiring them.

However, providers have learned from this experience and begun to think more critically about what should be reported and how the data should be used, he said.

State reporting requirements on the 27 "never events" put forth by the National Quality Forum have led to more focused patient safety efforts.

He added, however, that labeling these adverse occurrences "never" events was a misnomer that put unfair pressure on providers because many of these problems, such as decubitus ulcers, are not even largely preventable based on present science. "It's a politically driven misnomer," he said.

ROSEMONT, ILL. — Despite major patient safety strides during the past decade, health care providers need to create more accountability for medical errors and patient safety lapses in order to continue improving, according to Dr. Robert M. Wachter, professor and associate chairman of medicine at the University of California, San Francisco.

At the Joint Commission national conference on quality and patient safety, Dr. Wachter offered his perspectives on the status of patient safety in health care 10 years after the publication of the first Institute of Medicine report on the subject ("To Err Is Human: Building a Safer Health System"). He and Dr. Peter J. Pronovost of Johns Hopkins University, Baltimore, published an editorial on the topic shortly after the conference (N. Engl. J. Med. 2009;361:1401–6).

Balancing a culture of "no blame" with a culture of accountability remains a key challenge for providers.

"'No blame' is the right response for innocent slips and mistakes, which turn out to be most errors. But there need to be clear demarcations of blameworthy acts—not just gross incompetence, not just disruptive behavior, but also failure to follow reasonable safety standards," said Dr. Wachter.

Dr. Wachter cited the fact that average hand-washing compliance rates continue to hover at only about 50% as an example of the need for more accountability. "I don't believe that is fully a systems problem," he said. Part of the problem is that "there have been no penalties for transgressions."

Although accountability is essential, "my guess is it will go too far," Dr. Wachter added. "We've created an environment where people don't want to have to talk about errors. We're probably going to have to go too far and then come up with a sweet spot."

Dr. Wachter also commented on other aspects of patient safety:

▸ Regulation. Health care organizations need regulators to set standards, but the challenge is ensuring that these standards truly help organizations improve safety. Until the Joint Commission developed standards for reading back instructions, "virtually none of us thought of doing that on our own," he said.

At the same time, "it is extraordinarily difficult to have a set of rules and standards that apply equally in nuanced areas to organizations that are incredibly different in the way they do business, their financial resources, and their capacity," he said.

For that reason, "regulation is extraordinarily useful to get people moving, but it tends to run out of gas over time," Dr. Wachter said. To illustrate, he cited the Joint Commission's recent decision to remove adherence to medication reconciliation standards as a requirement for accreditation because organizations struggled to develop appropriate processes.

However, having an "outside organization creating rules and standards we must abide by was extraordinarily important in the first 5 years" after the IOM report, he said. Despite some glitches, "the Joint Commission has improved its processes tremendously."

▸ Reporting. "The admonition to report everything is silly," Dr. Wachter said. "Our mistake here was to not be thoughtful about what we are going to do with all of these reports" before requiring them.

However, providers have learned from this experience and begun to think more critically about what should be reported and how the data should be used, he said.

State reporting requirements on the 27 "never events" put forth by the National Quality Forum have led to more focused patient safety efforts.

He added, however, that labeling these adverse occurrences "never" events was a misnomer that put unfair pressure on providers because many of these problems, such as decubitus ulcers, are not even largely preventable based on present science. "It's a politically driven misnomer," he said.

ROSEMONT, ILL. — Despite major strides in patient safety during the past decade, health care providers need to create more accountability for medical errors and safety lapses in order to continue improving, according to Dr. Robert M. Wachter, professor and associate chairman of medicine at the University of California, San Francisco.

At the Joint Commission national conference on quality and patient safety, Dr. Wachter offered his thoughts on the status of patient safety in health care 10 years after the publication of the first Institute of Medicine report on the subject (To Err Is Human: Building a Safer Health System). He and Dr. Peter J. Pronovost of Johns Hopkins University, Baltimore, published an editorial on the topic shortly after the conference (N. Engl. J. Med. 2009; 361:1401–6). Balancing a culture of “no blame” with a culture of accountability remains a key challenge for providers. While it's true that “most errors are committed by caring, competent people trying hard to get it right … the system produces low-quality, unsafe, unreliable care partly because there's been no business case to do otherwise,” he said.

“'No blame' is the right response for innocent slips and mistakes, which turn out to be most errors. But there need to be clear demarcations of blameworthy acts—not just gross incompetence, not just disruptive behavior, but also failure to follow reasonable safety standards,” said Dr. Wachter. He cited the fact that average hand-washing compliance rates continue to hover at only about 50%. “I don't believe that is fully a systems problem,” he said. Part of the problem is that “there have been no penalties for transgressions.”

Although accountability is essential, “my guess is it will go too far,” Dr. Wachter added. “We've created an environment where people don't want to have to talk about errors. We're probably going to have to go too far and then come up with a sweet spot.”

Dr. Wachter also commented on the following aspects of patient safety:

▸ Regulation. Health care organizations need regulators to set standards, but the challenge is ensuring that these standards truly help organizations improve safety. Until the Joint Commission developed standards for reading back instructions, “virtually none of us thought of doing that on our own,” he said.

At the same time, it is difficult to have a set of rules that apply equally in nuanced areas to organizations that are very different in the way they do business, he noted. For that reason, regulation is “useful to get people moving, but it tends to run out of gas over time.”

However, having an “outside organization creating rules and standards we must abide by was extraordinarily important in the first 5 years” after the IOM report, he said. Despite some glitches, “the Joint Commission has improved its processes tremendously” and recently made an important step in the right direction with the creation of the Center for Transforming Healthcare.

▸ Reporting. “The admonition to report everything is silly,” Dr. Wachter said. “Our mistake here was to not be thoughtful about what we are going to do with all of these reports” before requiring them. Providers have begun to think more critically about what should be reported and how the data should be used.

State reporting requirements on the 27 “never events” put forth by the National Quality Forum have led to more focused patient safety efforts. “Until the state reporting system, our process [at UCSF] for doing root cause analysis was pretty haphazard and ad lib,” he noted. Now the institution holds a weekly 2-hour root cause analysis meeting attended by the same group of leaders.

▸ Information technology. Health care providers have developed a more robust, less naive understanding of the role of health care information technology in patient safety and now realize that it is not a panacea. Improvement efforts are not nearly as effective “if we just do the computer piece but don't educate people,” Dr. Wachter said.

Still, “even though we've got plenty of room to go, I think we should all be proud” of what has been accomplished in the past 10 years, he said.

ROSEMONT, ILL. — Despite major strides in patient safety during the past decade, health care providers need to create more accountability for medical errors and safety lapses in order to continue improving, according to Dr. Robert M. Wachter, professor and associate chairman of medicine at the University of California, San Francisco.

At the Joint Commission national conference on quality and patient safety, Dr. Wachter offered his thoughts on the status of patient safety in health care 10 years after the publication of the first Institute of Medicine report on the subject (To Err Is Human: Building a Safer Health System). He and Dr. Peter J. Pronovost of Johns Hopkins University, Baltimore, published an editorial on the topic shortly after the conference (N. Engl. J. Med. 2009; 361:1401–6). Balancing a culture of “no blame” with a culture of accountability remains a key challenge for providers. While it's true that “most errors are committed by caring, competent people trying hard to get it right … the system produces low-quality, unsafe, unreliable care partly because there's been no business case to do otherwise,” he said.

“'No blame' is the right response for innocent slips and mistakes, which turn out to be most errors. But there need to be clear demarcations of blameworthy acts—not just gross incompetence, not just disruptive behavior, but also failure to follow reasonable safety standards,” said Dr. Wachter. He cited the fact that average hand-washing compliance rates continue to hover at only about 50%. “I don't believe that is fully a systems problem,” he said. Part of the problem is that “there have been no penalties for transgressions.”

Although accountability is essential, “my guess is it will go too far,” Dr. Wachter added. “We've created an environment where people don't want to have to talk about errors. We're probably going to have to go too far and then come up with a sweet spot.”

Dr. Wachter also commented on the following aspects of patient safety:

▸ Regulation. Health care organizations need regulators to set standards, but the challenge is ensuring that these standards truly help organizations improve safety. Until the Joint Commission developed standards for reading back instructions, “virtually none of us thought of doing that on our own,” he said.

At the same time, it is difficult to have a set of rules that apply equally in nuanced areas to organizations that are very different in the way they do business, he noted. For that reason, regulation is “useful to get people moving, but it tends to run out of gas over time.”

However, having an “outside organization creating rules and standards we must abide by was extraordinarily important in the first 5 years” after the IOM report, he said. Despite some glitches, “the Joint Commission has improved its processes tremendously” and recently made an important step in the right direction with the creation of the Center for Transforming Healthcare.

▸ Reporting. “The admonition to report everything is silly,” Dr. Wachter said. “Our mistake here was to not be thoughtful about what we are going to do with all of these reports” before requiring them. Providers have begun to think more critically about what should be reported and how the data should be used.

State reporting requirements on the 27 “never events” put forth by the National Quality Forum have led to more focused patient safety efforts. “Until the state reporting system, our process [at UCSF] for doing root cause analysis was pretty haphazard and ad lib,” he noted. Now the institution holds a weekly 2-hour root cause analysis meeting attended by the same group of leaders.

▸ Information technology. Health care providers have developed a more robust, less naive understanding of the role of health care information technology in patient safety and now realize that it is not a panacea. Improvement efforts are not nearly as effective “if we just do the computer piece but don't educate people,” Dr. Wachter said.

Still, “even though we've got plenty of room to go, I think we should all be proud” of what has been accomplished in the past 10 years, he said.

ROSEMONT, ILL. — Despite major strides in patient safety during the past decade, health care providers need to create more accountability for medical errors and safety lapses in order to continue improving, according to Dr. Robert M. Wachter, professor and associate chairman of medicine at the University of California, San Francisco.

At the Joint Commission national conference on quality and patient safety, Dr. Wachter offered his thoughts on the status of patient safety in health care 10 years after the publication of the first Institute of Medicine report on the subject (To Err Is Human: Building a Safer Health System). He and Dr. Peter J. Pronovost of Johns Hopkins University, Baltimore, published an editorial on the topic shortly after the conference (N. Engl. J. Med. 2009; 361:1401–6). Balancing a culture of “no blame” with a culture of accountability remains a key challenge for providers. While it's true that “most errors are committed by caring, competent people trying hard to get it right … the system produces low-quality, unsafe, unreliable care partly because there's been no business case to do otherwise,” he said.

“'No blame' is the right response for innocent slips and mistakes, which turn out to be most errors. But there need to be clear demarcations of blameworthy acts—not just gross incompetence, not just disruptive behavior, but also failure to follow reasonable safety standards,” said Dr. Wachter. He cited the fact that average hand-washing compliance rates continue to hover at only about 50%. “I don't believe that is fully a systems problem,” he said. Part of the problem is that “there have been no penalties for transgressions.”

Although accountability is essential, “my guess is it will go too far,” Dr. Wachter added. “We've created an environment where people don't want to have to talk about errors. We're probably going to have to go too far and then come up with a sweet spot.”

Dr. Wachter also commented on the following aspects of patient safety:

▸ Regulation. Health care organizations need regulators to set standards, but the challenge is ensuring that these standards truly help organizations improve safety. Until the Joint Commission developed standards for reading back instructions, “virtually none of us thought of doing that on our own,” he said.

At the same time, it is difficult to have a set of rules that apply equally in nuanced areas to organizations that are very different in the way they do business, he noted. For that reason, regulation is “useful to get people moving, but it tends to run out of gas over time.”

However, having an “outside organization creating rules and standards we must abide by was extraordinarily important in the first 5 years” after the IOM report, he said. Despite some glitches, “the Joint Commission has improved its processes tremendously” and recently made an important step in the right direction with the creation of the Center for Transforming Healthcare.

▸ Reporting. “The admonition to report everything is silly,” Dr. Wachter said. “Our mistake here was to not be thoughtful about what we are going to do with all of these reports” before requiring them. Providers have begun to think more critically about what should be reported and how the data should be used.

State reporting requirements on the 27 “never events” put forth by the National Quality Forum have led to more focused patient safety efforts. “Until the state reporting system, our process [at UCSF] for doing root cause analysis was pretty haphazard and ad lib,” he noted. Now the institution holds a weekly 2-hour root cause analysis meeting attended by the same group of leaders.

▸ Information technology. Health care providers have developed a more robust, less naive understanding of the role of health care information technology in patient safety and now realize that it is not a panacea. Improvement efforts are not nearly as effective “if we just do the computer piece but don't educate people,” Dr. Wachter said.

Still, “even though we've got plenty of room to go, I think we should all be proud” of what has been accomplished in the past 10 years, he said.

CHICAGO — A five-factor scoring system that identifies women who nearly die from obstetric morbidity could potentially offer a more meaningful way to measure maternal health care quality between institutions, according to Dr. Whitney You of Northwestern University's Feinberg School of Medicine in Chicago.

Obstetric mortality has lost most of its value as a measure of maternal health care quality because it is so rare now in the United States, Dr. You said at the annual research meeting of AcademyHealth.

As an objective measure of near-miss obstetric morbidity, the scoring system could hold potential as an outcome measure for hospital case review as well as a reproducible maternal health measure for epidemiologic research aimed at identifying trends and risk factors, she said.

“I'm hoping to use it to figure out who is at greatest risk … where the disparity lies and why,” she said in an interview. Morbidity covers a range from mild fever to near death. “Where is that level where women are very ill, the next step before death?” she asked.

In their study, Dr. You and colleagues used ICD-9 codes to identify 815 women with a high potential for significant obstetric morbidity in a high-volume, urban, tertiary care center over a 2-year period (2001-2002). A maternal-fetal medicine specialist categorized cases according to clinical impression of degree of morbidity: no morbidity (23%), minor morbidity (52%), severe morbidity (19%), and near-miss morbidity (5%), Dr. You explained. The cases then were scored using the five-factor weighted scoring system, in which a score of 8 or more is considered a case of near-miss morbidity. (See table.)

Use of the five-factor scoring system revealed a near-miss obstetric morbidity rate of 4.2% (34 patients). The weighted scoring system showed a 63% sensitivity rate for near-miss morbidity, 99% specificity, positive predictive value of 71% and negative predictive value of 98%, according to results from a poster Dr. You presented at a meeting.

The study is the second to validate the Geller scoring system, developed by Dr. Stacie E. Geller of the University of Illinois at Chicago. “Most of the work has been done with a population at UIC. We wanted to see what would happen with a different population,” Dr. You said.

In Dr. Geller's original work, five clinical factors (organ system failure, ICU admission, transfusion of more than 3 units, extended intubation for more than 12 hours and surgical intervention) were grouped into several scoring system alternatives. A scoring system based on all five factors showed the highest specificity (93%), but even a four-factor system, which eliminated organ system failure, achieved a specificity of 78% (J. Clin. Epidemiol. 2004;57:716-20).

Additional studies can help determine whether other factors could be added to identify cases of near-miss morbidity missed in this investigation, Dr. You noted.

In this study, a single maternal-fetal medicine provider reviewed all the cases. Since then, an obstetric anesthesiologist and another experienced maternal-fetal medicine specialist have reviewed the cases as well. Dr. You and her associates plan to calculate sensitivity and specificity based on these additional reviews. “After we get that information, we can decide if it can be a good tool to use in other settings,” she said.

Adapting the scoring system to other types of institutions presents a key challenge.

“We need to figure out how it works in a rural setting or community hospital,” she said. These smaller facilities often refer severely ill patients to tertiary care institutions, “so they may never get a patient that needs multiple transfusions or intubation for an extended time. Our hope is to level the grading system, just because it's so hard to compare one hospital to another.”

Dr. You conducted this study while she was a National Research Service Award postdoctoral fellow at the Institute for Healthcare Studies under an award from the Agency for Healthcare Research and Quality.

ELSEVIER GLOBAL MEDICAL NEWS

CHICAGO — A five-factor scoring system that identifies women who nearly die from obstetric morbidity could potentially offer a more meaningful way to measure maternal health care quality between institutions, according to Dr. Whitney You of Northwestern University's Feinberg School of Medicine in Chicago.

Obstetric mortality has lost most of its value as a measure of maternal health care quality because it is so rare now in the United States, Dr. You said at the annual research meeting of AcademyHealth.

As an objective measure of near-miss obstetric morbidity, the scoring system could hold potential as an outcome measure for hospital case review as well as a reproducible maternal health measure for epidemiologic research aimed at identifying trends and risk factors, she said.

“I'm hoping to use it to figure out who is at greatest risk … where the disparity lies and why,” she said in an interview. Morbidity covers a range from mild fever to near death. “Where is that level where women are very ill, the next step before death?” she asked.

In their study, Dr. You and colleagues used ICD-9 codes to identify 815 women with a high potential for significant obstetric morbidity in a high-volume, urban, tertiary care center over a 2-year period (2001-2002). A maternal-fetal medicine specialist categorized cases according to clinical impression of degree of morbidity: no morbidity (23%), minor morbidity (52%), severe morbidity (19%), and near-miss morbidity (5%), Dr. You explained. The cases then were scored using the five-factor weighted scoring system, in which a score of 8 or more is considered a case of near-miss morbidity. (See table.)

Use of the five-factor scoring system revealed a near-miss obstetric morbidity rate of 4.2% (34 patients). The weighted scoring system showed a 63% sensitivity rate for near-miss morbidity, 99% specificity, positive predictive value of 71% and negative predictive value of 98%, according to results from a poster Dr. You presented at a meeting.

The study is the second to validate the Geller scoring system, developed by Dr. Stacie E. Geller of the University of Illinois at Chicago. “Most of the work has been done with a population at UIC. We wanted to see what would happen with a different population,” Dr. You said.

In Dr. Geller's original work, five clinical factors (organ system failure, ICU admission, transfusion of more than 3 units, extended intubation for more than 12 hours and surgical intervention) were grouped into several scoring system alternatives. A scoring system based on all five factors showed the highest specificity (93%), but even a four-factor system, which eliminated organ system failure, achieved a specificity of 78% (J. Clin. Epidemiol. 2004;57:716-20).

Additional studies can help determine whether other factors could be added to identify cases of near-miss morbidity missed in this investigation, Dr. You noted.

In this study, a single maternal-fetal medicine provider reviewed all the cases. Since then, an obstetric anesthesiologist and another experienced maternal-fetal medicine specialist have reviewed the cases as well. Dr. You and her associates plan to calculate sensitivity and specificity based on these additional reviews. “After we get that information, we can decide if it can be a good tool to use in other settings,” she said.

Adapting the scoring system to other types of institutions presents a key challenge.

“We need to figure out how it works in a rural setting or community hospital,” she said. These smaller facilities often refer severely ill patients to tertiary care institutions, “so they may never get a patient that needs multiple transfusions or intubation for an extended time. Our hope is to level the grading system, just because it's so hard to compare one hospital to another.”

Dr. You conducted this study while she was a National Research Service Award postdoctoral fellow at the Institute for Healthcare Studies under an award from the Agency for Healthcare Research and Quality.

ELSEVIER GLOBAL MEDICAL NEWS

CHICAGO — A five-factor scoring system that identifies women who nearly die from obstetric morbidity could potentially offer a more meaningful way to measure maternal health care quality between institutions, according to Dr. Whitney You of Northwestern University's Feinberg School of Medicine in Chicago.

Obstetric mortality has lost most of its value as a measure of maternal health care quality because it is so rare now in the United States, Dr. You said at the annual research meeting of AcademyHealth.

As an objective measure of near-miss obstetric morbidity, the scoring system could hold potential as an outcome measure for hospital case review as well as a reproducible maternal health measure for epidemiologic research aimed at identifying trends and risk factors, she said.

“I'm hoping to use it to figure out who is at greatest risk … where the disparity lies and why,” she said in an interview. Morbidity covers a range from mild fever to near death. “Where is that level where women are very ill, the next step before death?” she asked.

In their study, Dr. You and colleagues used ICD-9 codes to identify 815 women with a high potential for significant obstetric morbidity in a high-volume, urban, tertiary care center over a 2-year period (2001-2002). A maternal-fetal medicine specialist categorized cases according to clinical impression of degree of morbidity: no morbidity (23%), minor morbidity (52%), severe morbidity (19%), and near-miss morbidity (5%), Dr. You explained. The cases then were scored using the five-factor weighted scoring system, in which a score of 8 or more is considered a case of near-miss morbidity. (See table.)

Use of the five-factor scoring system revealed a near-miss obstetric morbidity rate of 4.2% (34 patients). The weighted scoring system showed a 63% sensitivity rate for near-miss morbidity, 99% specificity, positive predictive value of 71% and negative predictive value of 98%, according to results from a poster Dr. You presented at a meeting.

The study is the second to validate the Geller scoring system, developed by Dr. Stacie E. Geller of the University of Illinois at Chicago. “Most of the work has been done with a population at UIC. We wanted to see what would happen with a different population,” Dr. You said.

In Dr. Geller's original work, five clinical factors (organ system failure, ICU admission, transfusion of more than 3 units, extended intubation for more than 12 hours and surgical intervention) were grouped into several scoring system alternatives. A scoring system based on all five factors showed the highest specificity (93%), but even a four-factor system, which eliminated organ system failure, achieved a specificity of 78% (J. Clin. Epidemiol. 2004;57:716-20).

Additional studies can help determine whether other factors could be added to identify cases of near-miss morbidity missed in this investigation, Dr. You noted.

In this study, a single maternal-fetal medicine provider reviewed all the cases. Since then, an obstetric anesthesiologist and another experienced maternal-fetal medicine specialist have reviewed the cases as well. Dr. You and her associates plan to calculate sensitivity and specificity based on these additional reviews. “After we get that information, we can decide if it can be a good tool to use in other settings,” she said.

Adapting the scoring system to other types of institutions presents a key challenge.

“We need to figure out how it works in a rural setting or community hospital,” she said. These smaller facilities often refer severely ill patients to tertiary care institutions, “so they may never get a patient that needs multiple transfusions or intubation for an extended time. Our hope is to level the grading system, just because it's so hard to compare one hospital to another.”

Dr. You conducted this study while she was a National Research Service Award postdoctoral fellow at the Institute for Healthcare Studies under an award from the Agency for Healthcare Research and Quality.

ELSEVIER GLOBAL MEDICAL NEWS

CHICAGO — A primary care delivery model for older patients with multiple chronic illnesses increased physician understanding of the patients' clinical conditions and improved communication with patients and families, in a randomized trial at eight primary care practices.

The study findings, presented in a poster at the annual research meeting of AcademyHealth, add to a growing body of evidence that the model can improve outcomes, reduce costs, and enhance the quality of care for chronically ill older adults.

In the Guided Care model—an initiative developed by the Johns Hopkins Bloomberg School of Public Health, Baltimore—a Guided Care nurse works with three or four primary care physicians and a targeted population of 50-60 patients. The nurse receives additional training in patient education and coaching and in chronic disease management.

The nurse also develops a comprehensive care plan for each patient based on medical conditions, the home environment, and individual patient goals; monitors patients monthly; coordinates transitions between providers and sites of care; educates and supports caregivers; keeps an electronic health record; and provides physicians with detailed updates.

“Physicians don't have a lot of time to do these things for every one of their complex patients,” Jill Marsteller, Ph.D., noted in an interview. “The nurse helps very much in terms of having a close personal relationship with these people who are so complex, and can provide them with the guidance that they need to access community resources and help them keep track of all the different visits.” The nurse also summarizes the information for the physician.

The nurse visits the patient's home when Guided Care begins, making a connection that can yield valuable insights that might not come out in a medical interview.

A randomized trial of Guided Care has shown that this heightened attention to patients and support for physicians improves outcomes and increases patient satisfaction with their care. In the trial, 49 primary care physicians at eight community-based medical practices in the Baltimore and Washington region and 904 of their patients aged 65 years or older with multiple comorbidities were randomly assigned to Guided Care or usual care.

After 6 months, patients receiving Guided Care were twice as likely as patients receiving usual care to rate the quality of their care as high (J. Gerontol. A Biol. Sci. Med. Sci. 2008;63:321-7). After 8 months, Guided Care patients spent 24% fewer days in the hospital, and had 29% fewer home health visits and 15% fewer emergency department visits (Am. J. Manag. Care 2009;15:555-9).

The most recent phase of the study, presented at the AcademyHealth meeting, focused on physician satisfaction and perceptions after participating in the project for 1 year.

On a six-point scale ranging from “very dissatisfied” to “very satisfied,” physicians in the control group gave mean ratings of 4.25 and 3.94, at baseline and at 1 year, respectively, for satisfaction with patient/family communication, while physicians in the Guided Care group gave mean ratings of 4.03 and 4.40. On a four-point scale measuring clinical knowledge of patients, with responses ranging from “definitely not” to “definitely,” physicians in the control group gave mean ratings of 2.70 and 2.77 at baseline and 1 year, respectively, while physicians in the Guided Care group gave mean ratings of 2.85 and 3.17.

The study was funded by the John A. Hartford Foundation and the Roger C. Lipitz Center for Integrated Health Care of the Bloomberg School of Public Health.

CHICAGO — A primary care delivery model for older patients with multiple chronic illnesses increased physician understanding of the patients' clinical conditions and improved communication with patients and families, in a randomized trial at eight primary care practices.

The study findings, presented in a poster at the annual research meeting of AcademyHealth, add to a growing body of evidence that the model can improve outcomes, reduce costs, and enhance the quality of care for chronically ill older adults.

In the Guided Care model—an initiative developed by the Johns Hopkins Bloomberg School of Public Health, Baltimore—a Guided Care nurse works with three or four primary care physicians and a targeted population of 50-60 patients. The nurse receives additional training in patient education and coaching and in chronic disease management.

The nurse also develops a comprehensive care plan for each patient based on medical conditions, the home environment, and individual patient goals; monitors patients monthly; coordinates transitions between providers and sites of care; educates and supports caregivers; keeps an electronic health record; and provides physicians with detailed updates.

“Physicians don't have a lot of time to do these things for every one of their complex patients,” Jill Marsteller, Ph.D., noted in an interview. “The nurse helps very much in terms of having a close personal relationship with these people who are so complex, and can provide them with the guidance that they need to access community resources and help them keep track of all the different visits.” The nurse also summarizes the information for the physician.

The nurse visits the patient's home when Guided Care begins, making a connection that can yield valuable insights that might not come out in a medical interview.

A randomized trial of Guided Care has shown that this heightened attention to patients and support for physicians improves outcomes and increases patient satisfaction with their care. In the trial, 49 primary care physicians at eight community-based medical practices in the Baltimore and Washington region and 904 of their patients aged 65 years or older with multiple comorbidities were randomly assigned to Guided Care or usual care.

After 6 months, patients receiving Guided Care were twice as likely as patients receiving usual care to rate the quality of their care as high (J. Gerontol. A Biol. Sci. Med. Sci. 2008;63:321-7). After 8 months, Guided Care patients spent 24% fewer days in the hospital, and had 29% fewer home health visits and 15% fewer emergency department visits (Am. J. Manag. Care 2009;15:555-9).

The most recent phase of the study, presented at the AcademyHealth meeting, focused on physician satisfaction and perceptions after participating in the project for 1 year.

On a six-point scale ranging from “very dissatisfied” to “very satisfied,” physicians in the control group gave mean ratings of 4.25 and 3.94, at baseline and at 1 year, respectively, for satisfaction with patient/family communication, while physicians in the Guided Care group gave mean ratings of 4.03 and 4.40. On a four-point scale measuring clinical knowledge of patients, with responses ranging from “definitely not” to “definitely,” physicians in the control group gave mean ratings of 2.70 and 2.77 at baseline and 1 year, respectively, while physicians in the Guided Care group gave mean ratings of 2.85 and 3.17.

The study was funded by the John A. Hartford Foundation and the Roger C. Lipitz Center for Integrated Health Care of the Bloomberg School of Public Health.

CHICAGO — A primary care delivery model for older patients with multiple chronic illnesses increased physician understanding of the patients' clinical conditions and improved communication with patients and families, in a randomized trial at eight primary care practices.

The study findings, presented in a poster at the annual research meeting of AcademyHealth, add to a growing body of evidence that the model can improve outcomes, reduce costs, and enhance the quality of care for chronically ill older adults.

In the Guided Care model—an initiative developed by the Johns Hopkins Bloomberg School of Public Health, Baltimore—a Guided Care nurse works with three or four primary care physicians and a targeted population of 50-60 patients. The nurse receives additional training in patient education and coaching and in chronic disease management.

The nurse also develops a comprehensive care plan for each patient based on medical conditions, the home environment, and individual patient goals; monitors patients monthly; coordinates transitions between providers and sites of care; educates and supports caregivers; keeps an electronic health record; and provides physicians with detailed updates.

“Physicians don't have a lot of time to do these things for every one of their complex patients,” Jill Marsteller, Ph.D., noted in an interview. “The nurse helps very much in terms of having a close personal relationship with these people who are so complex, and can provide them with the guidance that they need to access community resources and help them keep track of all the different visits.” The nurse also summarizes the information for the physician.

The nurse visits the patient's home when Guided Care begins, making a connection that can yield valuable insights that might not come out in a medical interview.

A randomized trial of Guided Care has shown that this heightened attention to patients and support for physicians improves outcomes and increases patient satisfaction with their care. In the trial, 49 primary care physicians at eight community-based medical practices in the Baltimore and Washington region and 904 of their patients aged 65 years or older with multiple comorbidities were randomly assigned to Guided Care or usual care.

After 6 months, patients receiving Guided Care were twice as likely as patients receiving usual care to rate the quality of their care as high (J. Gerontol. A Biol. Sci. Med. Sci. 2008;63:321-7). After 8 months, Guided Care patients spent 24% fewer days in the hospital, and had 29% fewer home health visits and 15% fewer emergency department visits (Am. J. Manag. Care 2009;15:555-9).

The most recent phase of the study, presented at the AcademyHealth meeting, focused on physician satisfaction and perceptions after participating in the project for 1 year.

On a six-point scale ranging from “very dissatisfied” to “very satisfied,” physicians in the control group gave mean ratings of 4.25 and 3.94, at baseline and at 1 year, respectively, for satisfaction with patient/family communication, while physicians in the Guided Care group gave mean ratings of 4.03 and 4.40. On a four-point scale measuring clinical knowledge of patients, with responses ranging from “definitely not” to “definitely,” physicians in the control group gave mean ratings of 2.70 and 2.77 at baseline and 1 year, respectively, while physicians in the Guided Care group gave mean ratings of 2.85 and 3.17.

The study was funded by the John A. Hartford Foundation and the Roger C. Lipitz Center for Integrated Health Care of the Bloomberg School of Public Health.

CHICAGO – The Medicare Physician Group Practice Demonstration achieved modest cost savings and quality enhancements in the project's first 2 performance years, researchers reported at the annual research meeting of AcademyHealth. Data released in August reinforce that finding.

The project involves 10 large, geographically diverse physician group practices with a total of 5,000 physicians caring for 200,000 Medicare fee-for-service beneficiaries. The practices include multispecialty groups, integrated delivery systems, faculty groups, and a physician network.

During each year of the project, each group was retroactively assigned a population of Medicare beneficiaries, with an average of 20,000 patients per group (range 10,000-37,000). Each group was held accountable for total Part A and Part B expenditures for these patients.

Patients had complete freedom of choice in providers and were not required to receive care through the participating group practice. Only patients who received most of their outpatient evaluation and management for the year from the group practice were assigned to the group. Groups that kept increases in expenditures below 2 percentage points of their target growth rate shared up to 80% of the savings; Medicare retained 20%.

The group practices assumed all business risks associated with investments related to their participation, and there was no guarantee of savings.

“Savings are a function of the ability of the group to control growth in Medicare spending as well as changes in [health] status of their assigned population over time relative to their local market,” said John Pilotte, a senior research analyst at the Centers for Medicare and Medicaid Services. The groups were free to make whatever investments and enhancements they felt were necessary to reach their quality and efficiency goals.

In the first year of the demonstration, two participating group practices earned a total performance payment of $7.3 million and two lost a total of $1.5 million, Gregory Pope of RTI International in Waltham, Mass., a nonprofit research and development firm working with the CMS, reported at the meeting. In the second year, four groups shared a total payment of $13.8 million and one lost $2 million. Savings to Medicare totaled $677,000 and $1.6 million for the first and second years, respectively.

Results for the third year were announced in August; five physician groups will receive performance payments totaling $25.3 million as part of their share of $32.3 million of savings generated for the Medicare Trust Funds in that year, the CMS announced.

Quality was assessed by the groups' adherence to 27 measures as indicated by Medicare claims and clinical records data. The measures, developed by the CMS in collaboration with the American Medical Association and the National Committee for Quality Assurance, covered heart failure, diabetes, coronary artery disease, hypertension, and preventive care.

Two group practices complied with 10 of the quality markers in performance year one, while five groups complied with all 27 quality markers in the second year, said Musetta Leung of RTI International.

In the second year, all group practices met all of their quality targets for heart failure and coronary artery disease. Achieving the diabetes-related quality measures remained a challenge. Still, second-year performance data indicated significant improvements, she said.

In the third year, all 10 groups achieved benchmark performance on at least 28 of the 32 measures reported, according to the CMS. Two groups–Geisinger Clinic in Danville, Pa., and Park Nicollet Health Services in St. Louis Park, Minn.–achieved benchmark performance on all 32 performance measures.

Over the first 3 years of the demonstration, the physician groups increased their quality scores an average of 10 percentage points on 10 diabetes measures, 11 points on 10 heart failure measures, 6 points on 7 coronary artery disease measures, 10 points on 2 cancer screening measures, and 1 percentage point on 3 hypertension measures.

Additional research is needed to determine the keys to success, according to Mr. Pilotte of the CMS. “Trying to figure out how to make this work in a program that processes over 1 billion claims each year is not a small feat…. It takes a while to get these projects up and running both from our side of the house and the provider side.”

Although the group practices generally have sophisticated health information management systems and dedicated information technology leadership, “even that doesn't seem to be enough to control growth in expenditures…. There are probably other things going on that we haven't fully identified that are necessary to be successful.”

Overall, “the results show we're moving in the right direction, but bringing in expenditure growth under target is challenging under the existing target-setting methodology even for these large organizations,” he said.

One key lesson learned so far during the demonstration is that “leadership and champions within the organization are really important,” Mr. Pilotte said. “All of these groups have someone who is on point for monitoring and reporting the quality metrics to us every year and developing mechanisms … to be able to capture the information.”

Joyce Frieden contributed to this report.

CHICAGO – The Medicare Physician Group Practice Demonstration achieved modest cost savings and quality enhancements in the project's first 2 performance years, researchers reported at the annual research meeting of AcademyHealth. Data released in August reinforce that finding.

The project involves 10 large, geographically diverse physician group practices with a total of 5,000 physicians caring for 200,000 Medicare fee-for-service beneficiaries. The practices include multispecialty groups, integrated delivery systems, faculty groups, and a physician network.

During each year of the project, each group was retroactively assigned a population of Medicare beneficiaries, with an average of 20,000 patients per group (range 10,000-37,000). Each group was held accountable for total Part A and Part B expenditures for these patients.

Patients had complete freedom of choice in providers and were not required to receive care through the participating group practice. Only patients who received most of their outpatient evaluation and management for the year from the group practice were assigned to the group. Groups that kept increases in expenditures below 2 percentage points of their target growth rate shared up to 80% of the savings; Medicare retained 20%.

The group practices assumed all business risks associated with investments related to their participation, and there was no guarantee of savings.

“Savings are a function of the ability of the group to control growth in Medicare spending as well as changes in [health] status of their assigned population over time relative to their local market,” said John Pilotte, a senior research analyst at the Centers for Medicare and Medicaid Services. The groups were free to make whatever investments and enhancements they felt were necessary to reach their quality and efficiency goals.

In the first year of the demonstration, two participating group practices earned a total performance payment of $7.3 million and two lost a total of $1.5 million, Gregory Pope of RTI International in Waltham, Mass., a nonprofit research and development firm working with the CMS, reported at the meeting. In the second year, four groups shared a total payment of $13.8 million and one lost $2 million. Savings to Medicare totaled $677,000 and $1.6 million for the first and second years, respectively.

Results for the third year were announced in August; five physician groups will receive performance payments totaling $25.3 million as part of their share of $32.3 million of savings generated for the Medicare Trust Funds in that year, the CMS announced.

Quality was assessed by the groups' adherence to 27 measures as indicated by Medicare claims and clinical records data. The measures, developed by the CMS in collaboration with the American Medical Association and the National Committee for Quality Assurance, covered heart failure, diabetes, coronary artery disease, hypertension, and preventive care.

Two group practices complied with 10 of the quality markers in performance year one, while five groups complied with all 27 quality markers in the second year, said Musetta Leung of RTI International.

In the second year, all group practices met all of their quality targets for heart failure and coronary artery disease. Achieving the diabetes-related quality measures remained a challenge. Still, second-year performance data indicated significant improvements, she said.

In the third year, all 10 groups achieved benchmark performance on at least 28 of the 32 measures reported, according to the CMS. Two groups–Geisinger Clinic in Danville, Pa., and Park Nicollet Health Services in St. Louis Park, Minn.–achieved benchmark performance on all 32 performance measures.

Over the first 3 years of the demonstration, the physician groups increased their quality scores an average of 10 percentage points on 10 diabetes measures, 11 points on 10 heart failure measures, 6 points on 7 coronary artery disease measures, 10 points on 2 cancer screening measures, and 1 percentage point on 3 hypertension measures.

Additional research is needed to determine the keys to success, according to Mr. Pilotte of the CMS. “Trying to figure out how to make this work in a program that processes over 1 billion claims each year is not a small feat…. It takes a while to get these projects up and running both from our side of the house and the provider side.”

Although the group practices generally have sophisticated health information management systems and dedicated information technology leadership, “even that doesn't seem to be enough to control growth in expenditures…. There are probably other things going on that we haven't fully identified that are necessary to be successful.”

Overall, “the results show we're moving in the right direction, but bringing in expenditure growth under target is challenging under the existing target-setting methodology even for these large organizations,” he said.

One key lesson learned so far during the demonstration is that “leadership and champions within the organization are really important,” Mr. Pilotte said. “All of these groups have someone who is on point for monitoring and reporting the quality metrics to us every year and developing mechanisms … to be able to capture the information.”

Joyce Frieden contributed to this report.

CHICAGO – The Medicare Physician Group Practice Demonstration achieved modest cost savings and quality enhancements in the project's first 2 performance years, researchers reported at the annual research meeting of AcademyHealth. Data released in August reinforce that finding.

The project involves 10 large, geographically diverse physician group practices with a total of 5,000 physicians caring for 200,000 Medicare fee-for-service beneficiaries. The practices include multispecialty groups, integrated delivery systems, faculty groups, and a physician network.

During each year of the project, each group was retroactively assigned a population of Medicare beneficiaries, with an average of 20,000 patients per group (range 10,000-37,000). Each group was held accountable for total Part A and Part B expenditures for these patients.

Patients had complete freedom of choice in providers and were not required to receive care through the participating group practice. Only patients who received most of their outpatient evaluation and management for the year from the group practice were assigned to the group. Groups that kept increases in expenditures below 2 percentage points of their target growth rate shared up to 80% of the savings; Medicare retained 20%.

The group practices assumed all business risks associated with investments related to their participation, and there was no guarantee of savings.

“Savings are a function of the ability of the group to control growth in Medicare spending as well as changes in [health] status of their assigned population over time relative to their local market,” said John Pilotte, a senior research analyst at the Centers for Medicare and Medicaid Services. The groups were free to make whatever investments and enhancements they felt were necessary to reach their quality and efficiency goals.

In the first year of the demonstration, two participating group practices earned a total performance payment of $7.3 million and two lost a total of $1.5 million, Gregory Pope of RTI International in Waltham, Mass., a nonprofit research and development firm working with the CMS, reported at the meeting. In the second year, four groups shared a total payment of $13.8 million and one lost $2 million. Savings to Medicare totaled $677,000 and $1.6 million for the first and second years, respectively.

Results for the third year were announced in August; five physician groups will receive performance payments totaling $25.3 million as part of their share of $32.3 million of savings generated for the Medicare Trust Funds in that year, the CMS announced.

Quality was assessed by the groups' adherence to 27 measures as indicated by Medicare claims and clinical records data. The measures, developed by the CMS in collaboration with the American Medical Association and the National Committee for Quality Assurance, covered heart failure, diabetes, coronary artery disease, hypertension, and preventive care.

Two group practices complied with 10 of the quality markers in performance year one, while five groups complied with all 27 quality markers in the second year, said Musetta Leung of RTI International.

In the second year, all group practices met all of their quality targets for heart failure and coronary artery disease. Achieving the diabetes-related quality measures remained a challenge. Still, second-year performance data indicated significant improvements, she said.

In the third year, all 10 groups achieved benchmark performance on at least 28 of the 32 measures reported, according to the CMS. Two groups–Geisinger Clinic in Danville, Pa., and Park Nicollet Health Services in St. Louis Park, Minn.–achieved benchmark performance on all 32 performance measures.

Over the first 3 years of the demonstration, the physician groups increased their quality scores an average of 10 percentage points on 10 diabetes measures, 11 points on 10 heart failure measures, 6 points on 7 coronary artery disease measures, 10 points on 2 cancer screening measures, and 1 percentage point on 3 hypertension measures.

Additional research is needed to determine the keys to success, according to Mr. Pilotte of the CMS. “Trying to figure out how to make this work in a program that processes over 1 billion claims each year is not a small feat…. It takes a while to get these projects up and running both from our side of the house and the provider side.”

Although the group practices generally have sophisticated health information management systems and dedicated information technology leadership, “even that doesn't seem to be enough to control growth in expenditures…. There are probably other things going on that we haven't fully identified that are necessary to be successful.”

Overall, “the results show we're moving in the right direction, but bringing in expenditure growth under target is challenging under the existing target-setting methodology even for these large organizations,” he said.

One key lesson learned so far during the demonstration is that “leadership and champions within the organization are really important,” Mr. Pilotte said. “All of these groups have someone who is on point for monitoring and reporting the quality metrics to us every year and developing mechanisms … to be able to capture the information.”

Joyce Frieden contributed to this report.