What We’ve Learned from Horizontal and Vertical Integration of Physicians

flying cadeuciiAs health care reform rolls out, there is a growing focus on restructuring the health service delivery system in the hope of improving health care quality and “bending the cost curve.”

A key part of this focus has been on physician organization and, in particular, moving toward large, multispecialty physician groups or hospital-physician systems that can provide integrated, coordinated patient care (e.g., through “Accountable Care Organizations”).

In a recent chapter in Advances in Health Care Management’s Annual Review of Health Care Management, however, we and our co-author Jeff Goldsmith find that there is little evidence for the superiority of these integrated models in terms of patient care quality or cost-savings, and that the trends toward physician consolidation has been much less dramatic than is often thought.

Using data from a variety of sources, we find there are two separate phenomena at work in physician organization. At one end of the spectrum (bottom tail of the size distribution of physician groups), the majority of physicians continue to practice in small groups, although there has been some movement from really small practices (one to three or four physicians) to slightly larger groups (five to nine physicians).

Still, nearly two-thirds of office-based physicians continue to practice in solo settings, two-person partnerships, and small (usually single specialty) groups with five or fewer physicians.

At the other end of the spectrum (upper tail of the distribution), however, is a smaller number of very large and rapidly growing multispecialty physician groups, which are often owned by hospitals, health plans, private equity firms, or other non-physician sponsors.

These two stories of what is happening in the distribution of physician group size are described as “a tale of two tails.”

Despite the relative stability in the distribution of practice sizes, the upper tail accounts for an increasing percentage of practicing physicians and the most rapid growth in total physicians and physician visit volumes.

Key integration forms
There are two key forms of integration in physician markets. Horizontal integration occurs when individual physicians join group
practices or existing groups merge with each other. There are numerous theoretical reasons to expect that this type of integration might lead to improved quality and cost savings, including enhanced operating efficiency and economies of scale.

When groups consist of physicians across multiple specialties, there also may be gains from economies of scope, including improved coordination and quality of patient care, in-house or within network referrals, and capture of high-revenue services such as outpatient surgeries and imaging services.

We find some evidence that group practices are more productive than solo practices in terms of number of patient visits or gross revenues per physician, though scale economies appear to be reached quickly — at around 10 or so physicians in a group.

Beyond this, monitoring, coordination, and creating a cohesive culture become challenging. Scope economies do not appear to exist or are, at best, weak. The authors also consider the evidence supporting vertical integration, when physicians align with non-physician partners such as hospitals, universities/medical schools, and health plans.

Again, there are many theoretical rationales underpinning these relationships, including lowered transaction costs and improved efforts to monitor, manage, and coordinate patient care, increased network size and geographic coverage to handle risk contracting, and market power over buyers and suppliers.

In addition, these relationships may help physicians stabilize their incomes, manage malpractice, and improve the predictability of their caseload. Because of the supposed advantages, and based on numerous media reports of physician hiring, many analysts assert that physician employment levels have reached as high as 50-60 percent of all doctors.

Questions about effectiveness
The data show that hospital-physician integration is on the rise, but not as fast and nowhere near the levels that analysts propose. The percentage of physicians employed by hospitals has increased from 11% in 2000 to nearly 15% in 2008.

There are also major questions about the effectiveness of the physician employment model, as it has led to financial losses approaching $200,000 per year per physician, double the losses incurred during the early period of integration in the 1990s.

There is similarly no evidence of improved quality following integration. Finally, there is evidence that economic integration between physicians and hospitals does not automatically lead to functioning clinical integration and that, even after integration, a lack of alignment between physicians and hospitals threatens the success of these models.

Another form of vertical integration is between providers and health insurance plans. There are several well-known examples of these systems, including Kaiser Health Plan and other major integrated systems (e.g., Mayo Clinic, Cleveland Clinic).

The success of these integrated systems can be attributed to a number of features, in particular a notably physician-driven system, unified clinical and administrative cultures, a long history with sufficient time to develop this culture, and strong economic interdependence among their three arms (the physicians, the hospital, and the health plan).

Despite the success of these models, they have had difficulty expanding beyond their core markets due to physician resistance, difficulty ramping up enrollment, and employers’ preference to contract with one plan rather than offer a menu of options.

Areas of renewed focus
With the implementation of the Affordable Care Act, there is a renewed focus on horizontal and vertical integration of physicians. Insurers have purchased medical groups in efforts to cut costs by managing patient care and physician networks more tightly.

There is also a movement towards “virtual integration” which allows a physician to remain independent but exploit some of the advantages of group practice, including centralized administration, risk spreading, and leverage with health plans.

Though the mass of physicians remain organized into small, independent, and fragmented group practices, there is clearly flux in the physician market with growth in the number of large groups and increasing physician employment by hospitals.

There is no evidence to suggest how each tail will fare competitively going forward, but past experience suggests that achieving widespread cost savings and quality improvements through restructuring the delivery system alone will be challenging.

To be most effective going forward, policymakers ought to revisit the evidence on integrated systems and consider how they may be implemented and targeted (e.g., to populations who really benefit from coordinated care, such as those with chronic diseases).

Lawton R. Burns, PhD is professor and chair of the health care management department at The Wharton School of Business at the University of Pennsylvania.  He is also a senior fellow of Penn’s Leonard Davis Institute of Health Economics.  

Aditi Sen is a doctoral student in health care management at the University of Pennsylvania. 

This post originally appeared in the Penn LDI Voices Blog

10 replies »

  1. Actually, the 2013 Edition of AMA’s Physician Characteristics publication listed a little over a million TOTAL physicians in the US, of which 767 thousand are “patient care” docs and 134 thousand are “inactive”. Sixty thousand are “unclassified” and another forty thousand in non-patient care roles like administration. Sorry about confusion.

  2. I meant more comparatively speaking. From 2003-08, it was more a gradual transition and the numbers I have didn’t show a huge surge overall as much as several specialties including cardiology as you mentioned. Where I found it difficult to ascertain is when a practice technically didn’t sell its assets wholesale but stuck an agreement similar to the fashion you mentioned.

    Hospital-owned practice data (in terms of raw numbers) from HIMSS Analytics isn’t great but it has been good enough to get a sense of what is occurring the market.

    Thanks for the comment about SK&A and as for the ‘retirement’ issue I know what a huge problem that is from working for a vendor that had a huge legacy practice management customer base. There was a category for ‘deceased’ and another for ‘retired’ when tracking customer numbers.

  3. Not accurate. The resurgence of physician employment by hospitals began in 2003, actually slowed from 2008-2010, then surged again in 2011 and 2012. A lot of the surge came from cardiology, where major independent cardiology groups basically put themselves out to bid, and chose the hospital or system with the most generous (or reckless) income guarantees. They weren’t “acquisitions” in the sense that hospitals were paying “goodwill” to acquire practice assets, but rather transitions to multi-year contracts.

    We’ve had a lot of concerns about the accuracy and completeness of SK&A’s data on physician groups. And lots of luck trying to identify physician retirements. The AMA Master File surveys a third of its members annually, and the last survey found over 1 million “inactive” licensed physicians.
    The only way to know for sure how many have actually retired is to track claims frequency.

  4. The trend for hospital hiring physicians though didn’t take off really until PPACA was passed in ’10 and then it boomed for a couple of years. From what I have seen/heard from physician recruiters, hospitals are starting to realize that it makes sense to employ docs only a very selective and strategic basis otherwise their productivity takes a real nose-dive & hospital bleed cash ala the 90s.

    Getting reliable data on physician-practice information has always been tough because you typically had to use an amalgamation of sources (AMA Master file for practices with >5 physicians, MGMA or AMGA or larger group practices). Accenture tracks this and puts out a report on it although their are some limitations to their approach.

    Best source nationally is SK&A at the practice-level but it is very expensive. Billian’s has just released a product though that looks very promising and is cheaper too.

  5. Nice to see a piece by Dr. Burns. I am a graduate of the Health Care Management program at Wharton.

    This piece does not discuss other types of horizontal integration, such as IPAs and ACOs. I am the CEO of a clinically-integrated IPA in CT which has 400 physicians and is an ACO in the Medicare Shared Savings Program. We also have a growing number of ACO-type shared savings contracts with large commercial insurers, some of which are headquartered in CT.

    In most of these ACO-type deals our physician group takes responsibility for the care management and quality measures for a defined population of patients, usually based on the collective patient population our nearly 200 primary care physicians have with a specific payer. Insurers provide our group with claims information for this population, which we upload into Athena Communicator Enterprise, the population health management application developed by our partner Athena Health. We then are able to monitor compliance with specific quality measures (i.e.HbA1C tests and results for diabetics or pediatric immunization rates for kids) at an individual patient/physician/POD basis. We make this information available to all our physicians for their patients via the Athena web app. Our care coordinators work closely with individual physician practices to help them identify gaps in care and their most complex and high cost patients. We are usually paid a PMPM care coordination fee to help cover the costs of the application and that of our care coordination staff.

    These ACO contracts have an opportunity for shared savings, most of which is achieved by reducing avoidable admissions/readmissions, ED visits and high cost imaging and Rx, as well as better reimbursement for physician services.

    As a physician organization that is not owned by or part of a hospital system, we are not constrained to the level of hospital-owned groups when working to reduced high cost hospital services which may not be necessary in all cases.

    We are finding that commercial insurers are very interested in working with our group, which is independent of any hospital or hospital system. We are also working on plans to partner directly with large self-funded public and private employer groups, as well as to partner with other similar physician organizations to make a large network of clinically-integrated physicians an attractive contracting partner. We are also working to partner with independent community hospitals which can provide services often at lower rates than some of the multi-hospital systems.

  6. @Carol
    Good comments, Barry. I’m afraid of the porosity of digital records unless encrypted. I have had terrified patients come to my office begging that I destroy their records having to do with abortion procedures and STDs. Also, I’m guessing here but I would guess that almost any significant illness would want to hidden from a potential employer. Also, imagine someone running for political office. You don’t want your opponent to know you have type 2 diabetes melitis or a hundred other ailments.

    Accordingly, although I have used computers in labs forever–recall the PDP8 with 8 k of hard disk memory–something else might have to be done for personal records….specifically encryption. It’s too easy for a popular Silent Spring type expose or something on 60 Minutes to come out and spoil the public’s acceptance of the EHR. I think this could be predicted. It’s almost as if the bad handwritten progress notes were designed for the proper level of security for the chart (from the patient’sperspective.)

  7. I certainly agree that coordinated care is potentially most valuable for patients with chronic conditions, especially CHF and diabetes. I also agree that there are comparatively few economies of scale for both hospitals and large physician practices.

    From a patient’s perspective, I like the idea of electronic records, more easily affordable by large provider groups, eliminating the need to make me answer the same questions each time I visit a different specialist. I also expect coordinated care to be more likely to eliminate duplicate testing and adverse drug interactions since all doctors treating me would presumably have access to the same records in real time.

    For primary care doctors, I understand that they are a tiny piece of healthcare spending even if we include the cost of care performed by others to whom they refer patients. The administrative burden of dealing with insurers is also highest for PCP’s because the payment system favors doing procedures over evaluation and management. Presumably, the administrative burden is comparatively light for high billing cardiac and orthopedic surgeons.

    The downside of consolidation, especially within the hospital sector, is higher healthcare prices driven by enhanced local and regional market power. I’ve said for years that there needs to be special rules that cover how much hospitals and doctors can bill for care that must be provided under emergency conditions. These rules are especially needed by the uninsured and insured patients who are treated by providers who turn out to be out of network.

    I predict that as more people are covered by high deductible health plans and narrow network plans, in-country medical tourism will become more acceptable to patients especially for relatively expensive procedures. Boeing and Lowe’s already have contracts with the Cleveland Clinic to treat employees nationwide who need heart surgery and the CC will do it for a very competitive bundled price for the care episode. Also, as more care is delivered outside of a hospital inpatient setting, costs should come down and quite a few hospitals will have to downsize, close, or find acceptable alternative uses for the space. Finally, the Kaiser integrated delivery system model could become more acceptable to patients as they get used to narrow network insurance plans, especially if the IDS includes a well-known and highly regarded teaching hospital.

  8. In my opinion, good medical care relies on the relationships of the physician and patient, especially for Primary Care. The better those, relationships, the patients will have trust and confidence in the physician, and the physician will have knowledge of the patient as a person, their beliefs, culture, emotions, etc.
    We see great emphasis on providers being sensative to patient’s cultural needs, yet if we continue to have a fractured and inconsistent system, how can those needs and those relationships continue to benefit both patient and physician?

  9. Folks want us to integrate to get our prices down. Coordination of care is used as a reason but there is no research showing care suffering when docs are solo. Hospitals want us in groups so that they can negotiate with us about prices and to affect our uses of the hospital…and to improve its bottom line. Insurers want us in groups so that they do not have to contract with an anarchic gang, each with different demands and so they may exert monopsony purchasing. Politicians want to get our input factor prices down so that their dreams can come true and bring in votes.

    We would like to remind these people that we are only 20%of the health care dollar and that our pen will not bring down costs if it writes in Tiffany’s. We also believe that the best way to get our prices down is to allow us to compete proceedurally with each other. Imagine Permanente and the Cleveland Clinic in the same town. Pharma and administration have to compete also.