Informal Clinical Integration and Cardiac Surgery Payments

Abstract

Importance

To reduce the inefficiency and waste associated with care fragmentation, many programs are underway that target deeper clinical integration. However, these programs have led to only modest spending reductions. This may be because most programs focus on formal integration, which often bears little resemblance to actual physician interaction patterns.

Objective

To examine how physician interaction patterns vary between health systems and to assess whether variation in such informal integration relates to care delivery payments.

Design

Using national Medicare data (2008 to 2011), we identified beneficiaries who underwent coronary artery bypass grafting (CABG) and mapped interactions between all physicians who treated them—including surgical and medical specialists and primary care physicians—within a health system during their surgical episode. We then measured the level of informal integration in these networks of interacting physicians. Finally, we fit multivariate regression models to evaluate associations between episode payments made on a beneficiary’s behalf and the level of informal integration in the health system where he or she was treated.

Setting

Health systems (n=1,186) where Medicare beneficiaries underwent CABG.

Participants

248,073 Medicare beneficiaries age 66 and older.

Exposure

A health system’s informal integration level.

Main Outcome Measures

Price-standardized total episode and component payments.

Results

Informal integration levels varied across health systems. Compared to health systems with higher levels of informal integration, those with lower levels tended to treat more black patients (P<0.001) and more patients from urban areas (P<0.001). They were also less likely to have implemented electronic health records (P<0.001). After adjusting for these differences and other patient, health system, and community factors, higher levels of informal integration were associated with significantly lower total episode and component payments (P<0.001 for each association). When beneficiaries were treated in health systems with higher informal integration, the greatest savings were from readmissions and post-acute care services, predicted payments for which were 13.0% and 5.8% lower, respectively.

Conclusions and Relevance

Deeper informal integration is associated with lower spending. While most programs seeking to promote clinical integration are focused on health systems’ formal structures, policymakers may also want to address informal integration.

Inpatient surgery costs vary widely across health systems. Many point to the fragmented nature of surgical care delivery as a driver of this variation. Suboptimal coordination around the surgical episode can affect spending by increasing the likelihood that care team members provide duplicate treatments, tests, or services. Fragmentation also impedes physicians’ ability to identify patients’ postoperative needs early, resulting in emergency department visits and readmissions. To address care fragmentation, payers and policymakers have launched reforms like accountable care organizations (ACOs) and the patient-centered medical home (PCMH), which aim to deepen clinical integration among physicians.^1,2

However, the effects of ACOs and the PCMH on health spending have been modest. While some evaluations demonstrate decreases in costs, others report no effect or even increases.^3–6 A weakness of these programs and their evaluations is their focus on formal integration, understood as organizational structure, rules, and regulations.⁷ Studies of social networks in organizations suggest that informal physician interaction patterns—relationships arising from the shared care of multiple patients over time—may be more consequential than formal structure for health system performance.^8–10 Because health care reforms’ formal designs often fail to map onto actual practice, examining informal physician interaction patterns may help clarify the potential value of deeper clinical integration.

In this context, we analyzed surgical episode payments for Medicare beneficiaries undergoing coronary artery bypass grafting (CABG). Using ideas from network analysis, we developed a new measure—the informal clinical integration index—that characterized interactions among primary and specialty care physicians and examined how these interactions affected surgical episode payments. We hypothesized that greater informal clinical integration would be associated with lower episode payments. Our findings, which support this hypothesis, serve to inform health system administrators, policymakers, and researchers attempting to understand fragmentation between primary and specialty care physicians.

Study Data and Methods

Network analysts have developed techniques for characterizing interactions in social groups.¹¹ Use of these techniques has led to insights about the importance of informal interaction patterns inside formal organizations.^12–14 All networks share two building blocks: nodes and ties. Nodes are people; ties are interactions among those people. In our study, nodes are physicians who care for patients undergoing CABG. Ties are the patients each physician pair shares. Previous research finds that physicians who share patients are also more likely to share information.^15–17

Study Population

Beginning with the Medicare Provider Analysis and Review (MedPAR) file, we identified beneficiaries aged 66 and older who underwent CABG between January 1, 2008 and December 30, 2011. We chose this window because it immediately predates several national initiatives aimed at improving integration—including the Medicare Pioneer ACO Model, the Medicare Shared Savings Program, and the Center for Medicare and Medicaid Services’ Federally Qualified Health Center Advanced Primary Care Practice Demonstration—and therefore allows us to examine informal integration during a relatively stable period. We excluded beneficiaries who were not continuously enrolled in Medicare 6 months before and 60 days after discharge. Due to incomplete claims, we also excluded beneficiaries who had insurance through Medicare Advantage. After implementing these criteria, our sample includes 248,073 beneficiaries and 1,186 health systems.

Mapping Physician Networks

We identified relevant physicians using the Medicare Carrier file. First, we determined each beneficiary’s treating surgeon by isolating the surgeon who billed Medicare for CABG closest to the patient’s surgery date. Next, we identified each beneficiary’s primary care physician using a previously described algorithm.¹⁸ Finally, we located relevant medical and surgical specialists by extracting claims for services 30 days before and 60 days after the hospitalization for surgery.

Within health systems, we recorded an interaction between physicians if they billed for services for the same beneficiary around the beneficiary’s CABG episode.¹⁹ We mapped networks separately for each health system and year. Patients undergoing surgery at the same health system but in different years, or in the same year but at different health systems, experience different networks.

Characterizing Informal Clinical Integration

We based our index of informal clinical integration among primary care and specialist physicians on a measure from network analysis known as assortativity. Assortativity captures the degree to which ties occur between nodes with similar properties (i.e., physicians of the same specialty).²⁰ We use the reverse of assortativity, so higher values indicate deeper integration, and multiply by 100, so the coefficient ranges from −100 to 100. A network will have negative index values if physicians share more patients with colleagues in their own specialty; a network will have positive values if physicians share more patients with colleagues in specialties different from their own. Lower connectivity across specialties may mean communication among physicians overseeing different aspects of surgical care is weaker, which may in turn increase spending.

Calculating our index requires information on physician specialties. We used Medicare specialty codes to categorize physicians as primary care, medical specialty care, or surgical specialty care. We excluded radiologists and other specialists who are not directly involved with ongoing patient care. Details on the index are given in Online Supplements A and B.

Measuring the Efficiency of Surgical Care

To examine whether differences in informal integration may help explain surgical care spending, we extracted data on 60-day episode payments for beneficiaries’ surgical care using the MedPAR, Carrier, and Outpatient files. These payments reflect what Medicare actually paid for services rendered around CABG episodes. Although true costs may include more than payments, payments are an informative proxy. Following earlier studies,²¹ we decomposed payments into physician services, index hospitalization, readmission, and post-acute care components. We standardized payment values to account for regional price differences.²²

Statistical Analyses

For preliminary analyses, we stratified health systems into three equally sized groups (“low,” “medium,” and “high”) based on their level of informal integration. Index values were [−26.05, 2.24] for the low group (n=84,598), [2.25, 4.73] for the medium group (n=84,442), and [4.74, 50.00] for the high group (n=84,505). We then made comparisons among these groups using Kruskal–Wallis tests. At the patient level, we compared groups on age, sex, race, and level of comorbid illness (as measured by the Charlson index²³). In addition, we evaluated patient populations on socioeconomic factors, including income, education, and care access. At the health system level, our comparisons focused on size (i.e., number of patients and physicians), diversity of physician specialties (as measured by a Herfindahl-Hirschman index over the distribution of primary care, medical specialty care, and surgical specialty care physicians, subtracted from 1 to capture diversity), academic affiliation (using American Hospital Association data), and the proportion of patients undergoing emergency surgery. We also compared health systems’ formal structures, focusing on factors identified by organizational theorists as likely to influence interaction patterns. Our measures captured aspects of health systems’ technological (i.e., electronic health record [EHR] implementation), institutional (i.e., government or for-profit control), organizational (i.e., affiliation with other provider organizations) and geographical (i.e., number of providers in a health system’s physician network who practiced outside the health system’s region) structures. Finally, we looked for differences at the community level, defined as the health system’s Hospital Service Area. Using the American Community Survey and Dartmouth Atlas data, we compared primary care, medical specialist, surgeon, and hospital bed availability. We also compared communities on the size of their Hispanic, black, and overall populations.

Our next analyses used multivariate regression to assess whether payments varied with informal integration. The unit in our regressions was the patient, but we measured networks at the health system level. Therefore, we estimated multilevel models with health system random effects and clustered standard errors. Models also included year fixed effects. Our outcomes were price-adjusted episode payment components and our predictor was informal clinical integration. We controlled for confounders at the patient, health system, and community levels using the variables shown in Table 1. Our hypothesis will be supported if we find negative and statistically significant associations between informal clinical integration and price-adjusted episode payment components.

Table 1.

Patient, health system, and community characteristics across three levels of informal clinical integration.

	Low		Medium		High
	Mean	SD	Mean	SD	Mean	SD	P-value
Patient-Level
Charlson	2.01	1.75	1.99	1.73	1.87	1.66	<0.001
Age	75.00	5.93	74.94	5.87	74.66	5.72	<0.001
Race
White (proportion)	0.93	0.26	0.94	0.24	0.95	0.21	<0.001
Black (proportion)	0.04	0.20	0.03	0.18	0.03	0.16	<0.001
Female (proportion)	0.31	0.46	0.31	0.46	0.31	0.46	0.87
Health-System Level
Patients from outside the CBSA (proportion)	0.60	0.23	0.58	0.21	0.55	0.20	<0.001
Patients below poverty line (mean, proportion)	0.13	0.04	0.13	0.04	0.14	0.04	<0.001
Patients with bachelor’s degree (mean, proportion)	0.18	0.05	0.17	0.05	0.15	0.04	<0.001
Patients living in a rural area (mean, proportion)	0.25	0.19	0.28	0.18	0.36	0.18	<0.001
Patients with emergency admission (proportion)	0.45	0.18	0.46	0.18	0.45	0.18	<0.001
Academic hospital (proportion)	0.63	0.48	0.59	0.49	0.59	0.49	<0.001
Total patients	107.96	87.32	113.74	90.97	101.85	71.91	<0.001
Total physicians (log)	5.64	0.79	5.57	0.73	5.31	0.70	<0.001
Provider specialty diversity	0.62	0.02	0.62	0.02	0.62	0.03	<0.001
Formal structure
Technological (EHR, proportion)	0.32	0.47	0.32	0.47	0.33	0.47	<0.001
Institutional (government, proportion)	0.05	0.21	0.05	0.21	0.03	0.18	<0.001
Institutional (for profit, proportion)	0.08	0.27	0.07	0.26	0.07	0.26	0.09
Organizational (affiliated, proportion)	0.44	0.50	0.42	0.49	0.40	0.49	<0.001
Geographical (providers outside CBSA)	118.43	267.90	80.75	106.89	69.29	66.70	<0.001
Community-Level
Acute care hospital beds per 1,000 residents	2.41	0.62	2.32	0.54	2.32	0.54	<0.001
PCPs per 100,000 residents	71.78	19.69	68.49	15.80	66.11	12.67	<0.001
Medical specialists per 100,000 residents	52.41	18.82	48.59	14.91	43.23	10.19	<0.001
Surgeons per 100,000 residents	39.20	9.86	38.19	9.18	35.89	7.50	<0.001
Total resident population (in thousands, log)	13.53	1.15	13.27	1.02	13.03	0.90	<0.001
Total black population (in thousands, log)	11.35	1.92	10.86	1.81	10.46	1.76	<0.001
Total Hispanic population (in thousands, log)	11.20	1.72	10.75	1.62	10.27	1.40	<0.001
N	84598		84442		84505

Finally, to evaluate the strength of our findings, we performed sensitivity analyses, discussed in Online Supplement F. Our analyses were done using Stata SE Version 13.1. Statistical tests were two-tailed and used 0.05 as the Type I error probability. This study was approved by our university’s institutional review board.

Results

We found that the informal clinical integration index varied across health systems. Recall that the index captures the degree of interaction among primary and specialty care physicians. The lowest value observed on our index was −26.05 (relatively low integration); the highest was 50.00 (deeper integration). The average was 3.74, with a standard deviation of 3.13. A histogram is shown in Online Supplement C.

Figure 1 plots relationships among physicians at two heath systems, A and B. Physicians are represented by colored nodes. Colors correspond to specialties. Green are primary care physicians, yellow are medical specialists, and blue are surgical specialists. Red ties indicate relationships across specialties; gray ties indicate within specialty connections. The health systems serve comparable markets in the Midwestern United States. They have similar numbers of physicians (70 in A, 89 in B), and those physicians have similar numbers of ties.

Figure 1.

Informal clinical integration at two health systems.

Despite these similarities, the health systems differ in terms of informal clinical integration. Physicians are less connected across specialties in Health System A, which has an index value of −9.02. Among the 561 ties between physicians in this health system, about 59.1% are cross-specialty. By contrast, integration is higher at Health System B, which has an index value of 10.09. In this health system, about 72.7% of ties among physicians are cross-specialty. More information on these cross-specialty ties is given in Online Supplement D.

Table 1 compares health systems across the three levels of integration. Beginning with patient level factors, health systems with low and medium informal integration had patients with more comorbid illnesses (Charlson index: low group=2.01, high group=1.87; P<0.001). Demographically, these health systems also treated more black patients (black proportion: low group=0.04, high group=0.03; P<0.001) and more patients from urban areas (rural proportion: low group=0.25, high group=0.36; P<0.001). Associations at the health system level are also revealing. Health systems with lower informal integration tended to have more physicians (number [log]: low group = 5.46, high group=5.31; P<0.001), who were geographically dispersed (number outside CBSA: low group=118.43, high group=69.29; P<0.001), and less likely to use an electronic health record (proportion with EHR: low group=0.32, high group=0.33; P<0.001). Finally, at the community level, health systems with less informal integration tended to be located in more populous regions (population [thousands, log]: low group=13.53, high group=13.03; P<0.001), with more Hispanic (population [thousands, log]: low group=11.20, high group=10.27; P<0.001) and black residents (population [thousands, log]: low group=11.35, high group = 10.46; P<0.001). In sum, these descriptive findings suggest that health systems that treat more disadvantaged populations, particularly urban minorities, also tend to be less informally integrated.

Our next analyses examined associations between informal integration and payments. After adjusting for patient, health system, and community factors, we found significant associations, in support of our hypothesis. Our results are shown in Figure 2 by the integration tercile groups. Figure 3 compares predicted payments for the four components among health systems with high integration relative to those with low levels. Regression coefficients are shown in Online Supplement E. Although health systems with higher integration have better performance on all four components, savings are most pronounced for readmission and post-acute care. Here, we observe that health systems in the high integration group have predicted payments that are 13.03 and 5.82% lower, respectively, than the low group.

Figure 2.

Adjusted component payments for CABG across three levels of informal clinical integration.

Figure 3.

Relative change in component payments for CABG moving from low to high informal clinical integration.

To put these numbers in perspective, consider that roughly 250,000 CABG surgeries are performed annually in the United States. Assuming these procedures were done by health systems with high informal integration, we would expect savings of $130,500,000 on readmissions relative to what we would expect if the procedures were done by health systems with low integration. The corresponding expected savings on post-acute care is $108,500,000. For total episode payments (not shown in Figures 2 or 3), the expected savings are $640,277,500 annually.

Discussion

Our findings demonstrate variability across health systems in terms of how much primary care and specialty physicians interact around shared patients. This variability is associated with surgical care episode payments. In health systems with greater informal integration, we observe lower spending on heart surgery. We find that the greatest savings occur around payments for readmissions and post-acute care. These findings hold even after accounting for patient, health system, and community differences. Our results offer support for the idea that better informal integration around surgical care may improve coordination and thereby lead to greater efficiency.

Our study contributes to the understanding of integrated care delivery. To date, research has focused on the effects of formal integration—for instance, bringing hospitals and physician groups under the same umbrella—on outcomes. However, studies of social networks in organizations caution that formal structures often fail to map onto informal interaction patterns. Although formal organization helps support coordination across groups, there is no guarantee that it will bring people together. Qualitative research on ACOs suggests that understandings of integration vary among early adopters, and that although in some cases adoption appears to have increased informal integration, in others, changes have been minimal.²⁴ The informal integration index offers one way of differentiating among groups that have adopted similar formal approaches to integration, but that may still differ in terms of interaction among specialties.

Our study suggests that health systems may be able to improve their performance through deeper informal integration. The organizational literature offers many examples highlighting the importance of informal interaction patterns. Consulting firms like IDEO and Design Continuum are famous for helping their clients develop innovations. Studies suggest their success is due largely to frequent informal knowledge sharing among designers and engineers who specialize in diverse industries.¹² Interventions aimed at promoting informal integration have also been successful in settings where work is more standardized. Call centers, for example, have seen double digit improvements by making it easier for employees to communicate and share knowledge with their fellow employees.¹⁰

Within surgical care, identifying the best ways to promote informal integration is a promising area for future work. Based on research in other domains, administrators may see benefits from eliminating physical barriers between PCPs, medical, and surgical specialists (e.g., through colocalization of clinics).^25,26 Where these barriers cannot be eliminated, administrators may consider incentivizing physicians to use emerging health information technologies that promote collaboration, or organizing events that bring physicians from different specialties together (e.g., multidisciplinary case conferences), thereby growing metaknowledge of “who knows what” within their health systems.

Limitations

Readers should view our findings in the context of several limitations. Although our models control for many confounders, unmeasured factors may bias our results. For example, our models may not sufficiently capture differences in patients’ condition prior to CABG. In Online Supplement F, we show that our findings are robust to additional adjustments for patient complexity. However, future work may better measure disease severity by linking Medicare claims data to other sources (e.g., the Society of Thoracic Surgeons National Database) with more rigorous risk adjustments. Our findings may also be biased if more progressive health systems promote integration around surgical care and also take other (unmeasured) steps to reduce spending.

Administrative data allow us to examine differences across many health systems over time, which would be prohibitive with other methods. Nevertheless, our reliance on administrative data may omit some important relationships (e.g., curbside consultations) and care professionals (e.g., advanced practice providers) from our maps of health system networks, while including others that are less important. Although anesthesiologists are sometimes deeply involved with postoperative management of cardiac surgery patients, we exclude them from our physician networks because we cannot distinguish their care from providers who deliver anesthesia in the operating room. In Online Supplements B and F, we address some potential concerns regarding the construction of our network maps using statistical methods and simulations, the results of which add confidence to our findings. Validation studies also offer support for our approach. In one analysis,¹⁵ physicians were surveyed about their professional relationships. Responses from this survey were matched to Medicare claims. Surveyed physicians recognized up to 82% of claims-based relationships. Using a similar approach, a different study found that network measures based on claims were associated with perceptions of care team climate, as reported in surveys.²⁷ Assuming the structure of omitted relationships does not vary systematically from those observed, these findings suggest bias in our results should be minimal.

Implications

Notwithstanding these limitations, our study has several policy implications. Our findings suggest that health system administrators and policymakers may benefit from viewing formal and informal clinical integration as two distinct phenomena. Although programs like ACOs and patient centered medical homes may improve formal coordination, it is possible that their influence over physicians’ informal relationships is limited. Put differently, programs that aim to deepen integration through formal means may be acting on the tip of the iceberg, while leaving many informal, sub-surface connections untouched.

Our findings also suggest the possibility that informal integration may contribute to the success or failure of formal programs aimed at reducing fragmentation. Imagine that health systems A and B in Figure 1 adopted identical programs designed to incentivize coordination among primary and specialty care physicians. 1 year after implementation, evaluations demonstrated improvements at B but none at A. Although we may be surprised to see different outcomes at comparable health systems, these results make sense when we see that informal integration is initially far lower at A.

Finally, our index of informal integration may prove useful for health system administrators and researchers. The index can be easily calculated using administrative claims. Moreover, the index’s normalized -100 to 100 range helps facilitates comparisons across health systems. These features suggest our index may be valuable for as a diagnostic tool for identifying clinics, departments, institutions, or partnerships that are ripe for interventions aimed at improving care relations among specialists.

Conclusion

This study drew on insights from network analysis to develop a novel index for characterizing informal integration among primary care and specialty physicians within health systems. Greater informal integration was associated with lower episode payments for CABG.

Key Points.

Question

Is deeper informal clinical integration likely to help control spending in surgery?

Findings

In this study, the degree of informal interaction among primary and specialty care physicians around shared coronary artery bypass grafting patients was associated with lower payments from Medicare. Observed savings were greatest on readmissions and post-acute care.

Meaning

In addition to targeting the formal organization of surgical care, delivery reforms may benefit from attempting to foster deeper informal integration among primary and specialty physicians.