Valuenomics in Practice: Plugging the Leaks on Referrals

Missed our previous post? We suggest you start here,
with “Valuenomics in Practice: Redirecting Traffic to the ED”

Referral management is a broken system, and a significant source of avoidable costs for provider organizations trying to succeed under value-based care. Innovaccer’s Chief Medical Officer, Dr. David Nace, describes the evolution of the problem from experience:

"There’s been a sea change in referrals since I was in primary care practice twenty-five years ago. Back then, primary care docs and specialists used to collaborate. If I had a patient with an issue that was out of my league, I’d reach out to a specialist I thought highly of with a nice letter asking them to see my patient. Usually, they’d do so right away because of our personal connection. After the consultation, they’d send me back a letter thanking me for allowing them to see my patient and providing their evaluation and recommendations. Now, when I refer a patient, that’s the end of the story. I almost never hear back from the specialist. I don’t know what treatments the patient received, whether their issue was resolved, or where they were being sent next. It’s like they disappear."

The personal network Dr. Nace and other providers once relied on to make their referral decisions is largely a thing of the past. Maybe that’s okay. Dr. Nace based his decisions on relationships and educated perceptions, but not every physician had such strong connections to their peers or sufficient insights about their expertise and approach.

That said, today, referral decisions aren’t better. They’re much, much worse. They’re often made based on an outdated list of specialists with whom the physician might have no relationship at all—and there’s next to no collaboration between the PCP and specialist.

For a provider organization trying to succeed under value-based care, that presents three problems:

1. They’re not using quality or cost data to drive the referral decision
2. They can’t control or track where the patient goes once the referral is made
3. They don’t have the data to analyze and optimize the performance of their referral network

These operational gaps can cost providers thousands of dollars more per patient per referral than they should. And that, of course, adds up to millions in potentially unnecessary spending across a healthcare enterprise and billions for the US healthcare system as a whole

The Costs of a Referral Gone Awry

Providers operating under value-based care expect to rely on primary care physicians (PCPs) to help control costs and improve quality and experience. The problem is that PCPs have little direct control over costs, quality, or experience once they refer the patient to a specialist.

The bulk of those costs are largely in the hands of the specialist(s) downstream. Which specialists offer high quality care? Which are most cost-effective? Which ones have openings? Which are in-network for the patient? When the PCP makes their referral, they might as well be playing darts in a windstorm while wearing a blindfold. It’s very unlikely their throw hits the target.

Cost and quality can vary considerably across specialists. One cardiologist might be more likely to put in a stent rather than perform a bypass, despite a significantly higher rate of complications. One orthopedic surgeon might refer the patient to a chiropractor, while another might perform a hip replacement at a price considerably higher than the regional average.

Fig. 1: De-identified data shows the cost differential of a total knee arthroplasty between three specialists who receive referrals from the same PCP

It’s not that difficult to measure costs, but it’s also possible to measure quality. You just need to decide which data and what measures to incorporate. CMS’s Bundled Payments for Care Improvement (BPCI) initiative can provide guidance for certain procedures, conditions, or circumstances. But for some specialties it’s still a challenge. You might want to use days in hospital, readmissions, and patient satisfaction scores, for example, as part of your assessment, if those measures really make a difference on your performance.

Here’s an example we encountered. Examining data for referrals to orthopedic surgeons, we noticed that one surgeon performed significantly better from a cost and quality standpoint. After surgery, this ortho’s patients cost an average $2,000 less and had 35% fewer readmissions. When we dug into the reasons why, we learned that the ortho would refer patients to a chiropractor rather than a physiotherapist for their post-op surgical care, a decision, in this case, that reduced costs and improved outcomes.

A PCP doesn’t have time to seek out and analyze all that data. They just want a definitive answer for making a referral. What if they could look up a list of potential specialists and see scores next to each one that reflect the cost and quality data as a “rating” number for each (i.e., one rating for cost, one rating for quality—not too dissimilar to Yelp’s star ratings?

Network Leakage and Systemness

Then there’s the problem of network leakage. PCPs don’t know whether a specialist is available and can schedule their patient in a timely fashion. They don’t know whether their patient will follow through on the referral. They don’t know whether that specialist is in-network. And they don’t know where the specialist might refer the patient next, as that patient pinballs down the line. 

The ideal is a closed-loop referral. The patient receives a referral to a high-value specialist (with the right cost-quality score); the patient is able to make an appointment in a timely fashion and follows through on that appointment; and the primary care physician is informed of the results.

According to a 2018 NIH report, only 34% of referrals resulted in documented completed appointments. CMS puts that number at 50%. Patients may not follow through at all, or they may give up because the specialist does not have availability, does not respond, or is too costly because they are out-of-network. Specialists report back to the PCP less than 20% of the time.

Fig. 2: Closed-loop referrals help prevent network leakage and reduce costs by targeting high-value specialists with the best cost-quality scores

PCPs may not readily know whether the specialist is in-network, uses the same EHR, and operates under the same cost parameters. That problem only compounds once the specialist has seen the patient and refers them for tests, images, procedures, and interventions without knowing whether those are in-network, also.

The most innovative provider organizations are striving to become digitally integrated networks of PCPs and specialists. Even independent PCPs and specialists can be part of that network if they fit within its cost and quality parameters, and enable closed loop referrals.

This can create a winning situation for all stakeholders. The participating specialists get referrals and patient records. PCPs get reports on their patients. The provider organization has more control over costs and quality, even as it works with a large network of docs who may be operating under fee-for-service arrangements. And the patient gets the benefit of accessible, coordinated, high-quality care.

It’s not hypothetical. Here’s actual, de-identified data from a Valuenomics analysis we performed for a major payer, looking at one of their provider networks. In this case, using data and analytics technology to stratify specialists and manage PCP referrals could help save this payer more than $8 million a year.

• The average cost for total hip replacement across the top 5 physicians*: $9,000
   
  • If PCPs refer their patients to these top 5 physicians only, total cost will decrease by $600,000.
     
• The average cost for total knee replacement across the top 5 physicians*:  $9,500
   
  • If PCPs refer their patients to these top 5 physicians only, total cost will decrease by $960,000
     
• The total potential opportunity to reduce cost in DRG 470 alone: $1,560.000. Extrapolating across all major DRGs, there is a potential opportunity to save more than $8 million.

* Considering only orthopedic surgeons driving at least 2+ DRG 470 episodes

A Value-Based Approach

All of that takes technology. Think of it like Open Table for making dinner reservations. Data on reviews, results, performance measures, and availability all feed into one platform. The PCP and patient can easily see which specialists have the most value, when they have availability, and make that appointment.

The specialist gets timely information about the patient and closes the loop with the referring physician, who can discuss results with the patient and facilitate next steps. And the patient has a comprehensive longitudinal health record for future care episodes.

That doesn’t happen in healthcare today, but it should. And we’re working to create it, right now. For organizations starting to assume responsibility for total cost of care, outcomes, and experience, that kind of data and information flow is paramount for success.

The time is right, and the technology and data are available, to bring relationships, collaboration, and insight back to referral practices—with one profound difference: now we can know and influence whether we’re providing the best care at an optimal price.

The Cost/Quality Conundrum

As reported in the Annals of Internal Medicine, although there is broad policy consensus that both cost containment and quality improvement are critical, the association between costs and quality is poorly understood. Future studies should focus on what types of spending are most effective in improving quality and what types of spending represent waste.

Read the Entire The Valuenomics Series

• Introducing Valuenomics: the New Moneyball for Healthcare
• Why Healthcare Needs to Play Moneyball
• General Principles of Managing Risk
• From There to Here
• Healthcare’s Risky Business
• Wrestling with 30-Day Readmits
• Sniffing Out Low-Value SNFs
• Valuenomics in Practice: Redirecting Traffic to the ED