CEO SUMMARY: In its work for a federally qualified health center, Sonic Healthcare USA helped physicians use a data-driven approach to population health management that incorporated integrated financial and clinical analytics. Also, Sonic developed technologies that give ordering physicians clinical decision support and targeted patient engagement tools. It then developed a way to contact patients who had gaps in care. From its work with this health center, Sonic was asked to be more than a lab provider.
WITH THE ERA OF FEE-FOR-SERVICE PAYMENT SOON TO END, all clinical labs face a common question: If labs will not be paid a per-test fee, how will they generate adequate revenue to sustain lab testing operations?
Nothing less than financial survival is at stake. If payers consider lab testing to be a commodity, then only clinical labs with the lowest costs will survive, but they will do so only by accepting the lowest rates.
Stated differently, the labs that thrive will do more than just report accurate lab test results for the lowest fee. Rather, they will provide diagnostic services that contribute to improving patient care in measurable ways.
For this group of labs, payers will measure value in two ways. First, they will want diagnostic services that help physicians document improvements in patient care. Second, diagnostic services will help reduce the cost of care for each individual encounter or the entire episode of care or both.
The good news for hospital and independent laboratories that go down this path is that health plans and physicians are willing to pay them for this increased value, particularly in the form of sizable shared savings payments.
Such is the case in Texas and New York with innovative collaborations that Clinical Pathology Laboratories (CPL) and Sunrise Medical Laboratories, divisions of Sonic Healthcare USA, have implemented. In recent years, a large multi-physician group operating as a federally qualified health center (FQHC) engaged Sunrise Medical Laboratories to help it use lab test data to improve patient outcomes and reduce the cost of care for patients with diabetes and chronic kidney disease (CKD).
Sunrise used Sonic’s iMorpheus, its informatics system, to provide data to manage the patients and help close gaps in care for the FQHC, accountable care organizations, physicians groups, and other integrated delivery networks. For these services, Sonic Healthcare negotiated reimbursement in the form of outcomes- and value-based contracts and shared savings arrangements.
In September, Philip C. Chen, MD, PhD, Chief Healthcare Informatics Officer for Sonic Healthcare USA, gave a presentation on this topic at The Dark Report’s Precision Medicine for Health Network CEOs conference in Nashville. Chen described how community-based physicians struggle to adopt new technology for clinical lab testing.
“We can do all the sophisticated lab testing we want, but it’s still very difficult to get community-based physicians to actually use these services,” he commented. “It’s difficult unless these physicians can see the value of such services in terms of improved patient care and the ability to use such services to develop value-based payment.”
In a case study for the FQHC, Chen outlined how Sonic helps physicians use a data-driven approach to population health management. “Sonic uses integrated financial and clinical analytics and deploys technologies that give ordering physicians clinical decision support and targeted patient engagement tools,” stated Chen.
“Our goal is to go beyond simply being a provider of timely and accurate clinical lab test results,” he added. “One way we learned to deliver more value to ordering physicians was to develop tools to contact patients who had gaps in care.”
Chen explained that once Sonic deployed the patient-contact tools it developed for the FQHC, other payers, including an accountable care organization (ACO), became interested in the cost savings potential of identifying patients with chronic conditions and using Sonic’s patient-contact tools.
“The full set of tools Sonic provided enables physicians to develop contracting strategies that helps them and Sonic get paid in settings beyond fee-for-service,” Chen said. “The FQHC physicians using Sonic’s informatics systems were using data to support value- and outcomes-based contracting and to collect shared savings from payers.
50% of Spending
“In our work with organized delivery networks, such as accountable care organizations and integrated provider networks like they have in California, we saw that health plans were spending very little money for most patients,” Chen explained. Nationwide, about 5% of patients account for 50% of all spending.
“In our work with one California health plan, we analyzed the claims status of their 75,000 patients and tracked patient expenditures from one year to the next,” he said. “For this health plan, we showed the health plan that it was spending very little money on a very large percentage—88%—of its members. But for 1% of its members, the health plan was spending an average of $61,000 per patient per year!
“What’s striking about following these patients from one year to the next is that some patients move from being low-expenditure patients to being high-cost patients,” Chen commented. “About 62% of high cost patients in one year (2014) were costing the health plan very little in healthcare spending in the previous year.
“At Sonic, we wanted to know if we could identify those patients before they started costing a lot of money,” he added. “To answer that question, we had to know why they suddenly started costing a lot of money. Then—as a lab provider—could we identify an opportunity to stop them from moving into the high-cost category?
“By analyzing the diagnosis codes for that high-cost group, we could list the most expensive patients per capita,” Chen explained. “Those patients fell into 16 disease conditions.
“For this analysis, we removed those patients who had a one-time event that is not preventable through healthcare management, such as a car accident or hip replacement,” he noted. “That left those patients who had one or more of the 16 chronic diseases for which physicians can intervene.”
Ranked in order starting with the most costly, those 16 chronic conditions are:
• Renal failure*
• Chronic liver disease
• Congestive heart failure (CHF)*
• Chronic obstructive pulmonary disease (COPD)
• Ischemic heart disease*
• Rheumatoid arthritis
• Low back pain
• Morbid obesity*
• Alcohol/substance abuse
• Mental/behavioral health
*Conditions that are comorbid with other conditions.
Stratifying Patient Population
“Initially we looked closely only at diabetes and chronic kidney failure patients in both the Medicare and commercial claims population,” commented Chen. “For both conditions, we identified a very small number of people who had extremely high expenditures.
“In 2008, data from the United States Renal Data System showed that—for a set of 27-million patients in the general Medicare population—the median age was 75.6 years,” noted Chen. “Within this sample, CKD patients accounted for 8.7% of patients but 24.5% of costs and $49.7 billion in spending.
“In this same set of patients, congestive heart failure (CHF) patients represented 13.5% of the population but 35.8% of spending, or $72.6 billion,” he commented. “It was a similar story for diabetes patients who made up 23.6% of the population but represented 36.1% of Medicare spending, or $73.2 billion.”
Chen then described the incidence and costs of these same diseases for a three-million member sample of the commercial population, where the median age is 56.6 years, as follows:
• CKD patients represented 1.3% of the population and spending for these patients reached 7.8% of total spending, or $1.2 billion.
• CHF patients were 1.3% of population and spending for these patients reached 7.8% of total spending, or $1.2 billion.
• Diabetes affected 10.6% of the commercial population and spending totaled 21.5% of total spending or $3.4 billion.”
Sonic Healthcare Adds Value by Using Lab Data in Combination with Tools
WITHIN THE TYPICAL LARGE PRIMARY CARE PRACTICE, there are often gaps in care that can be identified by the clinical laboratory provider. This was the opportunity that Sonic Healthcare used to become a clinical collaborator with certain physician clinics in New York and Texas. Chronic diseases like diabetes were the focus of this effort.
Baseline statistics for diabetes among primary care practice
From its work with different primary care groups, Sonic Healthcare has learned the proportions of diabetic patients that typically don’t have a diabetes diagnosis code, have not been seen in more than 12 months, and have care gaps, as shown above.
Responses from automated patient engagement and pre-visit lab services
Based on physicians’ use of Sonic’s identification of patients who would benefit from getting care and its patient-contact tool, Sonic was able to encourage 44% of patients contacted to see their provider, thus helping to close those care gaps.
Opportunities for Labs
“These statistics demonstrate how much opportunity exists for clinical labs to deliver value that improves patient care and reduces healthcare costs,” noted Chen. “Improving the management of just these three diseases can have a profound impact on outcomes and the cost of care.
“Another client relationship involved a large primary care group caring for about 50,000 patients,” he continued. “Our analysis in iMorpheus produced interesting results. Along with our review of diagnosis codes we also reviewed the lab data for these patients.
“In this group, we identified a typical pattern that we see in primary care,” he observed. “Among 3,700 diabetes patients, almost 700 of them did not have a diagnosis code assigned.
“Why was the diagnosis code missing for these patients?” asked Chen. “Did the doctors forget to add the code after treating these patients? The answers were interesting and represented our lab’s opportunity to add value.
“For this group of physicians, we found that 27% of its patients had not seen a doctor in over 12 months,” stated Chen. “We also found that, from one practice to another, there is a range of about 20% to 35% of diabetes patients who have not seen a doctor for over 12 months.
Patients with No Claims
“For these patients, there are no claims, meaning they are actually the low spenders in the claims analysis, but they have the disease,” Chen said. “If they don’t show up for care, how do they get diagnosed?
“What frequently happens is that these patients see a doctor who suspects there is a problem and orders a lab test,” he explained. “But these patients are asymptomatic and so they don’t come back. That means there is no clinical encounter for the physicians to record the diagnosis in the EMR, even though their screening lab test results showed they have a problem. With no identifying code, they do not get followed or treated.
“But these patients are still sitting out there and the physician groups are responsible for the costs of their care,” noted Chen. “Over time, of course, those people with diabetes develop complications and they show up in the hospital. That’s when they jump from being low-cost to high-cost patients.
Diabetes Under Control
“On the other side, when we reviewed the data on those patients with diabetes who do show up regularly for routine physician care, we saw a significant number of them controlled their diabetes fairly well,” he added. “This observation challenges the premise that clinical decision support tools are effective to remind doctors about what they need to do. That premise may be incorrect. The key issue is not with the doctors, but with patients who do not show up for routine care.
“How do we address this failure-to-show-up problem?” asked Chen. “At Sonic, our answer was to create scorecards and a gaps-in-care roster that lists each physician’s patients with a chronic disease who are overdue for routine lab monitoring.”
In this way, Sonic is moving beyond simply reporting timely, accurate lab test results and is developing tools to help physicians improve patient care.
“The scorecard is useful because most physicians don’t know how many diabetes patients they have, much less who are overdue based on standard clinical guidelines,” continued Chen. “This scorecard and other tools we give them is a first step that allows them to identify those patients. But it doesn’t solve the problem if the patients don’t show up.
CASE STUDY: FQHC
“When we started this program about four years ago, our first client was a large federally qualified health center in New York with about 100,000 patients,” Chen said. “We analyzed their lab data and then sent the FQHC a list of 1,200 patients’ names who needed follow-up care. We said, ‘You need to call these patients because they need to be brought back for a follow-up visit with a physician.
“The chief operating officer of the center looked at me and asked, ‘Why do you expect me to call these patients back? Since you found them, why don’t you bring them back? I’m not looking for a laboratory; I’m looking for a healthcare partner.’
“That response was unexpected and it triggered an interesting discussion about the role of the laboratory in managing patient care,” Chen explained. “From that point, we expanded our ways to help.
“As a clinical lab, we don’t have the personnel to call patients back,” he said. “Therefore, we developed technologies that we can deploy—automated calls, voicemail, e-mail, and text messages—to alert patients that they are due for follow-up visits with their physicians.
“Our lab deploys these methods on behalf of the physicians,” he stated. “A first step is to record the physician’s voice. That way, patients hear their own doctors calling them to say they need to come back for an office visit.
“Here’s how our patient contact program works,” he added. “Each week we send an e-mail to the physician with their patient data. It shows them which patients need follow-up, based on diagnosis codes, prior lab results, and evidence-based care guidelines.
“This informs the physician about how many and which patients are overdue for the needed tests in accordance to the current clinical guidelines,” Chen continued. “The physician and the clinic staff can then check to see if those names on the list are still active patients.
“If a patient has moved away and is no longer active, the physician can opt the patient out,” he said. “Our service gives the physicians a three-click strategy. With one click they will authorize the patient list and select the outstanding laboratory tests for standardized routine care. With a second click, they authorize the ICD codes, and the third click authorizes the lab orders.
Patients Are Contacted
“When they authorize the lab orders, two things happen,” noted Chen. “First, the orders get sent to our patient service centers so that our PSCs can prepare to test these patients when they show up. Second, all the calls, e-mails, or texts go out to the patient.
“If the patient answers the call, we ask them if they’re ready to make an appointment,” he commented. “If yes, we transfer the patient to their provider to schedule a visit. Before we transfer the call, we tell the patients their doctor has ordered lab tests for them and so they should visit the lab before seeing the doctor.
“That process saves the patient an office visit,” Chen explained. “Usually when a patient sees a doctor, the patient will get a lab order and then go to the lab. But then the patient needs to see the doctor again a second time a week or two weeks later for follow-up management. With our system, the patient gets the lab test done before seeing the doctor, so that everything is done in just one visit with their physician.
“In addition to sending physicians a list of patients who need follow-up visits, we also stratify the list by putting the most critical patients on top,” he added. “Our algorithm is based on clinical laboratory data, which allows us to show how severe that patient’s disease markers are and how long it’s been since the patient’s last physician visit.”
“Our service includes an integrated stratification system that uses diagnosis codes to help us predict the likelihood of a high-risk event that could cost a lot of money or an event that might require an inpatient admission to the hospital,” he explained. “This stratification integrates Johns Hopkins ACG algorithms to identify those patients most likely to experience complications and high-expenditure events.
“We are not making a prediction. Rather, our tool is useful to prioritize the list,” added Chen. “Because all these patients need to come back, we put the most needy patients on top.
“In the first 39 primary care practices, we had data on about 200,000 patients and used this technology to follow them for more than six months,” stated Chen. “Among those 200,000 patients, there were about 14,000 who had diabetes.
“Across all the different practices, we found that 15% to 20% of diabetes patients that we identified were based on lab data and they did not have a diagnosis code,” he said. “Between 30% to 65% of patients had at least one laboratory care gap based on the clinical guidelines—such as an A1C test or micro-albumin test. Further, we saw that about 20% to 35% of patients had not seen their physicians for at least 12 months.
“After deploying this technology, 44% of people we contacted returned to the clinic, and half of them returned within 24 hours!” Chen said. “We track and document them with our technologies. For these patients, after we deployed the call, the next morning patients showed up at our PSC to get their blood drawn. This is a surprisingly high response rate, much higher than we anticipated.
“We attribute this response rate to the personalization of the contact by recording the doctor’s voice,” he commented. “This made a significant difference because providers told us their patients appreciated getting direct calls from their doctors telling them to come in. Also, the high response rate is partly due to the patient selection process. We call only those patients who are overdue for their care and most of the patients know they have a clinical need.
“Among the patients who did not come in, we identified that our data was not up to date on 35% to 45% of them. For them, we had what we call a ‘bad data issue,’ meaning we had the wrong phone number or the patient had moved away.
Only 7% of Patients Opt Out
“About 7% of patients opted out of the reminder service,” he said. “This number was lower than we originally estimated because, before we instituted this program, we surveyed about 1,500 people in the general population, not the population that has health problems. For that survey, 48% of people told us not to bother them with such contacts.
“That survey of the 1,500 people in the general population told us that we were on the right track by reaching out to those patients with clinical needs and by personalizing the contacts with the physician’s own voice, e-mail, or text,” he explained. “In this way, we used data beyond genomics to deliver this targeted intervention.”
Contact Philip C. Chen, MD, PhD, at 512- 439-1600 or email@example.com.