Home/Inspire/Data Intermediaries: Pulling Insights from Confidential Data

Data Intermediaries: Pulling Insights from Confidential Data

Over the past several years, I have had the privilege of providing actuarial support to health systems pursuing alternative payment models and value-based reimbursement arrangements. In some cases, the health system was approached by a managed care organization (MCO) as part of contracting. In other cases, the health system themselves wanted to develop a strategic approach to all future arrangements.

In our other Inspire white papers, you can read about the basics of value-based reimbursement, the importance of care management, analytics to support successful programs, and levers available to manage costs. In this article, I take a very focused view at certain data management issues arising from these payment models. I assume the reader has some basic familiarity with alternative payment models; if not, I suggest visiting our Inspire library to learn more.

Alternative payment models, value-based reimbursement, risk-based arrangements and similar terms are frequently used to describe different contracting arrangements where ultimately, the provider is accepting some transfer of risk. The concepts presented in this article are applicable in any situation involving provider risk. The examples are based specifically on hospitals and health systems as providers.

What’s the Problem?

With risk-based arrangements, health systems need more data than ever before – but in many cases, such detailed data isn’t available because of MCO concerns about confidentiality. An objective process is needed to meet both needs, that the health system and the MCO trust. At the same time, actuarial expertise is needed in order for the health system to quantify and manage the risk they have assumed under the value based contract.

In the remainder of this article, I will identify some common issues with risk-based contracting, provide examples of a major data problem that arises in practice, and demonstrate how a data intermediary may address that problem.

Key Issues

Most emerging payment models have some common characteristics, including:

  • A “performance” element, which typically includes measurement of the provider’s progress towards a baseline or target expectation
  • Full or partial delegation of care management duties by the MCO to the provider
  • Division of financial responsibility, covered services list, or some other documentation of the specific services covered under the arrangement

Considering these characteristics, the parties to the agreement will therefore need certain capabilities, some of which are listed below, from the health system perspective:

  • Credible and auditable performance measurement, where performance targets are frequently centered around specific cost and utilization metrics
  • Regular reporting of analytics to identify opportunities for improving performance
  • An understanding of what care is regularly being delivered outside the health system’s direct control, for which the provider has accepted some risk

And while I won’t cover it in this article, the health system also needs sufficient information to effectively perform the delegated care management duties and identify opportunities to improve care management. For more information on the importance of care management effectiveness, I refer you to David Axene’s forthcoming Inspire issue on this topic.

As shown in the examples below, data challenges arise when some portion of the arrangement involves utilization of non-system providers. Addressing the important capabilities listed above means sharing data, and in some cases the data-sharing becomes problematic.

Example: Capitation

In global capitation and certain partially capitated arrangements, some portion of care may be delivered outside the health system, which must be paid for out of the system’s capitation revenue. Let’s consider the following questions, from the health system’s perspective:

  • How do I develop an appropriate capitation rate, or know if the rate proposed to me by the MCO is reasonable, if I don’t know what providers outside my system charge for their services?
  • How can I monitor and track financial performance over time?
  • How do I report specific cost and utilization metrics by department and service level?
  • How do I identify specific opportunities to improve care management by bringing out-of-system care into the system, and how much money might I save?
  • Can I rely on MCO reporting to provide a clear picture of out-of-system cost and utilization?

These questions center around the critical need for the health system to understand what is happening outside their system. Ideally, the MCO would supply detailed claims and eligibility histories, including all dollars spent for all providers, for the past 12 to 24 months.

These questions center around the critical need for the health system to understand what is happening outside their system.”

In reality, the MCO has negotiated deals with every provider, and the confidentiality of those deals is of the highest priority for the MCO. With very good reason, the MCO does not want any provider to know what it has agreed to pay other providers.

Under a capitated arrangement, however, the health system may be financially responsible for the dollars paid to those other providers. How does the system answer the questions above, if the MCO cannot divulge those other costs?

Example: Multiple MCOs

I have worked with clients where there was a need to collect data across a group of competing MCOs, combine that data into a single dataset, and report results to the client and/or all of the MCOs. Questions from the MCOs included:

  • Can you demonstrate to me how you intend to keep my data confidential, not just from your client, but from the other competing MCOs?
  • What insights can you give me with respect to my position in the marketplace, without compromising confidentiality requirements?
  • Can you provide some reassurance that your interpretation of my data is correct?

In these cases, the MCOs were as eager to find a solution as the clients, because they were pursuing innovative care and reimbursement models. The success of such models hinged on the ability to share meaningful information.

Role of the Data Intermediary

In all of these situations, the solution was to identify a data intermediary. The data intermediary was charged with collecting and “sanitizing” information before providing data to the client for analysis. In some cases, the data intermediary was also responsible for reporting specific analytics which require analysis of detailed data.

This approach shouldn’t be conflated with data warehouses or electronic record systems. While such systems can be useful, the data intermediary approach is specifically attuned to the actuarial perspective on risk-based provider arrangements. The actuary, as part of data intermediary responsibilities, handles certain activities that require calculations on confidential data.

Actuarial expertise is typically central to successful risk-based arrangements, in terms of quantifying the risk, implementing risk mitigation tactics, and identifying actionable analytics (and the methodologies used for developing those analytics). In practice, the data intermediary approach is intended to provide the detailed data necessary for me to fulfil my role as the actuary, generate sanitized data for the health system, and address concerns about objectivity and auditability of reported results.

Structure & Process

Conceptually, the structure and process for a data intermediary is simple – (1)the MCO(s) supply detailed data to the data intermediary, (2) the data intermediary clears the data1, (3) the data intermediary prepares the dataset for the client, and (4) the data intermediary manages actuarial functions.

The data intermediary is most often an actuarial firm with significant expertise in risk-based contracting, data management, and actionable analytics. In practice, there are multiple questions to be resolved in planning, besides the standard procedure (frequency, data format, etc.):

  • Does the actuarial firm have any conflicts of interest? Many actuarial firms consult with both health systems and MCOs.
  • What agreements are needed between the MCO and the data intermediary?
  • How can we structure data layouts that don’t require significant resource investment from the MCO?

Actuarial functions frequently include development of rates for capitation or bundled payments; summarization of dollars spent outside the health system by service line; health risk scoring; and calculating performance analytics. Where multiple MCOs are involved, the actuary may provide MCO-specific analytics privately to each MCO, to allow them to compare their results against those for the whole population, and to validate the calculations performed by the actuary.

Performance metrics and analytics may be generated by the actuary as well; in other cases, the data intermediary produces files that can then be fed into analytics programs or used for in-house ad hoc analysis.

Ultimately, the actuary attests to the confidentiality of non-system provider data; compliance with agreements; the appropriate use of such data to develop work products; and the accuracy of calculations and methodologies.

Important Considerations

This approach is not without challenges.

  • MCO participation and capabilities. When I began working in a data intermediary role, I was fascinated by the wide range of capabilities for producing and reporting data at MCOs. MCOs face a variety of challenges: turnover in programming staff; historical changes in platforms; and prioritization of this effort as compared with the other demands placed on the programming team. In some cases, the MCO outright refuses to provide the historical data.
  • Paying for data intermediary services. Using a data intermediary is not an inexpensive proposition; in particular, the initial set-up and implementation can consume heavy resources depending on the reporting capabilities of the MCO. Health systems may want to consider an initial trial period and/or starting with a “bare-bones” structure. Efficiencies may be realized by integrating data intermediary and analytics activities.
  • Ultimate control of the process. Several of my clients needed data intermediary services at the onset of a risk-based arrangement, but have a long-term goal of bringing the services “in-house.” I have found that with sufficient documentation and training for both the client and the MCO, it is possible to bring many of these functions in-house.

I have had clients ask why they shouldn’t just go straight to in-house control. Some organizations can certainly take that route successfully. However, for the majority of my clients, an interim step is needed to establish the proper model by identifying, documenting, and testing the data intermediary functional tasks. Other clients may want to continue using data intermediary services indefinitely.

Collateral Benefits

Aside from the objective solution provided for data-sharing, I have found that exploring the data intermediary process (even if it’s not ultimately implemented) has delivered multiple benefits.

  • Partner readiness. Earlier in this article I described the variability of reporting capabilities among MCOs. Frankly, the ability to report simple datasets is an excellent way to evaluate the MCO’s readiness for the arrangement. While larger national and regional MCOs have excellent capabilities, smaller organizations will frequently pursue risk-based arrangements without necessarily building up the infrastructure to support the needed reporting.
  • Building trust between the health system and the MCO. It isn’t uncommon for MCOs and health system providers to disagree on reported results, especially if such results affect how much the health system is paid. Starting with a data intermediary establishes an objective process addressing concerns on both sides.
  • Focusing attention on what’s really important. A common affliction for practicing actuaries is that we may spend so much time collecting and clearing data, that we don’t have enough time to examine and interpret the actual results coming out of that data. The situation is similar when MCOs and health systems share data directly, requiring significant internal resources and leadership attention, which could be better directed toward analyzing performance and finding care management opportunities.

With all of this, I recognize that the decision to enter into a risk-based arrangement depends on more than just data sharing. However, data sharing can often become a major point of disagreement, and a barrier to the final agreement.


Health systems who take on risk need information that reflects data confidential to the MCO. The data intermediary serves a critical role not only in enabling the sharing of confidential data in a way that’s acceptable to all parties, but also in objectively enforcing and supporting the actuarial methodologies associated with risk-based arrangements. Furthermore, beginning the data intermediary discussion will help you understand the capabilities and readiness of your potential partner, regardless of how you choose to proceed.

1 I often use the term “data clearance” to describe what happens when a dataset is received. Data clearance includes acquisition of the data, general review of the structure and compliance with the layout specifications, reasonability testing, identification and resolution of visible data issues, and uploading for storage.

Download PDF

About the Author

Elaine Corrough, FSA, FCA, MAAA is a Partner at Axene Health Partners, LLC and is based in AHP’s Portland, OR office.