To estimate the direct and indirect costs of type 2 diabetes  mellitus (T2DM) in China in 2007 and project these costs for the year 2030, and to examine and compare the benefits of selected interventions. 

Annual direct costs of medical and nonmedical care and indirect costs of income losses were estimated through case calculation of data from a cross-sectional survey carried out in 4 major Chinese cities from March 2007 to

 The subjects were consecutively recruited T2DM outpatients and inpatients from 20 secondary and tertiary hospitals using selection probability proportional to size sampling. We combined the existing data from  cost-effectiveness studies into the case estimation to examine the benefits of the observed regime of interventions for preventing and treating diabetes.

 Annual direct medical and direct nonmedical costs per case averaged $1320.90 and $180.80, respectively. The mean annual indirect costs of T2DM and its complications were estimated to be $206.10. Based on case numbers in 2007

and projected case numbers in 2030, the direct medical costs of T2DM and its complications were estimated to be $26.0 billion in 2007 and were projected to be $47.2 billion in 2030. 

The results indicated that T2DM consumes a large portion of healthcare expenditures and will continue to place a heavy burden on health budgets in the future. Preventive intervention, screening, and treatment strategies may effectively decrease the incidence and complications of diabetes and therefore save costs. 

The economic burden of type 2 diabetes mellitus (T2DM) makes it an important clinical and public health challenge in China, requiring policymakers to determine both the current and future economic burdens of the disease to formulate effective plans for resource allocation.

T2DM consumes a large portion of health expenditures and will continue to place a heavy burden on healthcare budgets in the future in China.

The excessive costs are largely due to treatment of complications; therefore, the importance of preventing complications through the use of more effective treatment regimens is obvious. 

Preventive interventions, especially lifestyle intervention, in high-risk populations may effectively decrease the incidence and complications of diabetes and therefore save costs.

Nearly 200 million people worldwide (or more than 5% of the global adult population) have diabetes, and this number will increase to 333 million (or 6.3% of the global adult population) by 2025.

 Diabetes affects persons of all ages as well as their families, while also placing heavy economic burdens on national economies and healthcare systems. Globally, the annual direct costs of diabetes, which include costs incurred for patient transport and care in addition to direct medical costs (DMCs), for people age 20 to 79 years are estimated to be at least $129 billion and as much as $241 billion or possibly more.

 The direct costs of diabetes consume from 2.5% to 15.0% of annual healthcare budgets, depending on local prevalence and the sophistication of available treatments. In 2025, diabetes care costs are anticipated to account for between 7% and 13% of the world’s healthcare budget.

In China, an estimated 23.46 million people currently have diabetes

; that number is predicted to increase to 42.30 million by 2030.

 Between 1996 and 2006, the prevalence of type 2 diabetes mellitus (T2DM) increased rapidly in urban China, from 4.58% to 7.67%, and was much higher in major cities (6.1%) than in small cities (3.7%) and rural areas (1.8%).

 A recent study estimated the overall annual direct economic burden of diabetes among China’s urban population in 2002 to be in excess of $2.44 billion.

 In 2003, the estimate was updated to $2.29 billion for the whole population based on data from the 2003 National Health Service Investigation.

 The annual per patient DMCs of healthcare associated with T2DM patients with complications were estimated to be $1798, compared with $484 for those without complications.

 The treatment of chronic diseases is managed by a 3-level hospital system in Chinese urban areas. Under that system, patients chose to visit primary, secondary, or tertiary hospitals on their own, and may overuse healthcare services by selecting unnecessary high-level hospitals for diabetes treatment; it was reported that 997 (6%) tertiary hospitals nationwide consumed 54.73% of China’s total healthcare expenditures.

Many of the costs associated with diabetes and its complications are preventable.

 A large number of interventions (eg, intensive blood glucose and blood pressure control, use of lipid-lowering agents, screening for and treating diabetic retinopathy, active care of feet) are known to be cost-effective, and many of these interventions also are cost-saving,

 although controversy remains regarding the significance of the effect of some strategies (eg, intense glucose control for prevention of microvascular complications).

 In recent years, 4 major trials of T2DM prevention in China,

 have demonstrated that intensive lifestyle interventions involving a combination of diet and physical activity can delay or prevent diabetes among persons at high risk.

The economic burden of diabetes makes it an important clinical and public health challenge. To formulate effective plans for resource allocation, it is necessary to determine both the current and future economic burdens of the disease. The present study provides estimates of these burdens for T2DM patients in China. Cross-sectional data drawn from surveys conducted in 4 Chinese cities were used to help to determine the preventable costs under selected alternative interventions.

We conducted a cross-sectional study in 4 major cities—Shanghai, Beijing, Guangzhou, and Chengdu&mdash;located in eastern, northern, southern, and western China, respectively. The eAppendix (available at 

Subjects were eligible for the study if they (1) met the 1999 World Health Organization diagnosis criteria for T2DM; (2) had received T2DM treatment for at least 1 year; (3) were age 18 years or older; (4) had lived in the city for at least 2 years; and (5) were willing and able to give written informed consent and complete the questionnaire interview.

 The eligible subjects included outpatients and inpatients. Patients in the outpatient group were recruited consecutively from the hospitals’ outpatient clinics from March 2007 to May 2007. Patients were categorized as inpatients if they were recruited from hospital wards from March 2007 to September 2007 using selection probability proportional to size sampling. Enrollment of inpatients involved 3 steps. First, using data from a previously published study that determined the proportions of complications associated with T2DM from a sample of 299 patients living in 7 Chinese urban areas, we obtained the proportion of inpatients with various complications.

 Second, we calculated the sample size of each inpatient group stratified by type of complication, using the proportions in the previous step. Finally, we recruited the subjects consecutively for each group until the predefined sample sizes were met.

The direct costs of illness are those expenditures used to purchase medical supplies and services. In this study, direct costs are further classified as DMCs and direct nonmedical costs. Direct medical costs include all expenditures for treating T2DM and any related complications, including copayment, diagnosis, treatment, diagnostic testing, prescription drugs, and medical supplies. Direct nonmedical costs include costs for services such as transportation for the patient and family members to clinics and costs for taking care of dependents. Indirect costs refer to lost income of patients and their family members, and costs for hiring nurses or care providers.

A complication was defined as an infrequent and unfavorable evolution of a disease or a health condition that was diagnosed after the onset of T2DM.

Annual DMCs were estimated based on payment amounts for outpatient services and hospitalization; annual direct nonmedical costs were obtained from the patient’s self-estimate; annual indirect costs were estimated based on sick leave days taken by patients and by family members caring for patients, average daily income in urban China, and payment for hiring nurses or care providers. Specifically, for outpatients, the annual DMC was computed as the DMC per outpatient visit multiplied by the number of outpatient visits in the past 6 months multiplied by 2 plus annual hospitalization cost. For inpatients, the annual DMC was computed as the annual DMC associated with outpatient visits plus the DMC per hospitalization multiplied by number of hospitalizations in the past 12 months. The currency was adjusted to US dollars using the exchange rate as of June 15, 2007 ($1 = 7.6948 Chinese Yuan).

Study coordinators who majored in clinical medicine or preventive medicine were recruited from each study site and trained by the investigators. The subjects were interviewed face-to-face in the hospitals using a survey designed by the School of Public Health at Fudan University. The survey included questions concerning patients’ demographics, diabetes characteristics, existing complications caused by T2DM, and treatment history, as well as selfestimations of the cost of T2DM and its complications. Written informed consent was obtained from all study subjects. The Ethics Committee of the School of Public Health at Fudan University approved the study.

All data were entered into a Chinese database (EpiData version 3.1; EpiData Association, Odense, Denmark) and transferred into SPSS 16 for statistic analyses (SPSS Inc, Chicago, IL). Categorical variables were presented as counts and percentages.

General Characteristics of the Study Population A total of 2054 subjects were recruited from March 2007 to September 2007. After 14 subjects were excluded for having less than 1 year of T2DM treatment history or not providing cost information, 1524 outpatients and 516 inpatients were eligible for analysis (

). The 2040 subjects ranged in age from 18 to 97 years (average age 64.2 years). More than 90% reported having some kind of medical insurance; about 70% reported a personal income below $260 per month. Among outpatients, 39.4% did not suffer complications; the median number of complications was 1 (range: 0-8). Of the 516 inpatient subjects, only 5.8% did not experience complications; the median number of complications was 3 (range: 0-10).

Because the cost data were skewed, medians and quartiles were reported together with means. The DMC for each hospitalization averaged $2294.10; the DMC per outpatient visit averaged $40.70 (

). Care for T2DM was associated with 7.3 ± 7.2 outpatient visits per outpatient in the past 6 months, and an average of 1.7 ± 1.4 hospitalizations for inpatients in the past 12 months. The DMC was calculated by taking into account the DMCs of the current visit and the number of outpatient visits, plus annual hospitalization costs in the past year; annualizing these sums yielded an estimated mean annual DMC of $1030.10. The corresponding sum for inpatients was $2309 per hospitalization.

Mean direct nonmedical costs were estimated to be $180.80 ($165 and $227.80 for outpatients and inpatients, respectively) annually. Using the 2007 urban China average per capita daily income of $10.87, the annual indirect costs of T2DM and its complications were $206.10, or $99.40 and $515.70 for outpatients and inpatients, respectively.

Economic Burden Estimation and Projection

We applied data from this sample to the Chinese population as a whole. 

 summarizes estimated national annual DMCs for subjects who had differing numbers of complications. For subjects with only 1 complication, more detailed subgroups are provided.

Assuming a prevalence of 23.46 million T2DM patients nationwide,

 the estimated national DMCs for T2DM and its complications were $26.0 billion per year, or 18.2% of China’s total health expenditures ($142.5 billion in 2007).

 Using an estimated prevalence of 42.3 million T2DM patients in the year 2030, annual DMCs were projected to be $47.2 billion. Annual direct nonmedical costs for T2DM and its complications in 2007 were estimated to be $4.0 billion and projected to be $7.2 billion in 2030. Thus, the total direct costs in 2007 were $30.0 billion, which exceeded the 2003 estimate of $22.9 billion.9 Similarly, indirect costs in 2007 were estimated to be $2.2 billion and were projected to be $4.0 billion in 2030. In total, the estimated economic cost of T2DM and its complications was $32.2 billion for 23.46 million T2DM patients in 2007 and projected to be $58.5 billion for 42.3 million T2DM patients in 2030. Estimated Benefits of Intervention for Preventing Diabetes and Its Complications

We examined the benefits of the observed regimen of interventions to prevent and treat diabetes in China (

). The World Bank determined that lifestyle interventions for preventing T2DM could reduce the incidence among persons at high risk by 35% to 58%.13 (High-risk persons were those with either impaired glucose tolerance or T2DM, as estimated from the current cross-sectional study. Impaired glucose tolerance was defined as 2-hour glucose levels of 140-199 mg/dL [7.8/11.0 mmol] on the venous plasma glucose tolerance test.) Assuming that 95% of individuals would adhere to the program, the lifestyle interventions could result in $8.6 billion to $14.2 billion saved in T2DM health expenditures in 2030. Alternative scenarios assuming that China had the same share of indirect costs (as part of the total costs) as the United States

 yield potential total costs savings of $11.4 billion to $18.9 billion and $44.3 billion to $73.4 billion in 2030, respectively. Similarly, if using the World Bank—recommended metformin interventions for preventing T2DM in 17.2 million individuals at high risk, the incidence may be reduced by 25% to 31%, resulting in potential cost savings of $6.1 billion to $7.6 billion in 2030. If the higher US and Latin American and Caribbean shares of indirect costs were used, the interventions would save a projected $8.2 billion to $10.1 billion and $31.6 billion to $39.2 billion, respectively.

Estimated Benefits of Screening for and Treating Diabetes and Its Complications

The effectiveness and benefits of screening for and treating diabetes and its complications vary greatly (

). Screening strategies to prevent T2DM among the general population may save $1.8 billion in 2030, while more specific screening or treatment strategies (ie, blood pressure control, annual screening for microalbuminuria, annual eye examination, and foot care in high-risk persons) may save even more. For example, annual eye examinations targeting 0.6 million individuals with serious vision loss may save $55 million, while foot care targeting 0.2 million persons at high risk of ulcers would save $64 million in 2030. In addition, glycemic control in persons with a glycosylated hemoglobin (A1C) value greater than 9% would reduce microvascular disease by 30% per 1% drop in A1C, thereby saving $1.2 billion in 2007 and $2.2 billion in 2030.

Although numerous cost-of-illness studies have been conducted on the estimated economic burden of diabetes in various countries, this is the first quantitative study to systematically estimate the current economic burden and project the future costs of T2DM in China. In light of increasing diabetes prevalence globally and within China, cost-of-illness studies are urgently needed to increase awareness and facilitate lobbying for resource allocation.

In China, the most recently published analysis—known as the National Health Service Investigation&mdash;estimated the cost of diabetes in 2003 to be $2.29 billion.

 This estimate, however, was obtained by national statistics data and failed to consider the burden incurred by complications, which may have been underestimated if the comorbidity of the various complications of diabetes was not considered. Another study conducted by Chen et al in 2002 estimated the urban burden from a relatively small sample of 299 patients living in 7 Chinese urban areas.8 In the current study, we attempted to estimate the proportions of comorbidity of the various complications associated with T2DM. The magnitude of the difference between the published figure and our estimate could be largely due to the inclusion of costs related to complications and the increase of the number of cases. Our estimate of DMCs was based on a separate calculation of subgroups with different complications, which may have increased the accuracy of the medical cost estimate. In addition, we applied data from the current samples to both urban and rural patients, which may have extended the gap between the estimates and the actual health expenditure. However, the gap primarily was caused by the fact that rural patients have worse affordability difficulties than urban patients, other than the willingness-to-pay burden.26 It may be reasonable to estimate economic burden in the whole population using the data of our samples.

We found that the proportion of direct health expenditures consumed by T2DM had been as high as 18.2%, although the total annual DMCs ($26.0 billion) were far less than those of Germany (€30.6 billion)

 Prior studies reported the following costs related to T2DM as percentages of total public health expenditures: 7.4% in Spain,

 Because the costs incurred by T2DM in China are proportionally higher than those of developed countries, there is an urgent need for prevention programs that reduce the risks and control the complications of T2DM in China. The medical service is particularly unaffordable to the Chinese patients, who have to pay for treatment out of their own pockets with large differences under different insurance plans.

We found that indirect costs accounted for 6.9% of diabetes- related total costs in China. Our calculation of indirect costs was based on self-estimates of production losses from patients and their families. By using production losses and premature mortality, studies in other countries

 showed that indirect costs constituted a much higher proportion of costs than our estimate, which may be partly due to the higher proportion of unemployed patients in China. An increase in accessibility to care may increase the direct costs related to diabetes, and if care is effective, this may reduce the mortality and rates of disability related to diabetes and hence reduce the economic burden of T2DM.

Type 2 diabetes mellitus and its complications are preventable, and the costs of preventing and treating diabetes can be reduced. Recent studies in diverse settings have shown that lifestyle changes and metformin interventions were effective in preventing T2DM in individuals at risk, such as those with impaired glucose tolerance.

 We assumed that the effectiveness of these interventions was the same in China as in developed countries but that the costs of interventions and other diabetes care were higher in developed than in developing countries. The China Da Qing Diabetes Prevention Study reported a 51% reduction in incidence as a result of lifestyle intervention in China,

 which was consistent with the World Bank’s range. Using this assumption, we found that both lifestyle and metformin interventions for preventing T2DM in China could result in substantial cost savings in anticipated health expenditures in 2030.

We projected that diabetes screening strategies for the general population would save $1.8 billion in 2030 by preventing the onset of T2DM, while more specific treatment strategies such as blood pressure control may save even more. The cost savings would result mainly because of the treatments’ significant health benefits and relatively low intervention costs.

The findings in this study may be subject to several limitations. First, the study subjects were either patients recruited from outpatient clinics at secondary and tertiary hospitals or patients admitted to those hospitals, which may limit the ability to generalize the results and overestimate the economic burden of T2DM. To obtain a more accurate estimate of T2DM burden in the healthcare setting of China, it would be preferable to conduct a community-based study in the future. Second, the cost data, especially for direct nonmedical costs and indirect costs, were based to some extent on patient recall, which may be subject to recall bias. The DMC data, however, were based primarily on medical history combined with billing information, which increased the accuracy of those estimates. The cost data for the previous 6 months also may not accurately reflect the cost data for the previous 12 months. Moreover, it is difficult to differentiate T2DM-related complications from comorbidities, particularly among the elderly, and it is possible that some patients were misdiagnosed and therefore stratified incorrectly. Finally, another limitation is that the projected results were based on a study of urban patients, who may have different disease burdens than rural patients. Further study is needed to cover the rural areas. In conclusion, T2DM-related treatment consumes a large portion of health expenditures. The excessive costs are largely due to the treatment of complications. As diabetes prevalence in China continues to grow in the coming years, diabetes costs will continue to weigh heavily on health budgets. However, diabetes interventions among high-risk populations have the potential for considerable cost savings. As seen in this study, excessive costs increase proportional to the number of complications, so the importance of preventing complications through the use of more effective regimens is obvious.

We would like to thank Brian D. Smith, MA, for preparing the manuscript. We also thank the following hospitals for their significant contributions to data collection: 301 Hospital, China-Japan Friendship Hospital, People’s Hospital (affiliated with Beijing University), Peking University Third Hospital, Haidian Hospital, and Fuxing Hospital in Beijing; Ruijin

Hospital, Shanghai First People’s Hospital, Changhai Hospital, Dongfang Hospital, and Minhang District Central Hospital in Shanghai; Zhongshan University Third Hospital, Nanfang Hospital, Guangdong Provincial People’s Hospital, and First People’s Hospital of Guangzhou City in Guangzhou; and Huaxi Hospital (affiliated with Sichuan University), Sichuan Province People’s Hospital, Chengdu First People’s Hospital, and Chengdu Seventh Hospital in Chengdu.

the School of Public Health (WW, CF, BX), Fudan University, Shanghai, China; the School of Health Studies and Nursing (WPM), Georgetown University, Washington, DC; the Department of Epidemiology & Biostatistics (SZ), Peking University, Beijing, China; the Department of Epidemiology (RL), Sichuan University, Chengdu, China; and the Department of Medical Statistics and Epidemiology (WC), Sun Yat-sen University, Guangzhou, China.

This study was sponsored by GlaxoSmithKline (China) Investment Co, Ltd.

The authors (WW, WPM, CF, SZ, RL, WC, BX) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Concept and design (WW, WPM, CF, SZ, RL, WC, BX); acquisition of data (WW, CF, SZ, RL, WC, BX); analysis and interpretation of data (WW, WPM, CF, BX); drafting of the manuscript (WW, WPM, BX); critical revision of the manuscript for important intellectual content (WW, WPM, SZ, RL, WC, BX); statistical analysis (WW, CF); provision of study materials or patients (WW); obtaining funding (BX); administrative, technical, or logistic support (WW, SZ, RL, WC, BX); and supervision (WPM, BX).

Biao Xu, MD, PhD, MPH, Department of Epidemiology, Fudan University, 138 Yi Xue Yuan Rd, Shanghai 200032, China. E-mail: bxu@shmu.edu.cn.

. 3rd ed. Brussels, Belgium: International Diabetes Federation; 2006.

. Beijing, China: National Center for Disease Control and Prevention; 2006.

3. Wild S, Roglic G, Green A, Sicree R, King H. 

Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. 

4. Li LM, Rao KQ, Kong LZ, et al; The Technical Working Group of China National Nutrition and Health Survey.

 A description on the Chinese national nutrition and health survey in 2002 [in Chinese]. 

 Study on diabetes prevalence in urban China [in Chinese]. 

6. Xu L, Xie X, Wang S, Wang Y, Jonas JB. 

Prevalence of diabetes mellitus in China. 

Study on the epidemiological characteristics of diabetes mellitus and IGT in China [in Chinese]. 

 Assessing the impact of complications on the costs of type 2 diabetes in urban China [in Chinese]. 

 The study of economic burden of chronic non-communicable diseases in China [in Chinese]. 

Chinese Journal of Prevention and Control of Chronic Non-Communicable Diseases

10. Tang L, Chen X, Chen H, Zhao L, Hu S. 

The financing burden of treatment of diabetes II and its symptom in urban China [in Chinese]. 

 A summary analysis and some policy suggestion of the flows of total health expenditure in China [in Chinese]. 

12. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus.

 Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. 

Articles

DIABETES FOCUS

Treatment of type 2 diabetes mellitus and its complications places a heavy burden on healthcare budgets in China and will continue to do so.

Type 2 Diabetes Mellitus in China: A Preventable Economic Burden

To study the relative contributions of insurance product (commercial, Medicaid, Medicare), health plan, and physician group to quality of diabetes care. 

Cross-sectional observational study using data reported by Minnesota Community Measurement (MNCM) on care provided in 2005.

Individual performance rates for glycosylated hemoglobin (A1C) level <7%, lowdensity lipoprotein cholesterol (LDL-C) level <100 mg/dL, blood pressure <130/80 mm Hg, documented tobacco-free status, and aspirin use, as well as an all-or-none composite measure (Optimal Diabetes Care), were obtained from MNCM for 57 physician groups. Results included 7169 patients connected with 1 of 8 health plans and 1 of 3 types of insurance product.

All factors studied had a relationship to quality results. Of the factors, insurance product has the strongest relationship to A1C, LDL-C, and Optimal Diabetes Care. Health plan had the strongest relationship to tobacco-free status and daily aspirin, and physician group had the strongest relationship to blood pressure control. Physician group results varied more than those for health plan or insurance product on most measures.

All factors studied contribute to diabetes quality results. Reporting both physician group performance and health plan performance may offer a greater opportunity to improve care than reporting only health plan or only physician group results.

Physician group, health plan, and insurance product all affect diabetes quality of care. The relative contributions of these elements were examined in this cross-sectional observational study.

Reporting physician group performance in addition to health plan performance may stimulate greater improvement in patient care. 

Collaboration enhances a region’s ability to report physician group results. 

Incentives to focus on improving the quality of care community-wide might encourage collaborative efforts to report physician group performance.

A wide variety of factors are believed to contribute to the level of measured care quality for patients in the US healthcare system.

 Among these factors are the characteristics of the insurance product, health plan, physician group, individual physician, and individual patients.

Patient ethnicity, sex, and socioeconomic status have been shown to impact quality, as have provider group organizational structure, provider group culture, and the use of patient registries, electronic health records, and other practice supports.

 Health plan results vary by geography, for-profit versus not-for-profit status, patient ethnicity, enrollee socioeconomic status, and other factors.12-16 The Medicaid insurance product is associated with generally lower-quality results, whereas Medicare analysis demonstrates that for many measures, quality varies inversely with overall medical spending.

Publicly reporting healthcare quality results is becoming more common.

 The Centers for Medicare & Medicaid Services provides comparative quality-of-care results about hospitals, nursing homes, home care agencies, and health plans online.

 The National Committee on Quality Assurance reports quality-of-care results for health plans and by type of insurance product (commercial, Medicaid, Medicare).

 Regional organizations such as Minnesota Community Measurement (MNCM), Massachusetts Health Quality Partners, the Wisconsin Collaborative for Healthcare Quality, and the California Office of the Patient Advocate report quality-of-care results for physician group practices.

 Other programs report results by individual physician.

Although insurance product, health plan, and physician group have been demonstrated to affect measured quality of care, few studies have attempted to determine the importance of these elements relative to each other.

Studies comparing the relative importance of different levels of the healthcare system generally have found that smaller units within the healthcare system contribute more to quality-of-care results. For example, clinic site seems to have a stronger effect on patient experience of care than physician group and health plan; patient and physician factors appear to have a greater effect on diabetes results than clinic site; and the care delivery system has been found to be more related to clinical performance than the health plan.

 We decided to take advantage of a unique data set in Minnesota to compare the relative contribution of health plan, insurance product, and physician group to quality of care for diabetes.

We hypothesized a priori that quality-of-care results would vary across all 3 levels. We predicted that the variation would be the greatest among physician groups and that physician group would contribute more to quality-of-care results than either health plan or insurance product. We also predicted that the results of a composite quality measure using all-or-none scoring and individual quality measures would evidence similar relative contributions to quality. We expected that physician group practice management decisions, practice culture, and office systems would have a stronger effect on patient quality of care than any influences at a higher level in the healthcare system. Physician groups establish practice policies, make investment decisions about electronic health records, implement and maintain patient registries and other office systems, set performance expectations for staff, and interact directly with patients during encounters for medical care.

Health plans focus on improving quality results for their enrollees, often with strategies that supplement care provided by physician practices like telephonic disease management. Physician groups typically care for a mix of patients from most or all health plans within their region, so health plan quality results reflect the average quality of care provided to enrollees by many different physician groups.

Minnesota Community Measurement is an organization that provides standardized publicly reported measures of care quality across many different physician groups, 8 different health plans, and 3 different insurance product types in Minnesota. Data collected by MNCM provide a unique opportunity to address the relative impact of health plan, insurance product, and physician group on quality of diabetes care.

Seven Minnesota health plans and the Minnesota Medical Association initiated MNCM in 2002. Minnesota Community Measurement combines health plan data to measure and publicly report quality of care across multiple clinical domains in more than 70 primary care and multispecialty physician group practices in Minnesota.

The study was designed to assess how much of observed variation in care quality is attributable to health plans compared with insurance products and physician groups. The study also assessed whether using a single all-or-none composite measure of quality would yield results regarding variation and factor contribution similar to those obtained when assessing individual quality measures.

To simplify the assessment, we decided to limit our analysis to a single condition (diabetes) that has several well-recognized performance measures that have good evidence of a relationship to patient outcomes and for which an all-or-none composite measure has been reported.

 We analyzed comparative performance results for glucose control, blood pressure (BP) control, lipid control, tobacco-free status, aspirin use, and Optimal Diabetes Care (a composite of all 5 measures) as publicly reported by MNCM at www.MNHealthcare.org to address the study questions.

Minnesota Community Measurement uses health plan administrative billing data to identify patients of each physician group with various conditions (eg, diabetes, asthma, hypertension). The data set reflects patients enrolled in managed care plans, including not only patients enrolled in commercial health maintenance organization/point-of-service/preferred provider organization products, but also Medicare (cost and risk products) and Minnesota public healthcare programs (Prepaid Medical Assistance, MinnesotaCare, and General Medical Assistance products) from the 8 health plans that provide nearly all of the non-VA insurance coverage in the state. The data do not include the uninsured, patients who pay out-of-pocket, or patients covered by Medicaid/Medicare fee-for-service products. Data were collected by health plans in a standard way using a combination of Healthcare Effectiveness Data and Information Set (HEDIS) and MNCM specifications. Some measures were reported exclusively from administrative (health plan) data, whereas others (eg, diabetes) required on-site review of random samples of individual clinical records. Record reviews of a random sample of at least 60 adults with diabetes receiving care at each physician group were conducted by trained nurse reviewers following a standardized protocol.

Physician groups are defined by MNCM convention and are mutually exclusive. Groups do not share providers. One group, organized as a staff model group, has patients primarily from 1 health plan, providing care for approximately onethird of that plan’s enrollees. Other groups serve enrollees from multiple plans.

Physician group samples are identified through a series of steps. Individual health plans identify patients meeting denominator inclusion criteria per HEDIS specifications. Patients then are assigned to physician groups by each health plan using attribution logic based on evaluation and management and preventive care billing codes. These codes identify the frequency of patient visits to a physician group. Patients are attributed to the physician group they visited most frequently during the measurement year. If patients visited 2 or more groups with the same frequency, the patient is attributed to the physician group visited most recently. Validation processes are part of data collection, aggregation, and reporting. Detailed information about MNCM procedures is available at www.mnhealthcare.org.

Data available from MNCM consisted of deidentified patient- level diabetes performance data for care in year 2005 along with information to allow assignment of each patient to a specific health plan of enrollment, physician group providing care, and insurance product category of commercial, Medicaid, or Medicare. Patients dually enrolled in both Medicaid and Medicare were excluded from this analysis to simplify the comparisons.

Person-level diabetes performance measures consisted of glycosylated hemoglobin (A1C) level, systolic and diastolic BP, low-density lipoprotein cholesterol (LDL-C) values, documented tobacco-free status, and regular aspirin use for those over age 40 years. In addition to these individual measures, MNCM also reports an all-or-none composite measure, Optimal Diabetes Care, which is coded as positive for each patient meeting all 5 criteria: A1C ≤7%, BP <130/80 mm Hg, LDL-C <100 mg/dL, documented tobacco-free status, and regular aspirin use for those over age 40 years.32 The composite measure was coded as a failure for those not meeting all 5 measures or having missing data on any individual measure. However, those ineligible or contraindicated for aspirin use were coded as compliant. Patients with no test value within a year were classified as noncompliant. Raw data files consisted of the universe of diabetes patients sampled for chart review (n = 8401), reduced for analytic purposes by excluding patients dually enrolled in Medicaid and Medicare (n = 683), those lacking a physician group code (n = 524), and those with values outside of a clinically tenable range (eg, A1C <2%, LDL-C <10 mg/dL, diastolic BP <30 mm Hg) (n = 25). The resulting analytic file consisted of 7169 deidentified patient records.

All measures were analyzed as dichotomies to be in line with clinical goals desired in 2005 (A1C ≤7%, BP <130/80 mm Hg, LDL-C <100 mg/dL, documented tobacco-free status, and regular aspirin use for those over age 40 years) and were expressed as rates meeting the goal. Performance results were summarized by reporting the standard deviation of rates, minimum and maximum rates, and 25th/50th/75th percentiles for rates computed at the product, plan, and physician group level. Unadjusted variation by product, plan, and physician group is illustrated graphically in a bubble chart of raw rates (

Logistic regression assessed the contributions of insurance product, health plan, and physician group to the Optimal Diabetes Care composite and its 5 components for care in 2005. Product, plan, and group were treated as fixed effects and entered into models as sets of dummy coded contrasts. The P value from likelihood ratio tests comparing a full maineffects model (effects for product, plan, group) and a model with 1 set of contrasts deleted was used to assess whether the contribution of a particular set of contrasts (eg, for health plan) was significant. Model fit was assessed with changes in Akaike’s information criterion comparing a main-effects model with 2 classification variables to a full model. Because Akaike’s information criterion penalizes fit for the number of terms added to the model, it was used in this article to suggest which term, when added to the model, improved fit the most. The area under the receiver operating characteristic curve (C statistic) was reported for a model with all 3 classification variables in the equation (product, plan, physician group) to assess the overall ability of the model to predict care quality cross the different quality measures. The unique contribution to the C statistic of each classification variable when the other 2 variables were in the equation also was reported as a descriptive assessment of the unique predictive value of that variable above and beyond the other 2 variables. Because the C statistic was not penalized for model complexity and because each classification variable involved a different number of dummy codes (product had 2, plan had 7, group had 56), the magnitudes of these increments to the C statistic could be interpreted as relative contribution to prediction.

Although performance rates aggregated across product, plan, and physician group for individual measures ranged from 0.41 to 0.67, the rate for the Optimal Diabetes Care measure was only 0.08 because of the requirement of satisfying the threshold on all 5 measures for each individual in the denominator (

Variation in performance results is shown in 

 and the Figure by insurance product, health plan, and physician group. The bubble area in the Figure is proportional to sample size. A plot of model-adjusted rates was not notably different from the raw rates presented in the Figure and is not presented. In Table 2, for example, rates for the 57 physician groups for A1C ≤7% ranged from 0.19 to 0.67 (SD = 0.093) with a median rate of 0.47. The Figure illustrates that the low minimum rate of 0.19 was due to an outlier group with a rate considerably lower than other groups. Insurance product rates for A1C ≤7% ranged from 0.44 to 0.59 (SD = 0.081), and health plan rates ranged from 0.35 to 0.54 (SD = 0.065). Physician group rates ranged from 0.22 to 0.62 for BP <130/80 mm Hg, from 0.24 to 0.66 for LDL-C <100 mg/dL, from 0.32 to 0.86 for documented tobacco-free status, and from 0.40 to 0.90 for daily aspirin use. The plan with the largest number of patient-level observations had the highest rate among the 8 plans for 5 of 6 measures (Figure). In addition, the physician group with the largest number of patient-level observations among the 57 groups had among the 5 highest rates for 4 of 6 measures.

Variation as measured by the standard deviation was highest for the physician group for 5 of the 6 measures. The exception was tobacco-free status, for which health plan showed the highest variation. For all measures except documented aspirin use, denominators (number of patient charts) used to compute each rate in Table 2 for insurance product level summaries ranged from 1949 to 2666 with a median of 2554. For health plan, the denominators ranged from 220 to 2038 (median = 776), and for physician group they ranged from 52 to 965 (median = 63). The Figure illustrates that Medicaid patients had poorer quality on most but not all measures. This finding continued to hold in analyses that also considered the effects of plan and physician group.

Likelihood ratio tests all were significant (

), indicating that any variable when added last to the equation significantly improved the prediction of the measure examined. For all 6 measures, the A1C value was lowest when product, plan, and group were all in the model (data not shown) compared with any other main-effects model. The A1C increment column indicates that the addition of product to the model improved the model fit the most when examining A1C ≤7%, LDL-C <100 mg/dL, and the Optimal Diabetes Care measure. Plan improved model fit the most for tobacco-free status and daily aspirin use. Group improved model fit the most for BP <130/80 mm Hg. The total C statistic column indicates that tobacco status, aspirin use, and Optimal Diabetes Care were more accurately predicted by product, plan, and group (C ≥0.69) than the other measures (C ≤0.65). Product, plan, and physician group all showed statistically significant contributions to classification of goal status for all measures.

When we analyzed our data with the large high-performing physician group removed from the analysis, our findings did not change materially. Across the 6 measures reported in Table 2, standard deviations were reduced on average by 5% at the group level, 12% at the plan level, and 9% at the product level. The relative ranks of variation as summarized by the standard deviations in Table 2 remained the same, as did the relative ranks of A1C values in Table 3. Our conclusions remain the same.

All 3 variables are important. The conclusion one draws about relative importance varies depending on what is being examined. Our prediction that physician group would have the strongest relationship to quality-of-care results was true for only 1 of 6 measures: BP control.

For all measures, there was significant variation in the results by each grouping variable, even after considering the other 2 grouping variables (eg, by medical group after considering product and plan). All-or-none scoring lowered performance scores and dampened variation, but maintained the pattern of relative variation seen with A1C, LDL-C, and aspirin use.

Medicaid performance rates were lower than those of other insurance products for 4 of 6 measures: aspirin use, documented tobacco-free status, LDL-C, and Optimal Diabetes Care. Medicaid patients had the better BP control than patients enrolled in commercial or Medicare products, and Medicaid A1C control was intermediate. Medicare performance rates were the highest for all measures with the exception of BP; differences between insurance products were smaller.

Measuring and reporting quality are necessary to improve quality, but the resources for doing so are limited. Understanding the relative contributions of insurance product, health plan, and physician group can help determine where measurement and reporting resources will have the greatest impact. Our data suggest that measurement and reporting have the potential to stimulate quality gains at each of these levels, so it is important to report at each level if possible.

Reporting quality results by health plan and by insurance product is common. Reporting by provider group is not. Insufficient observations and the potential for atypical patients to distort results limit the ability and accuracy of individual health plans to report physician group quality-of-care results. Combining data from multiple health plans has enabled MNCM to report quality-of-care results for most physician groups in the state. More than 80% of primary care physicians practice in these groups.

Two major factors drove the development of these community- wide measures in Minnesota: the need for quality-of- care measures at the physician group level and costly inefficiencies related to regulatory reporting requirements for health plans. In the year prior to pilot-testing MNCM, Minnesota health plans sampled in excess of 12,000 clinical records to fulfill their diabetes quality reporting requirements. 

Despite this large number of clinical record reviews, results were unavailable for many physician groups. It was estimated that a sample sufficient to report performance of physician groups would require many fewer clinical reviews. Despite the apparent enhanced reporting capability and potential to improve quality, the MNCM model faces barriers to adoption in other regions of the country. Health plans commonly publicly report quality results, and competition for business is based partly on those results. Health plans less frequently collaborate to report physician group performance. Preference for improvement strategies capable of differentially impacting a health plan’s members and avoidance of strategies that “lift all boats” might be predicted in this situation. Indeed, many health plans sponsor telephone-based disease management programs to help differentiate their chronic care quality results, although some have suggested that physician group disease management services might be more effective and efficient.

 In other studies, Medicare results have demonstrated little evidence of effect from mandatory health plan reporting.

 A sole focus on health plan quality-of-care results may create a barrier to the development of physician group reporting or improvement efforts.

The data we present have several limitations. First, they are limited to diabetes. However, further investigation is possible as quality results are available for these physician groups for other conditions as well, including immunizations; breast, cervical, and colorectal cancer screening; an all-or-none cancer screening composite measure; and a range of other acute nd chronic condition measures. All are community-wide and well accepted by physician groups, who share ideas for improving care and compete for patients. Second, individual health plans were not identifiable in our data so we could not directly investigate possible explanations for the fact that health plan had a stronger effect than physician group on documented tobacco-free status and daily aspirin use. One health plan in our data had previously reported on a pay-for-performance program that was associated with significant increases in compliance with tobacco treatment guidelines.39 Some plans offered a subset of products, leading to unavoidable collinearity of product and plan in the analysis. This made it more difficult to disentangle the effects of product and plan. Third, because of uneven and sometimes thin distributions of data across combinations of product, plan, and group, the statistical models we used did not account for the possibility that the effect of physician group on performance varied by health plan influence. Plan influence on group characteristics is likely to vary with the proportion of the group’s patient covered by that plan.

One advantage of reporting community-wide quality-ofcare results is that it documents the variation in care quality across an entire community. Levels of diabetes quality of care in this community were better than levels reported nationally from 1988 to 2002. We hypothesize that if physician group reporting became as widespread as health plan reporting, it would be a positive force for community-wide healthcare improvement in other regions, but we see barriers to achieving that objective. We recommend consideration be given to creating incentives for health plans to collaborate to report physician group performance. Our results suggest such reporting could provide an extra stimulant to improve care.

Reporting physician group performance in addition to health plan performance may stimulate greater improvement in diabetes care.

Diabetes Care Quality: Insurance, Health Plan, and Physician Group Contributions

To assess the role of a Toyota production system (TPS) quality improvement (QI) intervention on appropriateness of perioperative antibiotic therapy and in length of hospital stay (LOS) among surgical patients.

Pre-post quasi-experimental study using local and national retrospective cohorts.

We used TPS methods to implement a multifaceted intervention to reduce nosocomial methicillin-resistant Staphylococcus aureus infections on a Veterans Affairs surgical unit, which led to a QI intervention targeting appropriate perioperative antibiotic prophylaxis. Appropriate perioperative antibiotic therapy was defined as selection of the recommended antibiotic agents for a duration not exceeding 24 hours from the time of the operation. The local computerized medical record system was used to identify patients undergoing the 25 most common surgical procedures and to examine changes in appropriate antibiotic therapy and LOS over time.

Overall, 2550 surgical admissions were identified from the local computerized medical records. The proportion of surgical admissions receiving appropriate perioperative antibiotics was significantly higher (

<.01) in 2004 after initiation of the TPS intervention (44.0%) compared with the previous 4 years (range, 23.4%-29.8%) primarily because of improvements in compliance with antibiotic therapy duration rather than appropriate antibiotic selection. There was no statistically significant decrease in LOS over time.

The use of TPS methods resulted in a QI intervention that was associated with an increase in appropriate perioperative antibiotic therapy among surgical patients, without affecting LOS.

We found that the use of Toyota production system methods as part of a nosocomial methicillin-resistant Staphylococcus aureus prevention initiative on a surgical unit resulted in

interventions to improve the quality of care in other areas.

Specifically, we found increased compliance with recommendations regarding appropriate prophylactic perioperative antibiotic use.

Changes were primarily driven by improved compliance with appropriate antibiotic therapy duration not exceeding 24 hours.

The Toyota production system (TPS) industrial engineering approach to improve manufacturing quality originated in the automobile industry and is being applied to quality improvement (QI) in healthcare delivery systems. In the TPS model, frontline work groups identify problems, experiment with possible solutions, measure the results, and implement strategies to improve quality, resulting in a “ground-up” rather than “top-down” approach to solving system problems.

Although the first article describing the use of the TPS in healthcare was published more than 20 years ago,

 numerous recent publications describe the use of the TPS model to improve quality or efficiency in the healthcare setting.

 Toyota production system methods are associated with improvements in (1) diagnostic accuracy of thyroid gland fine-needle aspiration and Papanicolaou tests,

 (3) elimination of dead stock in radiology,

 and increased medical record completion in the emergency department.

Beginning in 2001, a Veterans Affairs medical center (VAMC) instituted TPS methods to reduce methicillin-resistant Staphylococcus aureus (MRSA) infections on a general surgical floor. In addition to using the TPS process to facilitate implementation of MRSA prevention,

 the intervention evolved to address other areas for improvement in healthcare quality and efficiency on the surgical unit, such as developing and implementing an intervention to increase appropriate prophylactic perioperative antibiotic therapy, redesigning the supply cabinet and wheelchair distribution system, and improving the bar code scanning system for dispensing medications.

The objectives of the present study were to determine (1) whether the QI intervention for perioperative antibiotic therapy was associated with improvements in selection and duration of prophylactic therapy and (2) whether the overall MRSA prevention initiative was associated with a reduction in length of hospital stay (LOS) as a secondary outcome. We hypothesized that implementation of the QI intervention for perioperative antibiotic therapy would increase the proportion of admissions receiving appropriate prophylactic antibiotic selection and duration and that the overall MRSA initiative would be associated with a decreased LOS.

We used a quasi-experimental design to assess the association between implementation of a perioperative antibiotic prophylaxis QI intervention and appropriate perioperative antibiotic therapy over time. We also assessed the effect of the MRSA prevention initiative on LOS as a secondary outcome reflecting the overall efficiency of hospital care. This study received exempt human subject approval from the local institutional review board.

We used clinical and administrative databases to identify 2 local cohorts of surgical patients from the intervention VAMC and a national cohort of similar patients from all other VAMCs nationwide. Using the Veterans Health Information Systems and Technology Architecture (VistA) database, we identified a local cohort of patients who were admitted to the local surgical unit between fiscal years 2000 and 2004 with a primary International Classification of Diseases, Ninth Revision, Clinical Modification (

) surgical procedure code. To maximize the probability of patients’ spending time on the surgical intervention unit, we identified patients who spent their entire hospital stay on the unit and further refined the sample by including only those patients who had 1 of the 25 most common surgical procedures. This was necessary to ensure adequate numbers of patients within each surgical procedure type and to facilitate operationalization of appropriate perioperative antibiotic therapy. Once the top 25 procedures were identified for patients who spent 100.0% of their time on the intervention unit, we expanded our selection to include anyone who spent at least some of his or her admission on the surgical unit (

). Information on antibiotic therapy was only available from VistA and was assessed only in this local cohort defined using VistA. For the LOS analyses, we excluded patients with an LOS longer than 90 days because they were more likely to have a severe surgical or medical complication. Therefore, data from this initial local VistA cohort were used to examine perioperative antibiotic selection and changes in LOS over time.

To examine differences in LOS between our local VAMC and the other VAMCs nationally, we used the Veterans Affairs (VA) National Patient Care Database (NPCD) to identify a local cohort and a national cohort of patients who had an 

 surgical procedure code at the intervention VAMC for 1 of the same top 25 procedures. Finally, to identify the national study cohort, we used the same inclusion and exclusion criteria but extended our patient identification to all 116 remaining VAMCs nationwide.

The TPS intervention was initially implemented on a 36-bed general and subspecialty surgical unit at the intervention VAMC, a 146-bed tertiary care hospital for veterans. On this unit, the attending VA physicians are faculty members of the local university school of medicine and provide clinical training for surgical residents.

In November 2001, the surgical unit initiated a TPS intervention aimed to reduce nosocomial MRSA. The TPS team leader, a staff nurse (EEM), worked with unit staff to identify obstacles to identification and isolation of patients colonized with MRSA and to engage staff in problem solving for better MRSA infection control. The first interventions, initiated in spring 2002, involved training attending physicians, house staff, nurses, and laboratory technicians on proper hand hygiene, as well as systematic culturing of all admissions and isolating patients colonized with MRSA.

 Although the primary goal of this initiative was to reduce nosocomial MRSA, efforts evolved to address additional processes of care that interfered with the new MRSA prevention procedures. With increased recognition that appropriate antibiotic use in the operating room and on hospital floors may also help prevent the development of MRSA and other postoperative bacterial infections, the staff undertook the task of improving appropriate selection and duration of perioperative antibiotic therapy. Staff on the unit identified the following 4 key barriers to this process of care: (1) reliance on a computerized bar code medication administration system that was frequently nonfunctional, (2) lack of standardization of perioperative antibiotic orders, (3) reliance on paper-based rather than computer-based medication records, and (4) monthly rotations of new surgical residents responsible for writing antibiotic orders. These barriers were identified and addressed in an iterative process as guided by TPS methods.

Problems experienced by nursing staff in using the bar code medication administration system were addressed in late 2002 by purchasing a backup battery pack to ensure constant computer availability. To address lack of medical provider standardization in prescribing antibiotics, beginning in 2003 TPS staff enlisted the assistance of hospital pharmacists to establish evidence-based prophylactic antibiotic selection criteria for all common surgical procedures using the most recently published guidelines for specific surgical procedure types.

 This list of recommended antibiotics was approved by the head of infectious diseases (RRM) and the chief of surgery (MAW) at the intervention VAMC. By late 2003, guidelines for surgical antibiotic prophylaxis were available in the computerized VA order entry system and were printed on pocket cards distributed to all VA surgical attending physicians and residents. To address issues regarding appropriate duration of antibiotic therapy, a computerized standing order automatically discontinued antibiotics 24 hours after surgery. In addition, education was provided to new surgical attending physicians and residents at regular intervals.

Patient demographic and clinical data were abstracted from VistA and NPCD from the time of the index admission. The demographic characteristics were age in years, race/ethnicity (white vs nonwhite), sex, and admission source (nonscheduled or scheduled outpatient, nursing home or domiciliary, other VA or non-VA hospital, or other). We also assessed the number of comorbid conditions using the Charlson Comorbidity Index

 and the type of surgical procedure using 

Our assessment focused on 2 processes of care, appropriateness of perioperative antibiotic therapy and LOS. The analytic sample for the assessment of antibiotic therapy was restricted to surgical procedures performed within 1 day of admission, with no subsequent surgical procedures within 2 days. These criteria were used to limit the sample to patients who were specifically admitted for the surgical procedure and did not have more than 1 operation necessitating different types of antibiotic therapy.

Appropriate antibiotic therapy was defined as prescription of the recommended antibiotic agents consistent with the locally developed surgical prophylaxis guideline recommendations for 5 classes of surgical procedures (ie, neurosurgery, vascular, gastrointestinal and biliary, genitourinary, and orthopedic) for a duration not exceeding 24 hours. For carotid endarterectomy, antibiotics were not recommended by our hospital guideline; therefore, no antibiotic prescription was considered compliant. Operations of the skin were not assessed because of the heterogeneity of surgical infections and a lack of consensus on appropriate antibiotics.

We defined LOS as the date of discharge minus the date of admission and defined postsurgical LOS as the date of discharge minus the date of surgery. For patients who died in the hospital, LOS was censored on the date of death.

We summarized categorical variables using frequencies and continuous variables using medians and interquartile ranges (IQRs) because of the skewed nature of the frequency distributions. To compare surgical patients at the intervention VAMC with those at the other 116 national VAMCs, we used 

 tests for categorical variables and Wilcoxon rank sum tests for continuous variables. We used a 2-sided 

 <.05 to define statistical significance. All comparisons were performed using SAS 8.2 (SAS Institute, Cary, NC).

We used logistic regression to analyze appropriateness of perioperative antibiotic prophylaxis before and after implementation of the intervention, while accounting for the correlation of more than 1 surgical procedure for the same person (STATA 9.0; StataCorp LP, College Station, TX). Because the intervention was not applied to appropriateness of antibiotic therapy until the end of 2003, we used 2004 to define the first intervention year for these outcomes. To distinguish the effects of the intervention from any secular changes in perioperative antibiotic prescribing, we included a time variable in the models (year since 2000) and tested the differences in antibiotic compliance after the intervention (2004) compared with before the intervention (2000-2003) with adjustment for this linear trend.

To identify potential confounders, the research team reviewed all demographic and clinical variables that were available within the clinical and administrative databases and identified those variables that they believed could potentially confound the study outcomes. These variables included type of surgical procedure, admission source, age, and number of comorbid conditions. All of these potential confounders were then entered and retained in our models as control variables. An odds ratio greater than 1.0 indicates a greater likelihood of compliance during 2004 than in 2000 to 2003.

Because of the nonnormality of the data and the correlation of admissions for the same person or admissions from the same site, we analyzed LOS and postsurgical LOS using discrete survival analysis in SUDAAN 9.0 (Research Triangle Institute, Research Triangle Park, NC). We modeled LOS using a proportional hazards model that included covariate adjustment for the surgical procedure performed, admission source, age, number of comorbid conditions, and separate baseline hazards for each year. Hazard ratios (HRs) exceeding 1.0 correspond to a shorter LOS. In-hospital deaths were censored at the time of death. Models were also run that included patients with an LOS exceeding 90 days. These methods were used with the national data from the NPCD to test if the LOS at the intervention VAMC was similar to that at other VAMCs, as well as to test whether change in LOS after the intervention on the surgical unit was similar to the secular trends in LOS observed at the other VAMCs. The national analyses accounted for the correlation of admissions at the same site and admissions for the same person.

In the local cohort defined using VistA, we tested whether LOS changed after implementation of the intervention. The estimate of the intervention effect quantifies the effect of the intervention over and above a linear time trend, as estimated by including a linear term for time in the analysis. The analyses in the local cohort account for the correlation of admissions by the same person. For all LOS analyses, 2003 was considered the first year of the intervention.

Using VistA, we identified 5328 surgical admissions to the unit between 2000 and 2004, of which 2550 admissions had 1 of the 25 most common surgical procedures and an LOS of 90 days or shorter. Using the NPCD, we identified a similar local cohort of 2958 admissions at the intervention VAMC and 184,907 admissions at the 116 other VAMCs nationwide (

Most patients in the local study cohort identified using VistA were white (88.4%) and male (96.9%) and had a median (IQR) age of 66 (57-75) years (Table 1). Overall, 82.4% spent their entire hospital stay on the intervention surgical floor, and 8.2% underwent more than 1 surgical procedure. The most common types of surgical procedures were orthopedic (28.5%) and vascular (27.5%). As expected, the local VAMC cohort identified using the NPCD had demographic and clinical characteristics almost identical to those of the local VAMC cohort identified using VistA.

Compared with the national study sample, the local VAMC cohort identified using the NPCD was older (median age, 66 vs 65 years), and a greater proportion were white (89.4% vs 80.8%) (Table 1). The VAMC cohort had significantly higher rates of diabetes mellitus (30.4% vs 24.9%) and chronic pulmonary disease (14.9% vs 13.0%). However, the VAMC cohort had a lower incidence of malignant neoplasm (12.4% vs 16.1%). The local VAMC cohort also had a higher frequency of vascular and orthopedic procedures and a lower frequency of genitourinary procedures.

Appropriateness of Perioperative Antibiotic Therapy For the analysis of appropriate perioperative antibiotic therapy, we excluded 660 admissions for which the surgical procedure occurred after the first day of admission to limit the sample to patients more likely to be admitted for surgery. We also excluded 111 surgical procedures involving the skin because of the heterogeneity of surgical skin infections and a lack of consensus on appropriate perioperative antibiotic therapy for these infections. Therefore, appropriateness of antibiotic therapy was assessed in 1779 surgical admissions to the intervention VAMC. This subset (69.8%) of the local cohort did not differ from the full local cohort with regard to age, race/ethnicity, sex, or number of comorbid conditions (data not shown).

, the proportion of surgical admissions receiving appropriate antibiotic therapy was significantly higher (

 <.01) in 2004 after TPS initiation (44.0%) compared with any of the previous 4 years (range, 23.4%-29.8%). The overall improvement across this interval was attributable to improvements in compliance with duration of antibiotic therapy rather than improvements in appropriate antibiotic selection (

Overall, appropriateness of prophylactic antibiotic therapy significantly improved for neurosurgical, vascular, and orthopedic procedures between 2000 and 2004, but not all surgical procedures demonstrated a similar pattern of improvement (Table 2). The proportion of admissions receiving the appropriate selection of antibiotic therapy was significantly higher (

 <.01) for neurosurgical procedures in 2004 compared with linear trends from 2000 to 2003; the proportion of admissions receiving appropriate antibiotic agents was significantly lower for gastrointestinal and biliary procedures. The proportion of admissions who received antibiotic therapy for the appropriate duration was significantly higher for gastrointestinal and biliary, genitourinary, and orthopedic procedures in 2004 compared with linear trends for the corresponding procedures from 2000 to 2003.

, the median (IQR) LOS for surgical admissions (2000-2004) was shorter at the intervention VAMC (5 [3-9] days) than at the other 116 VAMCs nationally (6 [3-12] days). Although the overall LOS was shorter for surgical admissions at the intervention VAMC than at the 116 VAMCs nationally (HR, 1.2; 95% confidence interval [CI], 1.12-1.28; 

 < .01), there was no statistically significant decrease in LOS following full implementation of the TPS in 2003 compared with national rates (HR, 0.99; 95% CI, 0.91-1.09; 

Use of Toyota production system methods as part of a nosocomial MRSA prevention initiative on a surgical unit improved quality of care in other areas.


September 2009

Type 2 Diabetes Mellitus in China: A Preventable Economic Burden

Diabetes Care Quality: Insurance, Health Plan, and Physician Group Contributions

Toyota Production System Quality Improvement Initiative Improves Perioperative Antibiotic Therapy

Designing Programs for Populations with Chronic Care Needs: A Blanket or a Quilt?

Suboptimal Identification of Obesity by Family Physicians

Effect of a Managed Care Disease Management Program on Diabetes Care

Medicare Part D: Selected Issues for Plan Sponsors, Pharmacists, and Beneficiaries in 2009

Cost-Effectiveness of Roux-en-Y Gastric Bypass in Type 2 Diabetes Patients

Patient Adherence: A Blind Spot in Cost-Effectiveness Analyses?

Are There Unintended Consequences of Step-Therapy Programs?

Prior Authorization to Improve Testing of Glycosylated Hemoglobin in a Managed Care Setting

Patient-Centered Medical Home Demonstration: A Prospective, Quasi-Experimental, Before and After Evaluation