Mark R. Cullen, MD

 Hospital mergers and acquisitions are increasingly used as a strategy to facilitate value-based care. However, no studies have assessed health care utilization (HCU) and patient flow across merged institutions. We aim to evaluate patient population distribution, HCU, and patient flow across a recent hospital merger of an academic medical center (AMC), a primary and specialty care alliance (PSC), and a community-based medical center (CMC).

 This was a retrospective observational study.

 The study used 2018 adult demographic and encounter data from electronic health records. Patients’ parent health care institution was determined by the most frequently visited site of face-to-face visits. Differences in patient demographics and HCU (ie, emergency department [ED] visits, hospitalizations, primary care visits) were compared. Independent factors associated with utilization were identified using adjusted logistic regression models.

 A total of 406,303 adult patients were identified in the cohort. The PSC setting, compared with the AMC and the CMC, had significantly more female (62.7% vs 54.4% and 58.5%, respectively), older (mean [SD] age, 52.0 [18.1] vs 51.1 [17.8] and 49.2 [17.8] years), and privately insured (63.6% vs 51.3% and 56.0%) patients. A higher proportion of patients at the CMC (27.5%) visited the ED compared with patients at the AMC (10.8%). Approximately 1645 primary care patients (7%) at the CMC setting went to the AMC for specialized care such as oncology, surgery, and neurology.

 Hospital mergers are increasing across the United States, allowing AMCs to expand their reach. These findings suggest that patients mainly sought care at their parent health care institution, yet appropriately received specialized care at the AMC. These results provide insights for future mergers and guide resource allocation and opportunities for improving care delivery.

We assessed health care utilization and patient flow after a recent merger of community practices, a community hospital, and an academic medical center.

Variations existed in patient characteristics across the different health care institutions, with the academic medical center serving older, more complex, and non-White patients compared with the community practices and community medical center.

Patients sought care mainly at their parent health care institution and appropriately received specialized care at the academic medical center.

Continuous tracking of service utilization and patient flow is needed after hospital mergers and acquisitions to better deliver integrated, cost-effective care.

As health care becomes increasingly value based, health care organizations are working to efficiently use their resources, reduce costs, and improve quality of care. One of the strategies used to achieve these goals is through engaging in hospital mergers and acquisitions. Recent data have shown a rising rate of hospital mergers and acquisitions since 2010 when the Affordable Care Act was enacted. In 2017, there were 115 transactions of hospital and health system partnerships, and the number continues to grow.

 This approach can improve care quality and reduce health care costs through standardizing clinical operations across all hospitals of a health system and implementing best practices to deliver more cost-effective care.

 Several studies have examined the impact of merger and acquisition on health care delivery efficiency, costs, and quality of care.

 Their findings suggest that there is little evidence that mergers improve the quality of care,

 yet they reduce costs and improve hospital culture.

To our knowledge, no studies have assessed health care utilization (HCU) and patient flow across newly merged institutions. Understanding the patient population and patient flow through different institutions following a hospital merger is critical for understanding the particular needs of different communities and improving the efficiency of health systems. In addition, hospital management administrators and research units have frequently launched quality improvement projects and comparative effectiveness research; thus, understanding the patient population and resource utilization are necessary to guide the overall study design and participant recruitment.

Understanding the effect of hospital mergers on health services utilization and patient flow is essential as these acquisitions become more common. The purpose of this investigation was to quantify differences in facility utilization and describe patient flow across 3 merged institutions: an academic medical center (AMC), a primary and specialty care alliance (PSC), and a community-based medical center (CMC). Using the consolidated electronic health record (EHR) system supported by the AMC, we further describe a conceptual framework for classifying patients’ “home” health care institution and identifying their primary care setting. This work provides a baseline for understanding patient movement in newly merged and acquired health care systems using EHRs, which is necessary to identify and accommodate the specific needs of different communities.

This retrospective observational study included data from the 3 merged health care organizations: AMC, PSC, and CMC. The AMC provides a wide breadth of specialized and advanced health care services for complex illnesses and injuries and delivers primary and specialty care in more than 600 clinics; the PSC was formed in 2011 and includes more than 70 primary and specialty clinics located across California Bay Area communities. The CMC provides both inpatient and outpatient services in the East Bay’s Tri-Valley region and was acquired by the AMC in 2015. All health care settings in this merger and acquisition now share the same EHR system by Epic. The current EHR system at the CMC went live in early 2018 and replaced the previous paper-based system. Patients were identified from the clinical data warehouse, which included patient data from AMC, PSC, and CMC. The institutional review board of Stanford University approved the study.

The study included adult patients who had encounters of any type recorded in EHRs of the 3 organizations in 2018. Patients were excluded if they (1) did not have a face-to-face encounter (ie, hospital encounter, office visit) in 2018; (2) were 90 years or older, due to small sample size; (3) had died in 2018; or (4) were not California residents based on residential zip code.

Patient demographics were retrieved for each patient, including age, gender, race, ethnicity, and health insurance. Insurance types were categorized into private, Medicare, Medicaid, and other; both fee-for-service and managed care programs were included when applicable. The modified Charlson Comorbidity Index (CCI) score was calculated to categorize and represent patients’ comorbid conditions.

 The CCI contains 17 selected conditions and is designed for use with EHR data based on 

International Classification of Diseases, Ninth Revision, Clinical Modification

 diagnosis and procedure codes. A CCI score for each patient in 2018 was calculated.

HCU was obtained for each patient, including emergency department (ED) visits and hospitalizations; specialty and primary care visits; and face-to-face, telephone, and email encounters. For the face-to-face visits, we obtained the care settings (ie, AMC, PSC, or CMC), and the department specialties (eg, family medicine, cardiology, oncology) of these visits for each patient. Primary care visits were identified by the department specialties, including family medicine, family practice, general internal medicine, general practice, geriatric medicine, primary care, internal medicine, and coordinated care (ie, a primary care clinic that specialized in treatment of chronic or complex diseases). No-show or canceled appointments were excluded from the calculation of utilization rates.

Patient encounters were assigned to 1 of the 3 health care institutions: AMC, PSC, or CMC. Each patient was assigned to their “home” institution (ie, AMC, PSC, or CMC) based on the institution with the highest number of face-to-face visits in 2018. For patients with the same number of visits to 2 or more institutions, the following priority rule was applied, CMC > PSC > AMC, because an individual who visited a community health center was more likely to be from that community than other, farther locations. Once the patient’s home institution was determined, descriptive statistics were calculated comparing demographic and clinical characteristics across the 3 institutions.

Similarly, each patient’s primary care “home” was determined by the primary care location visited most. This algorithm was validated by manual chart review by a research nurse of 150 patients (50 from each institution) and confirmed using the patient’s residential zip code or, if not recorded, any location for physical therapy (PT). We used PT as an additional estimator in an absence of a zip code because the clinician consultants for the project stated that patients tend to choose PT locations close to their residence location.

Descriptive statistics were used to summarize demographic and encounter-specific information across the 3 institutions. We categorized age into 4 categories (ie, 18-34, 35-49, 50-64, and ≥ 65 years) for a better understanding of the population difference. Count and frequency were used for describing categorical demographic variables. Comparisons of age among the 3 health care institutions were performed using a 1-way analysis of variance (ANOVA). Chi-square tests of independence for categorical variables and Kruskal-Wallis nonparametric tests for continuous variables were used to test differences in demographic and clinical characteristics among groups. Post hoc pairwise comparisons were performed following ANOVA/Kruskal-Wallis with Bonferroni correction for multiple tests. Logistic regression was applied to assess differences in HCU (ie, ED visits, hospitalizations, and primary care visits) controlling for other known demographic (ie, age, gender, race, ethnicity, and health insurance) and clinical (ie, CCI) factors. Primary care visit was included as a factor for ED and hospitalization prediction. Patients from PSC were removed from regression analysis due to the large amount of missing data of the outcome variables. The level of significance was set at .05, and all statistical analyses were performed in Statistical Package for Social Sciences version 26.0 (IBM).

Patient Demographics and Comparison by Institution

A total of 677,115 unique patients were identified with at least 1 encounter in 2018. A total of 270,812 (40.0%) were excluded: 159,232 (23.5%) did not have any face-to-face visits, 54,052 were younger than 18 years, 6239 were older than 90 years, 3069 had died in 2018, and 48,220 came from outside California. Finally, a total of 406,303 patients were included in the study (

). Overall, the mean (SD) age of patients was 51.2 (17.9) years; 57.6% (n = 234,217) were women and 11.7% (n = 47,437) were Hispanic. Distribution of patient race was 49.7% White (n = 202,037), 19.5% Asian (n = 79,366), and 4.3% Black (n = 17,454). Fifty-six percent of patients were privately insured (managed care programs accounted for 53.8% of the privately insured individuals), 25.8% had Medicare, 8.5% had Medicaid, and 9.8% had other/unknown insurance. Based on the patient’s most frequently visited location in 2018, they were categorized into AMC (n = 220,410), PSC (n = 132,300), or CMC (n = 53,593). 

 presents demographic and clinical characteristics of patients by institution. AMC had the greatest number of racial and ethnic minorities, with 42.4% non-White (PSC, 36.2%; CMC, 39.7%) and 13.0% Hispanic (PSC, 9.6%; CMC, 11.3%) patients. PSC, as a community-based outpatient care institution, had the highest percentage of women (62.7% vs 54.4% for AMC and 58.5% for CMC) and the highest number of patients with private insurance (63.6% vs 51.3% and 56.0%, respectively) compared with the other 2 institutions. CMC had the youngest patient population (mean [SD] age, 49.2 [17.8] years vs 51.1 [17.8] for AMC and 52.0 [18.1] for PSC) and more White representation (56.0% vs 46.1% and 53.1%) than the other 2 institutions. All differences were statistically significant.

Health Care Service Utilization by Home Institution

 displays the health care service utilization in 2018 by the institution that provided the majority of care to the patient, which was designated the home institution. There was variability across different institutions for the complexity of patients. Patients whose home institution was AMC had the highest mean (SD) number of comorbid conditions (0.9 [1.8]), followed by CMC (0.5 [1.4]; 

 < .001). Also, AMC patients had the highest mean (SD) number of face-to-face visits (6.15 [10.14]) and email and telephone encounters (8.86 [13.74]) compared with patients seen most by the other 2 institutions (

 < .001). PSC had the highest percentage of patients with primary care visits (n = 81,911; 61.9%).

Patients whose home institution was CMC had a significantly higher mean (SD) number of ED visits (0.37 [0.82]) compared with AMC (0.15 [0.56]; 

 < .001). A total of 14,732 (27.5%) unique patients were seen at the ED of CMC in 2018. Results of multivariable logistic regression on HCU (

) were consistent with findings from univariate analyses; patients at CMC had higher odds of ED visits (adjusted odds ratio [OR], 1.97; 95% CI, 1.88-2.07) and hospitalizations (adjusted OR, 1.51; 95% CI, 1.45-1.57) compared with patients at AMC, adjusted for other covariates in the model. We observed a small number of PSC patients who had hospitalizations (n = 206; 0.2%) and ED visits (n = 926; 0.7%), possibly due to the missing data for this patient population.

Results of analysis of HCU data from patients whose home institution was AMC or CMC showed demographic and clinical factors associated with HCU (Table 3). Compared with White patients, Asian patients were more likely to have primary care visits (adjusted OR, 1.68; 95% CI, 1.65-1.72); also, the odds of ED visits were higher (adjusted OR, 1.18; 95% CI, 1.12-1.26) and of hospitalizations significantly lower for Asian patients (adjusted OR, 0.82; 95% CI, 0.79-0.86). In contrast, Black patients had higher odds of ED visits (adjusted OR, 1.66; 95% CI, 1.50-1.83) and hospitalizations (adjusted OR, 1.11; 95% CI, 1.03-1.19) than White patients. Additionally, Hispanic patients were less likely to have primary care visits (adjusted OR, 0.83; 95% CI, 0.80-0.86) and more likely to have hospitalizations (adjusted OR, 1.247; 95% CI, 1.18-1.32) than non-Hispanic patients. Similarly, patients insured with Medicaid had significantly lower odds of primary care visits (adjusted OR, 0.11; 95% CI, 0.11-0.12) compared with privately insured patients, but the odds of visiting the ED were higher (adjusted OR, 1.62; 95% CI, 1.52-1.72) and the odds of being hospitalized were greater (adjusted OR, 2.19; 95% CI, 2.09-2.30) than for privately insured patients.

Using the most frequently visited primary care location as the patient’s primary care institution, our approach achieved 96% accuracy after the chart review. The 

) presents the care utilization patterns across the 3 health care settings of patients who participated in at least 1 primary care visit. Overall, 240,166 patients (59.1%) were excluded because they were not seen by a primary care provider during the study period, leaving 166,137 patients (40.9%) in the cohort; their home primary care institutions were AMC (n = 61,055), PSC (n = 81,911), and CMC (n = 23,171). As shown in the eAppendix Figure, a majority of patients remain at their home institution for specialty care (AMC, 90.7%; PSC, 78%; CMC, 88.3%). We observed that 21.0% of primary care patients (n = 17,201) at PSC received specialty care at the AMC; the department with the greatest number of visits was oncology, followed by surgery, dermatology, ophthalmology, and cardiology. For primary care patients at CMC, a small proportion (n = 1043; 4.5%) received specialty care at the PSC and approximately 7.1% (n = 1645) received it at the AMC. Similar findings were observed for privately insured individuals in the managed care programs: 7.1% visited AMC and 5.5% visited PSC.

In this study, we describe the patients and patterns of their care across different health care institutions after a hospital merger and acquisition in the Northern California Bay Area. We observed variations in patient characteristics across the different health care institutions, with the AMC serving older, more complex, and more non-White patients compared with the PSC and CMC institutions. After the merger, the majority of patients continued receiving care from their home institution for routine checkups whereas a small proportion of patients sought highly specialized services, such as oncology or surgery, at the AMC. Interestingly, compared with those at the AMC, patients at the CMC had higher rates of HCU and were more likely to have ED, inpatient, or primary care visits. These findings provide insights of patient utilization after mergers between academic and community medical centers and can guide resource allocation and opportunities for improving care delivery.

In general, the patient population of the study is diverse and, compared with Census data from 2019,

 consists of more Asian patients and fewer White and Black patients. Within this refined area, the populations that the AMC and CMC serve are very different, with the AMC serving older patients who are sicker with more complex diseases. This finding is consistent with those of other studies

 because the AMC provides tertiary and quaternary care to more challenging patients and often functions as a referral center for the CMC. In addition, another study has reported that compared with CMCs, AMCs have a higher share of Medicare- and Medicaid-insured patients.

 This is also evident in our study, in which a higher proportion of AMC patients are insured by Medicare than CMC patients. In addition to the differences in the patient community it serves, the level of care offered by the AMC is comparable with that of other AMCs in the nation, providing both general acute care services and tertiary medical care for patients who are based locally, nationally, and internationally.

After the hospital merger, we found that patients generally received nonspecialty care at their home institute. Community medical centers serve as the primary sites for health care most of the time. Patients whose home institution is community-based “cross over” to receive care at another institution more than those whose home institution is an AMC. The patient movement across institutions is likely a result of the severity and complexity of disease conditions and the restricted care specialties provided in the home institution. Health insurance coverage might be another factor affecting utilization; inadequate health coverage may limit patients’ ability to visit multiple offices and providers. The results from this work provide an understanding of patient flow following hospital acquisitions that may be applicable to other hospital mergers and help with resource planning.

We found significant differences in ED visits and hospitalizations of patients across the 3 institutions, in that patients whose home institution was PSC had a significantly smaller number of ED visits and hospitalizations. These differences are likely due to the availability of health care services for acute and complex conditions within a specific region served by each institution. PSC outpatient clinics serve individuals of a wider region in the Bay Area where several other health systems are also accessible to patients. Thus, patients are likely to access EDs that are closer in distance and are hospitalized in other acute care facilities. It is likely, therefore, that ED visits and hospitalization records of these patients are not recorded in the EHR data set. CMC hospitals and clinics, however, cover patients in a relatively limited region with few other health care facilities; thus, it is likely that patients visited the CMC for an emergent health condition, and we are able to capture a higher proportion of ED visits and hospitalizations for patients whose home institution was CMC.

We acknowledge several limitations of the study. First, this study is just one example among many hospital mergers and acquisitions, and these findings may not generalize to other health care systems. However, the merger and acquisition of academic and community hospitals are becoming more common, and the results from this study are likely applicable to other mergers and acquisitions. Second, several other health systems are situated in the local area, which may affect the complete capture of HCU, such as ED utilization. This is particularly true at the AMC, which may treat out-of-state patients for specialty care. However, the complete capture of utilization is not unique to our health care setting, and therefore the results from this study are likely comparable with other AMCs. Last, the recent acquisition of hospitals may not have reached “equilibrium” yet in terms of patient service utilization. Continuous tracking of service utilization and patient flow is needed to guide future decision-making.

This study characterizes patient flow between newly acquired and merged hospitals and highlights movement from CMC to AMC for specialty care. This work demonstrates differences in patient demographics and HCU across different types of health care institutions now merged into a single health system. Patients were more likely to stay at their parent institutions and seek advanced cancer care or surgery at the AMC. It is also important to note the differences in patient populations at the individual institutions as hospitals respond to increasing pressures of health care reform and attempt to acquire a broader array of services to better deliver integrated, cost-effective care. This work provides novel information regarding patient movement among institutions following hospital mergers and acquisitions. 

 School of Medicine, Stanford University (RS, SB, MW, MRC, TS, TH-B), Stanford, CA.

 The study was funded by the Stanford University School of Medicine Dean’s Population Health Science Support.

 The authors 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 (SB, MW, MRC, TH-B); acquisition of data (TS, TH-B); analysis and interpretation of data (RS, SB, MW, TS); drafting of the manuscript (RS); critical revision of the manuscript for important intellectual content (RS, SB, MW, MRC, TS, TH-B); statistical analysis (RS, SB); obtaining funding (TH-B); administrative, technical, or logistic support (MRC, TH-B); and supervision (TH-B).

 Tina Hernandez-Boussard, PhD, School of Medicine, Stanford University, 1265 Welch Rd #245, Stanford, CA 94306. Email: boussard@stanford.edu.

1. 2017 in review: the year M&A shook the healthcare landscape. Kaufman Hall. January 2018. Accessed September 1, 2021. https://www.kaufmanhall.com/insights/research-report/2017-review-year-ma-shook-healthcare-landscape

2. US health services deals insights: Q2 2019. PwC. July 24, 2019. Accessed October 14, 2019. https://web.archive.org/web/20190816121257mp_/https://www.pwc.com/us/en/industries/health-industries/library/health-services-quarterly-deals-insights.html

3. Noether M, May S, Stearn B. Hospital merger benefits: views from hospital leaders and econometric analysis – an update. American Hospital Association. September 2019. Accessed September 1, 2021. https://www.aha.org/system/files/media/file/2019/09/cra-report-merger-benefits-2019-f.pdf

4. Noether M, May S. Hospital merger benefits: views from hospital leaders and econometric analysis. American Hospital Association. January 2017. Accessed September 1, 2021. https://www.aha.org/system/files/2018-04/Hospital-Merger-Full-Report-FINAL-1.pdf

5. Romano PS, Balan DJ. A retrospective analysis of the clinical quality effects of the acquisition of Highland Park Hospital by Evanston Northwestern Healthcare. 

. 2011;18(1):45-64. doi:10.1080/13571516.2011.542955

6. Schmitt M. Do hospital mergers reduce costs? 

. 2017;52:74-94. doi:10.1016/j.jhealeco.2017.01.007

7. Cutler DM, Scott Morton F. Hospitals, market share, and consolidation. 

. 2013;310(18):1964-1970. doi:10.1001/jama.2013.281675

8. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with 

. 1992;45(6):613-619. doi:10.1016/0895-4356(92)90133-8

9. Charlson ME, Pompei P, Ales KL, Mackenzie CR. A new method of classifying prognostic in longitudinal studies: development and validation. 

. 1987;40(5):373-383. doi:10.1016/0021-9681(87)90171-8

10. QuickFacts: United States. US Census Bureau. Accessed April 16, 2020. https://www.census.gov/quickfacts/fact/table/US/PST045219

11. Shin P, Alvarez C, Sharac J, et al. A profile of community health center patients: implications for policy. Kaiser Family Foundation. December 23, 2013. Accessed September 1, 2021. https://www.kff.org/medicaid/issue-brief/a-profile-of-community-health-center-patients-implications-for-policy/

12. PwC Health Research Institute. The future of the academic medical center: strategies to avoid a margin meltdown. University of Utah Health. February 2012. Accessed September 1, 2021. https://uofuhealth.utah.edu/hcr/2012/resources/the-future-of-academic-medical-centers.pdf

13. The differences between community and academic medical centers. The George Washington University. July 7, 2021. Accessed September 1, 2021. https://healthcaremba.gwu.edu/blog/the-differences-between-community-and-academic-medical-centers/

Articles

This study assessed health care utilization and patient flow after a recent merger of community practices, a community hospital, and an academic medical center.

Characterizing Patient Flow After an Academic Hospital Merger and Acquisition

 To examine the association between processes measures of diabetes care and time to progression for 4 diabetes complications: coronary artery disease (CAD), stroke, heart failure (HF), and renal disease (RD).

This retrospective study followed outcomes from 2003 through 2009 in a cohort of 1797 employees with diabetes who worked for a large US manufacturer and were enrolled in the same health insurance plan.

 Quality of care was measured by consensus standards for testing glycated hemoglobin, lipids, and microalbuminuria. Employees with diabetes who received all 3 measures of care in the baseline year (2003) were compared with those who received less complete testing. Cox proportional hazard regression models were used to assess potential associations between diabetes care and time to complications, controlling for potential confounders.

 Observed differences between the 2 groups in time to event were significant for 2 of the 4 complications: HF (hazard ratio [HR] = 0.39, 95% confidence interval [CI], 0.19-0.81; 

 = .0117) and RD (HR = 0.48, 95% CI, 0.24-0.95; 

 = .0339) and any of the 4 complications (HR = 0.66, 95% CI, 0.48-0.91; 

 = .0101). Differences in time to complication for CAD (HR = 0.70, 95% CI, 0.49-1.02;

 = .0635) and stroke (HR = 0.63, 95% CI, 0.38-1.07; 

 = .0891) showed the same trend but were not significant.

Employees with diabetes who received all 3 quality measures experienced fewer complications, risk-adjusting for other factors. These results provide support for the importance of care quality.

The proportion of patients with diabetes receiving standard processes of care based on systematic reviews is disturbingly low and in our fully insured population is just under 15%.

		The patients in our cohort who received standard processes of care developed fewer complications than those who did not, despite being sicker at the outset.

		Our results provide support for process measures and efforts to promote their application in practice.

		If even a modest portion of outcomes are causally associated with better processes of care, there is great potential for benefit.

In 2010, the estimated prevalence of diabetes among adults ranged from 5% to 13% across the United States and its territories,1  accounting for substantial morbidity, mortality, and associated economic costs.

Type 2 diabetes mellitus often develops 9 to 12 years before diagnosis

 and more than 50% of patients have at least 1 complication by that time.

 A significant body of research has established that effective treatments can significantly decrease the development and/or progression of complications.

However, diabetes care remains suboptimal and varied in the United States.

 In an effort to improve outcomes, The National Diabetes Quality Improvement Project (1997) developed a comprehensive set of process and intermediate outcome measures to assess quality of care for diabetes patients that are now considered standards of care.

These measures include annual  screenings for lipids and microalbuminuria, at least 2 measurements of glycated hemoglobin (A1C) at least 30 days apart, and annual flu vaccinations, dilated eye exams, and foot examinations.

 have used these performance measures to assess care provided to patients with diabetes. These   studies provide evidence of the substantial gaps between national performance measures for diabetes care and actual care received by persons with diabetes in the United States.

The present study examines 3 process measures of diabetes care: A1C tests, lipid tests, and urine screening tests for  microalbuminuria; and their association with 4 subsequent complications of diabetes: coronary artery disease (CAD), stroke, heart  failure (HF), and renal disease (RD); in a cohort of employees of a large national manufacturing company. This cohort provides an opportunity to study this issue in a geographically, ethnically, and socioeconomically diverse population with rich and uniform healthcare benefits. This study contributes to the literature by identifying systematic reasons for differences in care among privately insured Americans and examining whether adherence to recommended processes of care is associated with better outcomes.

The study setting was a large (approximately 36,000 employees) US-based manufacturing company with uniform healthcare benefits in 22 states that generated claims data during the years of the study. We examined the relationships between the 3 measures of the quality of diabetes care and the risk of major diabetes complications to determine whether employees with diabetes who received all 3 quality-of-care measures in a baseline year (2003) demonstrated a lower risk for developing any of 4 complications over the 6-year observation period (2004-2009). Baseline characteristics suspected to impact outcome risk and/or likelihood of receiving optimal care included in the multivariate model were age, sex, ethnicity, marital status, income, insulin use, comorbidities, and smoking. Due to data limitations, we were unable to assess the possible confounding roles of obesity and duration of diabetes. Data on annual flu vaccinations, dilated eye exams, and foot examinations were not available; thus, these elements were not included in our model.

This was a retrospective cohort study using administrative data on employees collected between January 1, 2001, and December 31, 2009. The majority of the employees held hourly manufacturing jobs; minorities and women were well represented. Although employees could select from a menu of health benefits, only a single preferred provider organization (PPO) network was available to  the entire population. Employees who did not opt for this PPO were excluded (3%). The panel was restricted to those employed by the manufacturer for some part of 2001 and 2002, and for all of 2003, the year during which quality of care measures were assessed. Health outcomes were assessed from claims between January 1, 2004, and December 31, 2009. 

During the “lead-in” period (2001-2002), diabetes status was ascertained from the claims data based on 

International Classification of Diseases, Ninth Revision

) codes for the condition on a submitted medical claim for at least 1 of the following: a hospitalization, emergency department visit, 2 office visits, or 1 prescription for a diabetes-specific medication. Participants were  included in the final data set if they were between 18 and 64 years old and employed throughout 2003 and for at least 1 month in the observation period (2004-2009). Employees with diabetes who had any 1 of the 4 end point complications coded in a medical claim in 2001 or 2002 and women with a claim containing an

code for gestational diabetes were excluded. Those with claims for 1 of the 4 complications in the baseline year (2003) were included in the panel, but were excluded from the analysis for that complication. A total of 1797 subjects were available for analysis.

Process measures of diabetes care (ie, 2 A1C tests at least 30 days apart and at least 1 test each for lipids and microalbuminuria) were assessed in the baseline year (2003).

 Each employee with diabetes was categorized into 1 of 2 mutually exclusive groups: those who received all 3 measures of care in 2003 and those who did not. Cox proportional hazard regression models were used to assess the associations between diabetes care and time to each of the 4 complications (CAD, stroke, HF, and RD) or any of the 4 complications.

The databases available for analysis have been described in detail in previous publications.

 Briefly, health data were obtained from the health insurance plan, extraction of onsite occupational health medical records available for about 40% of the locations, and other linked employer-managed administrative databases. Data included 

 codes from the Chronic Condition Data Warehouse list

 for disease diagnosis and National Drug Codes

 for prescription information. End point complications were identified based on 

 codes during a face-to-face exam or hospitalization. For end point data, the subset of 

 codes from the Chronic Condition Data Warehouse list for causes of kidney disease and heart disease unrelated to diabetes (eg, infectious or genetic etiology) were excluded. Demographic and health behavior data were obtained from the health medical record and human resources databases.

Vision claims were not available. Therefore, claims for routine eye exams and screening for diabetic retinopathy were not usable as measures of quality of diabetes care.

Sightlines DxCG Risk Solutions software (Verisk Analytics, Jersey City, New Jersey) was used to assign risk scores to the cohort during the baseline assessment and capture potential illness-related influences on treatment decisions. This actuarial program uses medical and pharmacy data with proprietary risk-adjustment algorithms to predict future health spending, and its risk scores are comparable to similar scores used by the Centers for Medicare & Medicaid Services and other insurance bodies for global risk estimates. A score of 1 in 2003 implies an actuarial prediction that an individual would consume the median amount of healthcare  or the population in 2004. Databases were linked by an encrypted unique identifier to ensure human subject privacy. This study was approved by the authors’ institutional review boards.

Hazard ratios for each outcome (CAD, stroke, HF, RD, and any of the 4 conditions) were calculated for quality of diabetes care (exposures) and covariates. Time to event, defined as the number of months from the start of the observation period (January 1, 2004) to the first medical claim for a complication, was used for all primary analyses. Employees with diabetes who no longer  received employer-provided health benefits were censored at the time of their last month of coverage. Records (n = 901) were  censored during the observation period for the following reasons: retirement, layoff or termination (n = 647), change from PPO to health maintenance organization or Consolidated Omnibus Budget Reconciliation Act (COBRA) coverage (n = 77), and death or  long-term disability (n = 30). For 119 subjects, the reason for loss of coverage could not be definitely established from the available  records. Differences between the baseline characteristics of censored and uncensored employees were assessed with x

 tests. The Kaplan-Meier method was used to identify crude time-to-event models for the associations between the exposure and outcome measures.

Cox proportional hazard models were used to evaluate bivariate and multivariate association, which enabled time-to-outcome assessment with risk adjustment based on socioeconomic and lifestyle risk factors, comorbidities, and severity of diabetes. The base model, calculating time to event as a function of quality of care, was specified with demographic covariates. Covariates including modifiable risks (smoking), clinical characteristics (insulin use and health severity risk score), and socioeconomic status indicators (individual income, job status [hourly or salary], and marital status) were then added to the base model. Finally, the fully adjusted model was used to assess the combined explanatory effects for all measured factors. Hazard ratios (HRs) for primary outcomes were calculated based on Cox proportional hazard analyses stratified by hourly job status, salary, sex, and insulin use. As a sensitivity test, kappa statistics, correlation analyses, and Cox proportional hazard analyses were used to assess outcome  differences based on 2 years (2003 and 2004) of continuous high-quality care (leaving 2005-2009 available for outcome observation). Other sensitivity tests included looking at each stratum of sex, job status, and insulin use in the full model. All analyses were conducted using SAS version 9.2 (SAS Institute Inc, Cary, North Carolina). 

A total of 1797 patients met criteria for inclusion. The cohort consisted primarily of white (69.2%), male (83.1%), married (74.2%), hourly employees (75.7%), a significant number of whom earned less than $35,423 per year (27.2%). More than half (55.1%) were between the ages of 18 and 51 years. The mean age for the total population was 49.0 years (standard deviation = 8.4). Smoking data were reported as never (14.4%), past/current (17.3%), and unknown (68.3%). The unknown group represents primarily locations for which health medical record data were unavailable. As a measure of diabetes severity, 20.1% of the population had insulin prescriptions during the baseline year. Health severity risk scores were also used to assess comorbidities, with 75.1% of the cohort receiving a score of 2.1 or lower (1 = average health risk spending probability, 2 = risk likelihood of double the average health expenditures). Regarding the quality of diabetes care process measures, a majority of patients received at least 1 A1C test (67.6%) and lipid testing (61.8%) in the baseline year. Other measures of care were less common: 43% of patients had at least 2 tests for A1C and only 24.0% were tested for microalbuminuria in the baseline year.

Characteristics of the baseline sample are shown in 

; the same subjects stratified by treatment status are shown in 

. At baseline, the group receiving all 3 process measures was smaller (n = 267) than the comparison group  (n = 1530). Differences between the 2 groups were assessed with 2-tailed

 tests for independent samples. Despite the sample size differences, there were no statistically significant differences between the 2 groups during the baseline period (2003) for demographic, socioeconomic status, or modifiable risk variables. However, there was evidence that the group receiving all 3 process measures was sicker at baseline. They were significantly more likely to be receiving insulin than the comparison group (27.7% vs 18.8%; 

 = .0007) and were also more likely to fall into the higher quartiles of health severity risk score (

 shows the characteristics of those who were censored compared with those who remained in employment throughout the   follow-up period. Although workers who left were older and less well paid on average at baseline, differences between censored employees and non-censored employees were not significant for race, marital status, occupational group (salary or hourly), insulin use, health severity risk score, and importantly, likelihood of receiving the treatment.

In total, 366 persons with diabetes (20%, n = 1797) had medical claims for at least 1 of the 4 complications with a mean time to complication of 29.1 months. The most frequent complication in the cohort was CAD (16.9%) with a mean time to complication of 26.6 months, followed by stroke (8.7%, 33.1 months), HF (5.8%, 29.7 months), and RD (4.9%, 38.1 months). Those getting all 3 process measures of care fared better for all end points. Hazard ratios for 2 of the 4 complications were significant: HF (HR = 0.39, 95% confidence interval [CI], 0.19-0.81; 

 = .012) and RD (HR = 0.48, 95% CI, 0.24-0.95; 

 = .034). The HRs for CAD (HR = 0.70, 95% CI, 0.49-1.02; 

 = .064) and stroke (HR = 0.63, 95% CI, 0.38-1.07; 

 = .089) showed the same trend but were not significant. Hazard ratios with CIs for each and any of the  4 end points are summarized in 

The HR for submitting a medical claim for any of the 4 complications was significantly lower for those receiving all 3 process measures (HR = 0.66, 95% CI, 0.48-0.91;

 = .01). Covariates associated with increased risk included increasing age 46-51 years (HR = 1.88, 95% CI, 1.36-2.61;

= .0001), 52-56 years (HR = 2.06, 95% CI, 1.47-2.89; 

 <.0001), and 57-64 years (HR = 3.09, 95% CI, 2.15-4.46;

 <.0001); health severity  risk scores of 2.1 or higher (HR = 1.91, 95% CI, 1.36-2.68; 

 = .0002); and smoking (HR = 1.44, 95% CI, 1.01-2.07; 

 = .047).  Differences in all other covariates were not statistically significant. The Kaplan-Meier estimates of cumulative hazards for all end points are depicted in 

We conducted 2 sensitivity analyses. The first used 2 years (2003 and 2004) of continuous process of care measures to reduce the likelihood of random misclassification. Results showed the same trends and similar point estimates (data not shown), although with shorter follow-up (2005-2009) these results were no longer significant. Likewise we tested for robustness within strata by sex, hourly versus salaried job status, and initial insulin use. Results revealed comparable point estimates for all strata with the exception of the women, in whom HR estimates for each complication hovered around 1. In general, the point estimates within strata were similar to those found for the full cohort, but with wider CIs.

This study compared 2 groups of employees with diabetes to assess how differences in process measures of care might have affected the onset of complications associated with diabetes. Significant differences in time to complication were observed between the 2 groups for 2 of the 4 complications  (HF and RD) and for any of the 4 complications. Remarkably, those employees getting optimal process scores were, on average, sicker at the outset based on greater insulin use and increased risk severity scores; hence, they would have been expected a priori to have worse outcomes.

Many studies have assessed whether interventions at the provider level improve processes of care and intermediate outcomes, but the effect of such process improvements on complications remains less clear because the outcomes themselves are rarely assessed.

 This study estimates the impact of process measures of quality care by analyzing data in a national sample across multiple physician groups administering diabetes care within the same health insurance plan structure. This study is the first to our knowledge confirming, in practice, the expectations for benefit from compliance with recommended process measures of care in which follow-up was long enough to observe the major clinical end points of interest.

That said, there remain significant limitations to our methods and our ability to draw causal inference from the results. We relied on claims data for recognition of our population with diabetes, for ascertainment of complications, and for assessment of the critical covariate of interest—execution of all of the quality of care diagnostic measures in the baseline year. Each step entails opportunity for  misclassification. Although the diagnosis of diabetes is readily inferred from claims with high sensitivity and specificity,

comparable data comparing claims with verifiable sources such as medical records or examinations do not exist for any of the major complications with the exception of RD,

which has been found to be identified with high specificity but less sensitivity than diabetes. That would not, however, bias our findings for patients who had been identified as having RD. Thus, it is possible that the end points misclassified cohort members in both directions and that some with complications already evident were not appropriately excluded. The impact of such misclassifications, assuming they were random, would likely drive our result toward the null, but there is no assurance that such errors were random.

Another concern is the impact on our conclusions of unobserved variables (eg, body mass index, duration of diabetes) as noted above. Although there is no strong prior reason to suspect that such variables were distributed in such a way as to confound our  result, more obese subjects might have received consistently poorer care and, independently, had a higher risk of bad outcomes. The impact of censorship may also have resulted in possible bias, as those faring most poorly might be expected to leave prematurely more often. It is reassuring that censorship was not associated with baseline characteristics of disease severity nor  with treatment group, but the size of loss—more than half of the cohort&mdash;introduces concern for unmeasured bias.

The employees in this study may not represent a generalizable population, as they work for a stable employer with rich and uniform benefits in the heavy manufacturing sector, both less common features than a generation ago. Our study panel requirements demanded a group stably employed during a period of years, further rendering it less representative of the larger workforce.

Perhaps most difficult to fully address is the possibility of endogenous differences among persons with diabetes leading some to both seek better care for their disease—picking better qualified doctors and/or advising them what tests to order&mdash;and take better are of themselves in other ways, collectively resulting in better outcomes. Nonetheless, it is noteworthy that in the primary test of  the hypothesis, the baseline characteristics created an uneven playing field, with more ostensibly sick patients having a higher rate of better care; thus, the deck was stacked against those with better care having better outcomes, an effect that began to be visually apparent in the third observation year (see Figure 5). It is also reassuring that our result survived the various sensitivity tests. We are exploring instrumental variables and other approaches to examine the causal pathway further, including assessment of the relationships between the process measures and intermediate outcomes such as medication change, adherence, frequency of exams, and the like.

Limitations notwithstanding, these results provide further support for process measures and efforts to promote their application in  practice, starting with the obvious; the proportion of employees with diabetes getting the standard of care based on systematic  reviews

 was disturbingly low (just under 15%) in our fully insured population. If even a modest portion of the association with outcome is causal, there is a considerable opportunity for benefit from driving better processes of care.

From Alcoa Aluminum (FJB), Pittsburgh, PA; Yale Occupational & Environmental Medicine (DG, MS, OT), Yale School of Medicine, New Haven, CT; Stanford University School of Medicine (IMC, MRC), General Medical Disciplines, Stanford, CA.

 Dr Bayer reports employment with Alcoa Aluminum, who provided data for this study. The other authors 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 (FJB, DG, MS, OT, MRC); acquisition of data (FJB, MRC); analysis and interpretation of data (FJB, DG, MS, IMC, OT, MRC); drafting of the manuscript (FJB, IMC, MRC); critical revision of the manuscript for important intellectual content (FJB, DG, IMC, OT, MRC); statistical analysis (FJB, DG, MS, MRC); provision of study materials or patients (FJB); obtaining funding (FJB, MRC); administrative, technical, or logistic support (FJB, IMC); and supervision (MRC).

This study was supported by National Institutes for Health grant 5RO1AG026291-04 and continuing support from Alcoa Aluminum. The funders had no role in the design of this study; collection, management, analysis, and interpretation of the data; conduct of this study; or preparation or approval of the manuscript. Alcoa Aluminum reviewed the manuscript prior to publication.

Adherence to process of care measures was associated with reduced risk of 2 diabetes complications or any of 4 complications in a national industrial cohort.

Process of Care Compliance Is Associated With Fewer Diabetes Complications

The cause(s) of suboptimal medication adherence in patients with chronic disease has been the subject of intense investigation.

 Reducing medication costs in order to improve patient adherence is common practice in value-based insurance design,

 however, absolute price reduction may have a marginal effect on overall adherence due to the large effects of non-economic factors.

To further explore, we report data from a natural experiment in a manufacturing workforce in which the imposition of marginal costs to otherwise rich and uniform drug benefits

 were staggered. As part of an overall assessment of the change, we use a difference-in-difference approach to evaluate the impact of price changes for 4 classes of drugs (angiotensin-converting enzyme inhibitors/angiotensin receptor blockers [ACEIs/ARBs], beta-blockers, statins, asthma controllers) among 2573 employees who had regularly filled prescriptions for at least 1 year prior to the benefit change (medication possession ratio >.80).

On January 1, 2004, one-third of the workforce—on average slightly younger, less well paid, and more likely female&mdash;were switched (based on plant location and worktype) from a small fixed copay to a coinsurance benefit with graded copays for generics (5% to 10%), brands with no generic option (20%), and choice brands (50%); the remaining workers experienced no year-to-year benefit change because of preexisting labor agreements. The 

 shows the out-of-pocket costs and medication adherence across the 2 years for this previously adherent group; the reported 

 values are for the difference-in-difference change between those with the new benefit versus no change in benefit. As shown, there was a disproportionate decline in adherence in all 4 drug classes, including the 2 where the out-of-pocket costs actually 

because of the coinsurance, possibly due to cross-price effects including the higher cost for outpatient visits which resulted in 11% to 17% less ambulatory utilization in the intervention group.

Multivariate analysis, adjusting for prior adherence, age, gender, income, race, and sex, shows an overall intervention effect of —.013 (

 <.05) or a 1.3% decline in overall medication adherence across classes. Elasticity of demand was about —.002 per $10/year change (

 <.001), in the range of a 2% to 3% decline for the more expensive drug classes, similar to that reported by Choudhry.

 Both effects were not included in the same model because of colinearity.

Overall the negative impact of the price change, at least among these previously adherent patients, was consistent with expectations but modest in absolute size for all classes except asthma controllers, the group that changed most in price. Although our multivariate model (data not presented) did not confirm either effect, we cannot exclude steeper consequences of price change among the less affluent, nor the possibility of interactions between price and other factors among those previously less adherent. The health impacts of these unintended consequences of design change among stably employed privately insured workers merits careful observation as the transition to higher marginal cost insurance schemes continues.

UNC School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC

Yale University School of Medicine, New Haven, CT

Stanford School of Medicine, Stanford, CA

From UNC School of Medicine (JJH), University of North Carolina at Chapel Hill, Chapel Hill, NC; Yale School of Medicine (DG), Yale University School of Medicine (MDS), New Haven, CT; Stanford School of Medicine (MRC), Stanford, CA.

The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

NICHD/NCMRR K12HD01097-06 Rehabilitation Medical Scientist Training Program; NIA 1R01AG026291-01A1 Disease, Disability, and Death in an Aging Workforce; The Network on Socioeconomic Status and Health; The John D. and Catherine T. MacArthur Foundation; Alcoa, Inc; and NBER.

 Mark R. Cullen, MD, Stanford School of Medicine, 1265 Welch Rd, MSOB x338, Stanford, CA 94305-5411. E-mail: mrcullen@stanford.edu.1. Goldman DP, Joyce GF, Zheng Y. Prescription drug cost sharing: associations with medication and medical utilization and spending and health. JAMA. 2007;298(1):61-69.

2. Farley JF, Wansink D, Lindquist JH, Parker JC, Maciejewski ML. Medication adherence changes following value-based insurance design. Am J Manag Care. 2012;18(5):265-274.

3. Choudhry NK, Avorn J, Glynn RJ, et al. Full coverage for preventive medications after myocardial infarction. N Engl J Med. 2011;365(22):2088-2097.

4. Busch SH, Barry CL, Vegso SJ, Sindelar JL, Cullen MR. Effects of a cost-sharing exemption on use of preventive services at one large employer. Health Aff (Millwood). 2006;25(6):1529-1536.

5. Einav L, Finkelstein A, Ryan S, Schrimpf P, Cullen MR. Selection on moral hazard in health insurance. NBER Working Paper 16969. http://www.nber.org/papers/w16969. Published April 2011.

The impact of price changes in drugs, at least among previously adherent patients, was modest in size except for drugs with large price changes.

Mark R. Cullen, MD

Articles

Characterizing Patient Flow After an Academic Hospital Merger and Acquisition

Process of Care Compliance Is Associated With Fewer Diabetes Complications

Drug Adherence After Price Changes in a Previously Compliant Population