Patient Panel Analytics

In a recent interview with a provider about how to gain some efficiencies in her practice, I asked how many patients she was caring for with a diagnosis of cancer in the past few years. After a “from the hip” answer, I showed her a report from one of her payers, and she became frustrated that the payer had a better summary of her patients than she could obtain from her own EHR.

Healthcare providers and management need to be empowered with tools to analyze information about their practice in which much effort is spent creating the data.

This podcast will demonstrate our Accountable Care Analytics Application’s ability to define patient panels and provide integrated summaries of patient information from clinical and claims data sources.

The Politics of Data in an ACO

Data sharingImagine the following scenario. You discover that you are the victim of identity theft, purchases have been made in your name, and your personal credit has been ruined. You are saved, though. You have paid a watchdog organization to monitor your credit, and they have information that clears your good name! So, when you apply for a loan with a bank, you request the credit monitoring agency to share the details of your prior credit problems and its resolution with the bank. But the monitoring agency will not share that information because that might help the bank understand your needs and negotiate a better price for their own credit monitoring service that they resell to their customers – i.e., you. The monitoring service won’t release your information.

Would you put up with this conflict of interest? NO!

In healthcare, we routinely tolerate a form of this conflict of interest, and in many different forms. Even though health insurers are not providing the patient care directly, these payers tend to accumulate a very useful holistic view of each patient’s history, including information from the perspective of what care was provided based on payment being demanded by many different care providers. There are numerous instances where, if this information was shared with other providers, it could positively impact the care management plan, doing so in a more timely manner, and increasing the likelihood of improving the quality of care the patient receives and possibly reducing the overall cost of care across an extended episode.

Here is an example- A patient is admitted to the hospital and receives a pacemaker to address his atrial fibrilation. After being discharged, the patient follows up with his cardiologist who has reduced the dose of digoxin, having diagnosed the patient with a digoxin toxicity. However, the patient attempts to save a few dollars by finishing their current prescription only to be admitted to the hospital a couple of weeks later for the toxicity. This is an opportunity where the care manager could have intervened based on the cardiologist’s toxicity diagnosis being submitted to the payer and no prescription was filled within a few days. The care manager could have helped the patient be more compliant with the cardiologist’s instructions avoiding an inpatient admission.

Healthcare provider organizations and payers (and in some cases regulators) are working together to break down these walls in an effort to increase value across the spectrum of care delivery and the industry in general. However, the sometimes conflicting vested interests of these interacting payers and providers can still be an obstacle, influencing the politics of information disclosure and sharing in the emerging environment of accountable care delivery models.

There is great diversity in the participating organizations that collaborate to make up an ACO. This is definitely not one size fits all. Viewed from the perspective of sharing risks across parties without the immediate concern about maximizing volumes, the integrated provider-based health plans, such as Kaiser Permanente, Geisinger Health System, and Presbyterian Healthcare Services, are already inherently sharing this risk and are reaping the rewards as a single organization. That’s great for the few organizations and patients that are already members participating in one of these plans.

Unfortunately, for other organizations there is still much more to be worked out regarding proactive sharing of data both within an accountable network of providers acting across care settings, and with the payer(s). Within the network, hospital systems usually have some of the infrastructure in place and they know how to routinely share data between systems and applications using standard data exchange conventions such as HL7 and CCD. In collaboration with HIE’s these systems can help facilitate active data distribution, and they very often provoke the organization to address some of the more common aspects of data governance. However, even when this routine “transactional” and operational data is being exchanged and coordinated, there is still a great unmet need for the ACO to buy or build a data repository for the integration and consumption of this data to support reporting and analytics across various functional areas.

Many organizations encounter further challenges in defining and agreeing on which are the authoritative sources of specific elements of data, what are the rights and limits on the use of these data, and how can these assets be used most effectively to facilitate the diverse objectives of this still-emerging new organizational model.

An even greater challenge for some ACOs is collecting the required data from the smaller participating provider networks. These organizations often have less capability to customize their EHRs (if they even have EHRs in place) and less resource capacity to enable the data sharing that is required. To get around this, some ACOs are:

  • Standardizing on a small number of EHRs- (ideally one, but not always possible) This provides the potential to increase economies of scale and leverage the shared learnings across the extended organization.
  • Manually collecting data in registries– Although not always timely, this addresses some rudimentary needs for population-focused care delivery and serves to overcome common barriers such as the willingness of a given provider to collect additional required data and complying with standards.
  • Not collecting desired data at all– While this seems hazardous, progress toward the overall clinical and/or financial goals of the ACO can still be positive, even if an organization cannot directly attribute credit for beneficial outcomes or improvements within the organization, and the ACO can avoid the overhead of collecting and manually managing that data.

Regardless of what data is collected and shared within the ACO, the payers participating often have the highest quality, most broadly useful longitudinal data because:

  1. The data is ‘omniscient’ – it represents, in most cases, all of the services received by (or at least paid for) that patient – provided a claim for those services has been submitted and paid by the participating payer.
  2. Some of this data is standardized and consolidated making it easier to manage.
  3. The data is often enriched with additional data residing in mature information systems such as risk models, and various disease-focused or geographic populations and segments.

Consequently, payer data very often forms the longitudinal backbone that most consistently extends across the various episodes constituting a patient’s medical history and is very important to the success of the ACO’s mission to drive up quality and drive down costs. Despite this opportunity for an ACO to improve its delivery of care to targeted populations, sharing of data is still achieved unevenly across these organizations because some payers feel the utilization, cost and performance data they have could be used to negatively impact their position and weaken their negotiations with the hospitals and other provider organizations.

While claims have traditionally been the de facto standard and basis for many of the risk and performance measures of the ACO, more progressive payers are also now sharing timely data pertaining to services received outside the provider network, referrals between and among providers, authorizations for services, and discharges, further enabling ACOs to utilize this information proactively to implement and measure various improvements in care management across the spectrum of care settings visited by patients under their care.

Collaboration between provider organizations and payers at a data level is moving in a positive direction because of the effort given to ACO development. These efforts should continue to be encouraged so as to realize the possibilities of leveraging timely distribution of data for better treatment of patients and healthcare cost management.

Moving from Volume to Value: How Do We Get There?

volume to valueListen to any healthcare pundit or industry observer longer than their opening paragraph and you’ll hear them use the current buzz phrase: Healthcare needs to move from volume to value.  See, there it is already.

We pretty much know where the volume comes from.  If a buyer, any buyer, agrees to pay an acceptable fee for each unit of a needed service, the service provider will soon recognize that the delivery of more units of those services to satisfied buyers yields more payments.  Simple enough, and healthcare service providers have responded as one would expect in this environment.  The fees paid for each unit of service motivate the healthcare provider to maximize throughput – to the limits of their capacity to deliver a satisfactory service – and this leads to an increased volume.  Yes, there’s also quality and necessity and regulation, but right now we’re talking about volume.

So, what about the value part?  What is going to drive value?  And value to whom?  And who is going to measure the value?  And decide how much to pay for it?  Unfortunately, several tenets of basic economics that ordinarily drive value (and operate in virtually every other transactional setting) are disrupted in the healthcare marketplace as payers and providers of every shape and size, employer- and other group-inspired benefit plans, preferred provider and referral networks, watchdog quality and safety groups – as well as the inherent complexity of the subject matter – all serve to distance the patient from a free and informed buying decision.

By the time the ultimate treatment decision is made, it has been framed and prodded by so many ancillary parties that the patient, sitting alone with the provider, can almost feel the other observers in the room; those who will decide after the fact, or who have decided well before the fact, whether this decision is appropriate and how each party in the transaction will be compensated, billed, measured, rewarded or penalized according to a growing litany of performance measures.  If the patient doesn’t feel all of this, the provider often does.

The complexity and confusion arise in part because each of the above fundamental elements has been intermediated – to one degree or another.  The patient – the ultimate target of the treatment – is not the only buyer.  Buying decisions affecting this single transaction were defined, negotiated and contracted months or years in advance, and are being monitored against a wide range of both clinical and financial measures before, during and after the single transaction between a given patient and a given provider.  The terms of these agreements can and do influence the chain of decisions that culminate in the choice of treatment and in the cascade of financial events that will promptly follow.  No wonder it’s confusing.

So then, how does value get defined?  Several common themes emerge as both payer and provider organizations strive to identify the appropriate fundamentals, define a useful and informative notion of value, and introduce that notion of value into the decision processes they share with their patients.

The common elements that lead to a determination of value seem to go something like this:

  • Who are my patients?  A fundamental question, but not always trivial to answer accurately or in a useful way that enables and extends visibility over a population.  Providers need to be able to identify each patient that is legitimately under their care and they must have access to a complete record of the care these patients have received as a baseline for measuring future performance.  Once this record is assembled the pattern of problems, interventions and care relationships can be discerned and used to both characterize and engage each patient.

Providers need to identify the core characteristics of their panel of patients so they can both tailor individual treatments and evaluate patient experiences and outcomes comparatively against similar patients they are treating or that are being treated by other providers.

In an accountable care world, if providers are assigned responsibility for patients retrospectively using a plurality of care or other statistical model, it doesn’t mean they have control over the care those patients are receiving. They can hardly be measured fairly on the outcomes those patients have experienced. They need to know the specific treatments these patients have received, their level of compliance, and what other providers they have seen, at what locations, and with what frequency. This increased understanding of their basic patient panel will begin to reveal the true nature of the relationships they (or others) have with these patients and will often constitute the first wave of relevant analytics into the value being delivered.

  • What outcomes are we targeting for these patients?  What are the care plans that will get them there?  How long have these targets and plans been in place, for which patients, and what results are we seeing?The segmentation analytics that was started with the patient panel can now be extended, as specific performance targets are defined for individual patients and the projection of these targets is aggregated into clinically coherent segments, yielding outcomes and results that perhaps for the first time give visibility and insight into how well the relevant population is being managed.

Care teams and practice management can now monitor the clinical, operational and financial performance measures of the segments that drive significant costs and consume substantial resources, enabling the exploration of new deployment models.

  • What is our baseline? For the patients’ and other payers’ expenditures for our services and for the actual costs we incur to deliver those services?  Virtually all risk-based contracts establish a baseline of expenditures using some form of statistical measure (e.g., weighted average) over a defined historical time frame.  Projecting the dynamics of patient mix, service mix, fee structures and delivery resources over the anticipated life of the contract provides a segment-able baseline for measuring and tracking contract performance and assessing value. From this foundation, organizations can apply complementary analytics so that under-performing practice areas or population segments can be localized and improvement programs can be appropriately focused and funded.  Over-performing segments can be examined and highlighted as potential sources of best practices targeted for broader dissemination.
  • Who is accountable, and for what?  As the payment structure for many conditions moves to more of an episode-based model, the deployment and coordination of care delivery resources takes on added significance.  Roles and responsibilities must be defined for the delivery of episode-focused clinical services across the network of care settings. Proactive coordination of transitions in care and the associated communications, hand-offs and follow-ups must be defined and written into performance contracts along with explicit adherence measures.

These metrics will begin to form the basis for concrete and measureable accountability models and will likely be a consideration when shared gains and losses are assessed retrospectively.  Evolution toward more proactive accountability models is likely to follow.

Accountability models based on the actual outcomes realized, as distinct from adherence to best practices, can be differentiated through analytics, enabling some flexibility for care redesign (potentially including patient choice) or other measured innovations undertaken by providers.

  • Are we correctly and accurately reconciling the various activities, billed services, payments, resource alignment and costs with our agreed-upon models for accountability?  This is non-trivial even within a single enterprise.  And now we have various ACO or ACA models where new participants are collaborating at levels they have never attempted before, and entering into risk agreements based on shared performance metrics.  Some organizations are experimenting with formal value stream maps where benefits and costs are explicitly modeled.  Others are punting any envisioned gains (or losses) to an aggregate ‘shared benefit’ to be ‘addressed later.’One key consideration is implementing at least some accounting (defining and tracking) of the revenues and the actual costs associated with care delivery to specific segments with different characteristics (e.g., populations, locations, groups, payers, service partners or venues).  These costs must not be (but often are) confused with the amounts the provider would like to charge; or the allowed amounts the payer will agree to; or the actual payment amounts received from all parties; or even the various provisional ratios used to approximate the real costs.  Accountability models will need to evolve much further if they are to offer any real operational decision-making value.

No one disputes that the changes underway in healthcare have the potential to be transformational, to varying degrees.  The complexity and diversity of the responses that will be required by various organizations is still taking shape and there are many variables that will determine the success that any given enterprise will achieve.

The core principles outlined here are being adopted and applied in diverse healthcare organizations to answer a few fundamental questions about the value they offer that, ironically, have been posed and answered for all time in other industries and economic settings.  Who is our customer?  What do they need or want?  Why are we the best organization to meet their needs?  How can we communicate the benefits and costs to all the parties who are involved in the decision to buy?  Can we deliver?  How can we measure these factors both as a baseline and on an ongoing basis so we can provide convincing evidence that we offer the best proposition of value to all concerned?  Healthcare organizations that can answer these questions and address the numerous issues that arise in their pursuit will have a leg up on everyone else and will both deliver the best value and enjoy the greatest success.